JHVEPhoto/iStock Editorial via Getty Images
Investment Thesis
As predicted (Intel: Tough 2023 Ahead), Intel (NASDAQ:INTC) is going to have a tough 2023, but it is going to be even tougher than expected. The confluence of the 7nm (now called Intel 4) delay and the economic downturn is clearly visible.
Nevertheless, the long-term investment thesis is arguably unchanged, as investing during a downturn could lead to strong returns down the road, in this case in preparation for the next upturn in conjunction with Intel’s return to tech leadership.
Q4 results, guidance, and discussion
Despite Intel saying it is going to be even more aggressive in pursuing its cost reduction, the Q4 employee count was actually flat from Q3 at ~132k, up over 20k since Pat Gelsinger became CEO.
Nevertheless, Intel is taking real actions in pruning the portfolio. After axing Optane last year, the next major business is the Barefoot Networks switching business, another casualty of Intel’s last decade strategy to become a data-centric company/conglomerate. Intel says the 7 units it has closed have saved over $1.5B.
Intel also expects to do a second SCIP (the first one was the $15B Brookfield deal).
PC
While Intel continues to maintain there could be an improvement in the second half, there does not seem to be a lot of evidence. Despite the very weak guidance for Q1, Intel continues to expect a PC TAM of over 270M units.
As Intel has been arguing since Q2, a big part of the financial underperformance has been the inventory correction, in Q2 described as a once in a decade event. At that time, a recovery was still expected by Q4. Instead, Intel said the under-shipping was even higher in Q4, and will be even higher still in Q1. For 2022, Intel’s sell-in was 10% below consumption.
Nevertheless, Intel said it saw its highest ASPs ever in Q4, and also grew market share in H2 2022, with further gains expected in 2023.
Intel said it has taped out the first Lunar Lake silicon, with production readiness (PRQ) in 2024. This is a bit puzzling since, as the successor of Arrow Lake, it was expected to be produced on Intel 18A, which suggests the test chips Intel was talking about for 18A (see below) includes one for Lunar Lake.
DCAI
Intel said there was a modest TAM growth in 2022, but this was offset by inventory burn, resulting in a modest shipment decline. For Intel, there were also market share declines. Intel expects Q1 and H1 server TAM to decline at an “accelerated rate”.
Intel said it expects to ship 1Mu of Sapphire Rapids by the middle of the year, which is a pace that seems roughly in line with Ice Lake. Emerald Rapids is still on track for H2 launch.
Other businesses
Intel saw record performance in 2022 for PSG (FPGAs/Altera), NEX, IFS and Mobileye. Mobileye in particular had a very strong Q4 with ~60% growth. PSG also grew 42%. Intel said it added a leading cloud solution provider for Intel 3. Intel also said it is now shipping the 0.5 PDK of 18A, which seems in line with what is required for a ramp in 2025.
Intel expects continued growth in these businesses, which compares to an overall semiconductor market that is expected to decline. Unfortunately, the issue is that these remain just a small part of the overall business, which is dominated by CCG and DCAI.
For AXG, Intel cited the launch of Sapphire Rapids with HBM (which isn’t even a GPU) for the sequential recovery.
Pat Gelsinger: “And I’d say all the products that we launched out of AXG, the Flex product line, the discrete graphics product line, the MAX product lines. All of those products are continuing forward, and we believe all of those will have strong ramps in their volumes, revenues and market impact as we go through the year.”
For IFS: “IFS achieved record quarterly revenue of $319 million, up 87% sequentially and 30% year-over-year on increased automotive shipments.”
Process technology
Intel provided following update, roughly in line with expectations.
Our progress against our TV road map continue to improve throughout calendar year 2022 and every quarter, our confidence grows. We are at/or ahead of our goal of 5 nodes in four years. Intel 7 is now in high-volume manufacturing for both client and server. On Intel 4, we are ready today for manufacturing, and we look forward to the Meteor Lake ramp in second half of the year. Intel 3 continues to show great health and is on track.
Intel 4 and 3 are our first nodes deploying EUV, and will represent a major step forward in terms of transistor performance per watt and density. On Intel 20A and Intel 18A, the first nodes to benefit from RibbonFETs and PowerVia, internal test chips and those of a major potential foundry customer have taped out with the silicon running in the fab. We continue to be on track to regain transistor performance and power performance leadership by 2025.
Nevertheless, some criticism is warranted. Some red flags around Meteor Lake became clear last year, and this has now been confirmed with Meteor Lake ramping in H2 2023, despite Intel simultaneously saying that the node is currently manufacturing-ready.
No analyst questions were asking about how to square this circle, nor why Intel is moving forward with a Raptor Lake Refresh. Supposedly, the reason for the delayed ramp is simply because Meteor Lake has become a H2 2023 product, given the annual roadmap cadence. Still, a similar comment could be made about Granite Rapids and Sierra Forest. With Intel 3 being manufacturing-ready in H2, this implies a launch should be possible in H1, which remains to be seen.
Financials
Revenue of $14B reset Intel back to levels last seen almost a decade ago. Gross margin was just 44%, impacted by high fixed costs and underloading charges. Notably, on the opex side, as evidenced by adding over 20k employees, R&D has approached the $18B run rate, over $4B more than pre-Pat Gelsinger levels. This is apparently despite exiting those 7 businesses.
Intel highlighted how the -$4B FCF of 2022 was within Q3 guide despite $3B of capital initiatives that shifted from Q4 to 2023. The FCF number was also relatively close to the initial guide of negative $1-2B, despite that the revenue was reduced from the $76B outlook to $63B. Intel cited working capital improvements besides spending reductions.
Guidance
The “highlight” of the earnings release was the $11B guidance. Intel managed to deliver a guidance miss for the ages in Q2… and managed to do it again (despite the already sharply reduced expectations) just two quarters later.
Nevertheless, to spare the bottom line, besides the $3B spending reduction plan, Intel is also doing some financial engineering (again), by extending its equipment depreciation from 5 to 8 years. While many observers will likely doubt that this is a coincidence, obviously the IFS business and even the chiplet strategy does extend how long a node will be in production (although arguably this was already the case anyway given the node delays).
Other notable items include the 39% gross margin guide and 30% tax rate.
Investor Takeaway
The downturn, which since Q2 has been exacerbated by the inventory correction, is more significant than Intel or most people likely had expected, and that is reflected in the results and guidance.
With regards to the investment opportunity, the stock has been adjusting in the last half a year to account for this new financial reality. Nevertheless, just like how the EPS has been able to contract by 80% or so due to lower revenue, similarly during the next upturn nothing prevents the EPS from rising by 5x again.
As such, with Intel’s investment period lasting through 2024, the stock remains a buy at its current deflated levels, on the expectation to start harvesting in 2025 onwards based on catalysts that include renewed process and data center leadership, and emerging foundry, graphics and Mobileye businesses providing additional growth.
Be the first to comment