Category: Semiconductor Stocks

TSMC: The Common Denominator to AI Stocks

Posted on by io-fund

TSMC is a leading foundry that manufactures the world’s most advanced chips. Most importantly, the company has over 90% market share in the manufacturing of advanced AI chips. The most advanced node in production today is the 3-nanometer process technology, which is primarily used by Apple for iPhones and MacBooks. Nvidia is also expected to use the 3nm for the new Blackwell architecture. The 5-nanometer process technology is popular for HPC and smartphone applications and will be moving to 3-nanometer and 2-nanometer process technology.

One of TSMC’s niches for success is its unique business model as a pure-play foundry. This approach, which focuses on manufacturing for its customers rather than designing or manufacturing semiconductor products under its own name, has been a significant factor in its growth. TSMC's customer base includes leading companies like Apple, Nvidia, AMD, Qualcomm, Marvell, and Intel.

We have revamped the Essentials Portfolio by including TSMC. The stock was previously in our I/O Fund Portfolio and we sold for +20% profits due to technical and geopolitical risk. Yet, we continue to monitor the company due to its strong revenue growth, profits, and cash flows. TSMC is benefiting from economies of scale and its leadership position in the foundry industry, best illustrated by the company maintaining strong profitability while increasing production. Many companies have struggled with rising costs, while TSMC has successfully navigated these challenges by controlling costs and negotiating better prices with its customers.

For our Essentials Members, we want to demonstrate that a successful entry is often planned months in advance. We will most likely enter in Q4 2024 to Q1 2025 for a long-term position. We think it’s a mistake to buy stocks without a strategy, and thus, we present our strategy on TSMC below. We crowned TSMC as the stock that got away for H1 2024. Our firm closed our TSMC position late last year for a +20% gain, when it was at $92.28. We decided to instead focus on stocks with heavier AI concentration with less geo-political risk.

Market Dominance

According to Trend Force research, TSMC is the leading global foundry in terms of revenue. It has a market share of 62.3% in Q2 2024. Samsung ranks a distant second with a market share of 11.5% in Q2 2024. SMIC ranks third with a market share of 5.7% and UMC ranks fourth with a market share of 5.3%.

Product and business model

Fabless semiconductor companies benefit from outsourcing their fabrication of chips to companies like TSMC. They hereby save the high costs of building and maintaining facilities for chip manufacturing. The fabless companies instead spend their resources on R&D for designing chips. There is a third category of semiconductor companies which are called integrated device manufacturers (IDMs), who design and manufacture chips like Samsung and Intel.

TSMC was founded by Morris Chang in 1987. He is known as the father of Taiwan’s chip industry and is credited to the concept of the foundry business. The company successfully developed a business model that ensures it will not compete with its customers as it does not manufacture under its own name. The company was able to win Apple’s business from Samsung since the latter is a direct competitor of Apple.  The company’s chips are mainly used in five platforms: smartphones, high-performance computing, Internet of Things, Automotive, and Digital Consumer Electronics.

Revenue

The analyst revenue estimates are trending higher reflecting strong AI demand.

Q2 2024 revenue grew by 32.8% YoY and 10.3% QoQ to $20.82 billion beating guidance of $19.6 billion to $20.4 billion, driven by strong AI demand and partially offset by smartphone seasonality.

Management guide for the third quarter is between $22.4 billion and $23.2 billion. This represents year-over-year growth of 31.9% at the midpoint, primarily driven by anticipated strong demand for smartphones and AI-related technologies.

Analyst consensus estimates are trending higher. They expect revenue to grow 38.8% YoY in the next quarter, up from the 32.5% YoY growth expected in mid-June.

Note: The below analyst consensus estimates will differ slightly from the actual reported numbers in the company IR due to the currency conversion. However, we use the estimates below to understand the expected growth rate trend.

TSMC’s Aug monthly revenue grew by 33% YoY and down (-2.4%) MoM to NT$250.87 billion. It recorded its highest ever August sales and second highest monthly sales this year suggesting strong AI and smartphone demand. July monthly revenue grew by 44.7% YoY and 23.6% MoM to NT$256.95 billion.

During the Q2 earnings call, management increased full year 2024 revenue guidance from low to mid-20% to slightly above mid-20% in US dollar terms. “Over the past three months, we have observed strong AI and high-end smartphone related demand from our customers, as compared to three months ago, leading to increasing overall capacity utilization rate for our leading-edge 3-nanometer and 5-nanometer process technologies in the second half of 2024. Thus, we continue to expect 2024 to be a strong growth year for TSMC. We are raising our full-year guidance and now expect our 2024 revenue to increase slightly above mid-20s percent in US dollar terms.”

Margins

Due to the economies of scale and its leadership position in the foundry industry, the company maintained strong profitability. Despite the rising costs, the company has mitigated these challenges through cost controls and negotiating favorable pricing with its customers. Management is confident of achieving a long-term gross margin of 53% and higher.

According to DigiTimes, TSMC has notified its clients that prices for its 3-nanometer and 5-nanometer process products will increase by 3 to 8% in 2025. In the last earnings call, management hinted that prices will increase due to cost escalation.

“To ensure a proper return from our investment, both pricing and cost are important. TSMC's pricing strategy is strategic, not opportunistic to reflect the value that we provide […] Today, we are investing heavily in leading-edge specialty and advanced packaging technologies to support our customers' growth and enable their success. If customers do well, TSMC should do well. For example, we are happy to see many of our customers' structural profitability improving in these past few years. At the same time, we face rising cost challenges due to increasing process complexity, a leading load, higher electricity costs in Taiwan, global fiber expansion in higher cost regions, and other cost inflation challenges. Therefore, we will continue to work closely with our customers to share our value. We will also work diligently with our suppliers to deliver on cost performance.”

The Q2 gross margin was 53.2% compared to 54.1% in the same period last year and 53.1% in Q1. It beat the management guide of 51% to 53%, primarily helped by better capacity utilization, cost improvement, favourable foreign exchange rate, and partially offset by margin dilution from N3 ramp. TSMC will see headwinds in the initial ramp phase before ultimately realizing higher margins once 3nm has scaled. The price increase and yield improvement is expected to improve the margins next year.

Management has guided Q3 gross margin to increase 1.3 percentage points sequentially to 54.5% at the mid-point due to the better capacity utilization, cost improvements, productivity gains. The margin is partially offset by the N3 ramp, N5 to N3 tool conversion costs, and higher electricity prices in Taiwan. Electricity prices in Taiwan increased by 17% last year and another 25% in April this year.

The Q2 operating margin was 42.5% compared to 42% in the same period last year and Q1. Due to the operating leverage the operating expenses were lower at 10.5% of revenue compared to 11.1% in Q1 that helped to expand the margins. Management guide for Q3 is 43.5% at the midpoint.

Net income was $7.66 billion or 36.8% of revenue compared to $5.9 billion or 37.8% of revenue in the same period last year. Return on equity was 26.7% compared to 23.2% in the same period last year.

EPS

EPS is expected to increase substantially moving forward.

Q2 GAAP EPS was $1.48, up 29.8% YoY and beat estimates by 4.2% due to better capacity utilization, cost improvement and operating leverage. Analysts expect GAAP EPS to grow 36.4% YoY in Q3 to $1.76 and $1.94 in Q4.

Cash Flows and Balance Sheet

The company’s financial stability is evident in its cash flow growth, which has more than doubled YoY.

Operating cash flow was $11.68 billion or 56.1% of revenue compared to $5.45 billion or 35% of revenue in the same period last year and 73.6% in Q1. The operating cash flows was lower in Q2 last year due to the income tax payment of $3.85 billion.

Free cash flow was $5.32 billion or 25.5% of revenue compared to (-$2.72 billion) or (-17%) of revenue in the same period last year and 43% in Q1. Capex was down (-22.2%) YoY to $6.36 billion in Q2. The foundry industry is capital-intensive and free cash flows could be lumpy as capex could vary each quarter. The company had negative free cash flows last year due to the income tax payment of $3.85 billion and also higher capex of $8.17 billion.

Management expects strong AI demand to continue and raised the midpoint of the capex for 2024 to $31 billion from the previous $30 billion.

Inventories were $8.39 billion compared to $8.35 billion in Q1. Inventory turnover days decreased seven days sequentially to 83 days due to higher N3 wafer shipment.

Cash and marketable securities were $63.05 billion and debt of $30.4 billion compared to $60 billion and $30.25 billion in Q1. The company paid $2.8 billion in dividends in Q2.

Revenue by Platform

As the leading foundry for AI accelerators, TSMC is riding the enormous wave of demand from Big Tech. The chipmaker’s high-performance computing (HPC) revenues rose 28% QoQ to $10.8 billion and accounted for 52% of Q2 revenue, up from 46% of revenue in Q1. It is above the 50% mark for the first time.

Smartphone revenues declined (-1%) QoQ due to seasonality and accounted for 33% of revenue compared to 38% of revenue in Q1. In the Q2 earnings call, management mentioned that they are witnessing strong AI and high-end smartphone-related demand, suggesting strong HPC and smartphone revenues in Q3.

Internet of Things revenue grew by 6% sequentially and accounted for 6% of revenue.

Automotive revenue increased 5% sequentially and accounted for 5% of revenue. Digital Consumer Electronics increased 20% sequentially accounting for 2% of revenue. Other revenue increased 5% and accounted for 2% of revenue.

Advanced Nodes

The Advanced nodes are defined as 7-nanometer and below. We discussed in our editorial on the advanced nodes and AI-related revenue reaching fresh records. Most of the AI chips produced by the company utilize 5-nanometer and 4-nanometer process technology. However, 3-nanometer revenue is expected to triple this year. Volume production for 2-nanometer is expected in 2025 and should have a meaningful revenue contribution in the first half of 2026.

We mentioned, “Currently, AI accelerators use TSMC’s 5nm process. Nvidia’s Hopper and Blackwell are built with a N4X process that is tailored for high-performance computer applications. This is a customized variant called “4N” that Nvidia uses, yet TSMC recognizes this as 5nm revenue in their earnings report. AI accelerators are expected to quickly move to smaller nodes to help lower power consumption. TSMC’s 3nm process is more energy efficient, and energy efficiency will improve further with the 2nm process.”N4X process that is tailored for high-performance computer applications. This is a customized variant called “4N” that Nvidia uses, yet TSMC recognizes this as 5nm revenue in their earnings report. AI accelerators are expected to quickly move to smaller nodes to help lower power consumption. TSMC’s 3nm process is more energy efficient, and energy efficiency will improve further with the 2nm process.”

Due to its leadership position, management has been optimistic about the long-term opportunity in the manufacturing of AI chips. C.C. Wei said in the Q1 earnings call, “In summary, our technology leadership enable TSMC to win business and enables our customer to win business in their end market. Almost all the AI innovators are working with TSMC to address the insatiable AI-related demand for energy-efficient computing power. We forecast the revenue contribution from several AI processors to more than double this year and account for low-teens percent of our total revenue in 2024.Almost all the AI innovators are working with TSMC to address the insatiable AI-related demand for energy-efficient computing power. We forecast the revenue contribution from several AI processors to more than double this year and account for low-teens percent of our total revenue in 2024.

For the next 5 years, we forecast it to grow at 50% CAGR and increase to higher than 20% of our revenue by 2028. Several AI processors are narrowly defined as GPUs, AI accelerators and CPU's performing, training and inference functions and do not include the networking edge or on-device AI. We expect several AI processors to be the strongest driver of our HPC platform growth and the largest contributor in terms of our overall incremental revenue growth in the next several years.”For the next 5 years, we forecast it to grow at 50% CAGR and increase to higher than 20% of our revenue by 2028. Several AI processors are narrowly defined as GPUs, AI accelerators and CPU's performing, training and inference functions and do not include the networking edge or on-device AI. We expect several AI processors to be the strongest driver of our HPC platform growth and the largest contributor in terms of our overall incremental revenue growth in the next several years.”

  • 3-nanometer process technology contributed 15% of wafer revenue, while 5-nanometer and 7-nanometer accounted for 35% and 17% respectively in Q2 2024.
  • 3-nanometer process technology contributed 9% of wafer revenue, while 5-nanometer and 7-nanometer accounted for 37% and 19% respectively in Q1 2024.

Advanced Packaging

The AI wave has also boosted the company’s advanced packaging business, particularly Chip-on-wafer-on-substrate (CoWoS) services. Taiwan Semi’s CoWoS capacity is expected to rise to 85,000 wafers by the end of next year. This lines up with recent estimates from Morgan Stanley, with analysts now expecting CoWoS capacity to reach 80,000 to 90,000 wafers per month by the end of 2025, up from a prior estimate of 70,000. The 2025 production capacity would suggest over 430% increase from 15,000 at the end of 2023.

Management said in the earnings call Q&A that the supply is expected to continue to be tight next year. They also mentioned in the Q2 earnings call they are working with OSAT (Outsourced Semiconductor Assembly and Test) partners to increase production capacity.

Gokul Hariharan

“How do you think about supply demand balance for AI accelerator and CoWoS advanced packaging capacity? And I think in your symposium you talked about 60% CAGR, component growth for CoWoS capacity in the next four, five years. So, could you talk a little bit about how much capacity for CoWoS would you be planning to build next year as well?”

C. C. Wei

“Gokul, I also try to reach the supply and demand balance, but I cannot today. The demand is so high, I have to work very hard to meet my customers' demand. We continue to increase, I hope sometime in 2025 or 2026 I can reach the balance. You're talking about the CAGR or those kind of increase of the CoWoS capacity. Now it's out of my mind. We continue to increase whatever, wherever, whenever I can. Okay. The supply continues to be very tight, all the way through probably 2025 and I hope it can be eased in 2026. That's today's situation.”

Gokul Hariharan

“Any thoughts on next year capacity? Are you going to double your capacity again next year for CoWoS?”

C. C. Wei

“The last time I said that, this year I doubled it, right? More than double. Okay. So next year, if I say double, probably I will answer your question again next year and say more than double. We are working very hard, as I said. Wherever we can, whenever we can.”

—End Quote

The company’s other advanced packaging technology, system-on-integrated chips (SoIC), is also in robust demand. According to TrendForce, the company is expected to increase the monthly capacity to 5,000 to 6,000 units by the end of this year, up 2.5x to 3x from 2,000 units in 2023. Furthermore, by the end of next year, it is expected to scale to 10,000 units. According to the Economic Daily News, the company has secured Apple as the second major customer of SoIC.

