Category: Cloud Infrastructure

Cloud Q1 Update: When Will the QoQ Decel Find a Bottom?

Posted on by io-fund

Cloud stocks have treated investors well with recurring revenue, resiliency during Covid, and some of the strongest examples of product-market fit available on the public markets. However, not even this can overcome the effects of lower budgets and cloud spending, which is the top driver in terms of year-over-year comparisons. Due to macro uncertainty, companies are optimizing their cloud spending.

We track the results of big tech companies and best-of-breed cloud stocks to gauge the sentiment of the cloud sector. We came up with our premium analysis in December when we said, “Cloud spending may turn out to be softer than industry surveys indicate, especially until inflation cools off. This is because surveys capture a perception while earnings results are the culmination of a 7.1% inflation rate, plus a softer Chinese market and a softer European market.”

We noticed the sequential decline in cloud stocks in December and saw a steeper decline during our analysis in March. Best-of-breed cloud reported a 71% slowdown in QoQ/YoY growth for Q4 guides and a 83% slowdown in QoQ/YoY growth for Q1 guides. There’s technically an improvement in the recent quarter, as the best-of-breed cloud stocks are guiding for a 72% slowdown in QoQ/YoY growth for Q2 guides.

However, the drastic slowdown will begin to run out of room until either one of two things happen:

  1. The low comps that cloud begins to lap in Q3 becomes a tailwind and companies are able to bounce back from the sudden drop off in growth that was reported last year, beginning in Q3.
  2. Or, the sequential growth will soon turn negative, which will result in a negative reaction in the market. The objective data below shows which stocks are most at risk for this happening.

Cloud Trends: Big Tech is the best proxy.

Big Tech Companies are more insulated than Best-of-Breed, yet offer a 360-degree view as to how the cloud industry is faring. We had said the following in our premium analysis in December.

“The Big 3 are the best proxy because their reports represent the layer in the tech stack that tends to be the most resilient in terms of churn. The switching costs are quite high for cloud IaaS services. The Big 3 also afford a more concentrated view by owning 66% of market share across three companies whereas SaaS is spread across thousands of companies.”

If the Big 3 are decelerating it simply makes it much harder for Best-of-Breed to accelerate given the Big 3 represents an appetite for cloud budgets and provides strong clues if we are in a period of expansion or a period of optimization and digestion.

Source: Company Results

Key Highlights from the Big Three Earnings:

Microsoft

  • Azure grew by 27% and 31% YoY in constant currency.
  • The growth is down 31% in the last quarter and 46% in the same period last year.
  • The company’s guide for the next quarter is 26% to 27% (in constant currency), an expected deceleration of 4-5% from the recent quarter.
  • Even though cloud growth is decelerating, there are some bright spots like AI and the guide for the next quarter includes roughly 1% from AI services.

The company’s priorities are related to the cloud, and subsequent AI spend. Satya Nadella said in the earnings call, “We continue to focus on three priorities. First, helping customers use the breadth and depth of the Microsoft Cloud to get the most value out of their digital spend. Second, investing to lead in the new AI wave across our solution areas and expanding our TAM.” Second, investing to lead in the new AI wave across our solution areas and expanding our TAM.” He further highlighted that Azure is taking market share, especially benefiting from the AI trend.

You can read our recent editorial on Microsoft here.

AWS

  • AWS revenue grew by 16% YoY to $21.4 billion.
  • The growth is lower than the 20% the last quarter and is a remarkable slowdown from 37% in the same period last year.
  • Management discussed in the earnings call that April revenue growth for AWS further decelerated to 11%. This is due to the ongoing tough macro environment, causing customers to optimize their cloud spending in the recent quarter.

The company’s CEO, Andy Jassy, highlighted that enterprise customers were being cautious. “In AWS, what we’re seeing is enterprises continue to be cautious in their spending in this uncertain time. Customers are looking for ways to save money however they can right now. They tell us that most of it is cost optimizing versus cost cutting, which is an interesting distinction because they say they’re cost optimizing to reallocate those resources on new customer experiences.”cost optimizing versus cost cutting, which is an interesting distinction because they say they’re cost optimizing to reallocate those resources on new customer experiences.”

Google Cloud

  • Google Cloud revenue grew by 28% YoY to $7.45 billion.
  • The growth is lower than the 32% growth reported last quarter and 44% in the same period last year.

The company’s CFO, Ruth Porat, said in the earnings call, “Our investments in product innovation, our go-to-market organization and our partner ecosystem delivered strong results as customers across industries and geographies increasingly rely on Google Cloud to digitally transform their businesses. That being said, in Q1, we continued to see slower growth of consumption as customers optimized GCP costs reflecting the macro backdrop, which remains uncertain. In terms of operating performance, we remain focused on driving long-term profitable growth in Cloud, while continuing to invest given the substantial opportunity.”we continued to see slower growth of consumption as customers optimized GCP costs reflecting the macro backdrop, which remains uncertain. In terms of operating performance, we remain focused on driving long-term profitable growth in Cloud, while continuing to invest given the substantial opportunity.”

More on Best-of-Breed Cloud Stocks

To help illustrate how the deceleration has continued for some best-of-breed names, we took a sample of the top-ranking cloud stocks on revenue growth, free cash flow, adjusted operating margin and/or valuations.

Among the best-of-breed cloud stocks, only MongoDB’s guide shows sequential growth. The company’s QoQ growth was 6% last year and is expected to be 7% this year. The largest deceleration was in HashiCorp, which grew 13% last year and is expected to have flat 0% sequential growth this year.

In the previous quarter, we stated the following:

“Among the best-of-breed cloud stocks, only ServiceNow’s guide shows sequential growth. The company’s QoQ growth was 7% last year and is expected to be 8% this year. The largest deceleration was in GitLab, with revenue that grew 12% QoQ last year, is expected to decline (4%) sequentially this year.

Overall, the category is slowing down sequentially (a rather drastic) 83% for Q1 guides compared to the previous year — from an average of 12% QoQ last year to 2% QoQ growth this year.

For example, best-of-breed cloud reported a 71% slowdown in QoQ/YoY growth for Q4 guides and is now guiding for 83% slowdown in QoQ/YoY growth for Q1 guides.”

We now see an improvement in the recent quarter, as the best-of-breed cloud stocks are guiding for a 72% slowdown in QoQ/YoY growth for Q2 guides – from an average 15% QoQ last year to 4% this year.

Source: YCharts/Seeking Alpha

In our writeups following last earnings season, we were wary of flat guides on YoY basis given cloud has decelerated ever since these companies went public. Due to the slowing revenue growth, companies must report perfect earnings reports. Otherwise, the stock is sold off after hours.

We had seen in the recent earnings season that companies like HashiCorp (lost 25% of its value), SentinelOne (lost about 35% of its value), Cloudflare (lost 25% of its value), and Snowflake (lost 10% of its value) as their guidance missed the Wall Street expectations. This is why we prefer to wait until the Cloud sector reports flat to accelerating growth on a QoQ basis. Due to the strength of the products, we believe this could be early 2024 (see below).   

Our Snowflake premium article in January 2022 explained the downside of the consumption-based pricing model, especially during increased macro uncertainty. “Consumption-based pricing has a few drawbacks. For example, its less predictable than subscription revenue and there isn’t a ‘floor’ on revenue, because if consumption declines then so will sales.”

We further highlighted this point in the risks. “Another risk is the company’s consumption billing model, which is inherently unpredictable. This can make growth lumpy and some quarters may disappoint the Street. Investors should expect increased volatility in growth from Snowflake in the near term as new customers ramp consumption.”

In the quarter following this premium note, Snowflake guided for slower growth during the announcement of the Q4 results in March 2022, which led to the stock being down 30% in after-hours. Similarly, the company missed the Q1 earnings, and Wall Street analysts grew concerned, mainly due to the company’s consumption-based model.

Price Action

There has been an improvement in the price action of cloud stocks in Q2 till date, particularly after May, as seen in the chart below. MongoDB is leading among the best-of-breed cloud stocks after announcing the strongest results in the group pictured above. HashiCorp and SentinelOne rank at the bottom, and are reporting steeper deceleration QoQ – 25% for SentinelOne and then a thin 0% QoQ growth for HashiCorp.

Source: YCharts

The I/O Fund’s Cloud Plan:

Generally speaking, the cloud category has weaker bottom lines and is unprofitable. Therefore, it’s our view that the cloud category will need to return to top line acceleration QoQ before it becomes an obvious buy. This is not true for all tech categories as some can lean on operational efficiencies to drive bottom line growth, and this is rewarded by the market.

Certainly, there are outliers each earnings season, yet the outliers last quarter failed this quarter – and vice versa. As discussed, Cloudflare and Snowflake did quite well last quarter yet missed their guidance this quarter. This is not due to the strength of their product rather how unpredictable budgets are in the current environment.

Per Cloudflare’s CEO: “In the first quarter, we continue to witness a challenging business environment, which deteriorated significantly in March when negative headlines emerged related to SVB, the broadening banking crisis and the worsening macroeconomic outlook. With intensifying business uncertainty, companies became increasingly cautious in more deeply scrutinized field, which impacted numerous areas of our business, including a material lengthening of sales cycles, delays in collections and the significant back-end weighting in the linearity for the quarter.”

Snowflake’s CFO said the following: “In Q1, consumption varied from month to month. We benefited from strong consumption in February and March. Starting in April, consumption slowed after the Easter holidays through today.”

For the most part, many cloud stocks are underperforming the Nasdaq YTD, whereas in the past, this category greatly outperformed the Nasdaq.

Source: YCharts

Conclusion:

In order for cloud companies to outperform the Nasdaq again, we will want to see a return to top line growth QoQ and YoY. The reason we track QoQ is because this shows deceleration quickly compared to YoY. The QoQ data above predicts MongoDB would be the strongest stock, and HashiCorp and SentinelOne would be the weakest stocks, and this matches the market’s reaction.

Below are the forward estimates for MongoDB. Assuming consensus is correct for this stock, the rebound would happen around October 2024 with an entry best timed in January 2024 when the company reaches a low point of 10.95%. One area of concern is the 29% in the current quarter won’t return until nearly two years later.

H2 2024 and 2025 look optimal. Should anything change before then on the fundamentals, we will keep you updated.

The I/O Fund Analyst Team contributed to this article

Recommended Readings:

Highlights from Google I/O 2023

Posted on by io-fund

Google recently held its annual developer conference Google I/O 2023. Google is a large real estate owner with arguably more data than any other tech company in the world. This advantage cannot be overstated when it comes to training large language models (LLMs). In addition to having a strategic advantage for future development of LLMs with data, Google can offer advertisers instant ROI.

The primary announcements from the event were:

  • Google drops the waitlist for Bard and announces new features.
  • Google launches new Large Language Model, PaLM2
  • Unveils its new AI-powered Search.
  • Google Cloud announces new A3 supercomputer VMs built to power LLMs.

Google drops the waitlist for Bard and announces new features

Among the more exciting announcements at Google I/O, the company dropped the waitlist for Bard and the chatbot is now available in 180 countries and territories. Bard supports English, Japanese, & Korean languages, and will soon support more than 40 languages. Google is also rolling out features such as better source citations, the ability to export content generated in Gmail and Google Docs, support for more visuals and an upcoming Google Lens integration to analyze pictures and write captions.

Background on Google’s Bard:

Earlier this year, Google’s stock (Alphabet) tumbled 7% when chatbot Bard was unable to complete a search with 100% accuracy. During the demonstration, Bard returned incorrect information about which telescope was the first to take pictures of a planet outside the Earth’s solar system. This was a minor mistake given how far large language models and generative AI has come, rather it was the timing that was a bit flawed as OpenAI’s ChatGPT, the chatbot powering competitor Microsoft Bing, had been dominating headlines since its November 30th launch.

Microsoft, being an opportunist, took it a step further and announced Bing would now be powered by a faster and more accurate version of GPT-3.5 one day after Bard’s failed demonstration: “We’re excited to announce the new Bing is running on a new, next-generation OpenAI large language model that is more powerful than ChatGPT and customized specifically for search. It takes key learnings and advancements from ChatGPT and GPT-3.5 – and it is even faster, more accurate and more capable.”

Both companies have been preparing for this moment for many years. Microsoft invested $1 billion into OpenAI a few years ago with a new $10 billion round announced last month. Meanwhile, Google acquired DeepMind in 2014. Google also previously developed conversational neural language models such as LaMDA, which is used by Google’s Bard for its conversational AI technology.

Despite the mishap with Bard, it would be a human-generated mistake to think Alphabet does not command a place of leadership right now in generative AI. Alphabet was one of the first tech companies to focus and invest on AI and natural language processing (NLP). We pointed out to our premium research members in July of 2022 that ChatGPT is based on transformer architecture that Google initially introduced in 2017 when we said:

“Transformers are becoming one of the most popular neural-network models by applying self-attention to detect how data elements in a series influence and depend on one another.

Sequential text, images and video data are used for self-supervised learning and pattern recognition, which results in more data being used to create better models. Prior to transformer models, labeled datasets had to be used to train neural networks.

Transformer models eliminate this need by finding patterns between elements mathematically, which substantially opens up what datasets can be used and how quickly.

Google first introduced transformer models in 2017 and transformers are used in Google and Bing Search. Transformers also led to BERT models, which stands for Bidirectional Encoder Representations from Transformers, and is commonly used for text sequences. Transformers are also used in GPT-3 (it’s the T in GPT) which improved from 1.5 billion parameters to 175 billion parameters. GPT-3 has the ability to report on queries it has not been specifically trained on.”

