The Data Center Delivered, the PC Broke: A $1.6B Guide Shortfall Resets Intel's Earnings Base, and the 28.6% Drop Is Not Yet a Bargain

Results vs. Consensus

Q3 2026 Scorecard

Year-Over-Year Comparison (Q3 2026 vs. Q3 2025)

Sequential Comparison (Q3 2026 vs. Q2 2026)

Revenue

Revenue of $16.51B grew 20.9% year over year, and almost all of that growth came from one segment. DCAI added $3.2B year over year while CCPG lost $438M. Sequentially the picture is starker: total revenue rose only $385M as a $1.06B DCAI gain was mostly consumed by a $780M CCPG decline. The 10-Q is unusually explicit about the mechanics. Server revenue rose on a 54% increase in ASPs against a 10% volume gain, while client revenue fell on a 14% volume decline that a 9% ASP increase only partly offset. Both segments carry higher ASPs: server mostly through a richer premium mix, with pricing a lesser factor, and client mainly through pricing actions to offset higher input costs. The client half of the company is now losing units faster than price can cover. An in-line revenue print understates how much the composition changed inside a single quarter.

Margins

Non-GAAP gross margin of 44.1% was the highest of the five quarters since Q3 2025 (40.0%, 37.9%, 41.0%, 41.8%, 44.1%), and the drivers were the ones the thesis needs: Intel 3 costs down about 15% from Q1, 18A costs down about 40% since Q1, and Granite Rapids crossing into margin-accretive territory. The 10-Q adds a caution on the year-over-year bridge: $328M of lower period charges helped, partly offset by higher inventory charges "as we aligned client product mix with customer demand." The problem is the next print. The Q4 guide of 38.0% gives back 610 bps, of which ~200 bps is Fab 28 scrap and the rest is under-absorption as client wafer starts come down, plus a mix shift toward lower-margin purpose-built silicon. That lands Q4 2026 almost exactly where Q4 2025 was (37.9%). In July the CFO called getting gross margin "comfortably into the 40s in every quarter" his top goal for 2026. Q4 misses that goal as reported, and only reaches 40.0% excluding Fab 28.

EPS

Non-GAAP EPS of $0.52 is operating quality: non-GAAP operating income rose $464M sequentially on margin and cost, and the tax rate held at the long-term 11%. The diluted share count is now the variable to watch. It rose to 5,307M from 5,104M in Q2 as the 242M shares sold in August entered for part of the quarter, and it reaches ~5.44B in Q4. That is 6.6% more shares than Q2 carried, which dilutes per-share growth for at least the next four quarters. GAAP EPS of $0.60 exceeded non-GAAP for a reason that will repeat: the Escrowed Shares liability owed to the US government marks down when the stock falls, and Q3 booked a $1.0B gain as the shares slipped from $128.32 at the end of Q2 to $123.00 at the end of Q3. The stock has since fallen much further.

Segment Performance

Operating Segments: Q3 2026

Segment revenue includes intersegment transactions; Intel Foundry revenue is mostly internal wafer and back-end sales, of which $411M was external. All Other declined year over year because Altera was deconsolidated in September 2025.

Revenue Drivers: Volume vs. Price

Key Operating Metrics

Data Center and AI (DCAI): Fastest Growth in Fifteen Years, Still Rationed

DCAI revenue of $7.32B rose 77.8% year over year and 16.9% sequentially, the segment's strongest growth in fifteen years. Operating income of $3.04B was up $565M sequentially and more than tripled from $964M a year ago, lifting the margin to 41.5%. Growth was broad across hyperscale, enterprise and neo-cloud customers, and Granite Rapids on Intel 3 is now the largest product family in the server line. The CFO said DCAI is on track to finish 2026 nearly $10B above where the company expected entering the year. Through nine months, DCAI revenue totals $18.6B, and the Q4 guide implies roughly $8.4B more.

"Despite our progress to increase capacity, demand for DCAI products continues to meaningfully exceed available supply." — David Zinsner, CFO

The composition deserves as much attention as the total. Server volume grew 10% year over year while ASPs rose 54%, mostly from mix. That is excellent for margin and a warning for share: management expects industry server CPU units to grow more than 30% in 2026, so Intel's unit growth is running well behind the market it describes.

Assessment: DCAI confirms the strongest pillar of the bull case: the CPU is not being displaced by AI accelerators, and the agentic-AI framing gives the segment a credible multi-year demand story. The constraint is now Intel's own throughput (Intel 3 wafers, EMIB substrates, and in Q4 the 10nm output lost at Fab 28), which the company says loosens only in the second half of 2027. Until then, Intel is monetizing scarcity through price while competitors take incremental units.

Client Computing and Physical AI (CCPG): The Memory Crunch Arrives

CCPG revenue of $8.10B fell 5.1% year over year and 8.8% sequentially and came in below management's expectations. Client volume fell 14% year over year, which the 10-Q attributes to lower demand and a competitive market. Management tied the demand side to memory prices that pushed system prices up while IT budgets tightened, and the pressure "intensified during Q3, particularly late in the quarter." Operating income of $2.29B (28.2% margin) held roughly flat sequentially, helped by lower inventory charges than in Q2. The positives are product-level: AI PC revenue grew 35% year over year, Wildcat Lake units rose about 70% sequentially and are sold out, and 18A now carries nearly a third of client CPU revenue.

