Intel: Mediocre Quarter

Oct. 22, 2021 11:16 AM ETIntel Corporation (INTC)AMD, ASML, NVDA, QCOM, TSM134 Comments35 Likes
Arne Verheyde profile picture
Arne Verheyde
9.07K Followers

Summary

  • Intel delivered a mediocre quarter in both the PC and data center. Intel had guided Q3 would be the worst quarter during the shortages, but investors were expecting more anyway.
  • I would urge investors to look at the bigger picture: remember how unanimously all analysts and investors said it would take years to turn the ship after Pat Gelsinger’s appointment.
  • The most important topic investors should currently be concerned about remains restoring Intel’s technology leadership, to which Intel is steadily executing.
  • Intel 7/4/3/20A/18A are all on or ahead of July timelines (while TSMC is delaying 3/2nm). Pat Gelsinger: "5 nodes in 4 years have never been done in history", but Intel is doing it.
  • Intel expects capex to grow to >$25B to drive a long-term revenue CAGR of 10-12%, but mid-term gross margin weakness; if technology leadership didn’t cost money, anyone could do it.

Intel logo en campus gebouw, op M.A.T.A.M Tech compound.

liorpt/iStock Editorial via Getty Images

Investment Thesis

Buy Intel (NASDAQ:INTC) while the market is being myopic on short-term results in a third-party supply constraint environment.

Pat Gelsinger, instead, is currently busy restoring Intel's former leadership. This may eventually translate into a premium valuation on the stock market. In doing so, Intel may displace those whom the market currently pays a premium valuation for: Nvidia (NVDA), TSMC (TSM) and AMD (AMD). This is a decade-long battle.

Q3 results

I'm not going to lie, Intel's Q3 results, although in line with guidance, were quite mediocre, and definitely below my estimates as laid out in my Q3 preview. In that regard, Q3 didn't disappoint as "moment of truth". (I will discuss the long-term guidance that Intel provided below.)

So let's start with the poor numbers. First, PC revenue declined QoQ, despite Q3 usually being a seasonally strong quarter. Some would argue that the COVID-19-induced surge in demand is waning. However, instead, Intel attributed this due to component shortages.

To be clear, Intel had guided for Q3 to be the bottom of the shortages, so investors should not be immensely surprised that this was a possibility. In that sense, the sell-off is actually a bit of a mystery: investors should have expected Intel to be supply constraint, so a sell-off implies that the stock for some reason was already baking in upside.

Nevertheless, investors may still remember that Intel made quite a bit of noise last quarter that it had started in-house production of substrates to deal with the shortages. So from that view, perhaps there really was a possibility of beating guidance: this is what I had assumed in my Q3 preview. However, the shortages are substantially broader than just substrates, with Pat Gelsinger citing Ethernet controllers and other stuff during the call.

Still, perhaps even more worrisome to investors than the PC miss was the stagnation on the data center side (compared to Q2): Intel had initially guided for a steady ramp in revenue during the year, and Intel admitted DGC performed below expectations. However, the story here is very similar, with the result being attributed to the shortages as well as China issues, a market where Intel has higher-than-average market share.

The third weak point in Intel's report was Mobileye. Here, also, given the current semiconductor shortages (which as many remember started in automotive) it is hard to analyze the result: Mobileye revenue has now stagnated for three straight quarters.

Fourth, I noticed that capex guidance seems to have decreased from ~$20B to $18-19B. Nevertheless, the 2022 capex is actually bumped to $25-28B and Intel boasted about how the $20B Arizona fab construction had started three months ahead of schedule.

Among all these "mediocre" results (one might still consider that before COVID-19, no one would have expected the PC to perform as well as it has done for the last year), there are still a few upside numbers. First, the quite severe decline in gross margin that Intel had guided to did not happen, with 56% compared to 53% guidance. (Nevertheless, instead of taking a moment to appreciate the upside gross margin performance, investors and others have already moved on to the next quarter, which again has been guided to for 53%.) Additionally, EPS and net income were actually up on the order of 60% YoY, although most upside was due to one-time events. Thirdly, IoTG continued its recovery and recorded record revenue, up 54% YoY.

Bigger picture

After digesting these numbers, let's step back for a second. Remember early this year when Pat Gelsinger became CEO? Legions of analysts and Seeking Alpha contributors argued that it would take many years (at the very least until 2024, since the roadmap was already set until then) to straighten the ship. So in that light, should a serious Intel investor really be heavily concerned about Intel's current shaky results in a third-party shortages environment?

