July 28, 2022
Press Release Commentary and Investor Presentation
Below are the points from the press release and investor presentation, outside the financial statements, that I found notable:
Press Release
• Revenue was down 22% year over year, to $15.3B and the company lost money on a GAAP basis, $(0.11) per share. Non-GAAP EPS was $0.29
• They cited adverse market conditions for the sharp drop in their Client Computing, Datacenter, and AI Groups. Their Network and Edge Group and Mobileye had record quarterly revenue, so whatever is wrong with Intel isn’t affecting every division.
• Large down revision of full year 2022 revenue guidance, from $76B to a range of $65B to $68B. They reiterated full year adjusted free cash flow guidance. This means either they are adjusting out whatever caused the revenue to drop, which seems unlikely, or they are cutting capital expenditures in 2022 to keep free cash flow whole, offsetting declining sales.
• From CEO Gelsinger, “This quarter’s results were below the standards we have set for the company and our shareholders. We must and will do better. The sudden and rapid decline in economic activity was the largest driver, but the shortfall also reflects our own execution issues…” It is positive that Gelsinger didn’t put it all on external factors, but his acknowledgement of their own execution issues is a departure from his past rhetoric which was almost universally positive, and over-the-top at times.
• “We are taking necessary actions to manage through the current environment, including accelerating the deployment of our smart capital strategy, while reiterating our prior full-year adjusted free cash flow guidance and returning gross margins to our target range by the fourth quarter,” said David Zinsner, Intel CFO. "We remain fully committed to our business strategy, the long-term financial model communicated at our investor meeting and a strong and growing dividend.” It seems a stretch that they will return gross margins to their target range in two quarters, after such a sharp drop. Their commitment to the plan laid out a few months ago doesn’t carry a lot of weight given they have stumbled so early in that multi-year roadmap.
• There will be more detail below on segment results, but they are mostly bad. CCG and DCAI, most of Intel’s revenue, were down 25% and 16%, respectively, year over year. Intel Foundry Services (IFS) was down 54% from a year ago. This is stunning since this business is just beginning.
Investor Presentation
• Reasons for the revenue and profit impact, in order given; deteriorating macro, inventory and supply chain disruptions, competitive pressures. I suspect competitive pressures were a major factor.
• The company claims Intel 4 will be ready for production in 2H 2022. They also state that Intel 3, 20A, 18A are on or ahead of schedule.
• The IFS partnership with MediaTek to make chips for smart edge devices is a sad statement, that this is the best product they could highlight for IFS. MediaTek is a second-tier company and smart edge devices are relatively simple compared to CPUs for servers, PCs and phones.
• They are seeing a near-term cyclical slowdown. 2022 PC TAM will be down 10%. They cite supply chain issues and customers reducing inventory.
• Intel massively missed their Q2 guidance; $15.3B in revenue was $2.7B below, 44.8% gross margin was 620 bps below, and $0.29 EPS was $0.41 (!!) below
• Operating income in their DCAI bus9iness was down 90% from a year ago, to $200M, or 5%. It was $2.1B at a 38% margin last year. OEM inventory reductions, mix-driven ASP declines, and competitive pressures were cited as the causes.
• While Network and Edge Group revenue increased 11%, the operating margin here was cut by more than half ($605M to $241M) to a margin of 10%, from 29%. Product mix, inventory sell-through, and investments in the future were cited as the reasons.
• FY 2022 outlook: revenue of $65B to $68B (from $76B last quarter), gross margin of 49% (from 52% last quarter), EPS of $2.30 (from $3.60 last quarter). Net CapEx will be $23B and adjusted free cash flow will be negative $1B to negative $2B.