"We expect CCPG revenue to decline greater than 20% sequentially as customers move quickly to reduce weeks of inventory and align purchases with their expectations for 2027 consumption." — David Zinsner, CFO

Assessment: Last quarter we called client a source of cash and wafers rather than a growth vector, and noted that its beat was price-driven. This quarter shows the cost of that: sell-in ran ahead of consumption, and Q4 and Q1 are now the correction. CCPG still produces more operating income than any segment except DCAI and funds the foundry build. A segment guided to fall more than 20% sequentially, into a 2027 PC market that management expects to shrink another 10%, removes the earnings cushion that let the company absorb Foundry losses.

Intel Foundry: Better Costs, Same Question

Foundry revenue of $6.70B grew 16.2% sequentially on better-than-expected Intel 3 and 18A output, and the operating loss narrowed by $236M to $(1.85B), a $468M improvement from a year ago. 18A yields improved at a double-digit rate in the quarter and are tracking toward industry-standard levels ahead of internal projections. External revenue rose to $411M from $293M, but both the CFO and the 10-Q attribute the gain mainly to a non-recurring customer program. The node roadmap delivered: the 14A 0.9 PDK went to select customers in October, as committed in July.

"These results represent meaningful progress, although we still have considerable work ahead of us to deliver a sustainably profitable Foundry business." — David Zinsner, CFO

Assessment: Foundry is improving as a cost center at about $200M-350M per quarter, which is real progress. As a merchant business it remains unproven: no external wafer customer has been named with a quantified commitment, the anchor named in April has gone unmentioned for two calls, the external revenue that did arrive is labeled non-recurring, and this call again gave no breakeven date. The 14A PDK milestone keeps the technology case on schedule; the commercial case is still a promise of announcements "in the next 2 quarters."

All Other: Altera Out

All Other revenue of $692M was down 30.3% year over year, which the 10-Q attributes to Altera's deconsolidation in September 2025, and essentially flat sequentially. Operating income of $179M compares with $100M a year ago.

Assessment: Not a thesis driver. The segment is now Mobileye plus a set of small businesses and contributes little to the earnings debate.

Key Topics & Management Commentary

Overall Management Tone: Management was more candid and less expansive than in July. The CEO's prepared remarks closed on disappointment at unmet customer demand, and the CFO conceded, when asked directly, that 2027 would probably not generate positive free cash flow. Confidence remained, but it moved out to server demand and the 2027-2028 roadmap and away from near-term numbers. The call was least convincing on external foundry customers, where the answer was again a non-disclosure policy and a promise to announce later.

1. The Guide-Down: Where the $1.6B Went

Q4 revenue guidance of $15.0B to $16.0B puts the midpoint about $1.6B below the ~$17.1B consensus and 6.1% below Q3. Management attributed the gap to two items. Fab 28 accounts for roughly $300M of revenue, nearly all of it 10nm server product. The rest is client: CCPG is guided down more than 20% sequentially, from $8.10B to something below $6.5B, while DCAI is guided up mid-teens to roughly $8.4B. Non-GAAP EPS of $0.25 is 42% below the $0.43 consensus, because the revenue shortfall falls on a fixed cost base and drags gross margin to 38%.

"Our guidance includes an impact of roughly $300 million to revenue and 200 basis points to gross margin based on our early assessment of the incident at Fab 28." — David Zinsner, CFO

The continuity point matters. In July, management said supply improvement was skewed to late Q3 and Q4 and that the company would still be "behind in the fourth quarter," which the Street reasonably read as a Q4 step-up. The server half of that promise is intact. The client half reversed in about eight weeks.

Assessment: The guide-down is about 80% a client-cycle problem and 20% an accident. Neither impairs DCAI, which is why the business case survives. Both are larger and longer than one quarter, and that is why the earnings base has to come down for 2027 as well as for Q4.

2. The PC Down-Cycle: Memory Prices Break the Client Market

Management now expects 2026 PC units to fall about 12%, with second-half units down more than 20%, and plans for another ~10% decline in 2027. The mechanism is memory: the AI build-out is absorbing DRAM capacity, system prices have risen sharply, and IT budgets cannot absorb the increase. OEMs are cutting inventory to match the softer sell-through, which hits Intel's sell-in harder than end demand. The CFO expects the inventory correction to fall mainly in Q4 and Q1, possibly stretching into Q2, with a normal seasonal decline in Q1 on top.

"In recent weeks, our customers began signaling a PC slowdown based on the memory prices and availability. Year-to-date system endpoint prices have increased significantly placing additional pressure on already limited IT budgets." — Lip-Bu Tan, CEO

The 2027 product response is Nova Lake, Intel's first 18A-P client product, which management says is ahead of schedule with design wins expected at twice Panther Lake's launch count and fills a gap in high-end desktop. The CFO called 2027 a "market share gain type of environment" for client.