Earlier this month, Google’s CEO, Sundar Pichai, gently reminded the AI community of how cutting edge Google’s research is when he stated, “Transformer research project and our field-defining paper in 2017, as well as our important advances in diffusion models, are now the basis of many of the generative AI applications you're starting to see today.”

BERT was designed to help Google better understand search intent, as despite billions of searches every day, about 15% of those searches are for brand new terms. This prompted Google engineers to develop a model that could self-learn.

The result is that searches results are more accurate by taking into consideration the nuances of language.

Google launches new Large Language Model, PaLM2

Google launched a new Large Language Model, PaLM2, that will power the updated Bard AI chat tool and more than 25 other new products & features including productivity software (Gmail, Google Docs), Healthcare and Security.  

PaLM 2 has the following capabilities:

  • Multilingual: The LLM is trained on more than 100 languages, which increases language proficiency
  • Reasoning: The LLM’s dataset has improved logic, common sense reasoning and mathematics
  • Coding: The LLM can generate code including programing languages such as Python, JavaScript and specialized languages such as Prolog, Fortran and Verilog.

Google Unveils its new AI-powered Search

The company has unveiled its new generative AI-powered search that will be subject to a waitlist. Google cites the example of the following search “what's better for a family with kids under 3 and a dog, bryce canyon or arches.” Previously, you had to break the question down into smaller ones, sort through the vast information available, and then put things together yourself. Now with generative AI, search will be able to better understand the question.

Generative AI will also provide a better experience for online shopping by instantly getting relevant information like reviews, images, and ratings. The new shopping experience is based on Google’s Shopping Graph, which has more than 35 billion product listings.

The company announced the ‘About this image’ feature, allowing users to identify fake images. It mentioned in its press release, “When the image and similar images were first indexed by Google, Where it may have first appeared, and Where else it’s been seen online (like on news, social, or fact checking sites)”.

Google launches new Large Language Model, PaLM2

The company launched the new Large Language Model, PaLM2, that will power the updated Bard AI chat tool and more than 25 other new products & features announced during the Google I/O 2023.

Its predecessor PaLM, launched in April 2022, was a 540 billion based parameter, and the company did not provide this detail for PaLM2. PaLM stands for Pathways Language Model. “What we found in our work is that it’s not really the sort of size of model — that the larger is not always better,” DeepMind VP Zoubin Ghahramani said in a press briefing ahead of the announcement. “That’s why we’ve provided a family of models of different sizes. We think that actually parameter count is not really a useful way of thinking about the capabilities of models and capabilities are really to be judged by people using the models and finding out whether they’re useful in the tests that they try to achieve with these models.”

PaLM2 is faster and more efficient than previous models. Some of the improvements highlighted by the company are that PaLM2 is trained for improved multilingual text, spanning over 100 languages, reasoning, and coding, including popular languages like Python & JavaScript. For example, due to the multilingual capabilities of PaLM2, it has helped Bard to expand to new languages. PaLM2 is available in four sizes: Gecko, the smallest, followed by Otter, Bison, and Unicorn. Other use cases include improved Workspace features while working in Gmail, Google Docs, and Google Sheets. PaLM2 can also be used for enterprise use cases like Med-PaLM2 in medical research and Sec-PaLM in cybersecurity.

The company also said that it’s working on a more powerful model called Gemini and it will also be available in various sizes so that it can be easily deployed to various products.

Google Cloud announces new A3 supercomputer VMs built to power LLMs

Google Cloud announced the A3 GPU supercomputer that can be used to train and run Artificial Intelligence and Machine Learning models. While the A3 GPU supercomputer is on a private preview waitlist, the previously announced G2 VMs are now in general availability. The G2 VMs are powered by the new Nvidia L4 Tensor Core GPUs. The company said that it is the first cloud provider to offer these new GPUs for serving generative AI workloads.

The A3 GPU VMs are made of eight Nvidia H100 Hopper architecture GPUs, 3.6 TB/s bisectional bandwidth between A3’s 8 GPUs via the Nvidia NVSwitch and NVLink 4.0, 4th Gen Intel Xeon Scalable processors, and 2TB of host memory.

The A3 supercomputer can deliver up to 26 exaFlops of AI performance, thereby improving the time and cost of training large machine learning models. The A3 workloads will be run on Google’s Jupiter data center networking fabric which the company states “scales to tens of thousands of highly interconnected GPUs and allows for full-bandwidth reconfigurable optical links that can adjust the topology on demand.”

Conclusion:

I would not be surprised if we exit 2023 with a reimagined way to use Search Engines. The iteration cycle here is likely to move quickly compared to AVs or the Metaverse, as there are real-world applications where AI can be applied without safety issues (AVs) or friction in terms of user adoption (Metaverse/VR headsets). Instead, the scale has already been built with Search being a viral, daily activity used by nearly every human on earth. AI advancements will simply improve what is already in place.

Cutting-edge chatbots can be quickly deployed on the search engines that already exist, and this is a substantial difference from other overhyped, early-stage technologies. Their accuracy may still need time, but they're probably not too far off from being deemed “reliable enough.”

Investors should expect that AI will become a winner(s)-take-all market. In time, the difference in how search and other applications operate in terms of user experience plus ROI for advertisers will help carve a larger lead.

Premium Members should check the forum for updates on our timing for an entry into the stock.

Recommended Reading:

Google Stock: Search Is On The Precipice Of Multi-Decade Disruption
Google Faces Biggest Lawsuit in Company History — What Companies Could Benefit

Highlights from Google I/O 2023

Posted on by io-fund

Google recently held its annual developer conference Google I/O 2023. Google is a large real estate owner with arguably more data than any other tech company in the world. This advantage cannot be overstated when it comes to training large language models (LLMs). In addition to having a strategic advantage for future development of LLMs with data, Google can offer advertisers instant ROI.

The primary announcements from the event were:

  • Google drops the waitlist for Bard and announces new features.
  • Google launches new Large Language Model, PaLM2
  • Unveils its new AI-powered Search.
  • Google Cloud announces new A3 supercomputer VMs built to power LLMs.

Google drops the waitlist for Bard and announces new features

Among the more exciting announcements at Google I/O, the company dropped the waitlist for Bard and the chatbot is now available in 180 countries and territories. Bard supports English, Japanese, & Korean languages, and will soon support more than 40 languages. Google is also rolling out features such as better source citations, the ability to export content generated in Gmail and Google Docs, support for more visuals and an upcoming Google Lens integration to analyze pictures and write captions.

Background on Google’s Bard:

Earlier this year, Google’s stock (Alphabet) tumbled 7% when chatbot Bard was unable to complete a search with 100% accuracy. During the demonstration, Bard returned incorrect information about which telescope was the first to take pictures of a planet outside the Earth’s solar system. This was a minor mistake given how far large language models and generative AI has come, rather it was the timing that was a bit flawed as OpenAI’s ChatGPT, the chatbot powering competitor Microsoft Bing, had been dominating headlines since its November 30th launch.

Microsoft, being an opportunist, took it a step further and announced Bing would now be powered by a faster and more accurate version of GPT-3.5 one day after Bard’s failed demonstration: “We’re excited to announce the new Bing is running on a new, next-generation OpenAI large language model that is more powerful than ChatGPT and customized specifically for search. It takes key learnings and advancements from ChatGPT and GPT-3.5 – and it is even faster, more accurate and more capable.”

Both companies have been preparing for this moment for many years. Microsoft invested $1 billion into OpenAI a few years ago with a new $10 billion round announced last month. Meanwhile, Google acquired DeepMind in 2014. Google also previously developed conversational neural language models such as LaMDA, which is used by Google’s Bard for its conversational AI technology.

Despite the mishap with Bard, it would be a human-generated mistake to think Alphabet does not command a place of leadership right now in generative AI. Alphabet was one of the first tech companies to focus and invest on AI and natural language processing (NLP). We pointed out to our premium research members in July of 2022 that ChatGPT is based on transformer architecture that Google initially introduced in 2017 when we said:

“Transformers are becoming one of the most popular neural-network models by applying self-attention to detect how data elements in a series influence and depend on one another.

Sequential text, images and video data are used for self-supervised learning and pattern recognition, which results in more data being used to create better models. Prior to transformer models, labeled datasets had to be used to train neural networks.

Transformer models eliminate this need by finding patterns between elements mathematically, which substantially opens up what datasets can be used and how quickly.

Google first introduced transformer models in 2017 and transformers are used in Google and Bing Search. Transformers also led to BERT models, which stands for Bidirectional Encoder Representations from Transformers, and is commonly used for text sequences. Transformers are also used in GPT-3 (it’s the T in GPT) which improved from 1.5 billion parameters to 175 billion parameters. GPT-3 has the ability to report on queries it has not been specifically trained on.”

Earlier this month, Google’s CEO, Sundar Pichai, gently reminded the AI community of how cutting edge Google’s research is when he stated, “Transformer research project and our field-defining paper in 2017, as well as our important advances in diffusion models, are now the basis of many of the generative AI applications you're starting to see today.”

BERT was designed to help Google better understand search intent, as despite billions of searches every day, about 15% of those searches are for brand new terms. This prompted Google engineers to develop a model that could self-learn.

The result is that searches results are more accurate by taking into consideration the nuances of language.

Google launches new Large Language Model, PaLM2

Google launched a new Large Language Model, PaLM2, that will power the updated Bard AI chat tool and more than 25 other new products & features including productivity software (Gmail, Google Docs), Healthcare and Security.  

PaLM 2 has the following capabilities:

  • Multilingual: The LLM is trained on more than 100 languages, which increases language proficiency
  • Reasoning: The LLM’s dataset has improved logic, common sense reasoning and mathematics
  • Coding: The LLM can generate code including programing languages such as Python, JavaScript and specialized languages such as Prolog, Fortran and Verilog.

Google Unveils its new AI-powered Search

The company has unveiled its new generative AI-powered search that will be subject to a waitlist. Google cites the example of the following search “what's better for a family with kids under 3 and a dog, bryce canyon or arches.” Previously, you had to break the question down into smaller ones, sort through the vast information available, and then put things together yourself. Now with generative AI, search will be able to better understand the question.

Generative AI will also provide a better experience for online shopping by instantly getting relevant information like reviews, images, and ratings. The new shopping experience is based on Google’s Shopping Graph, which has more than 35 billion product listings.

The company announced the ‘About this image’ feature, allowing users to identify fake images. It mentioned in its press release, “When the image and similar images were first indexed by Google, Where it may have first appeared, and Where else it’s been seen online (like on news, social, or fact checking sites)”.

Google launches new Large Language Model, PaLM2

The company launched the new Large Language Model, PaLM2, that will power the updated Bard AI chat tool and more than 25 other new products & features announced during the Google I/O 2023.

Its predecessor PaLM, launched in April 2022, was a 540 billion based parameter, and the company did not provide this detail for PaLM2. PaLM stands for Pathways Language Model. “What we found in our work is that it’s not really the sort of size of model — that the larger is not always better,” DeepMind VP Zoubin Ghahramani said in a press briefing ahead of the announcement. “That’s why we’ve provided a family of models of different sizes. We think that actually parameter count is not really a useful way of thinking about the capabilities of models and capabilities are really to be judged by people using the models and finding out whether they’re useful in the tests that they try to achieve with these models.”

PaLM2 is faster and more efficient than previous models. Some of the improvements highlighted by the company are that PaLM2 is trained for improved multilingual text, spanning over 100 languages, reasoning, and coding, including popular languages like Python & JavaScript. For example, due to the multilingual capabilities of PaLM2, it has helped Bard to expand to new languages. PaLM2 is available in four sizes: Gecko, the smallest, followed by Otter, Bison, and Unicorn. Other use cases include improved Workspace features while working in Gmail, Google Docs, and Google Sheets. PaLM2 can also be used for enterprise use cases like Med-PaLM2 in medical research and Sec-PaLM in cybersecurity.

The company also said that it’s working on a more powerful model called Gemini and it will also be available in various sizes so that it can be easily deployed to various products.

Google Cloud announces new A3 supercomputer VMs built to power LLMs

Google Cloud announced the A3 GPU supercomputer that can be used to train and run Artificial Intelligence and Machine Learning models. While the A3 GPU supercomputer is on a private preview waitlist, the previously announced G2 VMs are now in general availability. The G2 VMs are powered by the new Nvidia L4 Tensor Core GPUs. The company said that it is the first cloud provider to offer these new GPUs for serving generative AI workloads.

The A3 GPU VMs are made of eight Nvidia H100 Hopper architecture GPUs, 3.6 TB/s bisectional bandwidth between A3’s 8 GPUs via the Nvidia NVSwitch and NVLink 4.0, 4th Gen Intel Xeon Scalable processors, and 2TB of host memory.

The A3 supercomputer can deliver up to 26 exaFlops of AI performance, thereby improving the time and cost of training large machine learning models. The A3 workloads will be run on Google’s Jupiter data center networking fabric which the company states “scales to tens of thousands of highly interconnected GPUs and allows for full-bandwidth reconfigurable optical links that can adjust the topology on demand.”

Conclusion:

I would not be surprised if we exit 2023 with a reimagined way to use Search Engines. The iteration cycle here is likely to move quickly compared to AVs or the Metaverse, as there are real-world applications where AI can be applied without safety issues (AVs) or friction in terms of user adoption (Metaverse/VR headsets). Instead, the scale has already been built with Search being a viral, daily activity used by nearly every human on earth. AI advancements will simply improve what is already in place.