In that light, it seems analysts want too much to have their cake and eat it too: continue to beat Intel to the ground based on current results, while forgetting that they had previously already moved the needle to 2024 anyway for their expectations of Intel to show progress.

For example, is the Mobileye flat revenue because the Mobileye growth story is finished (despite the initial commercial robotaxi launch just a few quarters away and self-driving for consumers still in development), or due to the shortages? (I would bet that in Q3 alone, Mobileye won more design win units than Nvidia and Qualcomm (QCOM) have achieved over several years combined.)

Instead, what I would argue (as I have done a lot already): what investors should really be interested in for the time being is the technological side of the picture (before worrying about financials). This is similar to how technology mattered more than financials to AMD when AMD was at $2, and look where AMD is now. Without technology, Intel will never get anywhere (just like AMD during its own issues), so investing is the only path forward.

Is Intel catching up from its 10nm and product roadmap delays? Is Intel investing to restore its former technological leadership? Those I reckon remain the most relevant topics investors should be concerned about.

From that view, I would argue that Q3 more than delivered on that front. Instead of recapping what I have already covered in multiple articles over the last few months, I would suggest investors go to the Intel Q3 press release, skip the financials, and go to the "business highlights" section. Or from the investor presentation: "On or ahead of process tech timelines laid out in July". To quote two of the most relevant points:

Shared process and packaging roadmap updates for delivering five nodes within four years, putting Intel on a path to restore process performance per watt parity in 2024 and leadership in 2025 with key process innovations, including RibbonFET and PowerVia. Also introduced new advanced packaging technologies, Foveros Omni and Foveros Direct, for 2023.

Detailed Intel's biggest architectural shifts in a generation with the first in-depth look at Alder Lake, our first performance hybrid architecture with two new generations of x86 cores; Sapphire Rapids, our new standard setting data center architecture; our new discrete gaming graphics processing unit architecture; new infrastructure processing units; and Ponte Vecchio, our tour-de-force GPU architecture with Intel's highest ever compute density to accelerate AI, HPC, and advanced analytics workloads

This is also visible in a few other numbers from the report. First, R&D spending continued its steady increase since Pat Gelsinger became CEO, and reached another all-time high. Secondly, also employee count continued to increase as well. This is exactly what Intel bulls have been begging Intel to do for years while, instead, Intel was more busy with share buybacks.

Further analysis

I was busy writing a pretty scorching review of Intel's numbers. However, my view moderated as I was listening to the earnings call. There seem to be so many pieces going on that are not reflective of the underlying business potential. The shortages have been heavily discussed, and clearly, there is more to it than just substrates (which Intel calls "matched sets" to deliver a complete product). Additionally, management also called out the China regulatory issues, which Intel said it impacted Intel disproportionally given its high market share in that region. Intel didn't really quantify the impact, but it seems quite obvious that Intel's results could have been a lot better if not for all those third-party issues going on.

"We call it match sets, where we may have the CPU, but you don't have the LCD, or you don't have the Wi-Fi. Data centers are particularly struggling with some of the power chips and some of the networking or ethernet chips," Intel CEO Pat Gelsinger said in an interview with CNBC.

As an extension of this discussion, although some argue that is Intel reporting these subpar numbers during the biggest boom in semiconductors ever, Intel quite literally sold as many CPUs as they could, in light of the "matched sets" issue: if, for example, an Ethernet controller isn't deliverable due to the ongoing shortages, then one can't build a server or PC, and hence Intel won't see revenue. As Pat Gelsinger said, Intel in principle could have sold a lot more CPUs.

In other words, to solely attribute the Q3 results to competition from AMD is a pipe dream by AMD bulls at best (although Intel did admit competition was in line with expectations). Still, one has to be aware that it will take a while for Intel's market share position to improve, even if improved products are already launched: this is the reverse of how Intel on pure inertia has been able to largely maintain its market share while AMD was already selling vastly superior CPUs, during 2019-2020. As a case in point, despite the Ice Lake launch in Q2 (after all the delays), Intel is expecting to sell "just" 1M Ice Lake CPUs in Q4. Although this would already be as much as Q2 and Q3 combined, this compares to a TAM that should be on the order of 6-8M units per quarter. This means the majority of the CPUs Intel sells in the data center are still 14nm. Yikes (but of course not unexpected).

As such, although it is clear Intel is working very hard on a turnaround, as mentioned above it will take a significant amount of time for this to become visible, and only then it will become visible in the financials, and only then it will become visible in the stock price. In that order. This is similar to how AMD stock price for example bottomed to $2 during AMD's issues, despite that it was already known at the time that AMD had hired Jim Keller and was working on a substantially improved CPU architecture called Zen.