The gross margin for advanced packaging was previously lower than the corporate average. However, due to cost controls and economies of scale, it is now close to the corporate average.

Due to the strong demand TSMC is expected to assign the orders of the initial stage of CoWoS packaging, Chip on Wafer (CoW) to OSAT partner SPIL. This is the first time the company is outsourcing this process since the demand is high and previously WoS (Wafer-on-Subtrate) process was outsourced while keeping the higher margin CoW process in-house. This is line with the management comments in the earnings call to increase production by collaborating with OSAT partners.

Valuation

The company trades at a P/E ratio of 32.8 and a forward P/E ratio of 27.8. The P/S ratio is 12.4 and a forward P/S ratio of 10.9. In the last five years, the P/E ratio peaked at 41.8 in February 2021 and hit a low of 10.3 in November 2022. The stock is now trading above its five-year average P/E ratio of 24.1.

TSMC Technical Analysis

By Portfolio Manager Knox Ridley

Since the 2022 top, every time the S&P 500 makes a new high while semiconductors make a lower high, we have seen a correction follow. In June and July of 2024, most semiconductors put in a top and have since made a series of lower highs. When we compare this to the S&P 500 that is currently pushing to new all-time highs today, we are seeing the largest divergence between semiconductors and the broad market on record.  This does not guarantee volatility, but the history of this pattern, especially considering the importance of semiconductors in the new AI economy, warrants patience.

Like most AI related semiconductor stocks that we track, TSM appears to be setting up for one more drop in this on-going correction. All corrections, whether we are dealing with a multi-year secular bear market, or an intraday pullback, consists of 3 waves: The initial drop down (the A wave), followed by a corrective bounce (B wave), and then the final drop lower, which tends to be the most dramatic part of a correction (C wave).

Based on the pattern analysis, TSM, like most semis, put in a top in June/July, and is currently completing what appears to be the B wave bounce. Note the 3 waves higher off the August low. It looks like the B wave needs one more minor swing to the $190 – $207 region to complete. This should lead to the final C wave drop into the $138 – $125 region, which is where we will target our entry. If this plays out, it will complete multi-month correction pattern, which should lead to the next push to new highs. If TSM instead drops below $168 before making a swing high, then we will consider the B wave complete.

The alternative scenario that we are tracking is that the correction ended at the August lows. This would mean that the current pattern taking us to the +$230 region is an ending diagonal pattern, which needs to prove itself to us before we buy into it. Once this bounce completes, we will need to see a larger corrective drop that makes a higher low, or a direct breakout on heavy volume over the $207 region. If this plays out, we will likely pivot our buy plan to accommodate this scenario.

Conclusion

TSMC is the common denominator to the market’s biggest AI winners as it supplies Nvidia, Apple, AMD, Marvell, and Qualcomm. As the AI economy continues to grow, Taiwan Semi’s importance will only increase. The HPC segment recently soared to its highest levels ever, and although TSMC is lumpy at times within particular revenue segments, it is remarkably consistent on margins and earnings growth.

The company is negotiating better prices with its customers, is making cost improvements, and is maintaining strong margins and cash flows. We look forward to adding this stock to our portfolio in the coming months and will provide the play-by-play for our entries and adds to Essentials Members.

Royston Roche, Equity Analyst at the I/O Fund, contributed to this article.

Recommended Reading:

Micron FQ4 Earnings Preview: Attractive Valuation, Look for Non AI-Related Weakness

Posted on by io-fund

Micron will release tomorrow with management expecting revenue to grow 89.5% YoY to $7.6 billion at the midpoint, an 8-point acceleration from last quarter. The stock is up 10% YTD at the time of writing, below the 80% returns in mid-June. The stock has given up much of its gains due to the potential fears of a temporary slowdown in the non-high-bandwidth memory market and concerns that Micron’s upside is limited due to contracted pricing for the next 6-8 quarters. We covered this here stating: “The blemish in the report is the commentary that HBM is sold out for calendar 2024 and 2025, with “pricing already contracted for the overwhelming majority of our 2025 supply.”

The slowdown in non-HBM memory is expected to be from smartphone and PC customers who “have built some inventory,” according to comments from Micron’s management team at the Deutsche Bank conference in late August. Taken out of context, it could be a dire comment yet the overall tone was bullish, citing the increase in non-HBM inventory is to protect smartphone and PC OEMs from higher pricing and tighter supply due to the outsized demand for AI-related HBM. Essentially, the message is that AI-related HBM is going to crowd out non-HBM in the supply chain.

The not-so-bullish commentary from the DB conference is more economic related as Micron stated: “We have shared before and as is well known in the broad industry reports that consumer retail channels, industrial, automotive tend to be relatively weaker right now, as well as China has some weakness too.” China is important to Micron making up 25% of its revenue. The company had stated earlier at a Keybanc conference: “In China, economic activity and consumer buying patterns are weak or uneven at best, depends on the market.”

Previous commentary in the earnings call and also during Keybanc’s conference hint that Micron is expecting several hundred million of HBM revenue in fiscal ’24, multibillion HBM revenue in fiscal 2025 and “have HBM consistent with our DRAM share at some point in calendar '25.” However, the discussions regarding Q1 in the DB conference were a lackluster, stating: “So when we look at all of these factors and all of these trends, when we look at our FQ1 we think our bit shipments will be somewhat flattish to slightly up in FQ1 versus FQ4.” This has caused some analysts to downgrade the stock as the expectation is non-HBM is weighing on HBM if we will see bit shipments “only slightly up.”

When the analysts downgraded the stock, some of them cited an oversupply of HBM. We see no evidence of an oversupply of HBM, and in fact, Micron has stated the opposite, stating they are seeing “robust demand trends” and “these trends of AI and the tight supply environment we see that continuing in 2025 as well. And that's why we say that, we will have substantial revenue record in 2025, and of course, robust profitability in 2025 as well versus 2024.”

We will see if Micron’s call has any changes in tone on HBM-related inventory; it’ll be important to carefully distinguish what is causing the underperformance on bit shipments as Micron is trading at an attractive valuation – nearly 50% lower than its historic average.

The trim today is not Micron-specific rather Knox has not been liking SMH’s recent price action. We are sending a clear message to our members that we aren’t buyers of AI semis at this moment as we think there will be opportunities to buy lower. We aren’t heavy sellers either, although we have a strategy where this could be the outcome over the next 30-60 days. Real wealth is made during drawdowns, not by perfectly timing a top. We’ve covered this extensively in our webinars and have a thought leadership article on the free side coming out by Knox next week on the topic, as well.

Revenue:

FQ4 revenue is expected to accelerate to 90.5% YoY growth to $7.64 billion and decelerate to 75% growth to $8.27 billion in Q1 FY2025.

Last quarter, revenue grew by 81.5% YoY to $6.81 billion, up from 57.7% in FQ2 due to strong AI demand. The company reported a record high data center revenue mix with 50% sequential data center revenue growth.

  • DRAM revenue grew by 13% QoQ to $4.7 billion, helped by about a 20% price increase and offset by a decline in bit shipments in the mid-single digits.
  • NAND revenue grew by 32% QoQ to $2.1 billion, helped by an increase in bit shipments in the high single digits and a price increase of about 20%. Management expects DRAM shipments to be flattish and NAND shipments to increase slightly in FQ4.
  • Analysts expect FY2024 revenue to grow 63.3% YoY to $25.37 billion.
  • FY2025 revenue is expected to grow 50.2% YoY to $38.10 billion and 16.3% to $44.30 billion in FY2026.

Margins

Margins experienced a steep cyclical low and now appears to have bottomed. In FQ2, the company achieved its goal of positive adjusted operating margin a quarter ahead of expectations, primarily helped by the recovery in DRAM and NAND pricing.

Management expects gross margin expansion to continue, helped by price and also higher-value products like HBM, high-capacity DIMMs (dual in-line memory modules), and SSDs. The guide for FQ4 is 34.5%.

Management reiterated the sequential improvement in adjusted gross margin for the November quarter as they said in the KeyBanc forum, “On gross margin, sequentially August to November quarter, we expect gross margins to be up around 200 — or up around a couple of hundred basis points, consistent with what we've said before.”

  • FQ3 gross margin was 26.9%, up from (-17.8%) in the same period last year and 18.5% in FQ2. Management guide for FQ4 is 33.5%. The adjusted gross margin was 28.1%, up from (-16.1%) in the same period last year and 20% in FQ2.

The improvement in gross margin was due to higher pricing, product mix, and cost reductions.

  • Operating margin was 10.6%, up from (-46.9%) in the same period last year and 3.3% in FQ2. Management guide for FQ4 is 17.8%. Adjusted operating margin was 13.8%, up from (-39.2%) in the same period last year and 3.5% in FQ2.

Management guide for FQ4 is 20.6%. The operating expenses were at the lower end of the guidance due to cost controls and operational efficiencies. Management expects operating expenses to increase sequentially in FQ4 “due to an increase in R&D program expenses and a nonrecurring Q3 asset sale gain contemplated in our Q3 guidance.”

  • Net income was $332 million or 4.9% of revenue compared to a net loss of (-$1.9 billion) or (-50.5%) of revenue in the same period last year. Adjusted net income was $702 million or 10.3% of revenue compared to an adjusted net loss of (-$1.57 billion) or (-41.7%) of revenue in the same period last year.

EPS 

FQ3 GAAP EPS came at $0.30 and beat estimates by 1.2%. Adjusted EPS was $0.62, up from an adjusted loss per share of (-$1.43) in the same period last year. The company beat adjusted EPS estimates by 17.3%, which was helped by higher prices, higher margin product mix, and cost controls.

  • EPS guide for FQ4 is $0.61 at the midpoint and the adjusted EPS guide of $1.08 at the midpoint. Analysts expect adjusted EPS of $1.11 for FQ4 and $1.59 for FQ1.
  • Analysts expect FY2024 adjusted EPS of $1.23, up from an adjusted loss per share of (-$4.45) for the FY2023.
  • For FY2025 they expect adjusted EPS to grow 636% YoY to $9.04 and 34.5% YoY to $12.16 for FY2026.

Cash Flow and Balance Sheet

 FQ3’s operating cash flow was $2.48 billion or 36.4% of revenue compared to $24 million or 0.60% of revenue in the same period last year and 20.9% of revenue in FQ2. The cash flows have improved with higher revenue and profitability.

FQ3 adjusted free cash flow was $425 million or 6.2% of revenue compared to (-$1.36B) or (-36.1%) of revenue in the same period last year and (-$29 million) or (-0.50%) of revenue in FQ2. Capex was $2.1 billion in FQ3 compared to $1.4 billion in the same period last year and $1.2 billion in FQ2.

Management expects positive adjusted free cash flow in FQ4 despite a capex of about $3.0 billion. Capex would be $8.0 billion in FY2024, up from $7.0 billion in FY2023.

Management expects capex to rise around 35% of FY2025 revenue, i.e., comes to about $13 billion and the company is able to support it due to higher profitability. Mark Murphy, CFO, said in the FQ3 earnings call, “Record revenue and significantly improved profitability in fiscal 2025 will help support average quarterly CapEx in fiscal 2025 to be meaningfully above the fiscal Q4 2024 level of $3 billion. We expect CapEx around mid-30%s range of revenue for fiscal 2025, which will support HBM assembly and test equipment, fab and back-end facility construction, as well as technology transition investment to support demand growth.

As noted earlier, half or more of the expected CapEx increase in fiscal 2025 will be to support U.S. greenfield fab construction. As we have noted in the past, the CHIPS grants, ITC, and state incentives offset a significant portion of the U.S. fab CapEx investments.”

  • Inventory was $8.5 billion or 155 days compared to $8.4 billion or 160 days of inventory in FQ2. Management expects days of inventory to decline in FY2025. (This also runs counter to what some analysts downgrades are stating, which is there’s an oversupply when management comments point to the opposite).
  • Cash and investments were $9.22 billion and debt of $13.258 billion compared to $9.7 billion and $13.7 billion in FQ2. In FQ3, the company repaid $650 million in debt and paid $128 million in dividends. The company had also announced in early August that they might resume the stock repurchase program.

Business Units

Compute and Networking Business Unit (CNBU) grew by 18% QoQ and 85% YoY to $2.57 billion. This was an acceleration from 59% in FQ2.

DRAM data center revenue more than doubled year-over-year. We foresee this segment being strong yet there could be potential weakness in client-related compute.

Mobile Business Unit (MBU) grew by 94% YoY and down (-1%) sequentially to $1.59 billion due to a planned volume decline, which was partially offset by improved pricing. Look for potential weakness here given comments that the oversupply maybe coming from mobile and PCs.

Embedded Business Unit (EBU) grew by 42% YoY and 16% sequentially to $1.29 billion helped by record revenue in automotive. Look for weakness here given commentary about automotive, industrial and consumer-related end markets.

Storage Business Unit (SBU) revenue grew by 116% YoY and 50% sequentially to $1.35 billion with growth in all end markets. The company achieved record data center SSD revenue, which nearly doubled sequentially. There could be potential weakness in the client storage segment depending on PC demand.

Other noteworthy points to watch

HBM Revenue

The company reported HBM revenue of over $100 million in FQ3 and expects to generate several hundreds in FY2024 and multiple billions in FY2025. They also expect the HBM market share to match the overall DRAM market share sometime in CY2025, which they mention is in the low 20s. UBS expects HBM revenue to be $5.61 billion in FY2025 from the expected $603 million in FY2024.expects HBM revenue to be $5.61 billion in FY2025 from the expected $603 million in FY2024. Management said HBM revenue is accretive to FQ3 margins and expects this trend to continue.

HBM and NAND CAGRs

HBM’s bit growth CAGR is expected to remain strong and be above 50% for the next few years.

“Well, as we have said before, that we see the CAGR for HBM growth — in terms of bit growth CAGR to be well above 50% over the next few years. So certainly, HBM is a strong growth driver. And again, as we increase our mix of HBM going forward, it will, of course, be continuing to be accretive to our financial performance, including margins. And we are pleased that with the strong performance that we have we are sold out for '25 as well with overwhelming part of our output already committed in terms of pricing.”