Cutting-edge chatbots can be quickly deployed on the search engines that already exist, and this is a substantial difference from other overhyped, early-stage technologies. Their accuracy may still need time, but they're probably not too far off from being deemed “reliable enough.”

Investors should expect that AI will become a winner(s)-take-all market. In time, the difference in how search and other applications operate in terms of user experience plus ROI for advertisers will help carve a larger lead.

Premium Members should check the forum for updates on our timing for an entry into the stock.

Recommended Reading:

Google Stock: Search Is On The Precipice Of Multi-Decade Disruption
Google’s Antitrust Case: Why It’s Important

Amdocs (DOX) – Software Downstream from Big Telecom Capex

Posted on by io-fund

In February, we wrote about Big Tech capex shifting it’s spend to AI-related infrastructure. We've been tracking Big Tech capex since 2021 as a proxy for our semiconductor positions. Although the Big 3 are expected to be flat-ish YoY on absolute spend, these companies (plus Meta), are expected to increase AI investments in 2023. 

Below is a chart of the yearly Big Tech capex spending from 2016 to 2022.

During Covid, Big Tech capex was needed to support the surge in enterprises that migrated to public cloud and hybrid cloud architectures overnight. As offices closed, companies scrambled to provide work-from-home environments supported by the Big 3 cloud hyperscalers, plus these cloud environments directly funneled to security, operations, marketing and sales tech stacks offered by the hyperscalers.

In addition to this, hyperscalers offered instant scale and elasticity for consumer applications. Whether it was streaming content, online shopping and e-commerce sites or gaming, hyperscalers were ramping capex to support this surge in demand. 

Fast forward, and today, the capex spending continues to support the AI/ML ambitions for Big Tech. For example, Microsoft had to use tens of thousands  of Nvidia’s A100 graphic chips to power its AI.

This demand has had a domino effect and spurred capex in other sectors. 

Big Tech meet Big Tele: Can You Hear Me Now?

For example, this consumer demand has been an important driver behind capex Big Telecom (“Big Tele”). Big Tele refers to the large established US telecom companies AT&T, T-Mobile and Verizon. Together they have over 90% of the US market. AT&T has the highest at about 45%.

Big Tele are in the process of upgrading their networks from 4G to 5G. This will enable faster download speeds and increased connectivity between different types of devices. Below are Statista’s estimates of 5G capex since 2019.  This spending began in earnest in 2019 of which Big Tele has been the biggest spender. Similar to Big Tech, Big Tele capex has been multi-year at about 1/3 the size.

We evaluated stocks that could potentially benefit from Big Tele’s capex. One company we identified is Amdocs (DOX, $11B mkt cap). We have focused on software rather than hardware. In part, because after the hardware has been implemented, the software requirements tend to be recurring and the engagements multi-year. In the case of Amdocs, revenue reached an inflection point in 2021, two years after Big Tele 5G capex began.

Who are Amdoc’s customers and what are they looking for?

Amdocs is benefiting from Big Tele upgrading their networks to 5G because as they undergo these hardware upgrades, they need to upgrade their software needs. This includes moving their legacy systems/processes/billing to the cloud and utilizing methods to better monetize their subscribers through ancillary services and offerings. 

The majority of Amdocs clients are Big Tele and International telecom providers. Amdocs receives about 50% of its revenues from ATT and T-Mobile. Currently, Big Tele is focused on growing new revenue streams, cost reduction, and driving more efficient operations because of the ongoing trends of digital transformation, migration to the cloud and next-generation networks. Efforts all aimed at enhancing and monetizing the digital experience for the consumer, much like Big Tech.

Big Tele is investing in 5G and fiber rollout to meet the demand for increased bandwidth and innovation for digital services. This network modernization includes migrating their operational and business systems to the cloud and offering innovative new services for both enterprise and individual consumers that they can monetize.

5G will enable Big Tele to expand within existing and non-traditional business models. For example, Big Tele is partnering with leading suppliers to offer their customers a rich portfolio of offerings including media; entertainment, enterprise enablement; Internet of Things, and digital lifestyle services, all of which are driving Big Tele’s demand for multi-modal customer engagement capabilities and data.

One of the implementation challenges Big Tele faces is trying to rapidly introduce new cloud based applications while still operating legacy systems. Hence, Big Tele needs software providers that can provide modular expansion capabilities as it grows, to reduce these implementation risks.

In a presentation, at the Morgan Stanly Technology conference this month, Amdocs called Big Tele’s 5G initiatives as part of a wider megatrend which includes Cloud and Network Automation.presentation, at the Morgan Stanly Technology conference this month, Amdocs called Big Tele’s 5G initiatives as part of a wider megatrend which includes Cloud and Network Automation.

How does Amdocs help?

As a result, Big Tele is looking for software vendors such as Amdocs that can offer the right software and also provide managed services and end-to-end systems integration. Amdoc’s technological capabilities include individual products for commerce, catalog management, monetization, subscription management, Internet of Things, AI, services and network automation and network development and optimization.

For example, Amdoc’s eSIM Cloud enables Big Tele to launch Internet of Things solutions and monetize experiences on devices from Apple, Samsung, Microsoft, Google and other devices manufacturers.

Amdoc’s cloud based CES21 software suite enables Big Tele to build, deliver and monetize advanced services, leveraging their investments as a 5G standalone network, muti-access edge computing (MEC), software-defined networks (SDN), artificial intelligence (AI) and machine learning (ML). This technology is a visual software development approach that requires little to no coding skill on the part of the user, allowing the rapid development of applications with minimal dependency on IT and code developers. The suite also includes carrier-grade AI/ML based user cases to optimize the customer experience.

This is how Amdocs described their role in the most recent q123 call.

  • Amdocs is helping service providers to modernize and build agility in the 5G era by enabling the rapid launch and monetization of new 5G products
  • We see a growing number of service providers embarking on multi-year cloud migration journeys that Amdocs is supporting with our end-to-end suite of cloud platforms and services

Financial Impact

The “trickledown” effect from Big Tele’s 5G capex and demand for Amdocs software offerings can be seen through two key data points – orderbook and sales. Recall, Amdocs gets about 50% of its revenue from AT&T and T-Mobile.

Order book

There’s been a steady increase in the order book in the past few quarters, which provides strong revenue visibility. As of Q123, the 12-month backlog stood at $4b.

A portion of this consists of ‘Managed Services’ which are typically multi-year and have almost 100% renewal rates.

For example, of the 1.2B in Q1FY23 sales, 60% came from managed services which tend to be recurring. Overall, Amdocs estimates 75% of the total revenue is recurring.

Sales

Big Tele’s capex can also be seen in the inflection in Amdoc’s sales growth in 2021. For FY 2023, Amdocs has guided for a range of 6 – 10% sales growth. In FY 2022, sales grew almost 10%. This inflection point in 2021 occurred about 2 years after Big Tele’s 5G capex began.

What stock attributes do we like?

There are several stocks attributes we find attractive. In addition to the order book visibility and recurring revenues, the following stand out.

FCF generation

Currently, Amdocs has consistently generated free cash flow. The current FCF yield is almost 4% and pays a dividend of 1.8%.

Profitability – Gross Margins and Operating Margins

The inflection point in sales can also be seen in profitability. Gross margins have been steadily increasing since 2021 and through 2022 which was a difficult environment for many tech companies.

The same trend can be seen in GAAP OPM. Non-GAAP OPM have a similar trajectory and currently stand at almost 18%.

One of the attractions of Amdocs is its high portion of recurring revenue streams. Although not subscription based, the recurring nature has a similar impact on operating margins in terms of providing stability in different market environments. 

In 2022, Amdocs’s saw a steady improvement in margins at a time when other companies’ operating margins – such as cloud companies – were declining.  We discussed the differences of a subscription vs consumption model (Insert 12/10/22 blog link?)(Insert 12/10/22 blog link?)

EPS and Sales visibility

Amdocs order book visibility and recurring revenue can be seen in Amdocs positive and defensive earnings growth in 2022 (light blue bar). At a time when many tech companies were revising down estimates, Amdocs either beat or reported in-line earnings. Consensus earnings and sales forecasts – from 2023 to 2026 – paint a similar steady earnings profile.

Amdocs recently reported their q1fy23 earnings where they beat and revised up forecasts. Amdocs stated “Overall, our financial year is off to a strong start, positioning Amdocs to deliver consistent and profitable growth in fiscal 2023 within a global macroeconomic backdrop that remains challenging and uncertain.”. Amdocs reiterated their sales growth target of up to 10%.

Consensus has conservatively modeled 6% y/y sales growth, which provides an opportunity for upside surprise. In 2022, Amdocs had 10% sales growth.

Investment Summary

Taking all these factors into consideration. These are the investment attributes that we find attractive.

  • Well capitalized customer base – Big Tele is well funded and has embarked on strategic investments. Amdocs provides critical software to allow them to compete 
  • High switching costs – Once Big Tele begins to work with Amdocs, it makes it harder for them to switch to another provider given the type of mission critical support that Amdocs has provided. Amdocs has close to 100% retention rate
  • Revenue visibility – Amdocs has 12 month revenue backlog of $4.1B long which provides strong visibility into future earnings and is a positive reflection of the current demand environment 
  • Recurring revenue – Amdocs estimates that 75% of its revenue is recurring which means earnings will be more resilient in difficult macro environments
  • Steady and increasing profitability – given the strong visibility on top line growth, Amdocs has been able to focus on profitability. Operating margins have gradually increased through different market cycles. Amdocs has guided for 18% opm in 2023.
  • Financially strong – Amdocs generates healthy cash flow used to buy back shares and pay dividends. It targets 100% fcf conversion and has forecasted $700m in fcf which equates to a 6% fcf yield.
  • Consolidation beneficiary – When Big Tele acquires another competitor, Amdocs helps in the processes involved in migrating customer information, billing etc. For example, Amdocs should benefit from T-Mobil’s recently announced acquisition of Ryan Reynold’s Mint Mobile 

How does valuation look?

Currently, DOX’s valuation is not demanding based on 2024 EPS and trades at discount to the SPX. It has traded as high as 20x earnings in the past. Given the defensive nature of DOX’s earnings, strong balance sheet and FCF generation, we believe the market will pay a higher multiple for this type of business model and earnings visibility in the current economic environment.

We see a valuation potential of between $120 to $130. 

Amdocs Technicals

By Knox Ridley

Since the COVID low, DOX has been tracing a very large termination wedge pattern. The question remains – has it topped, or does it have one more larger swing before we complete the pattern? As long as any weakness holds the $82-$80 region, we could see a possible buying opportunity for the push into the $103-$110 region. If we break below the $82-$80 region, the odds favor the top being in. Like many stocks, DOX is marching towards a bigger top. We believe for the long-term investors, patience and lower prices will pay off handsomely for quality stocks.

Snowflake Q3 Earnings Report: Bold FY2024 Guide

Posted on by io-fund

The stock initially sold off about (13%) on lower product revenue, which went from 67% reported in Q3 to a guide in Q4 of 49% to 50%. This implies total revenue will follow as the two are closely related with product revenue making up 94% of total revenue. Technically, the product revenue guide implies a miss for Q4 as estimates were for 52.78% growth.

On the call, the stock regained some ground to finish at (6%) AH once management stated the full year 2024 product revenue would be 47%. Technically, this is also a miss as analysts had the number at 51% growth.

I'm not confident SNOW can resist further decel throughout FY2024; that's a bold guide with little to back it up.

It requires a lot of faith in management because they are essentially saying even though Q3 to Q4 dropped by 17% on product revenue growth (67% to 50%), they will only see a 3% drop for the following four quarters or an average decel of less than 1%.

It's also odd to get a FY guide right now as I looked at Q3 2022 call and it was not provided at that time. Yet, we are in even more uncertain macro backdrop than last year.

My concern is if the guide was intentionally provided to soften the blow for Q4’s product revenue meanwhile management has zero visibility to provide reliable guidance of this kind. I bring up zero visibility because at one point, management admitted to not having enough visibility to know how RPO will turn out for Q4, and yet they’re guiding early on FY2024 revenue.

The Q4 guide of 50% is only acceptable if there is no further decel. Otherwise, analysts will model a lower FY2024 and that would have hurt the price quite a bit. The motivation here is clear to me (give a full year guide that implies no deceleration) but how realistic is this? I discuss my thoughts in more detail below under Additional Notes.

Snowflake Financials Overview:

Snowflake beat on product revenue in Q3 with revenue of $522.8 million, and growth of 67%. Management had guided for $502.5 million, at the midpoint. Total revenue tracks in line with product revenue for revenue of $557 million and growth of 67%.

I’ve got a miss on Q4 in my notes as the product revenue growth of 49% to 50% will cause total revenue to fall short of estimates at 53.4%. MarketWatch also confirmed this is miss on product revenue with management guiding for $537.5 million at the midpoint while analysts were expecting $553 million.

The full year is in line with management guiding for $1.91 to $1.92 billion in product revenue, compared to analyst expectations of $1.92 for growth of 68%. Total revenue for the fiscal year is expected to be $2.04 billion for growth of 68%.

GAAP EPS was in line at ($0.63) although this is steeper than the ($0.51) EPS in Q3 last year. The adjusted EPS of $0.11 beat estimates of $0.04.

There were no surprises with the gross margins, GAAP and adjusted GM across total revenue and product revenue were all in line.