• Q3 2022 outlook: revenue of $15B – $16B, gross margin of 46.5%, EPS of $0.35
Financial Statement
Income Statement
Net revenue in the second quarter was $15.3B, down from $18.4B sequentially. Q2 of 2021 was $19.6B, so year on year revenue declined 22%. A list of macro reasons were given for lower sales (inflation, COVID-lockdowns in China, war in Ukraine, etc.) CCG saw lower desktop and notebook unit volume, though notebook ASPs were higher as the sales volume drop was concentrated at the lower end of the market. Server ASPs were lower because of mix and more competition. Somehow it took more COS to generate this lower level of revenue, compressing gross margin to $5.59B from $11.2B a year ago. This is a GAAP gross margin of 36.5%. I didn’t think I’d see a 3-handle on an Intel gross margin, even in the midst of a turnaround. Among the list of reasons given for the lower margins, it is notable that the company decided to end their Optane business and write-down the inventory. This is the disruptive memory technology co-developed with Micron Technology. Intel also increased inventory reserves and saw higher unit costs, both of which hurt gross margin. As Intel is in the midst of accelerating their process node transitions, developing a foundry business, and revitalizing their product portfolio, R&D expenses were 18% higher than a year ago, at $4.4B. They also spent $201M more on MG&A, a total of $1.8B. These higher overhead costs, which are necessary to execute Intel’s turnaround plan, led to an operating loss of $(700M). Intel has a negative GAAP operating margin. Stunning. The draw down in equity markets caused a $(90M) mark-down in investments. After interest expenses, the company lost $(909M). They had taxes provisioned from earlier to offset half of this, bringing the quarterly loss to $(454B), or $(0.11) per diluted share. Share count rose slightly from a year ago, to 4.1B.
Statement of Cash Flows
For the first six months of the year, Intel’s free cash flow is negative. Really negative. $7.1B negative. As will be a theme for the coming years, capital expenditures (CapEx) were more than twice depreciation and amortization, $11.8B compared to $5.5B. The reason for this is Intel’s construction of new fabs around the world, to establish their foundry business and grow their existing product portfolio. They are also transitioning five process nodes in four years. Technology migration is hugely capital intensive as well. Intel’s share-based compensation is considerable; $1.6B in the first six months of the year. Over the last seven years, Intel has reduced their share count by 20% at a cost of $44.1B in buybacks. 20% of the current market cap is $29B, so this has been a poor investment, judged today. Six months ago, when their market cap was $220B, it was about break-even. Buybacks and price appreciation offset the dilution of share-based compensation. I check this with companies that have significant equity grants to executives and employees because the cost of these are added back in to calculate operating cash flow. In the case where these grants are leading to dilution, operating cash flow will overstate the cash available to owners. This doesn’t matter much right now for Intel as free cash flow is negative, at $(7.1B), for the first six months of the year. Net income over this time was $7.65B, then $5.5B and $1.6B of D&A and share-based compensation were added back. $5.3B for investment gains and divestitures came off (Intel Ventures is large, active, and good at investing). Changes in short term assets and liabilities used $3.8B. Accounts receivable contributed $3.4B of cash in the last two quarters. Turning to investing activities, these were $2.5B negative in the first six months of the year, mostly because of $11.8B worth of capital expenditures. Purchase and sale of short-term investments was net neutral, though they churned $25B worth. Most of the CapEx was paid for by selling investments and from the divestiture of their NAND business to Hynix. These two items combined for $11.3B. Other investing consumed $1.7B, leading to the total of $(2.5B). Intel paid $1.7B in debt and $3B in dividends. It would be a great time for the company to be buying back their shares, if they could only afford it. They have bought back no stock in the last six months. In summary, selling off assets and investments in the last six months paid for Intel’s hefty CapEx bill. Operating cash flow was buoyed by other investments but still negative a billion dollars, and they paid off $1.7B in debt and paid out $3B in dividends. A long list of multi-billion-dollar line items netted out to a net change in Intel’s cash position of less than $500M, leaving them with $4.39B worth at the end of the period.
Balance Sheet
For the first half of the year, current assets are down $8B, to $50.6B. Cash, short-term investments, assets held for sale (their NAND business) and accounts receivable all shrunk, partially offset by a $1.4B increase in inventories and $3.1B in “other current assets.” All the capital expenditure happening to transform Intel increased PP&E from $63.2B to $71.7B. Current liabilities were flat in total and long-term debt was reduced by $1B, to $32.5B. Total debt is $35.4B, almost offset by $33B in cash and investments. Total shareholders’ equity is $101.2B, which is more than two-thirds of Intel’s current market cap. Measured by price-to-book, Intel has not been this low in eight years.
