Earnings Update on Nvidia and Snowflake
We have four more companies reporting in the next week – MRVL today, MDB, S and ASAN next week. Please check back on the forum tomorrow for thoughts on Marvell.
You’ll be getting key deep dives from us on the strongest companies to emerge from the macro-minded quarter. We take a week to thoroughly look through all the reports and to make any necessary changes. We are also scanning semis now for any up and comers. Royston is posting to this effect today. Lastly, we will have more crypto updates coming too. Knox’s trade alerts are provided to help gauge where we are in real-time with an earnings report.
The market is very jittery right now and is requiring perfection during a quarter impacted by a multitude of macro forces.
Our goal is to step aside or lean out of positions if there’s something inherent to the product affecting it for a few quarters but to also lean into positions where the market is not able to efficiently analyze when the transitory headwinds are broad based.
Nvidia:
You can view my pre-earnings report here.
I’ll start with Nvidia since that’s the easiest company for an investor to see the forest through the trees. A $500 million impact from China shutdowns is irrelevant to the long-term story. $400 million is from gaming. Here is how we know this is irrelevant to the long-term story:
“We expect strong sequential growth in Data Center and Automotive to be more than an offset by the sequential decline in Gaming.”
Fundamentally, it has what it takes to be our leading position for some time due to the groundwork it’s laying in AI. We haven’t done a deep dive on Nvidia in some time and in many regards, we are looking at a new and improved company since I last did my deep dives in 2019/2020.
For now, to be brief, the A100 GPU which we first covered two years ago here and again here is showing up in Nvidia’s earnings report in a big way right now. Jensen Huang made it clear on the call that inference is the harder piece over training and this has driven record revenue for Nvidia as the A100 combines both training and inference onto a single chip. In the previous call, he noted that revenue driven by inference use cases had tripled in the quarter.
“[Data center revenue] doubled year-over-year. and we're seeing really strong adoption of A100. A100 is really quite special and unique in the world of accelerators. And this is one of the really, really great innovations as we extended our GPU from graphics to CUDA to Tensor Core GPUs. It's now a universal accelerator.
And so you could use it for data processing for ETL, for example, extract, transform and load. You could use it for database acceleration. Many sequel functions are accelerated on NVIDIA GPUs. We accelerate Rapids, we accelerate which is the Python version a Data Center scale version of Pandas, we accelerate Spark 3.0. And so from database queries to data processing, to extraction, and transform and loading of data before you do training and inference and whatever image processing or other algorithmic processing you need to do can be fully accelerated on A100.”
Personally, I’m excited for when we get through the last of our earnings reports next week and I can focus on writing a full-length analysis on Grace CPUs and Hopper GPUs for our Members. It’s an analysis I’m very much looking forward to and I’ll also lay out why Nvidia will dominate automotive, as well. Additionally, the licensing of software will be another catalyst for Nvidia that has little coverage right now. This four-segment combo: data centers, automotive, software licensing and then professional visualization are not fully appearing in the earnings reports right now. Data centers clearly are, but I foresee a long runway for data centers while these other segments ramp. That’s what I plan to touch base on for our longer-term thesis.
For the near-term, the company is saying that gaming is solid and AMD said something similar.
“The underlying dynamics of the Gaming industry is really solid, net of the situation with COVID lockdown in China and Russia. The rest of the market is fairly robust and we expect the Gaming dynamics to be intact.”
However, Nvidia investors need to be prepared to ride out any turbulence from Ethereum’s merge to proof of stake. The good news (in my opinion) is that the automotive segment should be kicking in around that time.
“Our DRIVE Orin SoC is now in production and kicks off a major product cycle with auto customers ramping in Q2 and beyond.”
The Numbers:
Nvidia reported revenue of $8.29 billion, up 46% and ahead of estimates of $8.1 billion. EPS was $1.36 compared to analyst expectations of $1.29 EPS. Nvidia missed estimates for Q2 with $8.1 billion due to the $500 million impact from Russia and China, with $400 million attributed to gaming and $100 million to data centers.
Data center was very strong, up 83% a year ago and 15% higher sequentially. Gaming was up 31% YoY and 6% sequentially. Professional visualization was up 67% which is a deceleration from the triple digit growth from last year. Automotive was down 10% YoY but was up 10% sequentially. As discussed, we expect this to ramp in Q2 and even more so in Q3-Q4.
Nvidia is repurchasing stock with $2 billion repurchased this quarter with a repurchase program approved by the Board of $15 billion total.
Snowflake:
Snowflake has exposure to consumer due to evidence a handful of its largest customers reduced usage in the quarter. Here is the main comment regarding the current quarter’s revenue miss:
“Last year, we saw certain customers experience much higher than expected consumption — own businesses were growing extremely fast. Today, some customers face a more challenging operating environment. Specific customers consume less than we anticipated amid shifting economic circumstances, we believe are unique to their businesses, most notably consumer facing cloud companies.”
However, Snowflake passed our internal test of needing to have a strong bottom line. Here is what Snowflake’s increase in cash flow looks like compared to other leading cloud companies. Notably, some of this data is from the previous quarter but the trend line is the same.