Assessment: This is the risk the thesis did not carry as a named pillar, and it is now the largest driver of the estimate cut. Share gains from Nova Lake can soften a shrinking market but cannot offset a 10% unit decline plus an inventory flush. We add the client down-cycle as a new bear point at MATERIALIZING.

3. Fab 28: An Accident With a Two-Quarter Tail

An accident at the Kiryat Gat site in October caused casualties and injuries and halted production for about two weeks. Management expects a ~$300M revenue impact in Q4 and a similar amount in Q1, with most wafer scrap booked in Q4. The 10-Q is broader than the call: it expects the reduced volumes and inefficiencies to affect costs and results "in Q4 2026 and through the first half of 2027," and says the company cannot yet estimate the financial impact or insurance recoveries.

"Finally, early this month, an unfortunate and tragic accident led to an extended shutdown in our Israel factory. While the teams have worked quickly to address the situation, it will impact our Q4 10-nanometer server shipments." — Lip-Bu Tan, CEO

The episode also exposes a concentration risk that the 10-Q states plainly: a significant portion of current and expected revenue comes from Intel 7 products made in Israel, and the company is not insured for business interruption caused by war or political violence.

Assessment: The accident is a one-time event with a defined tail, and it removes server output in the quarter Intel can least afford to lose it. The more durable lesson is concentration: Intel 7 server supply runs heavily through one site in a region where the company itself names geopolitical risk. We add Israel fab concentration as a new bear point at EMERGING.

4. DCAI and the $200B Server CPU TAM

Management raised its demand framing again. Industry server CPU units are now expected to grow more than 30% in both 2026 and 2027, with supply constraints through 2027 and into 2028. The CEO put the server CPU TAM at $40B today and "more than $200 billion over 5 years," driven by agentic workloads that need far more general-purpose compute per AI deployment. Clearwater Forest, the first 18A server product, is shipping for revenue; Diamond Rapids on 18A-P is sampling for a 2027 ramp; and Coral Rapids on 14A follows in 2028. A planned multiyear collaboration with Oracle Cloud Infrastructure and requests from several customers for long-term agreements support the demand claim.

"By our estimates, server CPU TAM, largely driven by agentic workloads, could increase as much as 5x from where it is today, from $40 billion to more than $200 billion over 5 years." — Lip-Bu Tan, CEO

Assessment: A fivefold TAM claim is a ceiling scenario, not a forecast, and we do not underwrite it. The near-term evidence is strong on its own terms: a $7.3B quarter guided to roughly $8.4B, and an implied $33B-34B annualized exit rate that management called "a good starting point" for 2027. This is the part of Intel worth owning. The question is how much of that market Intel can supply.

5. Server Share: Selling Price While Competitors Take Units

The 10-Q puts numbers on a problem the call only described. Intel's server volume rose 10% year over year in Q3 while ASPs rose 54%. Against management's own forecast of more than 30% industry unit growth, that implies a meaningful loss of unit share in 2026, offset by mix and pricing. When asked about share against AMD and Arm, the CFO answered that share next year is "a function of supply," and the CEO set competitive parity at Coral Rapids in 2028, with leadership only in the generation after.

"In terms of Coral Rapids, we will be competitive to AMD. And the next-generation product, the tick-tock, we'll be ahead of the AMD on that." — Lip-Bu Tan, CEO

Assessment: Supply-limited share loss differs from demand-driven share loss, and customers who cannot get Xeon today may return when supply arrives. But sockets won by competitors in a 30%-growth year are hard to win back, and the CEO's own timeline puts product parity two years out. DCAI's revenue growth overstates Intel's competitive position, so we treat it as confirmation of demand rather than of share.

6. The Supply Ramp: Most Relief Arrives in the Second Half of 2027

The supply plan has four parts, and most of them mature late. Intel 3 yields and cycle times are improving, but the CEO is "demanding significant execution." Intel 3 capacity is being added in Oregon by converting 18A space, with output expected in the second half of 2027. Intel has arranged about 40% more substrate capacity with vendors, but "a lot of that doesn't hit until the back half of the year." Ireland tool installation, the 18A ramp at Arizona's Fab 62, and packaging expansion in New Mexico and Malaysia round out the plan.

"First, Intel 3 yield and cycle time continue to improve, but I am demanding significant execution from the team. Second, even as our Intel 3 wafer output ramps quickly, an industry-wide shortage of advanced organic substrates for our EMIB packaging is also impacting our ability to scale output." — Lip-Bu Tan, CEO

Assessment: The PC slump gives Intel some wafer room, since 18A capacity freed from Panther Lake can move to server. But Intel 3 wafers and EMIB substrates are both binding, and the main relief for both arrives in the second half of 2027. For modeling, the first half of 2027 is a supply-capped DCAI against a falling client market. That is a weak combination for earnings even with strong demand.

7. Gross Margin: 44.1% to 38%, and a Rule of Thumb for 2027

The Q4 margin guide falls 610 bps from Q3. Management attributed it to Fab 28 scrap (~200 bps), lower revenue against fixed costs as client wafer starts are cut, and a mix shift toward purpose-built silicon, whose margins are below the corporate average. For 2027, the CFO listed the offsets (Granite Rapids now margin-accretive, continued Intel 3 and 18A cost gains) against the headwinds (18A under-loading and lower-margin ASIC growth), and fell back on a rule of thumb: 40% to 60% fall-through of incremental revenue to gross profit.