During the Q&A, it was pointed out that the forecast for NAND CAGR growth was lowered from “the low 20s” to a new forecast of “growth in the high teens.” Management’s response was the following:

“And I'll also tell you that we basically revise the base here for the CAGR that we used. So this time, the CAGR that we used, we use the base year of 2023. And in 2023, as you know, we had bit demand growth in NAND that was higher, meaningfully higher than the CAGR. So that, of course, the larger base of 2023, just somewhat changed our outlook on the overall CAGR.”

AI PCs and Smartphones

The company expects to benefit from the AI PCs and smartphones. They expect AI PCs to have 40% to 80% more DRAM content than today’s PCs. Similarly, smartphones that are AI enabled will have about 50% to 100% more content compared to phones released last year. In the near-term, management has stated there could be an oversupply. We expect there to be more questions on the call about this comment from the DB conference. “PC and smartphone customers have built additional inventories due to the rising price trajectory, the anticipated growth in AI PCs and AI smartphones, as well as the expectation of tight supply as an increasing portion of DRAM and NAND output is dedicated to meeting growing data center demand.” There are additional China-related concerns for mobile and PCs.

Valuation 

On the top line, Micron is trading above it’s historic valuation at 4.85 compared to its five-year P/S ratio is 3.58. Notably, since the FY ends in August, the P/S ratio will be adjusted post-results and is trading at a 2.7 forward PS. The forward PS is the one that will reflect the valuation post-results.

On the bottom line, the forward PE Ratio of 10.5 will take effect post-results and is about 20% to 30% lower than the 3-year and 5-year median. These valuations are tricky because AI stocks are being re-rated. For example, Nvidia and AMD are trading about 50% lower than their averages, and yet are struggling to breakout. We think it’s important to weigh what the stocks have been trading at since the AI boom, or about early 2023. Micron has been trading considerably higher since the AI boom and we continue to watch the stock for the right entry over the next few months.

Conclusion

Micron is setting up to be an excellent buy over the next few months, yet we continue to watch the horizon to buy AI stocks at lower prices. Micron is perhaps the more complicated AI semi to time as it’s trading at an attractive valuation yet it’s also a bellwether of sorts for the economy. There could be a scenario where any AI bullishness is overshadowed by consumer-facing segments. Notably, the margins are expected to continue improving in FY2025 and we foresee Micron being more defensible than others when it comes to economic headwinds, albeit less defensible should we see China-related headwinds.

We are patient. We have time on our side. We will participate if Micron does well tomorrow, yet we also trimmed should Micron not do well tomorrow. Look for the I/O Fund to lean into defensibility while acknowledging inning one from AI (which started in 2023) has been a crazy-good run and our goal is to load up for inning number two in the coming months.

Royston Roche and Beth Kindig, Equity Analysts at the I/O Fund, contributed to this article.

Recommended Readings:

AI PCs Have Arrived: Shipments Rising, Competition Heating Up

Posted on by io-fund

This article was originally published on Forbes on Updated Sep 13, 2024, 06:42am EDTForbesForbes on Updated Sep 13, 2024, 06:42am EDT

Chipmakers Qualcomm, Intel and AMD are working to bring AI-capable PCs to the “mainstream”, delivering powerful neural processing units to PCs for on-computer AI operations. AI PCs are not only a consumer market, rather will also be driven forward by enterprises and developers seeking to upgrade their employee PCs. This being one of the biggest upgrade cycles in PC history, competition has heightened as Q3 comes to a close.

Intel believes the AI PC “promises to be a huge improvement for everyday PC usages,” as it “represents a fundamental shift in how our computers operate.” Lenovo executives believe that “AI enables a personalized user experience that, once adopted, will lead to significant productivity gains and foster greater innovation and creativity.” AMD’s executives have explained that they “see AI as the biggest inflection point in PC since the Internet with the ability to deliver unprecedented productivity and usability gains.” Qualcomm’s CEO has said that he believes the AI PC “is as significant as Windows 95. It is changing the experience, delivering groundbreaking AI capabilities, fundamentally changing how we interact with our PCs.”

AI PCs are expected to usher in a new wave of AI underpinned by increased productivity. AI’s trajectory will increase when more people can access AI-powered applications, which in turn, will help AI developers build larger ecosystems. Existing PCs are not yet powerful or efficient enough to run AI at the edge, yet the PCs currently being released with NPUs (neural processing units) will exceed 40 trillion operations per seconds (TOPS) and this will usher forth the performance upgrades needed to make on-device AI a reality for millions of consumers and enterprises.

AI PCs Drive Growth in Q2

AI PCs gained momentum with AI PC shipments doubling sequentially despite low-single digit growth in the broader PC market in Q2. Further growth is expected in the second half of the year as Intel and AMD prepare competing chips to Qualcomm’s Snapdragon X lineup.

Current AI PC projections:

  • For Q2, PC market growth was estimated in the low-3% range:
  • Canalys is placing growth at 3.4% YoY
  • Counterpoint Research seeing 3.1% YoY growth
  • IDC reporting is reporting 3% YoY growth with its preliminary numbers.
  • Total shipment figures varied slightly for each group ranging from 62.5 million to 64.9 million, representing QoQ growth of 9.1% for the broader PC market.

AI PCs were a predominant driver of growth sequentially, rising from approximately 7% share in Q1 to 14% share of shipments in Q2. On a unit basis, AI PCs shipments jumped 120% QoQ, from 4.0 million in Q1 to 8.8 million in Q2. For the first half of the year, that puts shipments at around 12.8 million units.

Full year forecast from Canalys projects shipments of 44 million AI PCs in 2024, implying that Q3 and Q4 will combine for 32 million units, or averaging nearly 50% QoQ growth in both quarters. IDC expects AI PC shipments to run closer to 50 million, with Gartner estimating 54.5 million earlier this year, though that would require ~300% growth from 1H to 2H.

Regardless of the exact number, the setup looks strong for H2 growth, although keep in mind, many competitors are competing for market share.

Sign up for I/O Fund's free newsletter with gains of up to 2600% because of Nvidia's epic run – Click here.Click here.

2024 is the Spark for 2025’s Boom

While 2024 is off to a strong start so far for AI PC shipments with Q2 showing triple digit sequential growth, the true growth story for AI PCs will be in 2025. This year is the initial spark for AI PCs to boom in 2025, when they are expected to quickly take significant market share.

AMD CEO Lisa Su cautioned that some investors may have mistakenly expected 2024 to be the big growth year, but that she believes that will instead be 2025. Su said at Goldman Sachs Communacopia and Technology Conference that she believes “we are at the start of a multiyear AI PC cycle. So again, you guys are always trying to go a little bit too fast. So, we never said AI PCs was a big 2024 phenomena. AI PC is a start in 2024. But more importantly, it's the most significant innovation that's come to the PC market in definitely the last 10-plus years.”

Should the broader PC market register growth in the mid-to-high single digit range to ~260 million shipments this year, AI PCs would take 17% share at 44 million units.

For 2025, shipments are forecast to rise more than 134% YoY to 103 million units, at the midpoint. Assuming high single digit growth in total shipments in 2025 to 280 million PCs, AI PCs would take close to 40% share at the midpoint of 106 million shipments forecasted.

A bar chart forecasting AI-capable PC shipments from 2024 to 2028, with separate projections from Gartner, IDC, and Canalys.

Source: Gartner/IDC/Canalys

Longer-term, growth is expected to remain robust. Intel’s executives expect AI PCs to account for more than 50% of the PC market in 2026, and as much as 80% market share by 2028. Canalys sees AI PC shipments top 200 million by 2028, taking up to 70% market share.

Commercial to Drive AI PC Growth

AI PC’s rapid adoption curve will be driven primarily by the commercial market, with AI PCs expected to be a catalyzing force ahead of the upcoming Windows refresh at the end next year.

There is indication the early majority will adopt AI PCs in 2025, and the late majority in 2026, with industry projections matching this narrative. This leaves time for consumers to participate, which so far has been a challenge for AI.

I covered this in the past, stating: “AI-capable PCs are expected to be an explosive trend through 2025 and beyond. The trajectory of AI will increase when more people can access AI-powered applications, which in turn, will help AI developers build a bigger ecosystem. Currently, there is a major bottleneck right now for AI applications to where client devices are not powerful enough or energy efficient enough to leverage AI capabilities at the edge.” For a deeper dive on AI PCs and industry commentary on the growth potential, refer to AI PC Stocks: Emerging 2024 And 2025 Story (io-fund.com).

AMD’s Su believes that while the industry is only in the “beginning” of the AI PC cycle, “next year, as we think about commercial PCs and commercial refresh cycle, we actually see AI PC as a driver of that commercial refresh cycle.” This is a view shared by IDC, which “believes the commercial market has the biggest short-term upside for AI in the PC industry,” with the consumer side “yet to be told in full.” This is a view shared by Lenovo’s executives, who believe the end of support for Windows 10 in 2025 “will necessitate a migration to Windows 11 for enterprises and commercial users, further driving demand for new PCs.”

Commercial market share of AI PCs is projected to reach almost 60% by 2027, per Canalys. This implies end-market shipments of nearly 100 million, with the remaining 40% share in consumer. Additionally, Gartner forecasts that 100% of commercial (enterprise) PC purchases will be AI PCs by the end of 2026, driven by the productivity gains realized by on-device AI, and as enterprise applications begin to take advantage of AI features. Lenovo also expects that “enterprises may increasingly require AI-enabled PCs to remain competitive.”

Every Thursday at 4:30 pm Eastern, the I/O Fund team holds a webinar for premium members to discuss how to navigate the broad market, as well as various stock entries and exits. We offer trade alerts plus an automated hedging signal. The I/O Fund team is one of the only audited portfolios available to individual investors. Learn more here.Learn more here.

Competition is Intensifying

Competition in AI PCs is quickly intensifying, with Qualcomm moving downstream to the $700 and below market with a new Snapdragon chip. The x-86 based competitors, Intel and AMD, see growth opportunities ahead, with Intel forecasting a surge in CPU shipments as it comes off the launch of its Core Ultra 200V (Lunar Lake) chips.

Nvidia and AMD reportedly are lining up powerful Arm-based CPUs to take on Qualcomm after its Arm exclusivity deal expires at the end of 2024. Apple is also rumored to be planning an M4-powered Mac refresh either by the end of this year or early 2025.

The chipmakers are competing on NPU performance, alongside efficiency:

Qualcomm’s Snapdragon X NPU offers 45 TOPS of AI performance, while CEO Cristiano Amon “claiming a performance-per-watt 2.6 times better than AMD and 5.4 better than Intel's Core Ultra 7 chips.”

Intel’s Lunar Lake chip offers up to 48 TOPS on the NPU, and Intel is claiming “1.4x AI performance over the Snapdragon X Elite running the Stable Diffusion tool in a GIMP plugin; faster overall core performance versus Ryzen and Qualcomm competition; and a 1.5x improvement over its previous generation in the performance of the integrated GPU.”

AMD’s Ryzen AI 300 series chips (Strix Point and Strix Halo) offer up to 50 TOPS performance from the NPU, the highest on the market so far.

Apple’s M4 chip offers up to 38 TOPS performance on the NPU, with the chip originally deploying on the iPad lineup with the Mac refresh rumored for this year or next.

Nvidia does not have an NPU competitor yet, as it believes its GeForce RTX GPUs offer significantly higher TOPS and more AI performance, meeting the bill for AI PCs. However, Nvidia and MediaTek are reportedly working on an Arm-based AI PC chip for a 2025 launch following the expiration of Qualcomm’s exclusivity deal.

Intel’s Shipments Ahead of Expectations, AMD Execs See Share Gains

Though we are on the precipice of going through a major shift to where Arm architecture will compete more directly with x86 architecture for PCs, Arm lost share in both desktops and notebooks in Q2, according to data from Mercury Research.

In notebooks, Arm lost 144 bp market share QoQ to 11.4% share in Q2. AMD’s notebook share rose 121 bp QoQ to 18.0%, while Intel gained 23 bp QoQ to 70.6% share.

In desktops, Arm lost 31 bp QoQ to 5.9% share, AMD also lost 83 bp QoQ to 21.6% share, while Intel’s share rose 113 bp QoQ to 72.5%.

This was reflected in Intel’s Q2 report, where management noted that its Core Ultra (AI PC chip) shipments exceeded expectations in that quarters and “more than doubled sequentially.” Core Ultra shipments have surpassed 15 million since December 2023, with Intel believing that it remains firmly on track to surpass 40 million AI PCs by the end of 2024 and more than 100 million cumulatively by the end of 2025.

Qualcomm has also said that its initial launch of Copilot+ PCs, powered by its Snapdragon lineup, is exceeding internal expectations, but management has not shed light on AI PC shipments or revenue. Analysts questioned management over the impact of AI PCs on fiscal Q4’s guide, with management stating that “it's too early to kind of have either a bullish assumption or a specific assumption on PC,” and it is more “about kind of the longer-term growth opportunity, and being very specific on sell-through in the short term is not really something that we have insight into.” Management plans to provide more information on the revenue ramp during its upcoming Investor Day conference.

Qualcomm also mentioned that “20 Copilot+ PCs from Microsoft, Dell, HP, Lenovo, Acer, ASUS and Samsung are now available across 20 countries and 47 retailers.” This is a fraction of what Intel is launching, with Intel saying that “Microsoft has qualified Lunar Lake to power more than 80 new Copilot+ PCs across more than 20 OEMs.”

Similar to its competitors, AMD has been quite bullish regarding the impact of AI PCs on the upcoming refresh cycle, as it eyes a growth opportunity in commercial PCs.

AMD’s Ryzen AI 300 series (featuring the industry’s fastest NPU with 50 TOPS) launched at the end of July to “to strong reviews,” with “more than 100 Ryzen AI 300 series premium, gaming, and commercial platforms on track to launch from Acer, ASUS, HP, Lenovo, and others over the coming quarters.” Management noted that “customer excitement for our new Ryzen processors is very strong, and we are well positioned for ongoing revenue share gains based on the strength of our leadership portfolio and design win momentum.”

This builds on CFO Jean Hu’s prior comments that AMD was “gaining share” in PCs, as CEO Lisa Su sees “clear opportunities to gain additional commercial PC share based on the performance and efficiency advantages of our Ryzen Pro portfolio and an expanded set of AMD-powered commercial PCs from our OEM partners.” AMD’s executives hold the view that the company is “underrepresented” in PCs, “but particularly in the commercial PC side.” Management believes they can do “above-typical seasonality” in the second-half of 2024, based on the timing and strength of its product launches.