GAAP operating margin of (37%) was better than previous quarters. Adjusted operating margin of 8% was also better than previous quarters and came in higher than the 2% guidance offered by management.

GAAP net margin of (36%) was also better than previous quarters.

Snowflake silences the debate on if stock-based compensation affects stock price. The company paid SBC of $229 million in the most recent quarter, or 43.8% of revenue. Yet, the market overlooks this high SBC margin and Snowflake trades at the highest valuation in the cloud universe.

Key Metrics:

RPO was in line at $3 billion with a slight acceleration from last quarter, at $2.7 billion. As stated on the forum prior to the earnings report, “For Snowflake, the market has accepted flat sequential growth on RPO as it really comes down to Q4 for Snowflake — this is where the majority of RPO growth happens sequentially. Last year, the company had sequential Q3-Q4 RPO growth of 30%.”

Net retention rate of 165% is lower than 171% reported last quarter but still a category leading number.

Within customer growth, Global 2K customers accelerated from 15% last quarter to 18% this quarter. The customers with TTM > $ 1million decelerated although still healthy at 94% growth.

Total customer growth was at 34% compared to 36% growth last quarter.

Additional Notes:

My contention is that with the 51% growth estimates — and now management’s guide for 47% — this assumes consistent consumption as we move into a possible recession.

The estimates pictured above help to reveal how little variation there is built into throughout FY204 estimates, and yet this year has shown us Snowflake is capable of wide variability. There are some flat quarters, such as Q1 to Q2, but inevitably there was a strong decel and it’ll require serious trust in management to assume no further material decel from Q4.

Here is what was discussed in the call:

“Sanjit Singh

This is Sanjit Singh for Keith. I wanted to go back, Mike, to some of the guidance framework that you laid out for us, particularly with respect to fiscal year '24, I think you talked about 47% growth. Is there any way you can sort of draw the bridge for us in terms of next quarter you're guiding to think about 49% at the high end; and then for the full year next year, approximately 47%. What sort of gives you the confidence that your Q4 exit growth rate is going to be durable going into next year?

Mike Scarpelli

Sure. Well, I'll say Q4 is — it is a quarter that has a lot of holidays in it, and we do think we've lived through COVID that people are traveling more. There is a big human component as well, too. So we all along have been forecasting that Q4, we'd see the impact of that, but we also have a number of significant customers that we have signed up, that we see them ramping up next year on Snowflake as well as some of the things we're doing with Snowpark with Python, we're starting to see traction in that as well, too. But we think that's going to be more of a 2024 impact.”

Product revenue is expected to grow 3% between Q3 and Q4. Last year, it grew 15.3% sequentially. So, even with the holidays being in Q4, the slowdown is much more prevalent this year than last year’s comps.

Notably, we covered Snowpark with Python last quarter and agree it’s a strong offering that investors should pay attention to. General availability went live Nov 7th.

There was a moment that Snowflake discussed October being weaker than expected, which further complicates a bold FY2024 guide.

“Gregg Moskowitz 

Congratulations on delivering very healthy product revenue in this environment. My question relates to Q4, where obviously the product revenue guidance was below where consensus was. And I'm curious, how much of this Mike is a reflection of a moderation in consumption in the month of November or over the last six weeks, as you said, in APJ and across the SMB as opposed to embedding more conservatism amid the existing macro uncertainty. Would you say it's tilted more towards one versus the other?

Mike Scarpelli

Well, the way we do our forecast is based upon historical performance, and we definitely did see a slowdown in the month of October, not that dramatic, but we typically would see week-over-week growth and we saw a number of weeks where it was pretty flat. I will say November is starting to tick back up again, and that's all factored into the guidance given the macro backdrop we have right now.”

Conclusion:

Thankfully Snowflake is not down 20%+ right now like many of its cloud peers. Yet, I can’t quite get comfortable with the FY2024 guide as it assumes no further decel from a company that decelerated quite a bit from the last year’s Q4.

As you saw today, we are not looking to go to the river with a stock that is priced 30% higher than its peers in a cloud-conscious market and that has now provided a guide that is hard to believe. Snowflake is capable of exuberant price action so we will us technicals for entries and exits.

Twilio Entry: Valuation is Attractive

Posted on by io-fund

Twilio is often categorized as cloud yet its business model is truly unique in every sense of the word. It’s not only the company’s gross margin of 50% versus the typical 70% that separates the company, but conversely, it’s also the lower sales and marketing expenses and lower R&D that make this company unique.

Here’s how Twilio compares to Top 10 cloud valuations on sales and marketing at 32% of revenue for sales and marketing (S&M):

Here’s how Twilio compares with Top 10 cloud valuations with 16% spent on R&D:

When you combine the two line items, Twilio makes up for its lower gross margin with lower S&M and R&D. These line items have been range bound with the percentages seen above since 2014 with the company going public in 2016 and some financials from 2014 being available. Therefore, this is not due to scale or any type of customer mix from the pandemic or some kind of economic pressure.

I’ve made the argument that I believe Twilio has established a defensible moat when I covered the stock for MarketWatch in July of 2020. This elusive moat is why I believe those line items track much lower for Twilio as a platform compared to cloud software companies. Here’s what the article said from two years ago:

Twilio works with 1,000 mobile carriers in over 150 countries. The logistics around this are a nightmare for most developers and also for potential competitors. Uber developed this in-house but it was a rare situation where SMS and voice APIs are integral enough to warrant the investment […]

In 2017, Uber had a $60 million annual run rate with Twilio and decided to source different vendors. Even at this level, the company did not build the capability in-house and still remained with Twilio for high-performance messaging. Eventually, Uber migrated to Twilio’s main direct competitor — MessageBird.

Despite chipping away at Twilio since 2011, MessageBird has about 15,000 customers compared with Twilio’s 165,000. The company has been able to do this by serving the developer market and building out an extensive API library and documentation in which telecom features are integrated through a few simple lines of code.”

*Twilio now has 275,000 active customer accounts and MessageBird has 25,000 customer accounts. Some of Twilio’s growth is from the acquisitions.

Twilio raised the number of registered developer accounts from 5 million to 10 million at the 2020 Investor Day. This puts Twilio second to companies like Apple and Nvidia in terms of its popularity with developers. Point being, there are puts and takes with this company – gross margin is low but operating expenses are also low from strong management execution.

Rather than Twilio’s gross margin weigh on the company, I believe it could be the GAAP profitability first and foremost (which is a line item where we are not seeing improvementwe are not seeing improvement in the near term) and then it was the contracting adjusted profitability margin and contracting free cash flow margin second (adjusted OM and FCF are reversing course, which is why we entered).

We also entered the stock because the forward revenue growth is actually quite high and puts the company into a higher valuation bracket than where it currently participates – but this is only important if we see some improvement in bottom line numbers as the GAAP profitability is where Twilio ranks low as a category peer.

The improvement in FCF and adjusted operating margin was not as easy to see in the previous earnings call because the company had only announced plans to be remote first resulting in lower real estate costs. The 11% layoffs announced in September help to complete the picture.

I do believe the end result will be Twilio becoming (narrowly) free cash flow positive, which I expand on below. Management has already stated they will be profitable on an adjusted basis “by early 2023” and analyst consensus agrees with this. Twilio has not missed a top line estimate or a bottom line estimate for four years (as far back as the history allows) and so the probability the company meets this adjusted profitability expectation in early 2023 is high.

It was primarily valuation that caused us to enter as the company is priced for the worst case scenario. Yes, macro can force valuations lower yet I believe Twilio should be in the 5.5X NTM revenue bracket, and instead, the stock is trading at 3X NTM revenue. The 5.5X NTM cohort still has messy GAAP operating margins yet many are FCF positive.

With quarter or two of a return to FCF positive, Twilio can contend with 5.5X up to 7-9X NTM revenue given its growth rate. Technically, this is still at discount as the organic growth rate would put the company in the 11X NTM revenue bracket. However, GAAP profitability weighs on Twilio and must be discounted, which we discuss below.

Twilio Financials

The company reported revenue growth of 41% in Q2 and is expected to report growth of 31% next quarter. Last quarter, the company reported revenue growth of 48%. This is down from 70%-80% growth in 2019 and down from 60%-70% during 2020-2021. The management has stated their goal is 30% organic growth.

The company’s organic growth was 33% in Q2 with $37 million from acquisitions and $44 million from telecom fees. In Q1, the company reported 35% organic growth.

Twilio’s organic growth is broken out because the company collects and then pays 10DLC A2P telecom fees and the company has also had a string of acquisitions (SendGrid, Segment, Zipwhip and Syniverse, for example).

Next quarter is key because the company will only break out A2P telecom fees for total vs organic as the acquisitions will have lapped one year. For comparison purposes, had this change occurred last quarter, then Q2 revenue growth would have been reported as organic growth of 38.5%.

The company’s net retention rate has slowly eroded over time, currently at 123% down from 127% in the previous quarter and down from 131% in the year ago quarter.

The company reported a GAAP gross margin of 47% and an adjusted gross margin excluding A2P fees of 54.3%. The GAAP GM was 48% last quarter and 47% two quarters ago yet the company previously had gross margins at or above 50%. Any slippage here is penalized.

Management stated the gross margin is lower because international is increasing in percentage of revenue at 35% compared to 33% in the year ago quarter. Here is what was said:

“This decline was primarily driven by the strength of our international messaging business. U.S. 10DLC carrier fees reduced gross margin by approximately 330 basis points in Q2. The continued success of our messaging business often serves as a critical foot in the door with our customers, opening up opportunities to add more value with our higher margin software products over time. We believe this will ultimately drive higher gross margins, allowing us to achieve our long-term non-GAAP gross margin target of 60%+.”

This quarter stood out for having less growth YoY in the company’s total gross profit at 34.5% YoY growth to $445M in gross profits compared to 59% YoY growth last year for $331M in gross profits in the year ago quarter.

Twilio stated goal is to reach 60% GAAP gross margin from a mix of the software product and the core product. Although this may take a couple of years to reach 60% GM, I believe with early signs of GM expansion, the stock will be rewarded.

The adjusted operating loss for the quarter was ($7M). For Q3, the company is guiding for adjusted losses of ($65M) at the midpoint which includes a one-time expense for employee sabbaticals of $35M. As stated, Q3 is expected to be the bottom for Twilio’s adjusted operating losses. Let’s hope it’s also the bottom for Twilio’s gross margin although international mix creates some uncertainty on where this could bottom.

The company reported adjusted EPS of ($0.11) compared to ($0.05) in the previous quarter and is flat from the year ago quarter. The adjusted EPS for Q3 is expected to be ($0.35) and will be ($0.13) EPS in Q4. As stated, by 2023, Twilio will be profitable on an adjusted basis.

Per the CFO on the Q2 earnings call:

“And you probably noted that we've taken some actions with respect to slowing down hiring except for some key areas. We're taking — we guided based on having a real estate charge, which I think reflects our kind of remote-first approach to the way that we're going to work going forward. And I think both of those things will drive profitability into next year. And I think irrespective of the macro environment, we're intending to be profitable next year no matter whether it has an impact on growth or not.”

Ultimately, Twilio’s Q2 earnings report weighed on the company due to a slipping bottom line. GAAP EPS was ($1.71) versus ($1.31) a year ago. Stock based compensation was at 25% of revenue for $242M, up from 18% of revenue last quarter.

Looking forward on GAAP EPS, there are one-time items that will weigh on Q3’s GAAP EPS. The first is a $100 million expense due to real estate closures as Twilio is moving toward being remote-first. This this was discussed in the prepared remarks: “We also announced our decision to become a “remote-first company” and made a significant reduction in our real estate footprint. We’ll continue to closely monitor the returns on our investments, and make adjustments as needed, in an effort to recognize further efficiencies as we scale.”

The company also recently announced layoffs, which will result in expenses of $70 million to $90 million. This one-time expense will also be primarily recognized in Q3 – so Q3 setting up to have a couple hundred million dollars of extra one-time expenses in the GAAP operating losses.

Twilio had negative free cash flow of ($78.5M) in the most recent quarter with a FCF margin of (8.3%) this is down from (4%) in the previous quarter although a slight improvement from the (11%) in the year ago quarter.

I believe what the market will be looking for is comments on free cash flow as the company LTM FCF was ($253 million) but with the two reductions — real estate from remote work and 11% headcount reduction, Twilio has to ability to become FCF positive very soon – in the next two quarters (Q4 and Q1) once the expenses pass and the reductions start to take effect.

Due to the market being forward looking, this is likely the line item that will help the stock’s price if management can appease The Street. I believe that’s the goal as Twilio has a median salary of $150,000 and at 11% reduction results in $150 million for the year that follows Q3. If Twilio targets a small percentage higher than its median, the company can get this number to $170 million on headcount alone. It’s also reasonable to assume Twilio will see $75 million reduction in real estate expenses. Twilio's lease ends in 2025 so the $100 million fee has to be able to offset three years left in a lease. Maybe it's closer to $50 million but somewhere in that area to justify the $35 million termination fee per year averaged over three years (is my thinking).

Regarding the headcount reduction, Twilio did state that it would come from the core product where they will push customers to be more self-serve. Due to the developer friendly documentation the company provides, they do not believe there will be any additional churn from this transition. From there, they will focus on new hires for the software team.

In addition, Twilio raised equity in February of 2021 and debt in March of 2021. The company currently has debt of $986M and cash and marketable securities of $4.4B down from $5.4B last year due to a $750 million equity investment in Syniverse. Syniverse was valued at $2.85 billion at the time of that Twilio became a large shareholder. Of this, the company has cash of $800 million down from $1.49 billion in the year ago quarter.