"The sequential decline in Q4 gross margin primarily reflects lower overall revenue against our fixed cost base as we adjust client wafer starts to current demand conditions." — David Zinsner, CFO

Panther Lake is not yet at the corporate average gross margin, though the CFO said it has made "huge strides."

Assessment: In July we said gross margin falling back below 40% would be one trigger for Underperform, and the CFO had set "comfortably into the 40s in every quarter" as his goal for the year. The reported Q4 guide of 38% crosses that line. We judge the trigger partly tripped rather than fully tripped, because 200 bps is a one-quarter scrap charge and the ex-Fab 28 figure is 40.0%. Q1 2027 is the test: a non-GAAP gross margin below 38% without a comparable one-time item would mean the 40s were a supply-scarcity peak, not a new base.

8. Capex, Free Cash Flow and the $23B Raise

Intel sold about 242M shares at $95 on August 12 for $22.6B of net proceeds. Cash and short-term investments now total $54.0B against $50.1B of debt, and the revolver has grown to $15B. The money is going into equipment: 2026 gross capex is on track for $20B, which, after $10.2B spent through nine months, implies roughly $9.8B in Q4. Firm equipment purchase orders rose nearly 40% sequentially, and capital commitments reached $19.6B at quarter-end from $12.8B in December. For 2027, capex will be "up strongly," and the CFO said positive free cash flow is unlikely.

"I think next year is an investment year for us as we look at the year. Probably not a dynamic where we're going to have positive free cash flow next year." — David Zinsner, CFO

He also committed to deleveraging in 2027 and repeated a target of at least a 20% return on each node, adding that Intel needs "to get through what is basically a knothole here in '27."

Assessment: The raise was well timed: Intel sold stock at $95 that now trades 30.8% lower, and it removes the financing risk we flagged in July. The trade-off is that the capital program is now fixed while the earnings base is falling. A 2027 that probably will not generate positive free cash flow, combined with a PC trough and a supply-capped server ramp, is the cash-flow bear point materializing on schedule. We move Bear #2 to MATERIALIZING.

9. Foundry: 14A on Schedule, Customers Still Unnamed

The technology milestones held. The 14A 0.9 PDK shipped in October as promised, and the CEO called it the company's highest-quality 0.9 PDK, with the 1.0 PDK expected in Q1 2027. 18A-P remains on track for launch across internal products by year-end, and the CEO cited inbound external interest. On customers, he said he has consistently maintained that Intel will not disclose them, and said "a couple of customers" would be ready to announce within two quarters, in an answer that began with advanced packaging. The marquee 14A relationship introduced in April went unmentioned for a second straight call.

"As is standard practice in the foundry business, I have consistently maintained that we will not disclose customer details. That being said, I remain highly encouraged by the progress we are making with multiple potential customers on 14A." — Lip-Bu Tan, CEO

Assessment: The non-disclosure policy is standard for a foundry, but it sits awkwardly with April, when management chose to name its 14A anchor. Leaning on the policy after that anchor went quiet does not prove anything changed, and it does leave investors with nothing to underwrite. The two-quarter announcement window is now the most specific foundry commitment management has made, and we will hold it to that. Bull #3 stays AT RISK.

10. Purpose-Built Silicon and the GPU D-Series

Purpose-built silicon, Intel's custom ASIC business inside DCAI, grew more than 50% sequentially and more than doubled year over year, and other DCAI revenue reached $1.4B. Management expects it to grow faster than DCAI in Q4 and to triple again over the next few years, with AI labs named as a new customer category. For the first time, the CFO addressed margins: they are "below the corporate average." Separately, the GPU D-series (formerly Crescent Island), an inference GPU, is drawing early interest, and the CEO said engagements are "progressing towards firm commitments."

"Purpose-built silicon remains our fastest-growing business segment, tripling over the last 3 years, and poised to triple again over the next few years." — Lip-Bu Tan, CEO

Assessment: The margin answer was one of our July watch items, and it confirms the expected trade-off: this is a gross-profit-dollar business, not a margin-accretive one, and management said it would accept some operating-margin pressure for the growth. That is a reasonable choice. It also means ASIC growth helps earnings less than its revenue growth suggests. The GPU is an option, not a model input.

11. GAAP Noise Reverses: The Escrowed Shares Gain

In Q2, a $12.5B mark-to-market loss on the Escrowed Shares liability turned a strong operating quarter into an $11.0B GAAP loss. In Q3 the same liability produced a $1.0B gain as the stock fell from $128.32 to $123.00 over the quarter, which is why GAAP EPS ($0.60) exceeded non-GAAP ($0.52). The liability stood at $13.7B on September 26, with 132M shares still in escrow. Intel's 10-Q states the direction plainly: gains follow future share-price declines.