Divisions
Client Computing and Data Center AI, which in recent years have made up most of Intel’s revenue and more than 100% of their earnings, saw terrible results in this period. Client Computing in the quarter was down 25% from a year ago, to $7.7B. Both notebooks and desktop computers saw large unit declines (38% and 19%, respectively,) because of sales drops in the lower end education and consumer sub-segments. ASPs offset this somewhat because of richer mix in notebooks. ASPs in desktop were about flat, which may mean Intel is seeing pricing pressure at the higher end, from competition. Client Computing is the easiest place for Intel to be attacked by competitors, because it is the most price-sensitive part of the market. DCAI revenue declined 16.4% from the same quarter a year ago. Server revenue because of lower purchase volumes was given as the reason, though the unit drop of 12%, being lower than the total revenue decrease, shows that pricing was also an important factor. The company said mix shift to Cloud and competition were the reason for lower ASPs. Intel historically downplays the effect of competition on their pricing, but it is clearly having a major impact on the company. Network and Edge was up 10% from a year ago, to $2.3B. Accelerated Computing and Graphics was flat while Mobileye increased by more than a third. Both of these divisions are tiny compared to the overall company. Intel Foundry Services (IFS) revenue dropped by more than 50%, because of lower sales of tools for making lithography masks. This seems bad, because IFS should be making money selling wafers, not fab equipment. But at least the decline in these tool sales is the reason for the revenue drop, and not a drop in wafer revenue. IFS is still in the investment phase. The most stunning segment results are in operating income. Because of the amount of operating leverage in the semiconductor business when you own your fabs, Intel’s drop in Client and DC revenues hammered their operating income. For the quarter, Client was down 75% to $1.1B. Data Center was down 90%, to $214M. Every segment except Mobileye saw steep drops in operating income, leading to an overall operating loss of $(700M), compared to a $5.5B operating gain in the same quarter a year ago. Part of this is the higher investments being made now to transform the company, but that could be covering up poor performance. It is hard to discern between the two. My view, from experience with Intel’s people and culture, is their investing is probably inefficient and their execution is sloppy. They were so profitable for so long that their culture became wasteful and unfocused. The Intel people I’ve worked with had an arrogance that blinded them to the need to regularly improve their performance. This wasteful and ineffective culture is the most important aspect of Intel’s turnaround, and the hardest.
The End of Optane
Though it was not a surprise to most, Intel announced in the second quarter that they decided to end their Optane product line. They will support current customer orders but will end the product. As a result of this decision, they impaired $559M worth of inventory in the quarter. This has an effect of lowering gross margin in Q2 by 3.6%. Optane is a memory product co-developed with Micron, beginning more than ten years ago. Micron called it 3DXP and ended the product in March of 2021. The write-down of this much inventory means the growth in Intel’s Inventory – from $10.8B six months ago to $12.2B at the beginning of July – is larger than it first appears.
Earnings Call Notes
Comments are all in reference to the non-GAAP financial figures
Pat Gelsinger (CEO) prepared remarks
• The company sold their drone business this quarter. I didn’t know Intel owned a drone business. They have divested six businesses since Gelsinger returned to the company.
• Intel sees long-term semiconductor demand driven by what they call “The Four Superpowers.” These are ubiquitous computing, cloud-to-edge infrastructure, pervasive connectivity, and artificial intelligence. The framework is accurate, I believe, but what a terrible name. Gelsinger repeated his prediction that the total semiconductor market will be $1T by 2030.
• PC TAM by units will decline 10% in 2022. They are shipping below consumption, so customers are consuming inventory. They believe the total PC TAM is sustainably above 300M annual units.
• They see data center growth on a mid-teens percent CAGR
• The company is planning to lose data center market share in the next few years, as they redo their portfolio. I like their honesty here.