Modeling Twilio’s GAAP Profitability

I thought it would be interesting to see a financial model for Twilio’s GAAP profitability given how many moving pieces there are — primarily, operating cost reductions yet an intention to ultimately increase headcount on the software side. Also, I thought it would be interesting to see the impact a 62% gross margin would have if/when Twilio gets there. You can see below this would result in GAAP EPS of ($0.84). This is an improvement but the company will not be GAAP profitable (by this estimate) until the company scales to much higher revenue. The gross margin will not be enough.

You can see the entire model on this Google Doc here. It was prepared by Arun Gopalakrishnan, a Financial Analyst that we are contracted with.

The following information was used for the model:

Expect sales & marketing expense to continue expanding albeit slowly "as we continue to expand our sales efforts globally" coupled with hiring freeze (Source: Q2 2022 page 31 Financials)

Expect R&D to increase as company continues to hire in growth areas and "focus on enhancing Twilio Segment and Flex products and strengthening platform infrastructure". R&D will house most of the increase in personnel costs (Source: Q2 2022 page 31 Financials)

We're focusing our hiring efforts on areas that we believe will unlock significant value and present strong opportunities for continued growth such as Segment, Engage and Flex, and we’ve frozen the vast majority of new hires and backfills outside of these areas. (Source: Khozema Shipchandler – Q2 2022 Earnings Call Page 9)

We also announced our decision to become a “remote-first company” and made a significant reduction in our real estate footprint. (Source: Khozema Shipchandler – Q2 2022 Earnings Call Page 9)

One-Time Expense of $35M in Q3 2022 for new sabbatical program for tenured employees & non cash impairment charge associated with the closure of several offices (Source: Khozema Shipchandler – Q2 2022 Earnings Call page 10)

Excludes One-Time Expenses, Non-Cash Charges, Capitalizes majority of R&D expense. Management is moving salaries to performance based salaries (non-cash) so not included in Non-GAAP.

Source: TWLO Profitability Analysis Google Doc Link Here

What is Twilio’s Thesis?

For Members who have been with our site for a few years, you’ll recall that I’m quite bullish long-term on Twilio. This has not changed, rather macro forced us to step aside and see how management weathers the new FCF-forward reality. My product thesis is best described in a 1-hour webinar from April 2021 but I’d like to take the opportunity to reiterate a few points.

An important catalyst to look for is when Google eliminates third-party data from Android and Chrome when the company officially launches Privacy Sandbox. This was expected to occur in 2023 yet has been delayed to 2024 although is being beta tested right now.

We want to watch Twilio closely around this time as its customer data platform is based off the importance of first-party data, yet the transition that Apple began with deprecating the IDFA and introducing App Tracking Transparency (ATT) will not be complete until Android and Chrome participate.

The results will be mobile identifiers and third-party cookies become obsolete by 2024 (note: I do wonder if this timeline will be moved up to end of 2023/very early 2024 if Google continues to see macro headwinds to its ad business).

Similar to how Twilio was a few years ahead of the mobile app economy, the management is very early is very early to the shift toward leveraging first-party data. Phone numbers and email addresses make up the most valuable first-party data available on the internet called personally identifiable information (PII). I believe the Twilio management team has been planning for this for 6 years since the 2018 SendGrid acquisition which combines their core product’s access to phone numbers with a trove of email data. This was a very smart strategic move that is not fully recognized yet until access to third-party data is eliminated, which will put immense pressure on companies to make the most of their first-party data. The caveat is the management team must execute.

My understanding is that very few companies will be able to help unlock the enormous value of first-party data as Twilio has a natural path to collecting phone numbers and emails on behalf of B2C companies.

Previously, I had written the following which I still believe to be true yet will take time to fully recognize (Google’s move being a lynchpin for Twilio competitors):

“Keep in mind, the power of data exponentially increases so each signal Twilio adds is not a linear increase in ROI from a marketer’s perspective – it compounds. In other words, by having four signals, Twilio can beat a competitor like Mailchimp with SendGrid because of the cross-channel data and touchpoints they offer*. The same could be said for Zendesk versus Twilio Flex as to why Flex might be stronger (cross-channel data points).

This is why, on a product-level, I see Twilio increasing its market share against competitors.”

Note: Twilio is not the marketer, the customer remains the marketer (Nordstrom’s, Kroeger’s, Chevrolet, Proctor & Gamble, etcetera) remain being the marketer. Instead, Twilio helps those companies’ market more their databases more effectively.  This is not the same asnot the same as third-party datathird-party data where Nordstrom’s (for example) mixes data that is not their own to market customers. This is done through third-party platforms with access to third-party data – which is a marketplace entirely enabled through mobile IDs and cookies.

Twilio is building the antithesis of a third-party ad exchange or marketing platform — and doing so well in advance of the first-party data era which officially began last year with Apple.

This was discussed on the Q2 earnings call:

Q: This is Abhinav on for Siti. I guess the first kind of question would be just with Google delaying recently their application of third-party cookies now, again, from 2023 to 2024, have you seen or do you expect this to change the demand environment a little bit around CDP's customer engagement more broadly and maybe that transition away from third party? And do you see some of the urgency kind of diminish for some of your customers that are coming in?

A (CEO, Jeff Lawson): Thanks for the question. Yes, I think what we see is the — it's taking a little bit of pressure off of companies. I mean, if you think about it, like 2023 is not that far away. So in some ways, it's actually pretty reasonable to give a little more time for such a big change in the Internet ecosystem to roll through.

That said, I mean we did a survey, and it was like — I think it was 70%, some number like that, of companies were not ready for this change. Now that creates demand for our products. But you also can't imagine that, that number of companies are going to magically like put the switch and completely flip their technology stack in just a month's time. And so I think it's giving a little bit of pressure release now.

But it still is a great environment for the CDP given that customers actually do have a lot more time to actually make these thoughtful changes, and we are already seeing some fantastic stories from the customer base emerge that really is giving our customers confidence that this first-party data approach is not just going to be tenable, like it's a big change for folks. But actually, it's going to be tremendously beneficial.”

Twilio also frequently discusses the higher ROI they are proving through case studies from using first-party data instead of only relying on third-party data for marketing purposes. Per the Q2 earnings call:

“A couple of the stories that we've shared publicly is Allergan using Segment was able to get a 41% reduction in their cost of customer acquisition, which is — those are amazing numbers. Another great customer story is from Domino's, who in Mexico was using Segment to build smarter customer audiences. And as a result of that, doing better ad buying, and they saw their return on ad spend increased 700%.”

In 2021, we had also discussed Twilio moving from being developer-centric to now also becoming sales/marketing-centric: “This also means Twilio is going through a pivot, not only on the product level but the customer level. Although Twilio will remain a developer-centric company, they will also need to appeal to the marketing and sales departments […] Developers will, of course, be the ones to implement Twilio’s APIs and products but there’s evidence that Twilio’s target customer is expanding to also include “marketing, sales and service leaders.”

This piece of the pivot (the new customer base being enterprises rather than developers) can be seen as either an obstacle or an opportunity. Twilio has certainly increased its serviceable addressable market and any success here will be seen in the gross margin long-term due to software sales.

5G is also a catalyst for Twilio. The company is quietly building more IoT products and will be able to tap into this market when adoption rises. I’ll leave that for a future deep dive when we revisit 5G from the consumer and enterprise standpoint. Marvell would give us the first signal this market is picking up since they are equipment related.

Risks:

A risk to consider is that Twilio’s gross margin could continue to soften from a higher international mix. Political ads will take effect in Q3 with the majority of spend recognized in Q4, which should bump up domestic spend. In Q4 of 2020, political spend was $23 million. The breakdown was not available for Q3 of 2020. Here is what was discussed in the Q4 earnings call:

“Fourth quarter non-GAAP gross margin was approximately 56% and was negatively impacted by 150 basis points from A2P fees. Twilio's gross margin, ex-Segment, was approximately flat quarter-over-quarter, aided by political traffic in the United States.”

If GM continues to soften, then we will look at how this affects the bottom line and make our decision based on both top line and bottom line margins.

Another risk to consider is Twilio’s revenue estimates are quite high in the current environment considering most ad/marketing companies are seeing reduced spend. However, with the current information we have, it seems B2C companies in some industries are less reluctant to switch off SMS compared to ads. Twilio said the following about their current trends:

“Now we did call out a couple of pockets of softness in Elena's section of the prepared remarks, where we commented that, for example, in crypto or social or on-demand related activities that we're seeing a little bit of slowdown. But on the flip, we are seeing some strength as well in financial services and IT-related spending.”

If Twilio misses on revenue, we will be forced to step aside as every management team needs to build trust right now with Wall Street given the bearish market. This is true for most of our stocks.

The third risk to consider is further acquisitions. We will close our position and step aside if this happens as the company has tested the upper limit of our appetite for M&A activity.

Additional Resources:

Earlier this week, we published a forum post about our entry in Twilio. There are additional posts from Members on Twilio that also articulate the pros/cons of this particular company and are well worth the read.

I wrote about Twilio’s gross margin in Q2 2022 here.

Twilio PDF 2021 Analysis
Twilio 1-Hour Webinar

Barron’s Podcast: What the Heck is Going on with Cloud Valuations

Posted on by io-fund

Earlier this week, I/O Fund CEO and Lead Tech Analyst Beth Kindig joined Jeremy Owens, Tech Editor, and San Francisco Bureau Chief of MarketWatch, on Barron’s Live. They discussed cloud valuations including those that are trading at 2X above Covid lows, what metrics matter when evaluating cloud companies, and what to watch for in upcoming earnings season — including a few comments on ad-tech. Barron’s Live. They discussed cloud valuations including those that are trading at 2X above Covid lows, what metrics matter when evaluating cloud companies, and what to watch for in upcoming earnings season — including a few comments on ad-tech.

Metrics and Valuations

As discussed in the podcast, the FOMC decisions have forced tech investors to look for cloud stocks that are expanding their margins and also have positive free cash flow. If you look at the best-of-breed companies that command the top 10 in valuations, the majority of them are free cash flow positive.

We had discussed with our premium research members back in May in a special report Compartmentalizing Cloud Stocks that “It’s true that cloud is deflationary but it’s also true that cloud can have profitability issues […] cloud is quite resilient in terms of growth, due to being deflationary, but those weak bottom lines may be questioned over time. Cash came easy over the past decade, and as cloud investors, we need to reframe our thinking on what constitutes an attractive cloud stock.”

Free cash flow is emerging as an important metric because cash gets rerated in a rising rate environment. As stated, not only were many cloud companies were not public during the previous rising rate environment of 2017 to late 2018 – but in addition to this, the previous rising rate environment was quite tame and we are currently in a more aggressive rising rate environment.

Along with free cash flow, GAAP operating margins are being closely examined. This has resulted in companies with high stock-based compensations being penalized during earnings.

The takeaway is that a best-of-breed company with a 15X or higher valuation must remain FCF positive or it will immediately lose its category high valuation. Revenue growth alone is not determining the top spots in this category any longer. This may seem obvious at first thought but we have found it’s better to close a stock at a higher valuation if it has contracting margins.

Sign up for I/O Fund's free newsletter with gains of up to 403% to get analysis like this delivered straight to your inbox every week. Sign up for I/O Fund's free newsletter with gains of up to 403% to get analysis like this delivered straight to your inbox every week. Sign up for I/O Fund's free newsletter with gains of up to 403% to get analysis like this delivered straight to your inbox every week.

The difference between Subscription and Consumption Models

Consumptions models occur in the Big Data and Analytics trend where data storage, processing, and analytic solutions are based on usage rather than on a recurring subscription fee. This trend is becoming popular because with consumption-based pricing model, revenue is uncapped. The consumption billing model does not have a ceiling on revenue, so if customer consumption rises, so does sales. There is what is meant by uncapped revenue potential.

We covered Snowflake’s Consumption Model in January of 2022 when we said in our free newsletter: “While Snowflake uses a “land -and-expand” sales strategy, it also uses a consumption billing model. For instance, Snowflake bills customers based on the amount of data they store and transfer and what resources they use. Accruing revenue based on consumption rather than a ratable subscription model decreases the predictability of quarterly revenue, but it leaves revenue uncapped. This provides revenue upside, because if consumption soars, then so will revenue.”

Some of the drawbacks, however, include the revenue growth being less predictable than subscription revenue. There also isn’t a floor on revenue because if consumption declines, then so will sales. Contracts help protect against this but are often only 1/3 of next 2.5 years of revenue.

The drawbacks were also discussed in the Snowflake’s Consumption Model article in January of 2022, “Another risk is the company’s consumption billing model, which is inherently unpredictable. This can make growth lumpy and some quarters may disappoint the Street. Investors should expect increased volatility in growth from Snowflake in the near term as new customers ramp consumption. However, management does expect revenue growth to smooth and become more predictable in the aggregate as customer consumption scales and matures on the platform.”

The lack of predictability is seen in Snowflake’s earnings history with Q1 earnings reporting revenue growth of 85% YoY to $422.4 million (beat estimates by 2.3%). However, the GAAP EPS missed by $0.02. The management had a hard time convincing the analysts in the earnings call that the company’s revenue was not discretionary and the consumption was lower due to shifting economic circumstances that impacted certain customers, particularly consumer facing cloud companies.