The topic of gross margins is further discussed in the next section, but here also the concerns are being hugely overblown. Sure, compared to Intel's standards these margins in the low 50s are rather low (although for 2021 they are guided at 57% still), but TSMC (which trades at a substantially higher valuation) hasn't reported higher gross margins than Intel. Ever.

Smart capital

With things like "IDM 2.0", one can't blame Intel that it hasn't tried to educate investors about its strategy. During Q3, Intel added yet another marketing term, which it called smart capital and refers to Intel being smart in how it spends its money, even though Intel will be spending a lot of money. It refers to three points:

  • Use of internal + external foundry capacity to satisfy capacity requirements (same as one of the three pillars of IDM 2.0);
  • The buildout of empty fab shells, which have the longest lead time, so these can be quickly filled with equipment to meet surges in demand (like during the current boom);
  • The use and being a beneficiary of government subsidies from the E.U. and U.S. Intel hinted that it could hike its capex spending beyond $28B if it got enough subsidies.

If one remembers how half a year some were ridiculing Intel for its $20B Arizona announcements compared to TSMC's $100B three-year plan, then clearly the current capex guidance is exactly in line with what TSMC is spending (especially if one considers that Intel is spending 100% of its capex on the leading edge, compared to 80% for TSMC). From that view, there seems to be a huge discrepancy in what the market is expecting TSMC and Intel to respectively achieve with similar spending/investments.

Long-term guidance (investor meeting)

Due to the CFO's retirement (which is causing the postponement of the investor meeting to February), Intel used the call to already give a preview of its long-term plan.

But I can a priori already understand why investors do not like these plans: Intel will be spending money to invest in the business. This means Intel is prioritizing the strength of its underlying business (read: technology) instead of focusing on whatever are investors' favorite metrics (usually EPS and P/E) that they can easily comprehend.

Nevertheless, the truth must also be said: in this case, investors are woefully wrong (and Intel is right). It is actually ironic: the market created a narrative last year of having to bail out the financial CEO Bob Swan in favor of an engineer-CEO to save Intel from the graveyard of dinosaur technology companies, in light of the 7nm delay. (The whole Dan Loeb saga.) But now that Investor Meeting is approaching and investors have to be educated about the spending implication of this plan to regain industry leadership, suddenly those investors are already going overboard again and selling their shares? If technology leadership didn't cost money, then anyone could do it. No one should have been surprised by this.

Pat Gelsinger: We broke ground on new fabs, shared our accelerated path to regain process performance leadership, and made our most dramatic architecture announcements in a decade.

With that preamble out of the way, let's review Intel's plan. Ultimately, Intel is investing on two fronts.

First, Intel is investing in manufacturing capacity for growth. Intel believes (like do most people in the semiconductor industry) that it is a growing market, and wants to make sure it can satisfy the potential demand that this growth would result in. So not just solving the shortages, but the increasing demand also after the shortages.

In particular, Intel expects to gain share and BOM (as the graphics business ramps) in the PC, but that remains its slowest business. More growth is expected from the data center and network and edge. Additionally, Intel is bullish about its new businesses (which are currently small): graphics, mobility and foundry.

These investments are exactly the same as what TSMC is doing. Similar to TSMC, Intel is heavily ramping its capex, and expect a further increase to $25-28B next year and expects capex may grow even further in the years after 2022, especially if the government subsidies kick in. As a result, Intel expects to grow at a 10-12% CAGR. (Just like the capital spending, this guidance would be just a bit below TSMC's near-15% guidance for revenue growth.)

Note that this growth forecast is a lot higher than Intel's growth in the previous decade. As such, the careful investor should note that Q3 in reality was an announcement for accelerated growth over time, and this hence increases the likelihood of a multiple expansion from Intel's historic low-teens valuation.

As such, this means that the market has misunderstood Intel's earnings call, and implies the market doesn't believe Intel can achieve its targets. Note that in the last few months, I have discussed mounting evidence that Intel has a reasonable likelihood of succeeding, and as such, I would argue that Intel stock is fundamentally mispriced.

Additionally, the market doesn't seem too pleased with the gross margin implications that this investment mode entails in the near term, with guidance for 51-53% over the next 2-3 years. Nevertheless, as mentioned above this is actually exactly the same kind of gross margins that investors are satisfied to see TSMC and ASML (ASML) deliver, so this point is immensely overblown.

In any case, this is where the second part of Intel's investments come into play, which Intel called "node compression". As mentioned above, Intel stated that all of its nodes that it has laid out (Intel 7, 4, 3, 20A and 18A) are all either on or ahead of schedule. Intel is going to deliver fives nodes in four years.