Assessment: With the stock now far below the quarter-end level, Q4 GAAP results will likely show a much larger non-cash gain, producing headlines that run opposite to the business trend. As in Q2, we adjust it out. The government stake is still a capital backstop, and its accounting cost now falls when the stock falls.

12. Long-Dated Options: Classiq and Ohio

Intel announced plans to acquire Classiq, an Israel-based quantum software company, to pair its software platform with Intel's spin-qubit hardware. The CEO said Intel would become a "full-stack revenue-generating quantum computing company" once the deal closes. In Ohio, the first major structural truss is installed, and management is evaluating "unique and highly differentiated opportunities" to accelerate the build-out, with more detail promised soon.

Assessment: Neither item affects the next eight quarters. Classiq is small optionality with no disclosed price, and any Ohio acceleration would add capex to a 2027 that is already unlikely to be free-cash-flow positive. An Ohio announcement tied to an external customer commitment would be a different matter, and that is the version worth watching.

Guidance & Outlook

The guide is a two-speed business stated in one number. DCAI is guided up mid-teens, with purpose-built silicon growing faster, and management said that without Fab 28 the server momentum would have carried straight through Q4. CCPG is guided down more than 20% as OEMs cut weeks of inventory. The CFO's framing for 2027 was "a growth year," led by data center and constrained in the first half, against a client market planned down about 10% in units with the inventory correction stretching at least into Q1. The CEO's framing was longer-dated: a server CPU TAM that could grow fivefold and a roadmap that reaches competitive parity in 2028.

"Looking ahead, we see the PC slowdown continuing at least through the first half of next year, but we remain optimistic about the opportunities across our broader portfolio, and we will continue making disciplined investments in capacity and working capital to increase supply and support future growth." — David Zinsner, CFO

Implied full year: Nine-month revenue of $46.2B plus the $15.5B Q4 midpoint puts 2026 at about $61.7B, roughly 17% above 2025's $52.9B and inside the $61-63B range we set in July. That range assumed Q4 would step up from Q3. The guide instead implies a second half of $32.0B against a first half of $29.7B, with the growth coming entirely in Q3. Non-GAAP EPS of $0.29, $0.42 and $0.52 plus the $0.25 guide sums to $1.48 for 2026.

Street at: Consensus sat at ~$17.1B revenue and $0.43 EPS for Q4, so the guide resets the quarter by about 9% on revenue and 42% on EPS. The larger revision is to 2027, where Q1 now carries a seasonal client decline, a second ~$300M Fab 28 impact and a supply-capped DCAI. A Q4 non-GAAP EPS run-rate of $0.25 annualizes to $1.00 before any recovery.

Guidance style: Intel guided conservatively all year and beat the revenue midpoint by 11.3% in Q1, 12.8% in Q2 and only 1.3% in Q3. This guide is different in kind. Management did not shave the quarter for safety; it described a known inventory flush and an incident in progress, and the CFO repeatedly called the Fab 28 figure an early assessment. We would not assume the usual cushion. If anything, the 10-Q's "through the first half of 2027" language suggests the Fab 28 risk is wider than the call's $300M per quarter.

Analyst Q&A Highlights

The Shape of 2027

The opening question asked management for guardrails on 2027, suggesting DCAI could add $10-15B while PC declined modestly. The CFO declined to guide but called 2027 a growth year led by data center, with the first half still supply-constrained and the client inventory correction extending at least into Q1. He closed with a longer-term view that the PC would go through a "renaissance" as agentic workloads move to client devices.

Q: "I mean it seems like DCAI could grow $10 billion to $15 billion year-over-year, but PC, I'm guessing, is maybe down a touch. So can you just maybe give us guardrails for the year and sort of how to think about the shape of the year?"
— Timothy Arcuri, UBS

A: "I would say at a high level, we see next year as a growth year. We feel like the momentum that we saw in data center this quarter, and taking away the impact from Israel, that momentum would have continued into the fourth quarter, we see that continuing next year. The only little nuance is the first half of the year, we are still a bit constrained."
— David Zinsner, CFO

Assessment: "Growth year" is a low bar after a 2026 that will finish near $61.7B with a falling Q4. The useful content is the shape: a supply-capped first half in DCAI coinciding with the worst of the client correction. That makes the first half of 2027 the earnings trough, and it argues against buying the stock on the expectation of a fast rebound.

Free Cash Flow in 2027

A follow-up asked directly whether Intel is committed to positive free cash flow in 2027 given rising capex. The CFO's answer was direct, with a light hedge: 2027 is an investment year, and positive free cash flow is probably not achievable. He tied the spend to ramping 18A, adding Intel 3 and preparing 14A for 2028 production, all at once.

Q: "But I guess the real question is, are you committed to generating free cash flow next year? Or do you think because of the CapEx burden that, that will be hard?"
— Timothy Arcuri, UBS

A: "I think next year is an investment year for us as we look at the year. Probably not a dynamic where we're going to have positive free cash flow next year. But it's because we see this as a tremendous opportunity to invest into the growth."
— David Zinsner, CFO

Assessment: A clear answer, and the most consequential one on the call for the equity. In July, the CFO said back-end investment could make reaching positive free cash flow in 2027 "a little bit more challenged." Now he expects that 2027 probably will not reach it. The financing is in place, so this is not a solvency issue, but it pushes the free-cash-flow inflection that a premium multiple needs into 2028 at the earliest.