Intel’s Margin Troubles and AMD’s Growth

While Qualcomm is bringing a formidable Arm-based competitor to the industry, management has provided little clues to the revenue or margin impacts from AI PCs. Intel is looking to make large strides by prioritizing shipments at the expense of margins (and shareholders), while AMD is posting the strongest YoY growth rates, though comps were still weak in Q2.

Qualcomm’s IoT revenue, which houses its PC segment, has posted YoY revenue declines since fiscal Q2 2023, or six consecutive quarters. IoT revenue in Q3 declined (8%) YoY to $1.36 billion, as the pace of declines slowed, from (32%) YoY in Q1 and (11%) YoY in Q2. Management has not stated the impact of PCs on IoT revenue growth, or its contribution.

Intel reported 9% YoY revenue growth in Client Computing in Q2 to $7.41 billion, decelerating from Q1’s 31% YoY growth. On a dollar basis, growth has declined for two consecutive quarters, falling from $8.84 billion in Q4 to $7.53 billion in Q1 and now $7.41 billion.

Bar chart showing Intel's Client Computing Revenue from Q2 2022 to Q2 2024, measured in billions of dollars.

Source: Company IR

Notebook revenue declined sequentially for Intel, despite shipping AI PC chips ahead of expectations. Desktop revenue gained sequentially, but not enough to offset the softness in notebook: desktop revenue rose just 2.7% QoQ to $2.53 billion, while notebook slipped (4.3%) QoQ to $4.48 billion.

While Intel is reporting sequential weakness, AMD is seeing growth recover and growth rates remain strong. AMD’s Client segment has registered YoY growth of 42%, 62%, 85%, and 49% respectively, though against weak comps. Revenue in Q2 was just under $1.5 billion, up 9.1% QoQ and reaching the highest level in eight quarters. However, Client revenue is still one-third below its peak levels since in 2022, suffering from the sharp inventory correction that hit the industry.

Bar chart showing AMD's Client Revenue from Q2 2022 to Q2 2024, with year-over-year growth percentages.

Source: Company IR

As this growth story unfolds, operating margins will be critical to track for both Intel and AMD. Intel’s margins and profitability took a hit from its decision to prioritize AI PC shipments, and management seems keen on keeping that a priority moving forward, suggesting margins may have not stabilized. For AMD, the Client segment has historically been a significant driver of operating income. For example, in Q2 2022, when revenues were above $2.1 billion, Client’s operating margin was 31%, compared to 6% now – essentially, there’s much more ground to cover on the profitability side versus the growth side.

Conclusion

AI PCs bring x86 and Arm to the battlegrounds, with Qualcomm’s Snapdragon lineup making an attempt at the first major inroad on x86 Windows-based CPUs. Intel and AMD see strong growth ahead for their x86 competitors, though Apple is expected to bring a major upgrade with M4 chips later this year. This is ahead of possible new Arm-based CPUs from Nvidia and AMD next year.

Shipments of AI PCs have only just begun accelerating, with Q2 showing triple-digit sequential growth but penetration rates remain low. AI PC adoption is projected to skyrocket towards 70% share of total PC shipments by 2027, with growth arising predominantly in commercial markets first.

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.

Recommended Reading:

Nvidia Stock Is Selling Off: It’s Not Because Of Blackwell

Posted on by io-fund

This article was originally published on Forbes on Updated Aug 30, 2024, 08:59am EDTForbesForbes on Updated Aug 30, 2024, 08:59am EDT

Our firm extrapolated supply chain data to conclude that Blackwell is in production at TSM and SMCI last week in the analysis: Nvidia Stock: Blackwell Suppliers Shrug Off Delay. The media was making much ado about nothing (and astonishingly, still is) despite crystal clear confirmation from Nvidia’s management team that all is well.

Given these delay rumors, it was widely expected that Nvidia’s management would provide some transparency in Q2 as to the status of Blackwell. I joined ‘Making Money’ on Fox Business Network shortly before Nvidia’s report, telling host Charles Payne that “we are getting bullish signals from the supply chain,” such as TSM’s HPC growth and Super Micro’s liquid cooling growth, and that I “fully expect Nvidia’s management team to calm any concerns about the outlook for Blackwell.”

Direct liquid cooling doesn’t lie as it’s intricately linked to the Blackwell launch, implying that Blackwell would indeed ship by Q4 – and Nvidia just confirmed that (multiple times) in Q2’s release:

“Blackwell production ramp is scheduled to begin in the fourth quarter and continue into fiscal 2026. In the fourth quarter, we expect to ship several billion dollars in Blackwell revenue.”

Later in the call, Jensen Huang stated: “There were no functional changes necessary. And so we're sampling functional samples of Blackwell — Grace Blackwell in a variety of system configurations as we speak. There are something like 100 different types of Blackwell-based systems that are built that were shown at Computex. And we're enabling our ecosystem to start sampling those. The functionality of Blackwell is as it is, and we expect to start production in Q4.”

We had published for our free readers going into the print that the valuation was stretched, and it would require fiscal year revisions to create room in the valuation. As you’ll see below, we got a few revisions today, which is paramount for the stock price. Will these upward revisions be enough to sustain the price? We look at this and more below.

Q2 Revenue Beats Estimates

Q2’s revenue of $30.04 billion increased 122% YoY and 15% QoQ, with management pointing out that “customers continue to accelerate their Hopper architecture purchases while gearing up to adopt Blackwell.” This marked a $1.3 billion beat to the consensus estimate for $28.75 billion. It also was a deceleration from 262% YoY growth in Q1, as Nvidia is now facing tougher comps against the vertical ramp of Hopper last year. GAAP EPS of $0.67 beat estimates by $0.03, and represented YoY growth of 168% and QoQ growth of 12%.

Nvidia guided for Q3 revenue of $32.5 billion, once again above consensus estimates, though it was only $700 million higher than the $31.77 billion estimate at the midpoint. This represents growth of 79.4% YoY at midpoint, compared to the estimate for 75.3% growth next quarter. Despite this being one of the ‘smaller’ beats in recent quarters, it’s a testament to the strength of Nvidia’s demand to guide for $2.5 billion sequential growth primarily based on Hopper demand with no contribution from Blackwell.

Revenue Growth

Source: I/O Fund

Sign up for I/O Fund's free newsletter with gains of up to 2600% because of Nvidia's epic run – Click hereClick hereClick here

Data Center Strength Visible with Blackwell on Tap

Data center revenue surpassed a $105 billion annualized run rate this quarter, up from $90 billion annualized last quarter, as Nvidia reported $26.27 billion in data center revenue, up 152% YoY and 16% QoQ. Nvidia said that “Hopper demand is strong, and shipments are expected to increase in the second half of fiscal 2025,” while Blackwell is on track to ramp in Q4 with “several” billions in revenue expected that quarter.

Notably, purchase commitments and obligations for inventory and capacity rose nearly 48% QoQ to $27.8 billion, including “new commitments for Blackwell capacity and components,” another signal that Nvidia is prepared to ramp in full-force come Q4.

In the segment, compute revenue was $22.6 billion, up 162% YoY, while networking revenue was $3.67 billion, up 114% YoY. In networking, Nvidia noted that InfiniBand and Ethernet drove growth in the quarter, and the 16% QoQ growth included “a doubling of Ethernet for AI revenue.”

Data Center Revenue

Source: I/O Fund

Nvidia’s Q3 revenue guide implies data center revenue above $28 billion to $28.5 billion, which we had modeled in our pre-earnings analysis earlier this week.

Delay Concerns Cleared, But Valuation Looks Stretched

Nvidia cleared the delay concerns for Blackwell, saying that they “shipped customer samples of our Blackwell architecture in the second quarter. We executed a change to the Blackwell GPU mask to improve production yield. Blackwell production ramp is scheduled to begin in the fourth quarter and continue into fiscal 2026,” with several billion in Blackwell revenue expected in Q4. Purchase commitments reiterated that Nvidia is serious about launching on schedule, and lining up the capacity and components to launch in full-force by the end of the year.

I spoke with Yahoo Finance on Thursday morning following the report, reemphasizing that the delay concerns were “completely thrown off the table last night. … Wall Street obviously is very closely tied to estimates, and we never saw revisions downward based on the so-called delay. … Nvidia beat, and they’re saying Blackwell is basically on time,” which is “not a concern — if anything, it’s extremely bullish.”

However, I cautioned on the valuation: “When you have a high-flyer like Nvidia, you get stretched at times. Going into the print, we warned our members that this valuation is looking a little toppy. What we need is for the fiscal year estimates next year to go up, so we’re in a waiting game for analysts to revise their estimates upward, which eventually they will, but until then the valuation is stretched.”

This morning, while I was being interviewed by Yahoo, we’ve already seen analyst estimates for Nvidia’s revenue revised higher following the report:

  • Fiscal 2025 revenue is now estimated at $124.8 billion, up 3.9% from the $120.1 billion estimate prior to Q2’s report.
  • Fiscal 2026 revenue is now estimated at $172.1 billion, up 5.2% from the $163.6 billion estimate prior to the report.

However, the true impact of Blackwell is yet to be seen in these estimates, with the only clues right now being Q3’s $32.5 billion guide and expectations for several billion in Blackwell revenue in Q4. From a long-term perspective, I explained on Yahoo Finance that the first “pathway for growth is to pay very close attention to Nvidia around the fiscal year guide,” while the “second-biggest moment of the year will be when Blackwell is shipping in volume. This will be the Q2 report, but we’ll get some signs in Q1 with that forward guide.” I believe that “early next year will be fireworks” for Nvidia, similar to Hopper’s moment in the fiscal Q1 report in May 2023.

Every Thursday at 4:30 pm Eastern, the I/O Fund team holds a webinar for premium members to discuss how to navigate the broad market, as well as various stock entries and exits. We offer trade alerts plus an automated hedging signal. The I/O Fund team is one of the only audited portfolios available to individual investors. Learn more here.Learn more hereLearn more here.

Eyes on Margins as Blackwell Ramps

Margins remained strong in Q2, with Nvidia reporting gross and operating margins at the high end and above its guided ranges. However, management guided for Q3 margins to contract slightly QoQ, suggesting that Q1 was the peak for both gross and operating margins with some pressure ahead as Blackwell gears up to launch in Q4.

  • GAAP gross margin was 75.1% in Q2, ahead of management’s guide for 74.8%. Adjusted gross margin was 75.7%, ahead of guidance for 75.5%. Per the CFO: “As our Data Center mix continues to shift to new products, we expect this trend to continue into the fourth quarter of fiscal 2025.” It’s likely she is referring to the higher cost of memory components, which we outlined in our pre-earnings analysis.
  • GAAP operating margin was 62.1%, ahead of the implied guide for 60.5%, indicative of the operating leverage power that Nvidia still commands in mid-launch cycle for Hopper with the H200s shipping now. Adjusted operating margin was 66.4%, ahead of the implied guide of 65.5%.
  • GAAP net margin was 55.3% down from 57.1% last quarter. This represents profits of $16.6 billion, up over $2 billion. This was a very large beat compared to the $14.3 billion guided.
Nvidia GAAP Margins

Source: I/O Fund

The chart above shows Nvidia’s margins, with the slight sequential contraction this quarter and next quarter visible. It’s no small feat to maintain GAAP operating margin >60% for four consecutive quarters while simultaneously undergoing the semiconductor industry’s most advanced and most rapid product release cycle. However, with management guiding for full-year gross margins to be in the mid-70% range, we’ll be keeping a close eye on how margins trend in Q3 heading into Q4 as Blackwell ramps — where the market is a tad concerned is gross margins, which peaked at 78.4% and will exit the year in the mid-70% range.

Conclusion

Our pre-earnings writeup expressed concerns about the valuation going into the print, and I think the selling on Thursday reflects the valuation. Our firm stuck our neck out over the past few weeks to bring quality information to our readers on how the supply chain for Blackwell is ramping. We were the first and only firm that I’m aware of to present actionable data that countered what other media outlets were reporting. To refresh your memory, media outlets stated Blackwell was delayed into Q1: “If the upcoming AI chips, known as the B100, B200 and GB200, are delayed three months or more, it may prevent some customers from operating large clusters of the chips in their data centers in the first quarter of 2025, as they had planned.”

In contrast, my analysis stated: “From the horse’s mouth, Nvidia’s own management team, it was stated during the GTC Financial Analyst Day in March that the very first systems will ship in Q4, but to expect constraints.”

Well, we have our answer – Blackwell is, in fact, shipping in Q4 and ramping in Q1. Purchase commitments up 48% QoQ help to reflect how serious the company is when it comes to the speed of ramping shipments.

Earnings reports are truly 50/50 – nobody can tell you what the market will do following a report. For example, we had high confidence Nvidia would beat, but there’s much more to consider than a beat. What’s important is to have a strategy. Our firm champions actively managing tech positions rather than buy-and-hold. Our plan is to trim Nvidia at key levels and attempt to buy lower. This is due to valuation concerns, but also importantly, many AI stocks are trading at stretched valuations. We’ve stated publicly a few times that Nvidia is a buy on dips, implying investors who are patient will find entries at lower prices.

Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.

Recommended Reading:

Nvidia Q2 Pre-Earnings: Hopper Thrives Yet Valuation is High

Posted on by io-fund

Nvidia is slated to beat this quarter and raise again next quarter. Capacity is increasing for Hopper and demand remains elevated.

Analysts are expecting demand for Hopper to absorb any minimal impact pushing Blackwell to Q1 will cause. It’s likely the B100s are canceled, which is the release that was expected this year. We wrote about why this could ultimately be bullish, if we assume the GB200s are oversubscribed. We believe the Blackwell delay will ultimately be immaterial to FY2026 revenue, and the point of last week’s analysis is there’s a much bigger picture to focus on.

With that said, I’m not loving Nvidia’s valuation. Nvidia is trading at its highest levels since the AI accelerator boom was priced in. We’ve seen some blow-the-doors off quarters during the period reflected below — which pushes back on the idea a beat/raise justifies this valuation. The big spikes downward are fiscal year adjustments, meaning if an investor feels compelled to buy at a high valuation, it’s better to do so in Dec/Jan than August.