On the call, Pat Gelsinger remarked that this has never been done in history before, but Intel is going to do it. This is why it is so crucial to see the R&D numbers go up, and to hear management talk about Intel investing in more nodes in parallel than ever before.

Meanwhile, what we've seen at Intel's competitors isn't going any faster, to say the least. On the contrary, Pat Gelsinger was already taking a small victory lap during a call by expressing how the progress to restore process leadership seemed to be going even faster than he had expected even a quarter ago. I couldn't help but wonder if this was not in part due to some of the things that have been expressed by Intel's competitors, most notably the slowdown in the pace at TSMC at 3nm and 2nm.

Hence, simply put, while Intel is accelerating, others are slowing down. As I see it, currently Intel is the company with the strongest execution in Moore's Law and it will be just a matter of time before this becomes visible and tangible. Of course, currently, it is the earliest of the early innings of this change in momentum, but this is nevertheless something investors may take note of.

Stock expectations

Quoted from a bit higher in the article:

Intel is working very hard on a turnaround, but it will take a significant amount of time for this to become visible, and only then it will become visible in the financials, and only then it will become visible in the stock price. In that order. This is similar to how AMD stock price for example bottomed to $2 during AMD's issues, despite that it was already known at the time that AMD had hired Jim Keller and was working on a substantially improved CPU architecture called Zen.

In other words, the stock market clearly doesn't care about roadmaps and that kind of stuff. It didn't care about roadmaps in AMD's case during its near-bankruptcy era, or in Nvidia's case during the crypto bust in 2018, and now it doesn't care about it in Intel's case. But as Pat Gelsinger said, currently the roadmap until 18A in 2025 remains on schedule. Maybe next quarter it won't be anymore (which is the risk), but knowing more than what the market cares about could result in substantial alpha down the road.

Summary

One just has to look at Nvidia (NVDA) and TSMC (TSM) to get a sense of what kind of premium the market likes to pay for leadership in the semiconductor industry.

As such, investors should be aware, more than anything, that this is exactly what Pat Gelsinger is investing in to transform Intel back into: the leader in the foundry space, in the data center, in graphics, HPC and AI, and also the leader in autonomous driving and robotaxis.

The market obviously isn't going to pay up for that until it actually sees those results. Hence, Intel is (currently) cheap. But clearly, the potential is significant.

The risk here obviously remains one of execution. Nevertheless, as I have been arguing lately, the tide may really be turning. Just look at where Intel was a year ago: Cascade Lake-Refresh on 14nm (a five-year-old architecture on a five-year-old process node). Look at where Intel will be in the near future with Sapphire Rapids and Alder Lake as much more credible products (that in several aspects are already a level ahead of AMD). And then look at where the industry is heading with RibbonFET in 2024. Meanwhile, TSMC is currently more busy with damage control about the 3nm delay and its 2nm in 2025 (which is clearly off schedule).

Hence, I would argue that with Pat Gelsinger at the steering wheel, Intel today is in a much more credible position going forward. So as I have been arguing for the last year, it doesn't matter that Intel's results aren't visible yet, as this is a bet that might be worth taking. The current low valuation could amplify the returns if it succeeds. (Even if Intel only partly succeeds or even fails, the chance of a revenue decline seems incredibly slim given all the industry and Intel-specific tailwinds.)

Earnings notes

Pat

  • Double-digit PC TAM despite shortages, gradual supply improvement. Strong products: new EVO, >70Mu TGL (fastest laptop ramp ever), Alder Lake shipping + launch, fastest launch across all categories.
  • DGC: long-term growth, 1Mu ICL shipments, >1Mu ICL shipments in Q4. SPR production in Q1. Network 5G moving to x86.
  • IoT: all-time record revenue, Mobileye lead in AVs, robotaxi in 2022.
  • Roadmap/IDM2.0: construction on/ahead of schedule, Arizona fabs 3mo ahead of schedule. Smart Capital: building shelves (empty fabs), external foundries (Meteor Lake/Ponte Vecchio), government subsidies. IFS: first packaging shipped for revenue, >100 potential customers, interest in Intel 16, 18A. Intel 7, 4, 3, 20A, 18A all on/ahead of July timelines. Meteor Lake power-up, performance within expectations. Architecture Day: Alder Lake (Golden Cove + Gracemont), Intel Arc Q1, Mount Evans IPU, Sapphire Rapids, Ponte Vecchio.
  • Intel On: new announcements.
  • Invest in future -> regain leadership.
  • CFO retirement -> Investor Meeting moved to February.
  • General picture of plan: graphics, networking, foundry, mobility + client and DGC -> invest for double-digit growth -> pressure FCF, maintain dividend, gross margins down for 2-3 years.
  • Added over 6k new engineers.