Sizing the Guide Cut and the Q1 Tail

One question asked management to split the roughly $1.8B gap to the Street figure the analyst was using between Fab 28, client and execution, and to say when each improves. The CFO sized Fab 28 at ~$300M in Q4 and a similar amount in Q1, with most scrap in Q4. For client, he said the Q4 level would reset the base and Q1 would then decline seasonally from it.

Q: "And in the context that the guide is about $1.8 billion below the Street, maybe you can kind of go through those 3 things, how big they are to the cut of $1.8 billion and when they get better, that would be great."
— Ben Reitzes, Melius Research

A: "But I think the current thinking is that we'll have about a $300 million impact to the fourth quarter. That's likely to be similar in the first quarter. We think most of the write-offs of wafers or scrapping of wafers is likely to all kind of happen in the fourth quarter."
— David Zinsner, CFO

Assessment: The answer converts a one-quarter guide into a two-quarter problem. With Fab 28 repeating and client declining seasonally from an already reset base, Q1 2027 revenue is likely below the Q4 midpoint unless DCAI growth offsets both. That is the number the Street will rebuild first.

Gross Margin Drivers Into 2027

A detailed question listed the moving parts: client under-utilization, rising depreciation and still-dilutive 18A and Intel 3 parts on one side, and server mix and pricing on the other. It asked how they net out. The CFO went through them in turn (Granite Rapids accretive, 18A under-loaded, purpose-built silicon below the corporate average) and fell back on a fall-through rule rather than a number.

Q: "I'm just trying to get a feeling of the gross margin evolution in this kind of an environment given all of those puts and takes. It just seems like there's a lot of them."
— Stacy Rasgon, Bernstein Research

A: "I think when you put all that together, and it's so many puts and takes, it's hard to quantify, I still lean towards the 40% to 60% fall-through. That's probably what we're shooting for. So if we have a growth year next year, we should see gross margin expansion given that kind of rule of thumb."
— David Zinsner, CFO

Assessment: The rule of thumb works when revenue grows, and it implies gross margin expansion only if 2027 revenue rises meaningfully above 2026. With client down and DCAI supply-capped in the first half, the first two quarters of 2027 may run below the 2026 average before the second half recovers. The answer also confirmed that ASIC growth will hold back the margin percentage by design.

Did Customers Over-Buy in Q2, and Will Memory Improve?

A follow-up called Q3 probably the worst third quarter for PCs, sequentially, in at least twenty years, asked management to confirm that customers had built inventory in Q2, and asked what gives it confidence that memory supply improves. The CFO agreed that Q4 and Q1 carry most of the correction, but he did not directly confirm a Q2 inventory build. On memory, he was blunt that 2027 will not improve.

Q: "So it does sound like customers probably built inventory in Q2. I guess can you confirm that? And then it sounds like that they're flushing out a bunch in Q4."
— Stacy Rasgon, Bernstein Research

A: "But make no mistake, we are not optimistic about next year in terms of memory. We think that it is going to impact the TAM. We think TAM will be down 10% on a unit basis next year."
— David Zinsner, CFO

Assessment: The pessimism on memory is credible and makes the 2027 plan easier to believe. The unanswered half matters too. If customers built inventory in Q2, part of the Q2 client strength that helped the July beat was borrowed from Q4, so the quality of that quarter's beat was lower than it looked.

Is $8.5B a Quarter a DCAI Floor?

A question annualized the implied Q4 DCAI figure of roughly $8.5B to the mid-$30Bs and asked whether that is a floor for 2027 and whether ASPs could expand further. The CFO accepted it as the right starting point, again flagged the tighter first half, and noted that purpose-built silicon, which sits inside DCAI, drives much of the Q4 growth.

Q: "And if I annualize that, that's like mid-30s billion in terms of a conceptual target for next year. So I'm curious, is that a doable floor? Can you do better?"
— Vivek Arya, Bank of America Securities

A: "Yes. I think that's a good starting point. I think that's the right way to think about it. What we would characterize is that the first half of '27 is likely to be tighter in terms of wafers and substrates."
— David Zinsner, CFO

Assessment: This is the most bullish data point management gave, and we accept it: a $33B-34B DCAI base for 2027, against about $27B in 2026, is real growth at a segment margin above 40%. The caveat is mix. A growing share of that base is lower-margin custom silicon, so DCAI profit grows more slowly than DCAI revenue.

When Foundry Customers Become Revenue

A question asked whether external customers would come on 18A-P or only on 14A, and when engagements would move from evaluation to tape-out and revenue so investors could model them. The CEO described packaging progress, inbound 18A-P interest and the 14A PDK schedule, and promised announcements within two quarters. He did not name a customer or give a revenue timeline.

