Intel: Massive 18A Foundry Prepayment

Summary:

• Intel has received a “significant” foundry prepayment for its 18A process node, signaling progress in its turnaround efforts as the very first tangible evidence of regaining process leadership.
• The company is accelerating the buildout of its new Arizona fabs, indicating that all pieces are coming together for a successful turnaround.
• The market is not giving Intel credit yet for these developments, as transforming from industry laggard to leader significantly changes the potential financial profile.

Investment Thesis

After nearly three years of having to hear that the roadmap was “on schedule” as the only sign of the turnaround progressing, Intel (INTC) finally delivered a bombshell moment by announcing that it has received a “large”, “significant” foundry prepayment for its 18A process node, which is the node it has long touted as returning to process leadership, and accordingly Intel is “accelerating” the buildout of its new Arizona fabs. This is the very first tangible evidence that Intel is successfully executing on the bull thesis.

Altogether, this implies all pieces are coming together for Intel to deliver on the turnaround thesis, which has the potential for significant investor returns as margins expand and revenue grows. As Intel is still near its multi-year low, none of this is baked into the stock, ever lowering the investment risk given these latest developments.

Background

My thesis for Intel in 2023 has been that this would be the year where the turnaround would be derisked, as discussed at the beginning of the year: Intel (INTC): Tough 2023 Ahead. As this is coinciding with the market downturn that has affected the stock (which isn’t looking much further), and indeed evidence of a successful turnaround is starting to mount as the year progress, the risk-reward is becoming very strong.

I discussed the most recent quarter here: Intel: Flawless Quarter, More Tailwinds Than Headwinds (INTC). The financial implications of the turnaround were discussed here: Intel: New World Order (NASDAQ:INTC).

Foundry prepayment

CEO Pat Gelsinger announced the news at a recent conference.

witeken on Twitter: “Deutsche Bank 2023 Technology Conference Pat Gelsinger, Intel CEO, will present, discussing Intel’s business and corporate strategy. pic.twitter.com/9fmaJXZGOm / Twitter”

To recap, after having fallen behind at 10nm and 7nm (original names, since renamed to 7 and 4), Intel appointed Gelsinger to lead the turnaround effort. Having worse transistors (than the competition) means lower competitiveness, leading to lower market share and degraded margins and pricing power. Now, while any CEO could have come up with the plan to regain process leadership (as had already been Intel’s goal), what Gelsinger added was the new foundry business. Since this is a significant market, this new business could in principle become a new major driver of (and contributor to) overall revenue growth.

In addition, foundries tend to generate their highest margins with their depreciated fabs, which contrasts with Intel’s historical practice of retooling its fabs every several years for the latest node. Basically, a node starts to ramp down just as the yield (fraction of working chips on a wafer) starts to mature and the fabs are depreciated. Overall, while Intel had dabbled before in the foundry space with some modest customer wins, there was no significant push and this effort eventually got mostly derailed by the process issues.

So when Gelsinger announced the foundry push in 2021, while the move made sense, sceptics were doubtful that Intel could succeed in the space as it has little experience and the concerns about process technology were still looming, making it hard to persuade customers to move away from TSMC (TSM). Even if it could solve these issues, it would take years before the business would start to generate meaningful revenue (given the length of design cycles), and lastly there’s the question of Intel competing against its foundry customers and vice versa.

In the nearly 2.5 years since this announcement, Intel has addressed every single one of these concerns. It hasn’t had any additional process slips in the last three years (unlike TSMC) and remains on track to return to leadership in 2025. It brought in capable leaders and signed agreements with relevant foundry ecosystem partners. Late last year it started running 18A test chips, to the acclaim of the likes of Nvidia’s (NVDA) Jensen Huang: Nvidia CEO Says Intel’s Test Chip Results For Next-Gen Process Are Good. It established the internal foundry model that treats even Intel’s own product/design teams on the same footing as any other foundry customer. And lastly, through its (advanced) packaging business (an area where Intel never lost its leadership) Intel already started to generate revenue fairly early. This business line also serves as potential entry for customers to later on also get their chips manufactured (not just packaged) at Intel.

The only two blemishes on this effort are that the initial leader of this business stepped down, and the falling through of the Tower acquisition, which would have positioned Intel as an end-to-end foundry from trailing to leading edge and even some specialty technologies.

Altogether, after having dealt with all possible concerns, it was only a matter of time before Intel would have been able to announce its first major customers. Finally (although there have been some earlier appetizers such as the MediaTek win for Intel 16 and Ericsson for 18A), this time has come.

While Intel couldn’t provide a specific name yet, what we did get was arguably equally valuable. As mentioned, Intel did not just sign up a customer for 18A, but already got a “large”, “significant” prepayment, and said it is now accelerating the build-out of the Arizona fabs. Gelsinger also took a stab at TSMC by saying that while others were having difficulties in Arizona, Intel is loving it.