George

  • PC below guide -> shipping + supply constraints.
  • Gross margin beat: higher desktop/premium mix.
  • EPS beat due to McAfee dividend + tax benefit + mix.
  • Component shortages. Excluding Apple: CCG up 10% YoY.
  • DGC below guidance due to shortages + cloud weakness.
  • IoT record, strength in industrial, retail.
  • PSG demand > supply ("significantly").
  • Q4 outlook: more modest DGC growth than expected (China + supply constraint).
  • Full year: revenue flat, PC TAM increase offset by Apple, DGC down due to competitive environment, lower demand from China, supply constraints.
  • Changes in accounting.
  • Long-term: revenue reflects strong TAM, growth driven by leadership product, >$74B revenue in 2022, revenue growth accelerates to 10-12% CAGR over 4-5 years, gross margins 51-53% over 2-3 years before increasing. Capex: $25-28B in 2022, with potential for further growth.

Q&A

  • Gross margin? Pivot point for Intel -> long-term growth -> massive opportunity. Next year: new manufacturing nodes (5 nodes in 4 years).
  • Confidence in double-digit CAGR? Markets: PC gain share + BOM (modest growth), stronger growth in DGC (stronger products), 4 new businesses: networking + edge, graphics, mobility, IFS.
  • Foundry? Fairly modest (takes a few years for a design) -> margins similar to market leader (TSMC), leverage R&D in process + architecture (new monetization) + Smart Capital strategy, "some of the largest customers in the industry" to be "on the best transistors in the planet", progress "even better than I would have thought".
  • DGC ASP? HCC mix (lower cloud than expected), bounce in networking, some unique issues in China (regulatory).
  • Gross margin? Catch up with capacity + build more flexibility to support higher revenue (to absorb higher depreciation), higher opex (R&D -> invest in more nodes in parallel) "node compression" -> better margins, pricing' "5 nodes in 4 years have never been done in history". "Closing gap on the industry more rapidly than expected than even a quarter ago".
  • Gross margin? Higher capital spending, node compression. Confident in growth, large markets, on track for product leadership.
  • Market share? Gain market share due to investments, Alder Lake gains share.
  • Data center, SPR volume, market share? Q1 production, ramp in Q2, high exposure to China ("ethernet controllers, etc." -> matched sets), strong backlogs.

Investor Takeaway

The topic that I argued investors should be most concerned about is technology, not the current results which are obfuscated anyway by the smoke and mirrors of semiconductor shortages and Chinese regulatory issues.

As I see it, the market is using the opportunity to give INTC stock a hard time because of the near-term results and even medium-term heavy investments. However, with someone as capable as Pat Gelsinger steering the ship (which is what technology enthusiasts had been begging Intel to appoint as CEO, for perhaps no less than two decades, as Intel was instead being led by either CFOs or internal politics), and given the results that are already visible on multiple fronts, Intel in perhaps half a decade from now could be a very different company.

A company with leadership technology, leadership gross margins, accelerated growth, and just as important for investors, a leadership valuation in the stock market (which is currently being reserved for TSMC, Nvidia and AMD). I made the analogy with AMD a few times: could even the strongest bulls have expected, half a decade ago, where AMD currently is?

In the most simple terms, investors (ironically) currently don't seem to like what is required to be the industry leader again (investments!), but Intel has promised double-digit growth acceleration, in the wake of an unprecedented five nodes within four years. If Intel delivers, then this growth in revenue should cumulate over time, and by definition become a very profitable income once Intel has restored its tech leadership.

Investors are currently running away like from a tsunami after hearing about the 51-53% gross margin forecast (which is still on par with TSMC, AMD and ASML), but this displays a fundamental misunderstanding of this market: Moore's Law (an economic law) has always rewarded the frontrunner in this ever-ongoing race. Intel is currently not the frontrunner, but there are tangible signs that the tide will be turning over the next few years. Are TSMC's 3/2nm issues and slowdown overblown, or is Intel this time finally going to deliver a product when it says it will: place your bets.

This article was written by

Arne Verheyde profile picture
9.07K Followers
With an engineering background, looking for companies with expertise to be well-positioned for growth and leadership.
Follow

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.

Recommended For You

Comments (134)

To ensure this doesn’t happen in the future, please enable Javascript and cookies in your browser.
Is this happening to you frequently? Please report it on our feedback forum.
If you have an ad-blocker enabled you may be blocked from proceeding. Please disable your ad-blocker and refresh.