Q: "Are you expecting external customers for 18A-P? Or should we assume that all the engagements will be on 14A? And when do you think Intel gets the confidence to move them from sort of the evaluation to the tape-out or tangible revenue column?"
— Vivek Arya, Bank of America Securities

A: "And then overall, I think we're going to have some -- stay tuned, we're going to have a couple of customers ready to announce in the next 2 quarters. And then, so clearly, we always want to wait for the customer. If they want to announce, we will respond to that."
— Lip-Bu Tan, CEO

Assessment: The same question drew the same kind of answer in July, with one improvement: a time window. Because the commitment came inside a packaging discussion, it may describe packaging customers rather than wafer customers, and only a wafer customer resolves the merchant-foundry question. We will score the January and April calls against this answer.

Server Share Against AMD and Arm

A question set management's 30% industry unit forecast against a competitor guiding to 70% revenue growth and fast-growing Arm designs, and asked how share would play out. The CEO pointed to the roadmap and parity at Coral Rapids. The CFO gave the more useful answer.

Q: "How do you see server market share next year? I think you talked about 30% unit growth. AMD has guided to 70% revenue growth. The ARM camp seems to be growing quite quickly."
— Joseph Moore, Morgan Stanley

A: "I'd say maybe one other addition is I really see market share next year being a function of supply. And so the more we can improve our supply on a relative basis, the better we'll do in terms of market share."
— David Zinsner, CFO

Assessment: The answer is candid and implies Intel loses unit share in 2027 for the same reason it did in 2026. Combined with the CEO's 2028 parity timeline, the share recovery case rests on supply arriving in the second half of 2027 and on Coral Rapids landing on time. Until then, DCAI's growth comes from a market that is growing faster than Intel.

What They're NOT Saying

1. A foundry breakeven year: Still no date. This call offered only "considerable work ahead of us to deliver a sustainably profitable Foundry business," in a quarter when the segment still lost $1.85B.
2. A named external wafer customer with a commitment: The CEO cited a non-disclosure policy and promised announcements within two quarters, but named no one. The 14A anchor introduced in April has now been absent for two straight calls.
3. The full cost of Fab 28: The call gave ~$300M of revenue in each of Q4 and Q1 and 200 bps of Q4 margin. The 10-Q says the company cannot yet estimate the financial impact or insurance recoveries and expects effects through the first half of 2027.
4. Whether customers over-bought in Q2: Asked directly to confirm a Q2 inventory build, the CFO described the correction but did not confirm the build. That matters for how much of the July beat was channel fill.
5. Server unit share: No share figure was offered. The 10-Q's 10% server volume growth, against a stated 30%-plus industry forecast, says more than the call did.
6. A 2027 capex number: "Up strongly," with equipment orders up nearly 40% sequentially and $19.6B of capital commitments, but no figure. An analyst on the call cited ~$30B as the Street's working number, and management did not correct it.
7. A Q1 DCAI view: The CFO declined to guide Q1 data center, saying he preferred to wait another quarter. With Fab 28 repeating in Q1, that is the line most at risk in the next guide.
8. The Classiq price: The acquisition was announced without terms.

Market Reaction

• Pre-print setup: INTC closed at $92.07 on October 29 ahead of the after-close release. It entered the print up 149.5% year to date and 122.7% over twelve months, but down 20.6% over the trailing 30 days and 34.7% below the June 22 52-week closing high of $140.94. The 52-week closing range was $33.62 to $140.94. The S&P 500 was up 12.8% year to date.
• After hours (October 29): Shares fell more than 20% in extended trading once the Q4 guide crossed, with reported declines of about 23%.
• Reaction session (October 30): The stock opened at $68.35, a 25.8% gap down, traded between $63.20 and $68.77, and closed at $65.76, down 28.6% ($26.31), on about 445M shares, 4.2x the 30-day average. It was the largest one-day decline in the daily price history we track back to December 2006, deeper than the 26.1% fall on August 2, 2024.
• Peers and tape: The S&P 500 fell 2.1% and the Nasdaq Composite 2.8%, so INTC trailed the index by about 26.5 points. The iShares Semiconductor ETF fell 8.3%. PC- and x86-exposed names fell hardest: AMD and Dell each lost 10.9%, Qualcomm 11.1%, Arm 10.0% and HP 9.6%, while Nvidia fell 4.2%.
• Since the Q2 print: From the $100.23 close before the July 23 release, the stock has fallen 34.4% while the S&P 500 gained 2.1%, and it closed 30.8% below the $95.00 price of the August equity offering.

Not a sell-the-news this time: a reset of the earnings base. In July, a blowout quarter failed to lift a stock priced for perfection. This quarter is a different mechanism. The market is repricing the earnings base: a Q4 EPS guide 42% below consensus, a Q1 that carries a second Fab 28 hit and a seasonal client decline, and a 2027 that the CFO expects will probably not be free-cash-flow positive. A fall of this size implies the market had been capitalizing a 2027 in which DCAI growth and client stability compounded together. The guide removed the second half of that assumption.

Positioning amplified it. The stock had more than doubled in 2026 before falling 20.6% in the 30 days before the print as investors questioned the scale of AI spending, so holders were already nervous and many were still sitting on large gains. The August offering added about 242M shares bought at $95, and those holders are now well under water. On the same morning, broad technology weakness and a heavy AI-spending print elsewhere in large-cap tech added to the selling, and PC-exposed names fell with Intel.