Here’s a bottom-line valuation that reflects 50 forward PE being a psychological hurdle since Jan 1, 2023, the period where we saw AI-driven beats/raises.

Where things get complicated is the FY2026 revenue estimates are simply too low. They are at $165B and the I/O Fund believes we will see $210B (or more) this time next year. That should create about 27% room. However, FY2026 is a ways off, as it begins nearly seven months from now in February. Therefore, Q4 offers a better risk/reward setup where we’d enter based on our understanding of the 27% upside that next fiscal year will bring.

Notably, to be balanced, this earnings report marks the moment when profits are expected to plateau (as opposed to skyrocket) and margins will slightly contract for the first time since Nvidia’s big AI moment.

I want to be crystal clear that I foresee Nvidia being the strongest mega cap stock for the next decade. Trading in this case sometimes does not work as emotions run high when stocks selloff, and some investors struggle to buy back when the market is in a state of fear (versus greed).

However, the I/O Fund is an actively managed portfolio, and so our style is distinct. The facts below point toward us trimming our position and attempting to buy again lower. Keyword is “attempt.” Knox’s Advanced Market Signals webinar this week is quite timely, being the first trading day after Nvidia reports. You can expect him to spend considerable time on this chart to inform our Members of our plan. He also briefly described the general outline here.

Revenue and Financials:

Revenue for the July quarter is expected to be $28.7 billion for growth of 112.6%. Management provided guidance of $28 billion for YoY growth of 107.3% at the midpoint. Q3 is expected to be $31.7 billion for growth of 74.8%.

Pictured Above: Nvidia reported peak growth in Q4 & Q1 and revenue will decline moving forward. Our firm covered this previously here.this previously here.

FY2025 Revenue growth is expected to be $121.3 billion for growth of 99.03%. Next year, growth is expected to be 38.7% for revenue of $168.2 billion.

We discussed earlier this year that Nvidia was hitting peak growth. Blackwell’s GB200 are the predominant catalyst for next year, along with an ongoing, rapid product release cycle.

Key Segments:

The data center revenue reported revenue of $22.5 billion, up 427% YoY and up 23% QoQ. Analysts are expecting between $105B and $107B in annual data center revenue.

This is what a rough idea looks like (back of the napkin math)

  • Q1 at $22.5B (reported)
  • Q2 at $25.5B – adding $3B
  • Q3 at $28.0B – adding $2.5B
  • Q4 at $30.0B – adding $2.0B

The chances Nvidia beats are high, for example: “Jefferies expects another "strong beat in July and strong guidance into October" from Nvidia with beats of about $1B for both results and guide.

Citi has also stated: “The firm sees $1B of upside to Street sales estimates but below the $2B beat in the last four quarters based on supply chain comments and Blackwell delay concerns. Citi expects Street estimates to go higher for the next two quarters post the earnings print and for Blackwell comments to reassure investors on a strong 2025 outlook for Nvidia.”

Margins:

You can read more about the importance of HBM3e and its integration into the H200s here.more about the importance of HBM3e and its integration into the H200s here.

This quarter, the H200s shipped with HBM3e memory, which are higher priced memory components that require a re-configuring of the pricing for Nvidia’s GPUs, and subsequently may also be affecting margins.

Back in February, Next Platform did a thought exercise to predict HBM3 represented $8,800 of the GPU price and HBM3e will represent $16,500 of the street price (Nvidia will pay a lower amount) if the H200s are priced at $41,000. Nvidia will push some of the costs onto the buyers, yet there is bound to be questions in the call to determine how these higher priced components are affecting margins. As of now, we are seeing some reports the H200s cost $39,900.

Notably, this quarter Nvidia is showing signs of plateauing on profit growth. We’ve seen historic quarter-over-quarter growth where Nvidia adds about $3 billion QoQ in profits. Yet, management is guiding for roughly flat QoQ growth, and in the case of net income, a nominal decline. This is a different discussion than YoY growth, which is quite high on profits, at well over 100%.

Gross margin last quarter was 78.4% and management has guided for a decline to 74.8%. The adjusted gross margin of 75.5% will also be lower than last quarter’s at 78.9% for adjusted gross profit of $21.14B. Last year, the adjusted gross margin was 71.2% for a 430 bps expansion YoY.

Management guided for GAAP operating margin to be 60.5% compared to 64.9% last quarter. The adjusted operating margin is expected to be 65.5% for adjusted operating income of $18.3 billion. The adjusted operating margin last year was 57.57% for a 293 bps expansion YoY.

Pictured above: Nvidia expected to report a decline in margins with last quarter marking the peak

Management guided for GAAP net margin of 51.1% for net profit of $14.3 billion. This will be lower than last quarter at 57.1% for net income of $14.8 billion. Yet, is up from 45.8% last year.

Nvidia is expected to report EPS of $0.64 for YoY growth of 137.8%. The company reported adjusted EPS of $0.61 last quarter for growth of 461.5%. Q3 is expected to report EPS of $0.71 for YoY growth of 76.8%.

FY2025 EPS is expected to be $2.75 representing growth of 111.8%. FY 2026 is expected to be $3.83 representing growth of 39.4%.

Cash:

Last quarter, the company report operating cash flow of $15.3 billion for a margin of 58.9%. The free cash flow margin of 57.3% resulted in FCF of $14.9 billion. The company has $31.4 billion on the balance sheet with debt of $9.71 billion.

The company utilized cash of $7.8 billion towards shareholder returns with $7.7 billion in stock repurchases and $98 million in cash dividends. The shares started trading on a 10-for-1 stock split basis on June 10th.

Conclusion:

Our process is to not become complacent around valuations. We closed heavyweight Microsoft on these concerns and closed best-of-breed CrowdStrike on valuation concerns.

We will notnot be closing Nvidia by any meansany means, but we do think a position that is 20%+ allocation and trading at a high valuation may have a mid-single digit to high-single digit trim in its allocation in order. Blackwell is coming, but it’s not here yet — we’ve got time.

Please note: Our style is (and should be) distinct from your style. We are not financial advisors and what is best for the I/O Fund may not be best for you. Per our terms and conditions, please consult your financial advisor before making any decisions.our terms and conditions, please consult your financial advisor before making any decisions.

We recently discussed an emerging ad-tech AI leader in our Advanced Market Signals Tier. Our proprietary custom-built screeners identified the stock. The company witnessed a strong acceleration in growth and margins due to its AI-powered advertising engine. To read the article, upgrade here.here.

Recommended Readings:

Nvidia Q2 Pre-Earnings: Hopper Thrives Yet Valuation is High

Posted on by io-fund

Nvidia is slated to beat this quarter and raise again next quarter. Capacity is increasing for Hopper and demand remains elevated.

Analysts are expecting demand for Hopper to absorb any minimal impact pushing Blackwell to Q1 will cause. It’s likely the B100s are canceled, which is the release that was expected this year. We wrote about why this could ultimately be bullish, if we assume the GB200s are oversubscribed. We believe the Blackwell delay will ultimately be immaterial to FY2026 revenue, and the point of last week’s analysis is there’s a much bigger picture to focus on.

With that said, I’m not loving Nvidia’s valuation. Nvidia is trading at its highest levels since the AI accelerator boom was priced in. We’ve seen some blow-the-doors off quarters during the period reflected below — which pushes back on the idea a beat/raise justifies this valuation. The big spikes downward are fiscal year adjustments, meaning if an investor feels compelled to buy at a high valuation, it’s better to do so in Dec/Jan than August.

Here’s a bottom-line valuation that reflects 50 forward PE being a psychological hurdle since Jan 1, 2023, the period where we saw AI-driven beats/raises.

Where things get complicated is the FY2026 revenue estimates are simply too low. They are at $165B and the I/O Fund believes we will see $210B (or more) this time next year. That should create about 27% room. However, FY2026 is a ways off, as it begins nearly seven months from now in February. Therefore, Q4 offers a better risk/reward setup where we’d enter based on our understanding of the 27% upside that next fiscal year will bring.

Notably, to be balanced, this earnings report marks the moment when profits are expected to plateau (as opposed to skyrocket) and margins will slightly contract for the first time since Nvidia’s big AI moment.

I want to be crystal clear that I foresee Nvidia being the strongest mega cap stock for the next decade. Trading in this case sometimes does not work as emotions run high when stocks selloff, and some investors struggle to buy back when the market is in a state of fear (versus greed).

However, the I/O Fund is an actively managed portfolio, and so our style is distinct. The facts below point toward us trimming our position and attempting to buy again lower. Keyword is “attempt.” Knox’s webinar this week is quite timely, being the first trading day after Nvidia reports. You can expect him to spend considerable time on this chart to inform our Members of our plan. He also briefly described the general outline here for Essentials.

Revenue and Financials:

Revenue for the July quarter is expected to be $28.7 billion for growth of 112.6%. Management provided guidance of $28 billion for YoY growth of 107.3% at the midpoint. Q3 is expected to be $31.7 billion for growth of 74.8%.

Pictured Above: Nvidia reported peak growth in Q4 & Q1 and revenue will decline moving forward. Our firm covered this previously here.this previously here.

FY2025 Revenue growth is expected to be $121.3 billion for growth of 99.03%. Next year, growth is expected to be 38.7% for revenue of $168.2 billion.

We discussed earlier this year that Nvidia was hitting peak growth. Blackwell’s GB200 are the predominant catalyst for next year, along with an ongoing, rapid product release cycle.

Key Segments:

The data center revenue reported revenue of $22.5 billion, up 427% YoY and up 23% QoQ. Analysts are expecting between $105B and $107B in annual data center revenue.

This is what a rough idea looks like (back of the napkin math)

  • Q1 at $22.5B (reported)
  • Q2 at $25.5B – adding $3B
  • Q3 at $28.0B – adding $2.5B
  • Q4 at $30.0B – adding $2.0B

The chances Nvidia beats are high, for example: “Jefferies expects another "strong beat in July and strong guidance into October" from Nvidia with beats of about $1B for both results and guide.

Citi has also stated: “The firm sees $1B of upside to Street sales estimates but below the $2B beat in the last four quarters based on supply chain comments and Blackwell delay concerns. Citi expects Street estimates to go higher for the next two quarters post the earnings print and for Blackwell comments to reassure investors on a strong 2025 outlook for Nvidia.”

Margins:

You can read more about the importance of HBM3e and its integration into the H200s here.more about the importance of HBM3e and its integration into the H200s here.

This quarter, the H200s shipped with HBM3e memory, which are higher priced memory components that require a re-configuring of the pricing for Nvidia’s GPUs, and subsequently may also be affecting margins.

Back in February, Next Platform did a thought exercise to predict HBM3 represented $8,800 of the GPU price and HBM3e will represent $16,500 of the street price (Nvidia will pay a lower amount) if the H200s are priced at $41,000. Nvidia will push some of the costs onto the buyers, yet there is bound to be questions in the call to determine how these higher priced components are affecting margins. As of now, we are seeing some reports the H200s cost $39,900.

Notably, this quarter Nvidia is showing signs of plateauing on profit growth. We’ve seen historic quarter-over-quarter growth where Nvidia adds about $3 billion QoQ in profits. Yet, management is guiding for roughly flat QoQ growth, and in the case of net income, a nominal decline. This is a different discussion than YoY growth, which is quite high on profits, at well over 100%.

Gross margin last quarter was 78.4% and management has guided for a decline to 74.8%. The adjusted gross margin of 75.5% will also be lower than last quarter’s at 78.9% for adjusted gross profit of $21.14B. Last year, the adjusted gross margin was 71.2% for a 430 bps expansion YoY.

Management guided for GAAP operating margin to be 60.5% compared to 64.9% last quarter. The adjusted operating margin is expected to be 65.5% for adjusted operating income of $18.3 billion. The adjusted operating margin last year was 57.57% for a 293 bps expansion YoY.

Pictured above: Nvidia expected to report a decline in margins with last quarter marking the peak

Management guided for GAAP net margin of 51.1% for net profit of $14.3 billion. This will be lower than last quarter at 57.1% for net income of $14.8 billion. Yet, is up from 45.8% last year.

Nvidia is expected to report EPS of $0.64 for YoY growth of 137.8%. The company reported adjusted EPS of $0.61 last quarter for growth of 461.5%. Q3 is expected to report EPS of $0.71 for YoY growth of 76.8%.

FY2025 EPS is expected to be $2.75 representing growth of 111.8%. FY 2026 is expected to be $3.83 representing growth of 39.4%.

Cash:

Last quarter, the company report operating cash flow of $15.3 billion for a margin of 58.9%. The free cash flow margin of 57.3% resulted in FCF of $14.9 billion. The company has $31.4 billion on the balance sheet with debt of $9.71 billion.

The company utilized cash of $7.8 billion towards shareholder returns with $7.7 billion in stock repurchases and $98 million in cash dividends. The shares started trading on a 10-for-1 stock split basis on June 10th.

Conclusion:

Our process is to not become complacent around valuations. We closed heavyweight Microsoft on these concerns and closed best-of-breed CrowdStrike on valuation concerns.

We will notnot be closing Nvidia by any meansany means, but we do think a position that is 20%+ allocation and trading at a high valuation may have a mid-single digit to high-single digit trim in its allocation in order. Blackwell is coming, but it’s not here yet — we’ve got time.

Please note: Our style is (and should be) distinct from your style. We are not financial advisors and what is best for the I/O Fund may not be best for you. Per our terms and conditions, please consult your financial advisor before making any decisions.our terms and conditions, please consult your financial advisor before making any decisions.

Recommended Readings:

Arm Stock: Buy Its Customers, Not The Stock

Posted on by io-fund

This article was originally published on Forbes on Aug 15, 2024, 05:09pm EDTForbes on Aug 15, 2024, 05:09pm EDT

Arm Holdings is the third-best performer of 2024 in AI-related semiconductor stocks with a 56% YTD return, behind only Nvidia and Taiwan Semiconductor. The stock is a market favorite as Arm’s heterogenous compute design has created a monopoly in mobile, primarily, yet Arm’s RISC architecture is also found in PCs, powers sensors and supercomputers. In total, over 280 billion Arm-based chips have been shipped dating back to the 1980s.

The latest Arm v9 architecture offers significant improvements in performance and efficiency, particularly for artificial intelligence (AI) applications. This has led to increased adoption by premium smartphone partners, and also with hyperscalers that are developing their own custom silicon for data center use.