• Gelsinger went through a long list of how customers have used Intel’s products in ways that are not a PC or server
• Regulatory approval to buy Tower has been granted in the US and three other geographies. This is one more approval gained in the last three months. They expect this deal to close early in 2023. In last quarter’s call, they said they wanted to close the purchase “as soon as possible.”
• Reiterated they plan to IPO Mobileye later this year, pending market conditions
David Zinsner (CFO) prepared remarks
• The OEM inventory correction was much larger than expected
• Free cash flow in the quarter was negative $6.4B
• He stated the two biggest opportunities he sees for Intel to improve to be cutting programs that don’t align with their strategy and reducing costs in both development and production. The key he called out for cost is their five nodes in four years plan being executed successfully. He stated again that his history in memory lets him see many places where Intel can improve on cost.
• FY 2022 outlook: revenue of $76B (was $65-68B), GM of 49% (was 52%), EPS of $2,30 (was $3.60). Total net CapEx will be $23B and adjusted free cash flow is forecasted in the range of ($1-2B).
• Q3-22 guidance: revenue of $15-16B, GM of 46.5% at the midpoint, EPS of $0.35, and a tax rate of 13%. They believe Q2 and Q3 to be the financial bottom for the company. This is a bold statement given how off-guard they were caught by the drop that hammered them this quarter.
Question and Answer
• Their confidence in Q2-Q3 being the bottom is because they think customers are consuming inventory today and that can’t continue at this rate through the end of the year
• The inventory adjustments seen this quarter were once-in-a-decade in magnitude. They didn’t pre-announce because they didn’t have as much information as they wanted. Not a satisfying answer.
• Gelsinger detoured from one question to say they have reversed the brain drain out of Intel and their employee engagement survey saw the best ever year-on-year improvement. This could be because they were so low previously, but also likely the new strategic plan and major investments being made have re-engaged their workforce.
• They are highly optimistic for the level of capital offsets they plan to receive (read: government grants). Zinsner quoted they are now planning for 4x the level they were forecasting at the beginning of this year.
• Mid-30% capital intensity is their forecast during their investment phase, which will settle down to mid-20% once the turnaround investment is complete
• They are not planning on any CHIPS Act money in 2022. That cash will start coming in during 2023, a combination of grants and tax credits.
• Stacy Rasgon’s question identified that their delivery of Sapphire Rapids was three to six months late to plan. A fair partial excuse is this product didn’t benefit from all the changes that have taken place since Gelsinger’s return.
• closing question asked how outsiders can check that the five nodes in four years plan is on schedule. Gelsinger said the technical updates they give at technical conferences will do this, along with independent validation of new nodes. These things he said will give some indication, but many of the characteristics of successful node ramps (yield, cash cost, etc.) cannot be validated outside Intel.
Summary
This was a terrible quarter for Intel, with the company widely missing guidance. One analyst asked why they didn’t pre-announce, given how bad these results were, to which the executives didn’t have much of an answer. It looks to me that they were completely blindsided by this. They called out macroeconomic reasons for the slowdown in end demand, and as the reason many customers decided to start drawing down inventory. But somehow most of the underperformance was in Client Compute and DCAI. This seems more like competitive pressure as macro issues should affect all businesses. It surprised me that only one analyst mentioned competitive pressure in the Q&A. Maybe they all know management won’t give a straight answer on this. To his credit, Gelsinger did say they anticipate losing market share in the data center segment in the next two years as they revamp their portfolio. This quarter shouldn’t have surprised me the way it did. Maybe it is because it happened sooner than I expected. Any big turnaround has initial excitement, then reality hits. The job of changing the culture at Intel from the arrogance and lack of execution that came from decades of success will take several years. Challenging times bring out weakness, so this market softening exposed how deeply the problems run at Intel. I’m still going to buy shares in the company some time during the next year or two, but I’m in no hurry. I think there is more blood to be spilled here before performance improves. Adding on top of that a recession and I’m waiting and hoping for an even better share price. Valued relative to book, this is the cheapest Intel has been going back to at least 2010. I think the softness in the economy and the coming (it always does) downturn in overall semiconductors will give me a chance to buy Intel even cheaper than today.
-S. Hughes (no INTC position)