The company’s CFO, Mike Scarpelli, said in the earnings call, “Consumption patterns may fluctuate from quarter-to-quarter. This variability does not detract from our long-term opportunity. Customer’s overall demand for Snowflake remains unchanged. This is supported by the contractual commitments they are making with us and their longer-term plans for adopting the data cloud across their organization.”

In the podcast, we also discussed how net retention rates are often higher for consumption models as spending ramps over time and is uncapped. It’s easier to re-accelerate here for that reason and it’s not the best apples-to-apples comparison for subscription NRR. The net retention rates for subscription-based companies are in the range of 130-140 range while Snowflake has remained in the 170 range.

Another metric is the remaining performance obligation (RPO). When customers sign onto the platform, they purchase consumption at specified prices, which gets recorded as remaining performance obligations (RPO). These contracts are for about 2.5 years. Although these key metrics are important, as mentioned earlier, what the market will reward or penalize most in a rising rate environment are operating margins and free cash flow.

Over the last two weeks, we've entered two bargain priced stocks on our premium site where the market may have gone too far, too fast — particularly those with an improving bottom line. Become a premium member to unlock real-time trade notifications on every entry and exit. Over the last two weeks, we've entered two bargain priced stocks on our premium site where the market may have gone too far, too fast — particularly those with an improving bottom line. Become a premium member to unlock real-time trade notifications on every entry and exit. premium site where the market may have gone too far, too fast — particularly those with an improving bottom line. Become a premium member to unlock real-time trade notifications on every entry and exit.

Ad-tech opportunity

In the interview, Jeremy Owens reminds me that I was the first person to warn him about how the Apple’s IDFA changes that would negatively impact Facebook’s revenue many years ago. It was a bold call at the time because I called the top for Facebook when it was a stock market darling in 2018. Despite the odds, it turned out to be accurate.

We discuss how ad-tech stocks are trading at historically low valuations with many 50% lower than where they have traded during times of economic uncertainty. The share prices of these ad-tech companies can grow over 100%. When the market senses a bottom is in — which I believe was either Q2 or will be Q3 — buyers will step back in to support higher valuations.

We discuss why CTV ads is the most investable trend in media right now.

What to look in the upcoming earnings season

Microsoft’s results are to be closely watched since the company is a bellwether for Cloud. Its suite of Cloud products drives down costs and it’s the most insulated cloud company. It benefits from cloud migrations and also the need for organizations to reduce costs.

Analysts in the earnings call are concerned that the enterprise sector is the next shoe to drop following consumers. The consumer cycle is very short, whereas for Enterprises, it depends on the renewal cycle and there is a period of negotiation. In addition to constrained enterprise budgets, many startups are not able to raise funding and are going out of business, which can weigh on cloud, as collectively startups are a sizable customer for cloud companies.  

The cybersecurity sector has reported exceptional fundamentals given the economic headwinds. Many companies have been reporting high growth rates and are cash flow positive. This sector also has no exposure to discretionary spending, which will help the category sustain long-term.

Bargain Cloud Stocks

We spoke about Best-of-Breed on this podcast yet we are currently building positions in companies that are undervalued and more of a “basement bargain” or “fire sale” valuation as we believe the market has not been entirely efficient with key stocks that have been penalized with low valuations. These stocks are 50% lower than their Covid low and have the potential to bounce back. In fact, one could argue there is more room for gains in these stocks than the best-of-breed companies which are within 30% of historic valuations for cloud stocks.

Sign up for premium to learn more.

Listen on Apple Podcasts | Listen on SpotifySign up for premium to learn more.

Listen on Apple Podcasts | Listen on Spotify

Timestamps:

00:00 Intro
00:44 Valuations
04:40 Consumption-based pricing
11:24 Snowflake vs MongoDB
13:15 Ad-tech
20:15 Upcoming earnings season
22:08 Cybersecurity
24:22 Best practices for retail investors

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Marvell Q2 2023 Earnings & CXL Memory Catalyst

Posted on by io-fund

Marvell’s management team did an excellent job of acquiring Inphi and executing. Typically, we avoid M&A for a year to allow the financials to merge, yet in this case, leaning into the acquisition was a good choice.

The Marvell management team’s execution skills are needed once again because Marvell has an opportunity to greatly increase its revenue and profits if management can execute in a new market one more time. The opportunity is a new architecture called CXL that disaggregates memory from the CPU. CXL is attracting a lot of attention at industry events, such as Hot Chips 2022, because it’s focused on optimizing one of the most expensive parts of the data center – which is memory.

Before we go into the 2023-2024 Marvell product road map, and why it’s key to the company’s future, I want to discuss the fiscal Q2 2023 earnings.

Fiscal Q2 2023 Earnings Overview

The market is concerned over Marvell’s data center guidance of 20% growth next quarter. This is a slowdown from the most recent quarter at 48% YoY growth and earlier quarters at >100% growth.

Chart: Data Center YoY Growth

At an estimated $600 million, it will also mean a sequential decline both from Q2 and Q1, which were at $643M and $640M, respectively. Marvell stated it’s the on-premise business weighing on their cloud data center business and supply issues (more below).

Notably, Q2 of last year was an important moment for the company when 56% sequential data center growth grew from $277 million to $434 million in the span of three months following the close of the Inphi acquisition in April 2021. From there, the company has sustained Inphi’s already high growth levels for over a year.

The company is now at an annualized run rate of $6 billion, which the CEO reminded analysts, was the target for October of 2023. The company met the target originally provided at the October 2021 Investor Day one year earlier than expected. Notably, this was six months after Inphi was closed so M&A not a factor here.

Marvell’s Segment Overview:

  • The data center represents 42% of revenue at $643 million and grew 48% year-over-year.
  • The carrier infrastructure segment, which is wired and wireless and reflects 5G growth, reported 45% YoY to $285 million.
  • Enterprise networking grew handily at 53% YoY to $340 million and is expected to grow at 70% next quarter. We break this segment down below.
  • Consumer was down (1%) to $164 million and is expected to be down (10%) next quarter. Marvell has exposure to the storage market and this can weigh on the more robust segments.
  • Automotive was up 46% YoY to $84 million and is expected to be up 40% YoY next quarter. We also break down this segment below.

Marvell Financial Overview

Marvell was reporting negative top line revenue when we first covered it in 2019 and Marvell took another hit on revenue during Covid before accelerating to the 50%-74% revenue growth range.

The current quarter’s top line revenue in Q2 was at 41% which is a deceleration from Q1 with 74% revenue growth. The company guided for 29% year-over-year growth, which was a slight miss as analysts were expecting 30.3% growth in the fiscal Q3 quarter. The company reported EPS in line with adjusted EPS of $0.57. The guidance on EPS was a slight miss, however, at $0.59 reported versus $0.61 adjusted EPS estimated.

Semiconductors make a tougher investment as analysts can’t model too far into the future beyond what management teams provide. That is why there were many questions looking for help with how to factor in the “acceleration” in the data center the Marvell team is expecting in Q4 and what this will mean for CY2023.

An analyst asked if they can assume 10% QoQ in the data center for $1.7 billion overall revenue and the CEO said it sounded “a little on the high side.” This has led to analysts modeling $1.65 billion in revenue in Q4, for 22.5% growth. Therefore, despite a single-digit acceleration in the data center segment, there will still be a top line deceleration, if today’s forecast does not change.

The company’s margins and cash flow are a bright spot, and I believe this is being overlooked. If we get an acceleration in the data center into next year, then Marvell is fundamentally a much stronger company than it was during the previous data center streak.

On a GAAP basis, the gross margin was at 51% in the most recent quarter, up from 35% in the year ago quarter and up from 46% in FY2022. The company is guiding for the same GM of 51% next quarter.

The GAAP operating margin has improved quite a bit YoY to 8.3% in the current quarter compared to (25%) in the year ago quarter. This is also an improvement from Q1 with GAAP OM of 4.80%. The adjusted operating margin “hit a record” at 36.5% and is guided for 37% next quarter. Stock based compensation was at $139 million in the most recent quarter.

Cash flow is also improving with operating cash flow at $332 million, or 22% of revenue. This compares to $194 million last quarter and $819 million in FY2022. However, the company carries debt of $4.6 billion and has $617 million of cash on the balance sheet. This is a 1.8X net debt to EBITDA ratio.

Therefore, there has been substantial improvement yet Marvell does have a weaker debt profile than a company like AMD or Nvidia.

Chart: MRVL, AMD, NVDA Financial Debt to EBITDA (TTM)

Source: YCharts

Note on Supply:

Marvell is aligned with AMD in that they believe supply chain issues will ease in Q4 and into 2023. Here is what Marvell said in the opening remarks:

“Therefore, for our overall data center end market, we project revenue in the third quarter to decline sequentially in the mid-single digits on a percentage basis. However, we expect our data center revenue in the fourth quarter to increase on a sequential basis, anticipating an improvement in supply and new product ramps in cloud.”

Here is what AMD said:

“The visibility with our customers, especially our large cloud customers’ second half of this year into next year is very good. And we’re planning really for the next four to six quarters, and that gives us good visibility” and later provided many references toward supply coming online in Q4, such as: “But overall, the 7% increase [in gross margin], I think, is very well supported given all of the new product ramps that we have going on in addition to some additional supply that’s coming in as we get into the fourth quarter.”

It never hurts to have two management teams agree on the larger broad-based issue. However, since those reports, we’ve seen analysts cast doubts on the effects of macro for the rest of the year: “[Mizuho analyst Rakesh] checks show hyperscale orders are seeing "pushbacks" but no cancels, with Q3 trending flat quarter-over-quarter and Q4 "potentially soft." Rakesh lowered estimates for AMD "with macro headwinds clouding the near-term outlook."

Marvell’s Products:

In six brief years, Marvell has pivoted away from consumer (storage) products as the revenue mix was previously 62% consumer/38% infrastructure to being 11% consumer/89% infrastructure today.

This was driven partly by hyperscalers building data center infrastructure and AI/ML driving the need for faster data speed. Inphi also contributed to this.

Data Center Segment

PAM Solutions:

Marvell offers 200-gig and 400-gig PAM-based electro-optics — and the company recently added 800-gig solutions. This market sees tailwinds from the need for more bandwidth as the electro-optics connect short distances and long distances to increase data rates. PAM4 has replaced NRZ data transmission with the benefit of doubling the bit rate.

Hyperscalers are going through an upgrade cycle that requires high bandwidth and port density. PAM4 connects networking ASICs and machines, like servers and AI machines. Digital-based PAM4 uses analog-to-digital converters to clean up the signal in the digital domain before converting it back to analog to transmit.

Artificial intelligence and machine learning drives demand for the 800-gig PAM to increase the speed of input-output and to process the data flows. This doubles the throughput (bandwidth) due to an 8x100Gpbs optical transceiver for inside and between AI clusters.

In the fiscal Q1 results ending in April, management had stated: “our first quarter results benefited from a ramp in volume shipments of our 800-gig PAM solutions at two large customers.” The company has also stated that their products will see increase demand with the release of more powerful CPUs.

COLORZ 400:

COLORZ allows regional data centers to be linked together in the same metro region to function as one single mega data center. COLORZ silicon photonics technology allows data centers in the same metropolitan region to function like a mega data center through a “network fabric.” This facilitates faster edge computing within an 80/120 km distance for 30-megawatt data centers as they will be linked together and function like a 120-megawatt data center.

Per the most recent press release:

“As artificial intelligence (AI), machine learning (ML) and high-performance computing (HPC) applications continue to drive greater bandwidth requirements, cloud-optimized 400G solutions are needed to support high-speed data center interconnections. These requirements can only be met through high bandwidth connectivity offered in a small, cost-effective form factor.  The Marvell COLORZ II 400ZR enables cloud data centers the ability to increase the speed of data movement while keeping the power and cost low.”

Another press release stated the company shipped 100,000 units.

Here is what was said on the call about how/why the growth in the data center can continue:

Harlan Sur

Good afternoon. Thanks for taking my question. On the cloud optical connectivity business, this is both inside and between data centers, the upgrade cycles have been this really great multi-year tailwind for the team.

And if I look into next year, I believe that there's still at least one of the top four US hyperscale titans that's going to start the 400-gig PAM4 transition. You still have China CSPs that need to fire. You've got multiple customers on DR that's going to fire as well. Historically, like these transitions, I don't think have been impacted by a slowing macro demand environment. They're viewed as, I think, very strategic.

But is that how your cloud customers are thinking about these upgrades and your views on continued upgrade momentum in this segment for next year? And just relatedly, is the Innovium team on track to drive $150 million in revenues this year?

Matt Murphy

Hey. Thanks, Harlan. Yes, I think the first part of it, you got pretty well in terms of the transition on 200 and 400 gig PAM4 inside the data center. And then, the new ramps we're seeing in 400 gig ZR for DCI between data centers.

What I'd add on top of that is — which has been extremely strong and also, in some ways, a little bit of a constraint we've seen in terms of being able to keep up is, the demand on 800 gig, which is happening right now really around, obviously, very advanced AI workloads.

That is an area where, if we could obviously produce more material, we would be shipping it in Q3. So that's also a positive trend. So you've got sort of the transition going on all the way up to 800 gig, and that continues to look pretty good.

NOTE: Innovium is an acquisition that closed in 2021 and at time of acquisition was expected to add $150 million in revenue for CY2022/FY2023.

Compute Xpress Link (CXP): 2024-2025 Data Center Catalyst

Marvell is launching a new product line called CXL, which will improve how data centers add memory. Right now, a server must be opened to add DRAM and the DIMM slots are limited in number and don’t pass service history or bit-error history, which is needed by hyperscalers.