The market is excited about Arm’s v9 architecture because it commands double the royalty rate with Arm receiving a higher percentage of the chip’s selling price when a manufacturer uses v9 designs. The estimated royalty rates for v9 are around 4%, compared to Arm’s blended 1.7% royalty rates for the prior generations – however, this is simply not enough to consistently accelerate Arm’s top-line growth to justify its valuation as $70+ billion more AI chips are sold this year. So even though Arm is leveraging its established royalty and licensing model through its extensive ecosystem to drive predictable future growth and a rather defensible bottom line, apart from the cyclical doldrums of the semiconductor industry, its growth story pales in comparison to some of its key customers.

Q1 Earnings Strong, Yet Q2 and FY25 Failed to Impress

Given the string of strong beat and raises from GPU leader Nvidia over the past several quarters, the market has been setting the bar high for AI-related stocks, including Arm. Despite beating Q1 estimates, Arm failed to meet high expectations as it guided for Q2 results below consensus.

Arm reported 39% YoY growth to a record quarterly revenue of $939 million in Q1. Looking ahead to Q2, Arm projected revenue between $780 million and $830 million, for flat YoY growth at midpoint, decelerating from Q1’s 39% print. Analysts had expected Arm to guide to $813 million in revenue for Q2, for YoY growth of 1%. Given the small growth rate, a miss feels odd given the trajectory of other AI stocks.

In Q1, licensing revenue rising 72% YoY and 14% QoQ to $472 million, offsetting a more than (9%) QoQ decline in royalties. Arm said that licensing “hit a record level as the proliferation of AI everywhere is driving more companies to make broad and long-term commitments to use Arm’s power-efficient technology in their future products,” while royalty revenues are benefitting from Arm v9, which commands higher royalties per chip.

With that said, Arm expects next quarter “to be the low point of the year due to the timing of revenue recognition from licensing,” while also being one of the “highest bookings quarters of the year.” Royalty revenue is also expected to accelerate from 17% YoY to the low-20% YoY range in the quarter.

Arm’s adjusted EPS guide also came in below consensus estimates for the quarter, with Arm projecting $0.23 to $0.27 in EPS, short of the $0.28 estimate. While these may seem like thin margins for a miss, Arm’s premium valuation offers little room for error.

For FY25, Arm guided for revenue between $3.8 billion and $4.1 billion, or $3.95 billion at midpoint, falling short of the $4 billion consensus estimate. This forecast points to YoY growth of 18% to 27%. Adjusted EPS was guided between $1.45 to $1.65, which at the midpoint fell short of the consensus estimate for $1.57.

Arm also guided down for royalty revenue growth, projecting royalty revenue growth in the low 20% range, compared to the mid-20% range previously. Licensing revenue is expected to increase in the mid-20% range, with Q2 expected to be the weakest quarter and Q4 the strongest.

Our firm has been quite vocal that IPOs are not worth the risk, stating “there is no riskier proposition than an IPO that is richly valued.” The liquidity event that an IPO becomes after its lockup expiration creates high risk for tech investors as individual investors are often up against a deluge of insider selling. The fact that Arm is an overpriced IPO that is missing estimates this early is a concern. GAAP operating margin also has contracted from 18.5% to 5.4% on a TTM basis, primarily from IPO-related expenses.

Sign up for I/O Fund's free newsletter with gains of up to 2600% because of Nvidia's epic run – Click hereClick here

v9 Growth Decelerated Quickly, But Projected to Rebound

As a primary driver of royalty revenue growth, v9’s revenue is important to track as it commands double the royalty of v8, and features in “virtually all high-end data center chips” and a majority share in smartphones. For example, v9 underpins Nvidia’s Hopper series chips, Amazon’s Graviton chips, Microsoft’s Cobalt chips, and many more. Arm also sees v9’s increased adoption (and increase in royalty mix) and the ramp of CSS-based chips in the second half of the year as growth drivers.

In Q1, v9 accounted for 25% of net royalty revenue, up from 20% last quarter, and up from 15% two quarters ago. Smartphones contributed to 40% of royalty revenue in FY24 and was the first to adopt v9. Notably, the smartphone market is recovering right now — smartphone revenues grew 50% YoY despite a single digit increase in unit sales, demonstrating how increased adoption of v9 can quickly impact revenue growth.

However, on the backs of a QoQ decline in royalty revenue in Q1, v9 revenue growth has decelerated dramatically, despite increasing its mix share by 500 bp QoQ. Growth decelerated from 46% QoQ in Q4 to 14% QoQ in Q1, reaching $116.8 million.

Arm Quarterly Revenue

On the backs of a QoQ decline in royalty revenue in Q1, v9 revenue growth has decelerated dramatically, but is projected to rebound. Source: I/O Fund

v9 revenue growth is expected to reaccelerate to above 30% QoQ in Q2, and remain at that level in Q3 as royalty revenue returns to QoQ growth, with ~9% QoQ estimated for Q2 and ~12% QoQ for Q3. Management said it expects the 500 bp QoQ mix increase “to be the continued trajectory” moving forward, implying Q2’s mix at 30% and Q3’s at 35%. This V-shaped recovery in v9 revenues arises from this consistent increase in mix along with a return to QoQ growth around the 10% to low-teens range. As such, v9 royalties are projected to rise to ~$153 million next quarter, up 31% QoQ, a 17 percentage point acceleration.

Yet despite the reacceleration in v9 revenue growth, it still contributes only a fraction of the end-market value being added from chips built on the design. Meaning, while v9’s revenue is increasing by just $37 million QoQ, end market customers such as Nvidia are selling $2 billion more QoQ in GPUs built on v9. I’ve said this since 2021 – despite Arm’s designs powering 99% of the mobile market, or the most important AI GPUs available today, the revenue gains generated from its royalties are nowhere near the same ballpark as the growth and revenue generated by its customers.

x86 Will Make for Strong AI PC Competition

Arm dominates in the mobile market, with more than 99% market share, but in some of Arm’s smaller end markets, such as microprocessors, the company faces steeper competition from x86-based players AMD and Intel.

According to data from Mercury Research, in the microprocessor unit market, AMD gained 68 bp QoQ to 19.2% market share in Q2, while Intel gained 37 bp QoQ to reach 71.1% market share. Arm, on the other hand, saw a 105 bp QoQ decline in market share, dropping to 9.7% share.

At a closer look, Arm lost share against AMD and Intel in both notebooks and desktop microprocessors, yet gained share in servers.

  • In notebooks, Arm lost 144 bp QoQ, falling from 12.8% share in Q1 to 11.4% share in Q2. AMD’s notebook share rose 121 bp QoQ to 18.0%, while Intel gained 23 bp QoQ to 70.6% share.
  • In desktops, Arm lost 31 bp QoQ to 5.9% share, AMD also lost 83 bp QoQ to 21.6% share, while Intel’s share rose 113 bp QoQ to 72.5%.
  • Servers were the only segment where Arm’s share grew in Q2, rising 41 bp QoQ to 6.7%. AMD’s share rose 40 bp QoQ to 22.5%, while Intel lost 81 bp QoQ to 70.8% share.

Notably, it’s premature to say how Arm-based PCs will do as more AI features are integrated into laptops. MacBooks switched from Intel’s x86-64 processors to Apple’s system on a chip (SoC) based on Arm 64 architecture. Arm offers much lower power consumption and generates less heat due to being a Reduced Instruction Set. The M2 is built on Taiwan Semi’s 5nm process with 100GB/s memory bandwidth and 24 GB of unified memory. When the M2 was released in 2022, Apple claimed 1.9X CPU performance at the same power. At the same performance level, Apple claims the M2 uses ¼ the power as x86.

From there, the M3 MacBook Air was released this year with Apple stating it’s 13X faster than an x86 Intel powered MacBook. The Arm-based system on a chip (SoC) combines a CPU and a GPU with a 16-core Neural Engine for what Apple is calling the “World’s Best Consumer Laptop for AI.”

Traditionally, Arm’s architecture was viewed as the best choice for mobile and Intel’s x86-64 as the best choice for the data center and PCs. However, we are on the precipice of going through a major shift to where Arm architecture will compete more directly with x86 architecture for PCs. This kicked off with Qualcomm’s Snapdragon Elite X, and will be further tested when AMD and Nvidia release Arm-based PC systems in 2025.

Every Thursday at 4:30 pm Eastern, the I/O Fund team holds a webinar for premium members to discuss how to navigate the broad market, as well as various stock entries and exits. We offer trade alerts plus an automated hedging signal. The I/O Fund team is one of the only audited portfolios available to individual investors. Learn more here.Learn more here.

Arm’s Growth Pales in Comparison to Customers

Arm’s value to the semiconductor industry is arguably indispensable, yet the stock trades at a premium valuation. The company is currently at highest top-line multiple in the entire semiconductor industry, even as its revenue growth pales in comparison to key chip customers.

Forward PS Ratio Comparison Chart

Source: YChartsYCharts

Arm trades at more than 33x forward revenue, a 40% premium to Nvidia’s 24x forward revenue multiple. The industry’s other most expensive names, and those with the most AI exposure trade at lower top-line multiples, with Monolithic Power at 19.6x forward revenue, and AMD at just 8.9x forward revenue.

Revenue Growth Estimate for Current Fiscal Year

Source: YChartsYCharts

Arm’s expected revenue growth of 23% for fiscal 2025 is not the most impressive in the industry. Astera Labs leads with 200% estimated revenue growth, although it’s coming off an extremely small base of less than $100 million per quarter. Yet, Nvidia is expected to nearly double its revenue with 97% estimated growth, essentially adding $60 billion YoY driven by GPU sales, on a very large revenue base.

Arm does not and will not share a similar hypergrowth profile from AI like Nvidia, which has seen ‘hockey-stick’ growth over the past six quarters, rising from more than $4 billion at the start of fiscal 2024 to potentially more than $25 to $26 billion in Q2 fiscal 2025. Arm, on the other hand, did not even add $300 million in revenue in that time frame.

For the entirety of 2024, Nvidia, AMD, and more of Arm’s key customers are likely selling more than $70 billion in additional AI chips this year, whereas Arm is only adding $740 million in revenue, or barely more than 1% of its customers’ sales increase.

While Arm’s royalty and licensing model offers more predictable revenue growth in a cyclical industry, it does not share the same operating leverage or margin profile to quickly grow into its premium top-line and bottom line multiple. Nvidia has visibly demonstrated its ability to capture much higher ASPs with each generation and drive significant operating leverage, rising from a 30% operating margin to 65% in four quarters while driving >600% bottom line growth.

Given this tremendous strength on the bottom-line, Nvidia trades at 43x forward earnings, with growth of 110% expected this year. However, Arm trades at nearly double that multiple, at almost 81x forward earnings, with just 23% growth expected. AMD and TSM both trade at much lower multiples for EPS growth in the 26% to 27% range.

Forward PE Ratio

Arm trades at the highest forward PE multiple among leading AI chip peers, at almost 81x forward earnings, with just 23% EPS growth expected. Source: YChartsYCharts

Conclusion

We’ve seen excitement over Arm’s AI opportunities and boosts to licensing fees and royalties, but our contention is that, similar to mobile, it’s better to own the AI leaders who license Arm’s technology. Investors have quite a few choices for AI stocks with stronger growth rates and cheaper valuations, rather than own Arm. Frankly, the valuation on Arm remains absurdly expensive, double or more than double the most-expensive chip stocks, including Nvidia.

There have been a few data points that have emerged since our last update on Arm in March 2024, and Arm’s Q2 and FY25 revenue guide also disappointed by falling short of analysts' expectations, a stark contrast to the billion-dollar beats and raises from Nvidia.

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.

Recommended Reading:

Alpha & Omega Q4: Revenue in Line, All Segments Rebound Sequentially

Posted on by io-fund

Alpha and Omega reported Q4 revenue in line with guidance, while margins rebounded at the high end of management’s guided range. Each of Alpha and Omega’s four key segments recorded sequential growth in the quarter, with management saying this “broad-based” rebound “confirmed the inventory correction is largely complete.”

However, management noted that PCs are “taking longer to recover than originally expected,” though sequential growth is still expected in the September quarter due to seasonal strength. Intel’s report last week showed that the challenged chipmaker is going all-in on AI PCs, making the decision to “more quickly ramp” its Core Ultra AI PC chips at the expense of margins and profitability.

Intel also said that its “Core Ultra volume more than doubled sequentially in Q2,” with shipments surpassing 15 million and remaining on track for more than 40 million by year end, and over 100 million cumulatively by the end of 2025. With Intel aggressively pursuing AI PCs, it’ll likely spark fierce competition in the space against Qualcomm and AMD, a positive for Alpha and Omega given its tie-ins and content on Intel and AMD’s AI CPUs.

Quite a few questions on the earnings call were about a graphics card customer that is using AOSL for accelerator cards, which increases the number of MOSFETs from 9 to 16 per driver up to 50 MOSFETs per driver. Given the customer has used AOSL for graphics cards in the past, and will now use AOSL for the upcoming GPU architecture, it’s likely the customer is Nvidia. There was also this announcement in June indicating Nvidia.

Notably, the company’s “midterm target model” is $1 billion in revenue with adjusted gross margin of 30%.

Revenue

Alpha and Omega’s Q4 revenue was largely in line with both management’s guide and analyst estimates. Notably, every key segment grew QoQ.

Q1’s forecast was above estimates and points to YoY revenue growth returning as early as next quarter. Management said that in Q4, they “saw relative strength coming from gaming, tablets, e-mobility, A.I., and home appliances, while the PC segment is taking longer to recover.”

  • Q4 revenue was $161.3 million, down (0.1%) YoY but up 7.5% QoQ as all four key segments rebounded sequentially. Management had guided to ~$160 million in revenue for the quarter, with analysts expecting the same.
  • FY 2024 revenue was $657.3 million, a decrease of (4.9%) YoY.
  • Q1 revenue was guided to $180 million, +/- $10 million, for flat YoY growth and QoQ growth of 11.6%. This came in above the consensus estimate for $174.4 million.

Alpha and Omega is now expected to potentially return to top-line growth in Q1, and accelerate through Q3 FY2025 to the low double digit range. Exiting the prolonged inventory correction, which dented growth significantly in FY 2023, clears up one major hurdle to revenue growth, though the extended PC recovery timeline is something to pay closer attention to through the September quarter.