What the price now discounts. At $65.76, the equity is worth about $349B on 5.31B shares. With $54.0B of cash and investments against $50.1B of debt, enterprise value is close to that figure. That is about 44x 2026 non-GAAP EPS of $1.48 and about 45x the midpoint of our reset 2027 range. The drop removed the excess, not the premium.

Street Perspective

Debate: A Two-Quarter Inventory Flush, or a 2027-Long Earnings Reset?

Bull view: The client weakness is an inventory correction concentrated in Q4 and Q1, Fab 28 is a one-off, and DCAI keeps growing through both. By the second half of 2027, client sell-in matches consumption again, Nova Lake gains share and substrate capacity releases server supply, so earnings snap back and the drop is the entry point.

Bear view: Memory constraints persist through at least the first half of 2027, the PC market shrinks again, and the inventory flush is followed by a smaller market rather than a recovery. Combined with a supply-capped DCAI, rising depreciation and a full share count, 2027 earnings land well below where models stood a week ago, and estimate cuts arrive for several quarters.

Our take: The bear framing fits the first half of 2027 better, and the bull framing fits the second half. Management itself plans for a smaller 2027 PC market and a tighter first half, so the sequence is cuts first, recovery later. That argues against buying ahead of the January guide, which will show Q1.

Debate: Is DCAI's Supply Constraint Pricing Power or Lost Share?

Bull view: Demand exceeding supply lets Intel raise prices and lift mix, which is why DCAI margins reached 41.5%. Customers want long-term agreements, a large cloud provider announced a multiyear collaboration, and when Intel 3 and substrate capacity arrive in the second half of 2027, units follow at those higher prices.

Bear view: Intel's server volume is growing at a third of the industry rate. Competitors are filling the sockets Intel cannot supply, Arm designs are gaining at hyperscalers, and parity at the product level is two years away. Scarcity pricing fades when supply loosens, while lost sockets stay lost.

Our take: Both are true at once, which is why we treat DCAI as confirmation of the demand thesis but not of a share recovery. The margin is real now. The share position is weaker than revenue growth suggests, and the timing of supply relief makes 2027 a holding year for share rather than a regaining one.

Debate: After a Drop of This Size, Is the Reset Priced?

Bull view: The stock has given back a large part of its 2026 gain in a single session and trades well below the August offering price. The core asset (a supply-constrained server CPU franchise growing above 70%), a funded balance sheet and a 14A roadmap on schedule are intact, and estimate cuts are quickly priced. Buying after a capitulation day has worked for this name before.

Bear view: The multiple did not compress, because estimates fell further than the price. At more than 40x a 2027 that probably will not be free-cash-flow positive, with a Foundry still losing nearly $2B a quarter and no named external wafer customer, the stock still discounts a successful turnaround. More cuts follow the January guide.

Our take: The stock is closer to fair than at any point since we downgraded, but not cheap. The January 2026 sell-off bought a guide miss that management could beat with supply it was already bringing online. This sell-off buys a client down-cycle and a supply ramp that both run into the second half of 2027, at roughly 3x book value. We would rather wait for the January print than try to time the bottom here.

Model Update

Valuation: The estimate cut is larger than the price cut. Our 2027 framework falls by roughly 40% at the midpoint, to $1.10-1.80 from $2.25-2.75, because the July framework assumed a Q4 step-up that has become a step-down, and because a full share count, a client trough and a supply-capped first half all land in the same year. At $65.76, INTC trades at about 44x 2026 EPS of $1.48 and about 45x the 2027 midpoint, against about 37x the old 2027 midpoint at the pre-print close. Our fair-value band moves to roughly $50-75 from $85-110. The low end applies ~45x to the bottom of the 2027 range, a premium multiple for a trough year in a supply-constrained market. The high end applies ~30x to a normalized 2028 of about $2.50, when supply relief, Nova Lake and 14A risk production should all be visible. The stock sits inside that band, a little above its midpoint.

Thesis Scorecard Post-Earnings

Overall: The business thesis is unchanged and the earnings thesis is weaker. The two pillars that describe what Intel is becoming (18A execution and DCAI demand) both held, and the 14A roadmap hit its milestone. What changed is the earnings base those pillars sit on: the client business that funded the turnaround has entered a memory-driven down-cycle, an accident removed server output for two quarters, and 2027 is now unlikely to be free-cash-flow positive. Bull #3 still lacks its proof, and two new bear points enter the scorecard. The stock repriced in a single session, but our estimates fell by more.

Action: Hold. Do not add into the drop and do not chase it lower. Holders who trimmed after July are well placed and should keep a core position. New capital should wait for the January print, which will show Q1 with its Fab 28 and client effects, the 2027 capex figure and whether the first foundry announcement has arrived. A named external wafer customer, Q1 client sell-in that stabilizes against consumption, or a price in the low $50s with DCAI intact would move us to Outperform. A Q1 guide with DCAI falling sequentially, Fab 28 effects extending past the first half of 2027, or no customer announcement by the April call would move us to Underperform.