Memory pooling allows memory to scale independently from processors by taking memory for a task and then releasing the memory. The new fabric removes the need for local DRAM, which adds a bit of latency from 100ns to 140-160ns, however, there’s a possibility of adding a CXL accelerator to be more “cache coherent.”

The CXL switch will be used to accelerate protocol-level processing across ethernet, DPUs, SmartNICs and solid-state drive controllers (SSD).

What Marvell is proposing with CXL is a new server architecture to “dynamically assign memory resources between servers.” The result is boosted memory bandwidth and also the ability to enable memory pooling. The company sees a future where a new architecture will separate compute, memory and I/O racks with the interconnect being CXL. Partially-disaggregated racks are expected to deploy in 2024-2025.

Marvell is at the forefront of the shift toward “disaggregate memory from the CPU” because it currently supplies the optics that this new fabric will disrupt. Inphi is the leader in silicon optics, PAM-4, and the encoding of PAM-4 for PCIe 6.0.

2024 seems like a long ways off yet the market will be paying attention to this In Q2/Q3 2023.

Here’s an excerpt from the call:

“As you recall from our discussion last quarter, we see CXL as the next big evolution in cloud data centers that will enable us to increase our reach into the memory ecosystem and presents a multibillion-dollar SAM expansion opportunity for Marvell.

This includes a host of new products such as CXL expanders, cooling devices, switches and accelerators and the potential to embed CXL IP and a broad range of our data center products. Events and presentations at FMS strongly validated our excitement around CXL. This is the hottest topic at FMS with standing room-only presentations by many leading industry participants.

The Marvell booth, we demonstrated the industry's first CXL memory pooling solution, addressing the challenges related to memory scaling and cloud data centers. While the industry is still in the early stages of CXL adoption, we are working on closing significant opportunities right in front of us at key customers and envision a strong design win pipeline.”

Why Marvell for CXL?

There are a handful of companies going after the CXL opportunity. Marvell could be front runner as the company already works closely with memory OEMs by supplying HDD controllers, SSD controller and preamplifiers. The company also has an aggressive PCIe roadmap with the company shipping Gen 5 sockets whereas most SSD device are shipping Gen 4 solutions. Marvell is already investing in Gen 6, which in turn, attracts more Tier 1 memory OEMs.

Marvell acquired Tanzanite, a developer of advanced CXL technologies. The company plans to expand to CXL expanders, cooling devices, switches and accelerators.

The company has stated this will drive “a multibillion-dollar PAM expansion opportunity driven by CXL overtime.” (Note: Marvell is referring to PAM, their premiere product)

We will focus on this more next year. You can listen to a recent tech talk here on CXL. The presentation is located here. This is an article about Microsoft’s interest in CXL with a statement that “50% of their server costs are taken up by DRAM.”

Carrier Infrastructure:

The OCTEON processors and platform is an Arm-based compute architecture for embedded applications, such as wireless networking equipment including 5G, including switches, routers, firewalls and monitoring solutions.

The OCTEON DPU is used with SmartNICs and security accelerators with a 5nm design that delivers to the infrastructure industry the same processing node as consumer smart phones and high performance computing and shipped in 2021. The most recent release from last year was the OCTEON 10 DPU and Prestera carrier switches which combined consumes 50% less power than competitors (according to Marvell).

Marvell’s processors help 5G networks meet latency and bandwidth demand while also allowing the networks to upgrade as cellular standards evolve. Marvell also offers customized solutions, which is ideal for Tier 1 customers who can combine their IP with Marvell’s Arm v8 processors and accelerators.

Recently, Dell and Marvell partnered to develop a server-class accelerator card for 5G base stations based on Marvell’s arm-based OCTEON Fusion processor. The hardware accelerators deliver more processing power including processing solutions for smart radio heads to support massive MIMO antenna rays.

We wrote about MIMO a few years back in a reference guide: “Massive Multiple Input and Multiple Output (MIMO) sends the data through multiple data streams called layers, which increases parallelism and throughput. MIMO helps avoid lost signals with multipathing, which allows the base station to send multiple copies of the same signal for increased redundancy. 

Note: The antenna array is one fundamental change to 5G infrastructure. The initial 5G rollout will use existing cell towers, however, newer, dedicated 5G network infrastructures will require many more antennas than used in previous generations. Read more.”

The distributed unit (DU) shares the load with the radio unit by running L1 functions on the RAN protocol. Marvell has been a proponent of OpenRAN with the O-RAN platform, which is an open protocol and open platform that allows Marvell’s hardware to be used with various software vendors. Facebook (Meta) is a partner with Facebook Connectivity.

DPU processors, or digital processing units, are gaining traction for 5G transport, 5G RAN intelligent controllers, edge computing and cloud data center workloads. These hardware accelerators enable high speed connectivity and can improve packet processing rates by 5X. DPUs are ideal for power sensitive edge applications. Marvell’s strength in DPUs is one reason it may be able to stave off competition, which in the narrow field of 5G base stations includes Qualcomm/HPE and Analog Devices. Beyond 5G, Marvell has other competitors for DPUs such as AMD/Pensando and Nvidia.

Regarding 5G, over 7 million of the Octeon processors have been used in 3G, 4G and 5G base stations with Tier 1 customers. In the past, we reported that Samsung and Nokia use Marvell, and supplying these particular companies was a tailwind when Huawei was blacklisted. More recently, Marvell has stated they have design wins with four of the top five global OEMs and next-tier OEMs building base station equipment. These design wins are based on the 5nm platform.

Marvell uses TSMC for the 5nm OCTEON DPUs and this is an advantage because Marvell has the 5nm now and is able to move quickly on a 3nm release.

Notably, 5G has been a long time coming but I do believe it will reward investors over the next few years. Technavio has a CAGR of 67% for 5G equipment through 2025. The growth trend of 5G/edge computing is not one that we plan to complacent on as it will provide the next leg up for substantial capex spending similar to data center capex spending.

Enterprise Networking:

Marvell sells ethernet switches and ethernet PHYs to IT managers and networking equipment manufacturers. The company uses DSP technology for CAT5e ethernet cables to supply data rates up to 5Gbps with support for CAT6 and CAT6a.

Management discussed on the call that the main driver for this market right now is wireless, specifically WiFi 6 as the wireless rate line is now faster than the wired rate. The call also pointed toward content per port going up in the transition to multi-gig. According to the CEO, “it's not like 10%, 20%, 30%. It's sort of multiples on a per port basis of where it was before.”

Increased enterprise share and content gains from wired and wireless enterprise networking drove 53% YoY revenue growth and 19% QoQ revenue growth.

Automotive:

Similar to the networking that Marvell supplies enterprises and the data center, Marvell also supplies auto manufacturers with ethernet PHY transceivers, camera bridges and switches for in-vehicle networks. This is used for things like collision detection, lane warnings, and autonomous driving.

Marvell believes Ethernet will be the backbone for connected and autonomous vehicles to connect the electronic control unit (ECUs), cameras, sensors, and central compute devices. The Ethernet device is called Brightlane.

ON Semi has partnered with Marvell on use cases such as pairing a standardized protocol, such Ethernet PHY, with ON’s portfolio of ultra-dynamic range image sensors.

Automotive was up 46% to $84 million, yet was down 6% sequentially. Management cited supply issues rather than demand. Marvell counts eight of the largest 10 OEMs worldwide and 36 OEMs total. The company believes revenue growth will be 40% next quarter.

Note on Consumer Market:

Marvell sells hard disc drives (HDD) and solid state disc (SSD) controllers. This is a weaker segment, declining 1% YoY and 8% sequentially to $164 million. For next quarter, Marvell expects revenue to be down 10% YoY and flat sequentially.

Conclusion:

There is a new, powerful trend on the way that is on par with the cloud computing trend. This trend of edge computing will rely on distributed computing rather than centralized processing. Both will exist and rely upon each other but edge computing will have a stronger growth trend when it breaks ground (by virtue of being new/rapidly expanding). Much of this will be in sync with the 5G buildout.

Marvell has the potential to be a strong stock during this buildout as the company provides the base station hardware, supports MIMO antenna rays, beamforming, and accelerates 5G transport and controllers which results in high-speed connectivity.

The company also provides electro-optics and silicon photonics for increased data rates and a network fabric for edge computing. The edge is defined as many things, but what all definitions can agree on, is that the edge needs superior connectivity/networking. Electro-optics, silicon photonics, DPUs, SmartNICs and ethernet in the data center are a warmup for Marvell supplying edge servers and edge devices. As this occurs, the demand for Marvell’s product suite will increase.

In addition to this, Marvell is thinking outside the box by focusing on restructuring memory while most companies are focused on more powerful chips. CXL drives down costs on DRAM and is likely to rapidly adopted by hyperscalers once it becomes available. There’s no guarantee that Marvell will be the one to win the contracts but it’s certainly a front runner.

Quick Update on Snowflake and Nvidia

Posted on by io-fund

Snowflake exceeded top line expectations, as outlined on our forum here. Management focuses on product revenue versus total revenue and GAAP metrics are a bit buried under the more accessible adjusted non-GAAP metrics. However, Snowflake delivered what the market needed to see – which was 83% product revenue growth compared to 72% growth expected. The guide for Q3 was in line with expectations while FY2023 was slightly raised from $1.893B at the midpoint for 66% growth to $1.910B at the midpoint for 67.5% growth.

The top key metric to note was a re-acceleration in customers with TTM product revenue above $1 million, which was at 112% this quarter at 246 customers, up from 98% growth last quarter. This is an important forward-looking metric as it takes 9 months to fully onboard new customers. Another reason this key metric holds more weight is the consumption model means the upside is uncapped, whereas with SaaS, the monthly amount has a ceiling (usually). You can read more about the consumption model here.

Net revenue retention rate is down 300 basis points sequentially but it up 200 basis points year-over-year and this re-acceleration is what’s important. RPO growth of 78% is the lowest in reporting history, down from 82% last quarter and 122% in the year ago quarter.

Analysts did note on the call that there are tougher comps for RPO coming down the line in Q4 (quarter ending in Jan 2023). It was in this quarter that RPO moved from $1.8 billion to $2.6 billion. Right now, it stands at $2.7 billion, so not too much H1 growth over the past six months. This is simply something to note. Of this RPO, 57% is recognized over the next 12 months, or $1.5 billion.

Here is what was discussed:

Brad RebackBrad Reback

Hi, thanks very much. Mike, I know you mentioned the 3Q consumption comp. You also have a really, really difficult 4Q RPO comp. But given your commentary, should we expect a healthy end to the year given that renewal pool? Thanks.

Mike ScarpelliMike Scarpelli

Yes. We expect we will have a big increase in RPO. But I’m not guiding to it. You’ll have to wait and see. I’m never going to guide RPO.

Total customer growth was at 36% growth compared to 40% growth last quarter. As noted above, it’s the TTM > $1M that matters.

There is no denying that on a GAAP basis, Snowflake is largely unprofitable. The company’s GAAP operating margin was at (42%) compared to the adjusted operating margin of 2%. The operating losses of $207 million this quarter increased from $189 million in GAAP operating losses last quarter. Stock based compensation increased from $164 million in the year ago quarter to $209 million in Q2 2022.

Free cash flow fluctuates with $54 million in free cash flow this quarter. This is up from ($12) million in free cash flow last year. The free cash flow margin is 11% and the company raised its adjusted free cash flow guide for the year from 15% of revenue to 17% of revenue. The company has $5 billion on the balance sheet.

Moving Forward …

The top catalyst for Snowflake (in my opinion, and was discussed on the call) is Snowpark going into production with Python. It’s not open to general availability yet.

Here is what was discussed on the call:

Frank SlootmanFrank Slootman

Yes, I will start, and maybe Christian can finish. Python is – so Snowflake for Python is red hot, and people are jumping that for us to declare it GA, which is something and we have customers that are really wanting us to let them use it in production now some of the largest customers that we have. So, pressure is on because the demand is there. The thing about the Iceberg Open Table formats that really completely open Snowflake up to be – for Snowflake cables to be used by anybody and everybody that can support that format. We are seeing incredible results in terms of performance of like executing against that file format. So, these are all very, very, very promising developments for us. And I think that the pressure is on for us to declare these things generally available because people are trying to rip them out of our hands right now.

Mike ScarpelliMike Scarpelli

Yes. As we said at our Summit conference, we expect those to be GA at the end of this year. So, a meaningful contribution to consumption will happen next year.we expect those to be GA at the end of this year. So, a meaningful contribution to consumption will happen next year.

Here is what we’ve said in the past:

“Snowpark offers the ability to migrate business logic with popular programming languages Python, Scala/Java Virtual Machine or Java. The library and DataFrame API allow querying and processing data without having to move data to where the application code runs. This extends programming functionality for ML model training and allows data processing to run natively in the data cloud. 

Prior to Snowpark, code deployment required separate infrastructure. Building applications that interact with Snowflake’s virtual warehouses minimizes processing time and lowers the learning curve/broadens adoption of complex data pipelines by removing the need to move or copy data into other systems to overcome working with SQL.

The recent announcement of adding Snowpark for Python is key because of Python’s widespread popularity among developers. With the Snowpark Accelerator, Snowflake is courting developers to build more applications and this is likely to help Snowflake maintain a competitive advantage with a newer class of machine learning startups.”