What is not pictured above is price to free cash flow, which SNOW does not rank well on this metric coming in around 600 compared to CRWD and DDOG around 80. However, if FCF continues to improve so will the bottom-line valuation. The main point of the chart above is to illustrate the progress Snowflake has been making during a time when access to cash is tightening. Snowflake has stock-based compensation weighing on GAAP operating margin yet the FCF margin is hard to deny in terms of being best in class.
Adjusted FCF was 43% – this was positively impacted by the timing of Q1 collections. Normalized over the year, FCF is guided at 16% of revenue. We want to stick with cash efficient companies bc as the CEO of Snowflake pointed out, private companies are becoming cheaper in terms of valuation and a war chest of cash that can be leveraged while private tech valuations are low will help the company come out stronger than its peers.
The company has best in class net retention rate of 174.
In terms of being profitable on an adjusted basis and this profitability increasing in a predictable manner, Snowflake also stands out.
One drawback is how management interacts with the analysts on the call. They can come across impatient and meanwhile SNOW has a fairly hard to understand business for financial analysts to wrap their heads around. Ultimately, management was trying to convey they are not discretionary despite the slight miss while analysts kept poking around trying to get an admission that Snow could be discretionary. Here is the most pointed answer management provided (which may have fallen on deaf ears as questions on macro impacts to core business model were relentless):
“I mean, the reality is that, our type of workloads become very heavily grafted into core business processes. And that, by the way, is also one of the reasons why it's — we've talked about this for the last two years in the calls, how difficult it is to move workloads to Snowflake, because these workloads are so heavily grafted into operational processes.
So these things are not going anywhere. They're not optional. They're not like, what do I feel like doing today? That, by the way, there are workloads like that, that's far more on the data lake side, where essentially you have a massive repository of files. And you may have data scientists that are just sort of fishing falls out of the lake and trying to decide to do some interesting stuff for that. That sort of thing is highly discretionary, but that's not the focus of our business.
[….] But in our world, as I said, it is so embedded into core business processes. It's not something that you can just sort of shut off for a while until things get better.”
And since this happened to be the sticking point — which is analysts grappling with whether Snowflake’s consumer-related miss means the company will see future headwinds related to macro compared to other more enterprise-only cloud companies, I’ll provide another moment on the call where analysts drilled into this point.
Brad Zelnick
Great. Thank you so much guys. So instead of trying to dream up the 16th way of asking you about the macro and the impact it's having, I want to maybe put that aside for a sec, and I mean you're delivering amazing growth at scale, and certainly, that shouldn't be lost. But if we just think about pricing and maybe competitive dynamics, I think you guys have had strong discipline when it comes to pricing, very ROI-focused. But is there any reason to believe pricing is an obstacle for adoption? And maybe any evidence that you have to believe your competitors are seeing the same things you're seeing? And perhaps — I don't know if it's changes the win rates or customers even stratifying their consumption across other alternatives to save money, anything that you can help us to appreciate what's going on along those dimensions would be helpful. Thank you.
Frank Slootman
Yes. This is Frank, Brad. Our business is not commoditized, which is sort of another way of characterizing your question. There's certainly people in our world that are trying to commoditize the business. But customers are trying to do very difficult things, also very amazing things. So what they're paying for credit is already incredibly optimized. It's incredibly competitive. This is physics and economics, right? And there aren't many places to hide in terms of what we charge for.
The Numbers:
Royston did a post-earnings writeup here.
Per our last analysis:
“Snowflake is steadily improving its margins from 58% gross margin a year ago to 65% gross margin in the recent quarter. The company has improved its GAAP operating margin from (90%) a year ago to (40%) in the recent quarter. The company has a positive adjusted operating margin of 5% and has stated they will end the year with a positive 1% adjusted operating margin. They have to deliver on this promise to maintain a category-high valuation.”
In the most recent earnings report, Snowflake reported GM at 65% in the most recent quarter and operating loss of (44.5%), an improvement from the (90%) OM in the year-ago quarter. GAAP EPS was (0.53) this quarter compared to ($0.70) in the year-ago quarter. Shares outstanding increased from 291 million to 314 million.
We know the top line is decelerating and we noted this in the last analysis. We had said the following:
“The company is expected to report revenue of $412 million, representing growth of 80%. The previous quarters the company reported revenue growth of 101% in Q4, 109% in Q3, and 104% in Q2. For the fiscal year 2023 ending in January, the company is expecting revenue growth of 67% for revenue of $2.03 billion. Analyst consensus shows revenue of $3.17 billion, or growth of 56% for fiscal year 2024.”
I have in my notes a slight fiscal year miss yet some analysts have Snowflake coming in as-expected on FY2023 for the $1.9 billion guide. Regardless, I think the FY2024 growth is key to keep it above the 50% mark while reaching consistent adjusted profitability.
Conclusion:
The analysts on the call are doing what we were doing in our last analysis, which is due diligence on the cloud companies most likely to survive a recession. This is why the analysts continually hammered management on this very minor miss. It’s not about the miss in terms of dollar amount (don’t care about that too much) rather the question is if the miss translates to Snowflake being discretionary.
In terms of our position sizing and allocation, we need more information from MongoDB as we will allocate more to the stronger company between these two.