Margins

Margins came in at the high end of management’s guided range, with CFO Lifan Yiang saying the QoQ expansion in adjusted gross margins, which flowed down the line, “was mainly driven by the improved factory utilization.”

  • GAAP gross margin in Q4 was 25.7%, up 200 bp QoQ from 23.7%, but down 190 bp YoY from 28.2%. Adjusted gross margin was 26.4%, up 120 bp QoQ but down 210 bp YoY.
  • For Q1, management guided GAAP gross margin at 25.0%, and adjusted gross margin at 26.4%, signaling that margins are likely to have found a bottom in Q3.
  • FY24’s GAAP gross margin was 26.2%, down 270 bp YoY. Adjusted gross margin was 27.2%, down 300 bp YoY. Management has remained positive about reaching a long-term target of >30% for gross margins.
  • GAAP operating margin was (0.9%) in Q4, well ahead of the implied (5.2%) margin based on management’s expenses guide, and increasing from (7.0%) last quarter. Adjusted operating margin was 2.0%, up from (0.7%) last quarter but down from 4.3% in the year ago quarter.
  • For Q1, GAAP operating margin is implied to be (1.1%), due to the guide for a slight QoQ decline in GAAP gross margin. Adjusted operating margin is implied to be 4.2%, a solid QoQ increase of 220 bp.
  • GAAP net margin was (1.7%), improving from (7.5%) last quarter. Adjusted net margin was 1.6%, improving from (0.8%) last quarter. GAAP EPS in Q4 was ($0.09), ahead of estimates for ($0.31), while adjusted EPS of $0.09 was ahead of estimates for $0.04.

Cash and Debt

Cash flow on this company (and most small caps) needs to be watched closely. FY2023 ended with cash flow margin of (-13%).

  • Operating cash flow was $7.1 million in Q4, for a 4.4% margin. This included $4.5 million in repayments of customer deposits. Management expects to refund ~$8.4 million in customer deposits in Q1.
  • For FY24, operating cash flow was $25.7 million, increasing 25.6% YoY.
  • Q4 free cash outflow was (-$0.21 million) with Capex at $7.2 million
  • Cash and equivalents totaled $175.1 million.
  • Debt totaled $38.4 million.

Key Segments

Computing

Computing revenue decelerated to 37.6% growth in Q4. The segment rebounded 4.4% QoQ to approximately $71.6 million in revenue. Management said that this was “slightly below our original expectation for mid-to-upper single digit growth,” primarily impacted by PCs.

Management said they saw “relative strength from tablets, A.I, and graphics cards in the quarter, offset by a slower PC market recovery. Notably, tablet revenue was a record high, and the contribution from A.I. and datacenter related applications continued to grow.”

Looking ahead, Alpha and Omega expects Q1 to see mid-single digits QoQ growth on a seasonal PC pickup, combined with strength from tablets, graphics, and AI accelerators. CEO Stephen Change shed more light on some of the opportunities ahead: “We are working on multiple opportunities leveraging our existing relationship with a key graphics card maker, as well as our product portfolio including new multiphase Vcore controllers and power stage solutions for advanced computing. We are also seeing some ramp in September from a leading power supply maker that is a key supplier to the same A.I./graphics customermultiple opportunities leveraging our existing relationship with a key graphics card maker, as well as our product portfolio including new multiphase Vcore controllers and power stage solutions for advanced computing. We are also seeing some ramp in September from a leading power supply maker that is a key supplier to the same A.I./graphics customer.”

Consumer

Consumer revenue declined (35.5%) YoY but rebounded 19.7% QoQ, as the segment looks to have marked a bottom in Q2, with revenue accelerating from ~$23 million quarterly to more than $28 million quarterly. Management said that it “is now clear that the inventory correction in gaming is behind us and a seasonal build is underway.”

Looking ahead to Q1, management expects “low double-digits sequential growth in the consumer segment driven by strong seasonal pick-up from wearables and continued strength in gaming, offset by slower home appliances.”

Communications

Communications revenue rose 59% YoY and 2.1% QoQ, driven by the “seasonal pick-up from a Tier 1 U.S. smartphone customer, offset by sequential declines from Korea and China OEMs.” This was ahead of expectations for flat QoQ growth.

For Q1, management is expecting QoQ growth to accelerate to the double-digits on “seasonal strength ahead of new smartphone launches in the US and increasing demand from China smartphone OEMs.” Management added that they see two tailwinds for growth, one from rising mix of premium phones, and the other from increasing BOM content in smartphones, stemming from increased battery charging currents.

Power Supply and Industrial

Power supply and industrial revenue declined (33.7%) YoY but increased 11.3% QoQ. Management said this was ahead of expectations for mid- to high-single digit QoQ growth due to “strength in the e-mobility segment for e-bikes and e-scooters and DC fans for applications in areas such as datacenters.”

For Q1, QoQ growth is expected to accelerate to 15% to 20%, driven by quick chargers and “strength from AC-DC power supplies tied to the seasonal build in PCs.”

Earnings Call:

AI Accelerator Cards

Per an announcement in June, AOSL is expected to become a new GB200 supplier for Nvidia and to start shipments by the end of this year. Although AOSL does not call out Nvidia specifically on the call, the announcement helps us infer that it’s likely Nvidia being spoken about on the calls. The only other option is that it’s AMD.

AOSL sells MOSFETs that go into bus converters for DC to DC power conversion. AOSL works with a leading power supply company that, in turn, supplies the unnamed AI/graphic customer plus the company supplies the unnamed AI/graphic company directly. AOSL supplies the MOSFETs powering each GPU, and for the upcoming platform of this customer, AOSL will additionally sell the total solution controller and the multiphase controller. The number of MOSFETs is expected to triple from 9 to 16 up to 50 MOSFETs per GPU.

Here was some discussion on the call about this opportunity:

“David Williams

With that, I guess, Stephen, I wanted to ask — yes, I wanted to ask a little bit just on the graphics card and some of the datacenter accelerator and GPUs. We've talked about this before and starting to see those revenues, but I'm trying to understand what is it the magnitude of maybe that could be over time? And maybe is there a way to size, understanding there's different flavors or varieties of those products, but is there a good way to think about what your content can be and maybe where you're at within that maybe qualification process? Any color around that would be, I think, incredibly helpful. Thank you.

Stephen Chang

Sure. Yes. So an entry into artificial intelligence programs, a lot of it actually is built upon where we have already been with our graphics cards. And accelerator cards actually aren't that different from a graphics card in the sense that you are basically powering a high-performance GPU in both cases. And — but with datacenter, that performance requirements are being driven even higher.

[…] . So to quantify some of that, so for example, in graphics card, you can have anywhere from 9 to 16, something in that range of number of driver [MOSFETs] per GPU. But when you move to an A.I. accelerator card, that number actually jumps up to even up to 50 power stages to power that GPU.

And those are the solutions that we are shipping today in our graphics card/AI customer in their existing platforms. And we are working with them on being — on transitioning over to the new platform that they will be launching soon.”

-End Quote

There was also discussion on the call that AOSL’s opportunity is expanding with the new platform as the company will be integrated from the ground-up. The CEO stated they are engaged at the manufacturing process and engaged with the power supplier for the OEM.

Conclusion:

It was interesting the PC opportunity was not discussed in the call today, rather the call was almost entirely about the unnamed AI/graphics company. Perhaps this is because the June announcement helps to expand AOSL’s immediate opportunity considering the new platform for this unnamed customer is expected to ramp in H2 and into next year.

There is a contract expiring in early 2025 that we plan to go back and dig up information on this 24-month contract to make sure it’s not presenting an outsized risk. The contract, signed in February 2023, licensed AOSL's SiC tech and provided engineering and development services for 24 months, with a total fee of $45 million. $18 million was paid upfront by the customer, with AOSL also receiving $6.8 million in March and July 2023, with the remaining balance to be paid upon achieving certain milestones. Given that AOSL has already collected a majority of the contract's revenue, it does not look to present outsized risk here.

On the topic of risks, there is high customer concentration with AOSL as there is with most small cap semiconductor companies. For example, last quarter, the top two customers accounted for 71% of revenue.

AOSL is a stock where we are using technical analysis to its fullest. Our goal is to participate in opportunities where small cap suppliers expand their serviceable addressable market (SAM) as AI systems increase in complexity. However, the only reasonable way to do so is to adhere to our stops. You can expect updates from Knox as we go along.

Damien Robbins and Beth Kindig, Equity Analysts for the I/O Fund, contributed to this analysis

Recommended Reading:

Alpha & Omega Semiconductor Earnings Preview: Signs of recovery ahead?

Posted on by io-fund

Alpha & Omega Semiconductor will announce its Q4 FY2024 results on August 07th. The midpoint of the management revenue guide for FQ4 is $160 million, representing a YoY decline of (-0.9%). Analysts expect revenue growth to resume from the December quarter.

Management has been quite transparent throughout the inventory corrections in the PC space in 2022 and the growth opportunities it sees in the computing segment, not just in PCs but also extending to graphics and AI accelerators. Management is seeing signs of a gradual rebound in the coming quarters and is approaching a recovery phase of the next cycle. Until we see actual improvement in the fundamentals, we would treat the company as a momentum play.

Revenue

Analysts expect FQ4 revenue to decline by (-0.9%) YoY to $160.03 million. Revenue will further decelerate to (-3.5%) YoY to $174.37 million and then return to growth, accelerating to 4.2% and 11.9% in the subsequent quarters.

FQ3 revenue grew by 13.2% YoY to $150.1 million, in line with analyst estimates. Management expanded on seasonality, saying that while the “March quarter is historically our seasonally lowest revenue quarter due to the technicality of consumer spending, the year-over-year growth indicated the strength of our recovery from the inventory corrections.”

Margins

  • Gross margin improved 50 bps YoY but was down 290 bps sequentially to 23.7%. Adjusted gross margin improved 10 bps YoY but was down 280 bps sequentially to 25.2%. Management attributed the sequential decline to lower utilization and ASP, partially offset by a better mix.
  • Management guide for the next quarter is 26.3% at the midpoint. Over the longer term, management remains optimistic about a return to 30% adjusted gross margin as it works towards hitting its $1 billion revenue goal:

Q, Analyst Craig Ellis: “What are some of the bigger gives and takes that we should be aware of for gross margin and really the pace of expansion and what do you need to see to be confident that gross margins can move back to that 30% level and then at some point higher? Thank you.

A, CFO Yifan Lang: At this point and I mean we still think on our mid-term target model and when we reach the $1 billion in revenue, we expect to get to 30% gross margin on the non-GAAP basis level.”

  • Operating margin improved 390 bps YoY and declined sequentially by 630 bps to (-7.0%). Management guide for the next quarter is (-5.2%). Adjusted operating margin improved 150 bps YoY and declined by 580 bps sequentially to (-0.7%). Management guide for the next quarter is 1.6%. Higher payroll taxes at the start of the year also negatively impacted the operating margins.
  • Net loss came at (-$11.2 million) or (-7.5%) of revenue compared to (-$18.9 million) or (-14.3%) of revenue in the same period last year. Adjusted net loss came at (-$1.2 million) or (-0.8%) of revenue compared to (-$5.9 million) or (-4.4%) of revenue last year. GAAP loss per share came at (-$0.39) and beat estimates by 18.2% and adjusted loss per share came at (-$0.04) and beat estimates by 71.4%.
  • The analysts expect adjusted EPS to decline (-77.2%) YoY to $0.04 in FQ4.

Cash Flow and Balance Sheet

  • Operating cash flow was $28.2 million or 18.8% of revenue compared to $11.6 million or 8.8% of revenue in the same period last year. The operating cash flow included $9.9 million in repayments of customer deposits. Management expects to refund about $4.5 million of customer deposits in the June quarter.
  • Free cash flow was $20.5 million or 13.7% of revenue compared to (-$11 million) or (-8.3%) of revenue last year. Capex was $7.4 million compared to $22.7 million in the same period last year. Capex was higher last year due to the Oregon fab expansion. Management guide for FQ4 capex is in the range of $6 million to $8 million.
  • The company repurchased 287,000 shares of employee restricted stock units vested for $6.7 million.
  • The company had cash of $174.4 million and debt of $41.2 million compared to $162.3 million and $44.1 million in the previous quarter.

Key Segments

Computing

Computing segment revenue grew by 80.4% YoY and declined by (-4.3%) sequentially to $68.7 million. It was the largest segment and represented 45.8% of the total revenue. The growth was led by graphics cards, tablets, and AI accelerators and partially offset by seasonal decline in notebooks. Management expects the computing segment to grow in the mid to higher single digits with continued strength in graphic cards, AI accelerators and tablets.

Consumer

The consumer segment declined by (-47.1%) YoY and up 0.3% sequentially to $23.6 million. The inventory correction in gaming continued in the March quarter. However, the rate of decline slowed, and management expects double-digit sequential growth in the segment, which will be helped by the end of the inventory correction in gaming.

Communication

The communication segment grew by 39.2% YoY and declined (-7.4%) sequentially to $26.8 million. Strong sales from the Korea and China-based OEMs were offset by a seasonal decline in shipments to the Tier 1 U.S. smartphone customers and a slowdown in networking. Management expects a strong sequential rebound in shipments to Tier 1 U.S. smartphone customers as they prepare for the fall launch, while they expect a sequential decline from Korea and China OEMs.

Management said “even with a sequential decline, our China OEM business remains strong and up significantly year-over-year. Overall, we estimate the Communication segment will be flat sequentially in the June quarter, which is notably higher year-over-year, because of our BOM content and market share increases.”

Power Supply and Industrial

The Power Supply and Industrial segment revenue declined by (-6.5%) YoY and (-29%) sequentially QoQ to $24.8 million on continued inventory corrections in quick chargers in addition to sequential declines in AC-DC power supplies, power tools and solar. Looking ahead to the June quarter, management expects sequential growth in the mid-to high-single digits as the quick charger inventory correction ends alongside strength in e-mobility.

Other Key Points to Watch

AI PCs and AI Mobile

In our recent deep dive here, we discussed that we will be closely monitoring the Computing segment that is expected to benefit from the strong demand for AI PCs.