Nvidia …

Nvidia remains one of our highest convictions and we’ve laid out those reasons in great detail. We will provide an earnings overview soon but you’ve likely already heard through the pre-announcement that gaming was down 44% sequentially. The company’s guidance also missed by $1 billion with $5.9 billion guided versus $6.9 billion expected.

Nvidia is undeniably the highest priced semiconductor, as well, and the one issue investors face when a company misses on EPS (noted in the pre-announcement) is the valuation gets richer overnight. Nvidia has been trading in the 30-35 forward P/E ratio range, yet is now in the 46 forward PE ratio range.

It’s likely we trim here a bit here and re-allocate (perhaps to Snowflake tomorrow). The position is large so the trim is not for lack of conviction, we can promise you that. We also won’t hesitate to buy back again.

Chart: Nvidia Forward PS and PE Ratio

 

Datadog Q2 2022: Lower Commitments YoY; Same Revenue Guide

Posted on by io-fund

Datadog is a heavy hitter on revenue growth and offers a rare, balanced bottom line. We believe this is because Datadog rides the coattails of digital migrations to AWS, Azure and Google Cloud. As cloud migrations continue – evidenced by growth across The Big 3 – the number of applications and containers to monitor has grown and the complexity has also grown within cloud-native environments. We previously covered this here.

The company is recognized as a Leader in observability and sits comfortably above competing platforms offered by the Big 3 in Gartner’s Leader Quadrant. This is important because Datadog serves the dominant trend of multi-cloud by offering flexibility for companies that prefer to work with more than one cloud vendor (Google, Azure or AWS) while not compromising on quality of observability and security products.

There’s a lot to unpack in the report, but ultimately, we feel it was a strong report and that as more cloud earnings come down the line, Datadog will ultimately stand out among its peers.

Q2 Overview of Financials

Datadog beat on Q2 revenue reporting $406 million for 74% growth compared to a consensus of $381 million, or 63% growth expected. The guide for Q3 came in as expected at $410 to $414 million compared to $410.7M estimated.

The company reported GAAP EPS of ($0.02) with the market expecting ($0.07) EPS. Adjusted EPS also beat at $0.24 reported compared to $0.15 EPS expected.

The GAAP operating margin was at (1%) compared to 2.87% last quarter. Notably, this was a slight improvement YoY with GAAP op margin at (4%) in the year ago quarter although lower sequentially from +3%. As we said with Microsoft and Google, margins that are flat/unchanged are a win right now. This resulted in ($3) million in operating losses compared to a ($10) million operating losses in the year ago quarter.

The company’s operating cash flow of $73 million was down sequentially from $147 million yet was up YoY from $52 million. The free cash flow in Q2 was $60 million compared to $130 million last quarter and $43 million in the year-ago quarter. The company has $1.7 billion in cash on its balance sheet, which is unchanged from last quarter and up $300 million from the year ago quarter.

Stock based compensation more than doubled to $82 million in the recent quarter, up from $34 million in the year ago quarter. SBC this year across two quarters is at $149 million.

Growth of larger customers above $100K ARR grew 54% which is down but not too meaningfully given macro. Last quarter growth was at 60% YoY, and in the year ago quarter, ARR > $100K was up 60%. DBNRR was over 130% for twenty consecutive quarters.

As discussed below, the strength to Datadog is the land and expand, or the upgrades. The company continues to do well here with 79% of customers were using two or more products, up from 75% a year ago. 37% of customers using four or more products, up from 28% a year ago and 14% of our customers were using six or more products, up from 6% a year ago.

The company stated they are aggressively hiring, which makes Datadog an outlier in that regard, and is often a forward-looking indicator.

Datadog guided in line, but the market was expecting the company to raise guidance considering the last two quarters were sizable beats. Q3 guide was for $410 to $414 million with consensus at $410.68 million. The company reiterated the $1.62 billion guide for FY2022, or 58% growth. Adjusted EPS guide for Q3 also came in as expected while full year EPS guide was slightly lower than expected, at the midpoint, with adjusted EPS of $0.74 to $0.81 compared to $0.80 adjusted EPS expected.

Why the Market Got a Little Nervous in the Pre-Market

Although Datadog was able to provide revenue guidance consistent with previous guidance, analysts wanted more visibility into full year guidance as Q4 sits lower than usual if we assume $1.62 billion this year.

Total deferred revenue was $467 million in Q1 with current deferred revenue of $455M and non-current deferred revenue of $12.8 million. In the current quarter this softened to $458.5M in total deferred revenue with $444 million current and $14.5 million non-current. While non-current fluctuates, this was the first decline in current deferred revenue over the past two years. This information was available pre-market.

On the earnings call, Datadog reported Billings and RPO growth that was lower than overall revenue, whereas in the past, these two key metrics typically exceeded revenue: “And pro forma for those adjustments, billings growth year-over-year was in the mid-50s. Remaining performance obligations, or RPO, was $881 million, up 51% year-over-year. Current RPO growth was in the mid-50s year-over-year, and contract duration was slightly lower than the year ago quarter.”

This is down from 85% growth in RPO last quarter and billings growth of 103% last quarter. In the year ago quarter of Q2 2021, RPO was up 103% and billings was up 69%.

The company stated this was due to these three reasons:

1. Due to the land and expand model where customers come in the door and then upgrade:

“As we said in previous quarters, billings and RPO growth can fluctuate significantly and vary from revenue growth, whether higher or lower due to the timing of invoicing and duration of customer contracts. To illustrate this, we note that billings growth for the first half of the year of 2022 was 72% year-over-year.” billings growth for the first half of the year of 2022 was 72% year-over-year.”

2. They believe that even though customers are less likely to commit due to macro, they will still spend in-line and their guidance reflects their expectations: “In addition, we observed that some customers aren't changing their level of usage growth but are being more conservative in their commitments, which impacts billings and RPO growth but not revenue growth.”customers aren't changing their level of usage growth but are being more conservative in their commitments, which impacts billings and RPO growth but not revenue growth.”

3. There was an accounting pull forward to where some customers were billed in Q1 this year yet were billed in Q2 last year: “As in previous quarters, we had some differences in the timing of billings of a few large customers, which were billed in Q2 last year but were billed in Q1 this year.”differences in the timing of billings of a few large customers, which were billed in Q2 last year but were billed in Q1 this year.”

Here is one question on the call about RPO and Billings:

“But if you can help us reconcile your RPO growth. I'm sure there are company specific things that pertain to how you see of your growth on a year-over-year basis, sequential growth basis. How do we look at that in the context of what's happening with the hyperscalers. And they did put up even during an uncertain time, tremendous backlog growth, whatnot. So is there something specific to Datadog? Maybe it's the 18% or so, high teens percentage exposure to consumer discretionary.”

The CEO defended Datadog by saying their growth has exceeded the hyperscalers: “In Q2, we did a lot better than the hyperscalers. So we're growing a lot faster than all of them combined. And they've decelerated actually more than we've done in relative basis. So we actually feel good about the ratio there.”So we're growing a lot faster than all of them combined. And they've decelerated actually more than we've done in relative basis. So we actually feel good about the ratio there.”

The CFO had the following color to add: “So remind everybody that with our land and expand where we start getting used by clients, they scale up the growth and when they get to a certain point through, this has been going on for the whole business model. They go to an increased commit. Because of that, there's variability in the billings and RPO that net-net, over time, on average, go towards the ARR growth. Again, remember, we mentioned that the ARR growth is the best metric. And the way to look at that is that you look at the revenues […]

And I think we said we basically put in there that in the first half of the year, the growth of this was in the 70s, pretty close to revenues. the growth of this was in the 70s, pretty close to revenues. Why? Because there was timing of billing in the first quarter relative to the second quarter Because there was timing of billing in the first quarter relative to the second quarter that moved the first quarter up and the second quarter down, but it really doesn't have much effect on the drivers of our business.”

The other question on analyst’s minds is basically this: Is Consumer Discretionary isolated as the only vertical that is reducing budget and cloud usage oror is Consumer Discretionary simply the first in a row of dominoes with other verticals soon to follow?

Datadog’s answer to this (for their business) is that they are seeing decreased usage in larger customers and not from SMBs. The analysts were surprised by this due to the ongoing reports SMBs are reducing their spend. There was more than one question but here’s an example:

“Good morning. David, I think everyone is still a little confused. You're seeing an inversion with what's happening with SMB and large enterprise. Many companies are kind of weakness in SMB not at large. Can you explain why you think you're seeing this inversion? And are you embedding a more conservative view in the back half?”

The most direct answer was this: “[SMBs] have the same exposure but simply the difference is, what's your time to say, $5,000. Probably not. If you're a much larger customer, it's worth your time to sell $500,000. And that's what we see with those optimizations.”

Datadog didn’t give much improvement in terms of full year guidance.

Here’s what I have right now:

Q1: $363M or 83% growth
Q2: $406M or 74% growth
Q3: $412M at midpoint or 53% growth
Q4: $439M if we factor in FY2022 $1,620 guide or 34.6% growth. Note: Analysts are modeling 45% growth for $471M for Q4.

Management is keeping it safe by remaining with the full year guidance they previously with the word “conservative” mentioned a lot. This is what was stated on the call:

“I wanted to talk a little bit about some of the trends you're seeing in the business and particularly with respect to the guide. I guess the first question is, as the quarter progressed, when did you start to see some of these slower usage trends in some of these verticals? If you could give a comment on that? [..] And then, David, in terms of the guidance in terms of how you were framing it, can you give us a sense of what you're sort of assuming in the back half with respect to Q3 and Q4? Is it some of the trends that you're seeing in July, did that improve or stabilize or worsen?”as the quarter progressed, when did you start to see some of these slower usage trends in some of these verticals? If you could give a comment on that? [..] And then, David, in terms of the guidance in terms of how you were framing it, can you give us a sense of what you're sort of assuming in the back half with respect to Q3 and Q4? Is it some of the trends that you're seeing in July, did that improve or stabilize or worsen?”

The CEO stated, “We saw that start really in late April, May and June. So as we got deeper into the quarter. I should say that this is – if you're thinking of what happened in terms COVID, this is not a sharp pullback as we have seen at that time. But we saw it's just, for some customers still growth, but slower growth for certain types of customers and others than what we would have seen historically”

The CEO later clarified the following on the APM product: “That's a great question. So for APM, there's actually part of APM that looks like logs […] And that's the part on which we've seen some slower growth. It's still growing, but both are actually still growing healthily, but I would say, slower than they were in recent quarters for these types of customers.”

The CFO stated, “On guidance, as you know, we have always been conservative in our guidance by using lower organic growth and other metrics than we've seen historically and continue to maintain that philosophy. I would note that if you look at the raise here and the percentage of the beat that was passed through into the raise from Q2, it is lower, more conservative than we have done in previous quarters. And the reason for that is the macro uncertainty where we can't be as confident about what happens given the macro uncertainty.we have always been conservative in our guidance by using lower organic growth and other metrics than we've seen historically and continue to maintain that philosophy. I would note that if you look at the raise here and the percentage of the beat that was passed through into the raise from Q2, it is lower, more conservative than we have done in previous quarters. And the reason for that is the macro uncertainty where we can't be as confident about what happens given the macro uncertainty.

So I would say there, if you want to take that, there were some incremental conservatism put into this. But I'd remind everybody that we've always been quite conservative in using assumptions that are lower than the past when we give guidance.”there were some incremental conservatism put into this. But I'd remind everybody that we've always been quite conservative in using assumptions that are lower than the past when we give guidance.”

Here is management’s track record in that regard. Notably, this track record was accomplished even given the uncertainty of Covid.

 So, for next quarter, we know we need to make sure that Q4 guide comes up to match analyst expectations.

The Gartner Magic Quadrant for APM and Observability came out recently in June of 2022. Datadog’s products are comfortably ahead of what The Big 3 offer.

Conclusion:

I'm guessing by the time we get all cloud Q2 reports, DDOG will be a leader in cloud for both top line and bottom line. The market clearly wants perfection for the higher valuations in cloud and I think DDOG gave us what will be seen “as close to perfection as possible" as we go along. The primary remaining blemish was the company did not raise full year guidance to imply healthy/consistent growth in Q4.

DDOG's report was a bit murky for less defensible cloud companies as they pointed to a slowdown in usage in May and June. DDOG is more defensible due to a few things — it sits at a critical piece in the stack (last to be cut even if there is softness in commitments, revenue growth can still occur as planned), is well diversified across many verticals and is the best direct correlation to cloud IaaS growth (ref. our previous analysis one example is here).

I've felt for some time that if we see AWS, Azure, etc., start to slow then maybe time to look at our (very high) DDOG allocation, otherwise, our goal is to ride on these coattails. DDOG could still have some nice growth once cloud infrastructure slows — it's simply one data point I use when we determine the allocations since it’s so closely correlated.

Knox trimmed 1% from Snowflake because SNOW is certainly more exposed to usage trends as is Confluent. I know Confluent had a nice quarter today, however, it appears their growth is slowing to mid-30% by Q4 and analysts are in agreement with this. Perhaps the company will raise guidance as they go along.

Takeaway:

We already cut ASAN to reduce exposure to discretionary spending. We also cut TWLO due to discretionary cuts we saw from ad-tech and other signs although I suspect we will be back into Twilio by H2 2023. Our goal is to not "call earnings" rather to reduce exposure and add back in when the skies are bit clearer plus we want to allocate to any companies showing strength for 2022. Snowflake could do well but there is exposure to discretionary spending and the company bills on a usage basis – we covered this here.