This signals that Intel has signed up a top 10 fabless company, as only such a customer is likely large enough to warrant such a prepayment. Note that since it is unlikely any company would suddenly shift its whole business to a new foundry in one step (with IFS becoming its primary supplier, hypothetically), this only makes it more likely that a big fabless company has signed up to shift some portion of its leading edge requirements to Intel.

Of course, even if this customer is shifting all of its next-gen capacity to Intel, then that also only add to the thesis that Intel is a worthy foundry. Put differently, if a foundry customer is willing to bet its whole (or at least part of its) business on Intel’s 18A, then surely investors might follow suit in investing in Intel.

The two most likely candidates seem to be Nvidia (NVDA) and Qualcomm (QCOM). Both companies are part of the 18A RAMP-C program, although IBM (IBM) is as well. Intel initially announced Qualcomm as a 20A customer, but has long since shelved plans for 20A as a foundry node. So while there have been reports/rumors of Qualcomm no longer using 20A, this has actually long since been more or less confirmed by Intel itself.

Bottom line, although Intel is still only commercially shipping just one of its five nodes in four years, this plan was always back-end loaded with basically four nodes remaining in around 18 months. In the overall development cycle, this means 18A is basically near the latter stages, as indicated by the test chips that have been running since late last year. As the node to recapture process leadership and as such also Intel’s first leading edge foundry node, having signed up what only seems to be a major customer finally tangibly validates the turnaround thesis that investors have been patiently banking on for the last few years.

Basically, all the pieces are now coming together, successfully, for Intel to complete the turnaround and start a new chapter in its history back at the bleeding edge of the industry. This is a significant milestone that simply cannot be understated. And given how low the stock price is compared to the Bob Swan tenure, this arguably has major investor implications just as well for strong returns.

Other announcements

Intel said revenue was trending above the midpoint, which is a sequential increase from Q3 but still sharply down from levels of years past.

With regards to 20A, Intel said it has taken up its volume expectations for Arrow Lake, which Gelsinger indicated was contrary to rumors. He likely referred to recent rumors about Intel cutting its N3 orders from TSMC, as the graphics tile was supposed to get an upgrade to N3 with Arrow Lake. Note that this is not the first time that Pat Gelsinger has explicitly rebutted rumors, which he previously said (regarding Meteor Lake) Intel would prove “firmly” wrong.

He hinted at more announcements regarding 20A and/or Arrow Lake at Intel Innovation, as this node is (despite coming after Intel 4 and 3) as discussed just around the corner. Intel 18A is “looking good” and “solidly” on track, and the team also walked him through the following two nodes the day before the conference, so clearly the pipeline/roadmap is full. With rumors of TSMC forming a N2 “task force” (which apparently TSMC confirmed), Intel has a chance to extend its lead only further.

Discussing 20/18A key innovations, Pat Gelsinger was keen to point out that (potential) foundry customers were fond of its PowerVia industry-first backside power delivery network, which he put on equal footing to RibbonFET, and which Intel has also been very keen to point it has multi-year leadership with. “Ours is a work of art”. In particular, Gelsinger pointed out the area gains foundry customers were available to achieve because of it.

With regards to the financials and the stock implications, Gelsinger once again reiterated his view that the market is giving Intel zero credit for its possibility to achieve a TSMC kind of gross and operating margin structure because of the above as well as the increased cost discipline. While even Intel’s model only partially builds in this goal (20% FCF margin), investors can easily do the math themselves in the blue sky scenario of growing towards $100B revenue with 40% operating margin, although such a case should be seen as a perhaps decade out target. “If we’re gonna have the best transistors, we’re gonna do very well.”

Investor Takeaway

At the end of his rhetoric, Pat Gelsinger pointed out that given the progress in the last 2.5 years in rebuilding Intel, “you have to be much less skeptical”. While certainly Intel doesn’t have a perfect track record under his leadership, many of the things that do not have progressed as smoothly, such as the data center roadmap, can be mostly attributed as fallout from prior management, as Gelsinger took over while the ship was still sinking.

Overall, I agree that indeed the pieces are neatly coming together, as 18A has remained on the schedule announced in mid-2021. Essentially, for the last two years the main bull argument has been that there have not been further delays, which implies Intel will reclaim leadership with 18A in 2025 (since as a TSMC customer itself Intel knows TSMC’s targets). Now, the narrative has massively changed as what seems like a major foundry customer has put real money behind its adoption of 18A, providing tangible outside evidence of a successful turnaround for the very first time.

So as the turnaround is officially progressing from R&D projects and test chips to multi-billion dollar fabs (by the end of next year, to be sure), investors who have remained skeptical have the opportunity to enter with an acceptably low risk at still very soft stock levels.

---

Analyst’s Disclosure: I/we have a beneficial long position in the shares of INTC either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

---