We had noted, “Moving through the rest of 2024 and into 2025, Computing segment’s growth is likely to be driven by two primary factors – AI PC growth aided by Intel’s Meteor Lake platforms and AMD’s FP8 and FP11 next-gen chips, as well as a rather robust server and AI accelerator point-of-load (POL) roadmap.

For example, AOSL’s power delivery content is increasing significantly on Intel’s Meteor Lake chips, compared to the Arrow Lake predecessor, which should translate into a healthy BOM increase. Intel is targeting a significant 8x increase in Core Ultra AI PC chip shipments (across its Arrow Lake, Meteor Lake and Lunar Lake platforms) — from 5 million since December 2023 to at least 40 million by year-end 2024. This is estimated to grow 50% in 2025 to 60 million Core Ultra chip shipments.”

Intel CEO Patrick Gelsinger said in the recent Q2 earnings call that they shipped 15 million AI PCs since the Meteor Lake launch in December. “We have now shipped more than 15 million Windows AI PCs since our December launch, multiples more than all of our competitors combined, and we remain on track to ship more than 40 million AI PCs by year end, and over 100 million accumulative by the end of 2025.” They also mentioned Lunar Lake, the next generation AI PC, achieved production release ahead of schedule in July. Arrow Lake will follow it. Next, Panther Lake is expected in the second half of 2025.

“The AI PC will grow from less than 10% of the market today to greater than 50% in 2026. We know today's investments will accelerate and extend our leadership and drive significant benefits in the years to come. Our efforts will culminate with the introduction of Panther Lake in the second half of 2025.”

Lisa Su, CEO of AMD, provided positive insights into the Client segment in the recent earnings call “Our view of this is the AI PC is an important add to the overall PC category. As we go into the second half of the year, I think we have better seasonality in general, and we think we can do, let us call it above-typical seasonality, given the strength of our product launches and when we are launching. And then into 2025, you're going to see AI PCs across sort of a larger set of price points which will also open up more opportunities.”

Outlook

Management mentioned that they saw signs of a gradual rebound in the coming quarters and are approaching a recovery phase of the next cycle. We would closely watch management comments on the recovery of the PC and smartphone market. The CEO mentioned in the last earnings call, “Looking beyond, we anticipate the second half of this year will be stronger than the first half as customers gear up for new product launches in smartphones as well as PCs.

Looking beyond 2024 to the growth phase of the next cycle, AOS is transitioning from a component supplier to become a comprehensive solution provider, enabling us to go deeper with increasing BOM content and penetrating new products and verticals.”

GB200 Supplier

According to the analyst Ming-Chi Kuo, Alpha and Omega Semiconductor, a current supplier of MOSFETs and power ICs for HGX/DGX H100, is expected to become the new GB200 supplier. If the news is confirmed, it could be another positive catalyst for the stock.

Valuation

The P/S ratio peaked in March 2022 when the quarterly revenue was close to $200 million, GAAP profitable, and before the PC inventory correction. The company is currently trading at a P/S ratio of 1.5 and a forward P/S ratio of 1.4 and higher than the average P/S ratio of 1.2.

Conclusion

AI PCs may emerge as one of the stronger growth trends in consumer AI devices over the next few years. AOSL’s tie-ins to Intel’s Meteor Lake CPU and its position in the development stages for AMD’s FP11 platform position it for strong growth in this space. The company is also seeing signs of recovery in its segments. However, until we see actual improvement in the fundamentals, we would treat the company as a momentum play.

Royston Roche, Equity Analyst at the I/O Fund, contributed to this article.

Recommended Reading:

Lam Research FQ4 Earnings: Margins Recover Yet DRAM Declines

Posted on by io-fund

Lam Research followed in the footsteps of WFE peer KLA with a top and bottom line beat in fiscal Q4, while guiding fiscal Q1 revenue slightly above consensus. Margins rebounded in fiscal Q4 with management forecasting fiscal Q1 GAAP operating margin to expand from Q4’s print. China’s revenue contribution dipping sequentially, which was a main focus on the call.

Lam slightly boosted its WFE outlook for calendar 2024, now seeing WFE spending in the mid-$90 billions, compared to its view for the low to mid-$90 billions in the March quarter. Management said the increased view was primarily driven by domestic China shipments, though recent news with the US targeting more restrictions on China’s ability to access HBM may play a negative role down the line, if implemented.

Lam sees foundry, logic, DRAM and NAND to all rise YoY in calendar 2024. Management noted that memory’s WFE will be “biased” towards tech upgrades, meaning the recovery still may not yet be fully under way as memory’s contribution dipped QoQ.

Overall, we are likely to close Lam as we are not comfortable with high exposure to China, combined with the decline in DRAM.

Revenue and EPS:

Lam returned to topline growth in its fiscal Q4, reporting YoY revenue growth of 20.6%, though QoQ growth was marginal, at just 2.1% QoQ. Lam’s fiscal Q1 guide points to revenue growth in the high-teens, or approximately 16.4% YoY at midpoint.

  • Q4 revenue of $3.87 billion beat estimates by $40 million, with Lam reporting YoY revenue growth after five consecutive quarters of declines.
  • FY24 revenue was $14.91 billion, down (14.5%) YoY.
  • For fiscal Q1 (September 2024 quarter), Lam guided for revenue of $4.05 billion, +/- $300 million, for YoY growth of 16.4% and QoQ growth of 4.6%. This was slightly ahead of analyst estimates for $4.02 billion.
  • This quarter, GAAP EPS of $7.78 beat estimates by $0.42, representing YoY growth of more than 30% and QoQ growth of 6%.
  • Adjusted EPS of $8.14 beat estimates by $0.55, representing YoY growth of more than 36% and QoQ growth of nearly 4.5%.
  • FY24 GAAP EPS was $29.00, down (12.7%) YoY.
  • For Q1, Lam guided GAAP EPS of $7.79, +/- $0.75, for approximately flat QoQ growth and 17% YoY growth.
  • Adjusted EPS was guided at $8.00, +/- $0.75, for a QoQ decline of (1.7%) but YoY growth of nearly 17%.

Margins:

Lam’s margins expanded sequentially down the line, as operating leverage improved while gross margins remained flat QoQ. This came despite China’s revenue contributing shrinking 300 bp to 39%, with management previously citing China customer mix as a gross margin tailwind. As stated, there were many questions about China’s revenue being smaller this quarter.

  • GAAP gross margin was 47.5% in Q4, flat QoQ but up 200 bp YoY. Management guided to a gross margin of 47% next quarter, and an incremental headwind in December.
  • Adjusted gross margin was 48.5% in Q4, down 20 bp QoQ but up 280 bp YoY. Both figures came at the high end of management’s guided range for the quarter.
  • For Q1, GAAP gross margin was guided at 46.9%, for a YoY and QoQ contraction of 60 bp. Adjusted gross margin was guided at 47.0%, a 150 bp QoQ and 90 bp YoY contraction.
  • GAAP operating margin was 29.1%, a 120 bp QoQ and 250 bp YoY expansion. 
  • Adjusted operating margin was 30.7%, a 40 bp QoQ and 340 bp YoY expansion — this was the highest adjusted operating margin since Q2 2023.
  • For Q1, GAAP operating margin was guided to be 29.4%, for a QoQ expansion of 30 bp while remaining flat YoY. Adjusted operating margin was guided to be 29.5%, a 120 bp QoQ and 60 bp YoY contraction at midpoint.
  • GAAP net margin was 26.3%, up 80 bp QoQ and 130 bp YoY, and reaching the highest level since Q2 2023, as improved operating margin aided bottom line growth. Adjusted net margin was 27.6%, up 60 bp QoQ and 260 bp YoY.

For the full year, GAAP gross margin was 47.3%, up from 44.6% in FY 2023; adjusted gross margin was 48.2%, up from 45.3% in FY 2023 as gross margins remained strong on China customer mix.

FY 24’s GAAP operating margin contracted 110 bp to 28.6%, as FY 23’s first half strength more than offset back half weakness.

Cash and Debt:

  • Operating cash flow was $862.4 million in Q4, for a margin of 22.3%. OCF declined (23.2%) YoY, with the quarter seeing a ($260 million) detrimental impact from changes in operating assets and liabilities.
  • Free cash flow was $761.7 million, for a margin of 19.7%, a nearly 1300 bp QoQ contraction, primarily as a result of the decline in OCF.
  • Inventory was $4.22 billion, down slightly from $4.32 billion in the prior quarter.
  • Cash and equivalents totaled $5.85 billion.
  • Debt totaled $4.98 billion.

Key Metrics:

Interestingly, Lam reported DRAM and NVM revenue contribution shrinking QoQ, while logic/other revenue surged QoQ.

As a percentage of systems revenue, which was $2.17 billion:

  • Foundry accounted for 43% of revenue, down from 44% last quarter.
  • DRAM accounted for 19% of revenue, down from 23% last quarter. Two quarters ago, DRAM was 31% of revenue. 
  • NVM accounted for 17% of revenue, down from 21% last quarter.
  • Logic/other accounted for 21% of revenue, up from 12% last quarter.

Combined, memory accounted for 36% of systems revenue, down from 44% last quarter.

Geographically, Lam saw China revenues moderate QoQ, while Taiwan and US revenue contribution rose substantially QoQ.

  • China accounted for 39% of revenue in Q4, down from 42% last quarter, and also reaching its lowest share of revenue in fiscal 2024.
  • Korea accounted for 18% of revenue, down from 24% last quarter, potentially as one of the leading factors in memory’s weaker revenue contribution on a QoQ basis.
  • Taiwan accounted for 15% of revenue, jumping from 9% last quarter.
  • The US accounted for 10% of revenue, rising from 6% last quarter.

Earnings Call:

DRAM Decline is Odd

The decline in DRAM is a concern as demand is surging for HBM3 and HBM3e, driven by memory being the most critical component in the upcoming generation of GPUs and ASICs. We expected more from Lam in this report considering HBM3 and HBM3e is the catalyst for memory stocks Micron, Samsung and SK Hynix, who Lam supplies.

Management stated that “The decline in the memory segment was mainly attributable to DRAM. DRAM came in at 19% of systems revenue compared with 23% in the March quarter as investments in mature nodes declined in the June quarter.” Theoretically, this decline in mature nodes should be offset by growth in HBM.

Per a previous Lam analysis in April Lam was expecting DRAM to triple YoY: “Lam is expecting its “HBM-related DRAM and packaging shipments to more than triple year-on-year and outpace WFE growth in this segment by a significant margin” in 2024.

CEO Timothy Archer added that in HBM, Lam is seeing “very, very strong demand. I think that whether or not at some point, it's shipping above peak, I think that this AI market is continuing to evolve at a very, very fast rate. And all we're focused on right now is ensuring we are building out our own capacity and capabilities. And ensuring that we maintain that technology leadership that's allowing us to hold 100% market share of the TSV formation in HBM.”

We were not the only ones expecting more from the DRAM segment, as one of the first questions on the call asked for more discussion on why DRAM isn’t showing up in the report. The answer was not clear, rather answering with what their technology can solve rather than addressing what is likely a competitive issue, where another supplier is taking the business (my best guess).

Question
Timothy Arcuri (Analysts)

[…] Can you just talk about — I know your leverage to the advanced packaging part of the HBM dollars being spent but that's still a pretty small piece of it. So can you just maybe give a chance to kind of discuss some of the view that you're not very levered to DRAM and give us a sense of maybe where you're investing and where you think you can gain share in DRAM.”

Answer
Timothy Archer (Executives)

[…] The other side of it, a lot of the excitement around DRAM is related to HBM. And there, as you commented, we play extremely well with our strong position in both TSV, etch as well as the TSV electroplating. And I think that we don't see any change in that strong position going forward. So we get the benefit both from the scaling and architectural changes that are occurring in DRAM going forward and from the advanced packaging and HBM related expansion. And all of these — on both of those sides are multiplied by the fact that you get fewer bits per wafer. And so everybody recognizes you're going to need a lot more DRAM wafers processed going forward. And ultimately, that translates into more equipment from LAM.

-End Quote

Another analyst asked for clarification on if DRAM should “at some point, come back strongly.” The CEO stated that “we see DRAM demand for DRAM equipment continuing to grow through 2025 and probably well beyond that.”

China is the Opportunity; but also a Major Risk

As stated, China revenue was at 39% down from 42% last quarter. To help illustrate the importance of China to Lam, the word was stated 33 times in the earnings call compared to HBM being stated 14 times. Not only is China a risk politically, but the revenue can be lumpy for Lam.

Source: Lam’s Quarterly Earnings Slides

There was a solid question on the call that helps investors understand the risk in terms of losing a customer due to restrictions, and why Lam may be lagging some of its peers

Question
Atif Malik (Analysts)

Doug, if I look at the 2023 year-over-year China sales growth among the big 5 equipment makers. All of them are up quite well. ASML is up like 250% and the U.S. peers are up in teens or 20%, but you guys were down 11% total China sales in 2023. And this year, you're expecting China sales to be up. So I'm just trying to understand the dynamics last year. Were this just a function of maybe NAND spending and the NAND project not being active or are there competitive elements in China that are working against you?

Answer
Douglas Bettinger (Executives)

Atif, I'll remind you that perhaps our largest customer got restricted when the regulations came out, our NAND customer in China. That customer was pretty strong in '22, went away in '23. So the year-over-year comparisons you're making, you've got to factor that in. And then the strength we're seeing '23 to '24 is a different mix entirely, really not any NAND in China to speak of, at least not domestic China. I don't know if that helps you, but make sure you're thinking about that.

Conclusion:

The decline in DRAM could be short-lived, to where the segment bounces back quickly next quarter. However, being in the midst of such strong HBM growth, it feels odd to see DRAM decline in light of the strong commentary we saw in the first half of the year.

In addition, Lam’s exposure to China is problematic. Theoretically, it would be harder for Lam to replace China revenue than a memory company that is sold out of supply 6-8 quarters out, or a GPU design company that is also seeing outsized demand. For stocks that have China risk, we’d rather own TSMC, for example, or more of Micron if we seek exposure to memory, or even Nvidia for the clear capex raise we got from Microsoft and Meta.

Our plan is to close Lam and to give the I/O Fund team the task of re-allocating it to a stock with less risk, and with its AI-related segments reporting more growth.

Damien Robbins, Equity Analyst for the I/O Fund, contributed to this analysis.

Recommended Reading: