Intel Q3 2026 Earnings: Can Foundry Growth Power a Turnaround?

 

Intel reports third-quarter results after the US close on October 29, 2026, with the conference call at 5:00 p.m. Eastern (2:00 p.m. Pacific), corresponding to roughly 5:00 a.m. Singapore time on October 30. Results had not been published as of this article's verification point.

 

One figure in the foundry expansion story is widely misread. Intel Foundry's second-quarter segment revenue of $5.8 billion, up 31% year over year, gets cited broadly as evidence of progress in the foundry business. But only $293 million of that segment revenue came from external customers, roughly 5%, with the remaining 95% or so representing Intel's product groups purchasing internally from the foundry.

 

Converting that, external foundry revenue equals roughly 1.8% of the company's total revenue of $16.1 billion. Over the same period the segment posted an operating loss of $2.1 billion, roughly 7.2 times its external revenue.

 

That does not mean the foundry business is making no progress. External revenue grew from $22 million a year earlier to $293 million, roughly 13.3 times, an extremely fast rate. But in judging whether results can support the foundry expansion thesis, the $5.8 billion segment figure supplies almost no information, while the $293 million line supplies all of it.

 

This article covers the second-quarter baseline, where third-quarter guidance sits relative to consensus, the change in the pace of margin improvement, and the metrics to check line by line tonight.

 

 

 

Foundry external customer revenue was $293 million, roughly 5% of the segment's $5.8 billion. The remaining 95% or so is intersegment, leaving external revenue at roughly 1.8% of total company revenue.

 

The foundry segment lost $2.1 billion, roughly 7.2 times its external revenue. The loss narrowed by $1.1 billion year over year, with operating margin improving from negative 71.7% to negative 36.2%.

 

Roughly 84% of the data centre segment's operating income is offset by the foundry loss. That segment earned $2.5 billion against the foundry's $2.1 billion loss.

 

Consensus sits above the upper half of company guidance. Third-quarter guidance is $15.8 billion to $16.8 billion with a $16.3 billion midpoint against consensus of $16.49 billion, and guidance of $0.38 EPS against consensus of $0.40.

 

The year-over-year pace of margin improvement has narrowed sharply. Second-quarter non-GAAP gross margin rose 12.1 points year over year while third-quarter guidance implies a rise of just 2.0 points.

 

The second quarter's $11.0 billion GAAP net loss came from a non-cash item. It was driven chiefly by a mark-to-market adjustment on escrowed shares tied to the CHIPS Act Secure Enclave agreement, roughly $2.45 per share, against non-GAAP net income of $2.2 billion.

 

2026 capital expenditure is planned above $20 billion against roughly $19.7 billion of net proceeds from the August share offering. The raise equates to roughly one year of capital expenditure.

 

 

Results come after the US close on October 29 with the conference call at 5:00 p.m. Eastern (2:00 p.m. Pacific) the same day, which is roughly 5:00 a.m. Singapore time on October 30.

 

One source conflict needs noting: a third-party source records the call time as 2:00 p.m. Eastern, while the time corresponding to the company's investor relations page is 5:00 a.m. Singapore time, meaning 5:00 p.m. Eastern. The two differ by three hours, so anyone joining the call should rely on the company's investor relations page.

 

The relative position of guidance and consensus sets the bar for a beat.

 

 

Consensus sits above company guidance on both revenue and EPS, meaning merely reaching the guidance midpoint would register as a miss against market expectations, and Intel needs to land in the upper half of its own range to meet them.

 

The market has grounds for that positioning. Second-quarter revenue of $16.1 billion came in $1.8 billion above the April guidance midpoint, roughly 12.6%; non-GAAP gross margin of 41.8% ran 2.8 points above guidance; and non-GAAP EPS of $0.42 came in $0.22 above guidance. Consecutive substantial beats against its own guidance explain why expectations have moved up, while also compressing the margin for error.

 

On the share price, the stock closed at $93.74 on October 27, up 3.61% on the day. Across 47 analysts, the average target is $117.71, implying roughly 25.6% upside, with a high of $200.

 

The dispersion among analysts deserves separate note. New Street Research upgraded to Buy with a $125 target, while JPMorgan upgraded from Underweight to Neutral with a target of $85, roughly 9.3% below the October 27 close. The high target of $200 is 2.35 times that $85 figure. Dispersion of that width indicates no market consensus on the eventual value of the foundry business rather than disagreement about the current quarter.

 

For the full baseline on second-quarter results and third-quarter guidance, see our earlier piece on Intel's second-quarter earnings and guidance.

 

 

This is the misreading this article considers most in need of correction.

 

 

The shares and intersegment figure above are this article's calculations from disclosed data rather than company disclosures.

 

Segment revenue and external revenue measure two different things. Segment revenue includes the internal transfer price Intel's product groups pay the foundry, which creates no new external demand and represents only an internal accounting division. So "foundry revenue up 31%" principally reflects shipment growth in Intel's own products rather than third-party customer adoption.

 

What actually measures foundry expansion is the $293 million line. Its growth rate is very high, roughly 13.3 times from $22 million, but off a tiny base, standing at roughly 1.8% of total company revenue.

 

Another ratio worth recording concerns losses against external revenue. The segment's $2.1 billion operating loss is roughly 7.2 times its $293 million of external revenue. That ratio is not a unit economics measure, since the segment simultaneously serves internal demand and carries the company's manufacturing assets and advanced node investment. But it illustrates one thing plainly: the scale of the external foundry business and the scale of the losses being absorbed for it currently differ by roughly an order of magnitude.

 

Viewed from another angle, the loss is narrowing fast. Operating loss fell $1.1 billion year over year with operating margin improving from negative 71.7% to negative 36.2%, a 35.5 point improvement. At that pace the loss is narrowing faster than external revenue growth can explain, indicating the improvement comes mainly from higher utilisation and internal shipment volume.

 

 

The company disclosed the following node progress in the second quarter. The 18A node has successfully ramped with improving yields and cycle times, and 18A-P risk production has progressed. For 14A, risk production is planned for the second half of 2027, high-volume manufacturing for 2028, and process design kit version 0.9 was delivered in October 2026.

 

What that timetable implies is that external customer revenue on 14A cannot reach scale until 2028, and the October 2026 design kit delivery is a precondition for customers to begin designing rather than an order. So any external customer news tonight needs sorting by type, since design collaborations, letters of intent and signed capacity commitments sit far apart financially.

 

 

Placing the three segments' results side by side makes the company's current profit structure clear.

 

 

Data Center and AI is the engine of this growth cycle, with revenue up 59% year over year and 24% sequentially, operating margin rising from 16.1% to 39.5%, and operating income rising from $0.6 billion to $2.5 billion.

 

But roughly 84% of that $2.5 billion is offset by the foundry's $2.1 billion loss, a ratio calculated here from the two segments' disclosed figures. The three segments' operating income sums to roughly $2.7 billion, broadly consistent with the $2.8 billion of non-GAAP operating income the company reported.

 

What this structure means is that the company's current profitability actually rests on the client computing business, while nearly all of the data centre segment's profit absorbs the foundry's loss. That is not a criticism of the strategy, since manufacturing capability is the precondition for a foundry business and advanced node investment necessarily shows up as losses. But it means two things need watching this quarter: whether data centre growth of 59% holds, and whether the foundry loss keeps narrowing at roughly $1 billion a quarter. Should the former slow while the latter stalls, consolidated profit would come under pressure quickly.

 

 

Second-quarter GAAP net loss was $$11.0 billion with diluted loss per share of$$(2.16), against non-GAAP net income of $2.2 billion and EPS of $0.42.

 

The difference comes chiefly from one non-cash item: a mark-to-market adjustment on escrowed shares tied to the CHIPS Act Secure Enclave agreement, worth roughly $2.45 per share, which on the share count works out to roughly $12.5 billion in magnitude. That item is a mark-to-market measure whose movement depends on the price of the relevant shares rather than operating performance.

 

Operating income shows the improvement more clearly: non-GAAP operating income was $2.8 billion in the second quarter against a $0.5 billion loss a year earlier. GAAP gross margin was 40.4%, up 12.9 points, and non-GAAP gross margin 41.8%, up 12.1 points.

 

So any statement that Intel lost $11 billion requires first establishing which basis is in use. On an operating basis, the second quarter marked a significant turn to profit.

 

 

Several previews call this quarter a margin test, and placing the two year-over-year figures together shows why.

 

 

The absolute level is rising, from 41.8% to 42.0%. But the year-over-year improvement narrows from 12.1 points to 2.0 points, shrinking to roughly one sixth.

 

The reason is the base. The prior-year margin sat at a low level, producing a large year-over-year improvement in the second quarter, and as the comparison base rises the remaining improvement has to come from genuine cost structure optimisation rather than a low base. Put another way, the easy part is over.

 

That also explains why margin is the most important item this quarter. Per third-party previews, the margin trajectory depends on product mix, factory utilisation and the cost of advanced node investment, of which the first two are cyclical and the third structural. Meeting the 42.0% guidance would indicate that utilisation gains are offsetting the drag from advanced node investment, while falling short would indicate 18A ramp costs running above expectations.

 

 

The company raised its 2026 capital expenditure plan to above $20 billion, roughly $3 billion above earlier expectations, citing AI-related server demand.

 

On funding, the company completed a share offering in August 2026. Per our earlier compilation, the deal was initially proposed at $15 billion then upsized to $20 billion, pricing roughly 210.5 million new shares at $95 each for expected net proceeds of roughly $19.7 billion before any exercise of the underwriters' option on a further 31.6 million shares, closing on August 12. Against roughly 5.044 billion shares outstanding as of July 17, the base offering implies roughly 4.2% dilution. Proceeds are for general corporate purposes including capital expenditure and working capital. For the relationship between that offering and the foundry expansion, see our analysis of Intel's financing and foundry expansion.

 

Placing the two figures side by side: roughly $19.7 billion of net proceeds equates to roughly one year of capital expenditure. That means if capital expenditure stays above $20 billion a year while operating cash flow does not cover it, further external financing or asset disposals would still be required.

 

One cash flow figure needs explaining. A third-party source gives second-quarter operating cash flow of $7.0 billion, but the same source labels negative $8.4 billion as both "adjusted free cash flow" and "second-quarter foundry cash flow," definitions that conflict, so this article does not cite that free cash flow figure. An accurate comparison of operating cash flow against capital expenditure should come from the company's quarterly filing.

 

 

First, the foundry segment's external customer revenue. It was $293 million in the second quarter. This is the only direct measure of foundry expansion, and the $5.8 billion segment figure does not carry equivalent information.

 

Second, non-GAAP gross margin against the 42.0% guidance. Note that year-over-year improvement has already narrowed from 12.1 points to 2.0 points, so the tailwind from a low base is largely finished.

 

Third, revenue against consensus of $16.49 billion rather than the $16.3 billion guidance midpoint. Consensus sits above the midpoint, so merely reaching the midpoint would register as a miss on market terms.

 

Fourth, Data Center and AI growth alongside the pace of foundry loss reduction. The second quarter delivered 59% growth and an $1.1 billion reduction respectively, with roughly 84% of that segment's $2.5 billion profit offset by the foundry loss.

 

Fifth, whether fourth-quarter guidance and the full-year capital expenditure plan are adjusted. The current plan is above $20 billion.

 

Sixth, any external foundry customer news needs sorting by type. Design collaborations, letters of intent and signed capacity commitments sit far apart financially, and 14A high-volume manufacturing is scheduled for 2028.

 

On MEXC, review the contract rules, leverage tiers and position limits for the Intel stock perpetual, noting that books typically thin considerably around earnings, so check trading sessions and risk limits before submitting an order.

 

 

For James Mitchell, discussion of this company carries a recurring statistical error, and it happens to sit on the most important metric.

 

"Intel Foundry revenue up 31% to $5.8 billion" is technically correct and informationally close to empty. Roughly 95% of that segment revenue is the internal transfer price Intel's product groups pay the foundry. That money moves from one of the company's pockets to another and constitutes no new external demand. What genuinely comes from third-party customers is $293 million, roughly 5% of segment revenue and 1.8% of company revenue.

 

So when someone asks whether results can support the foundry expansion thesis, the first step is establishing which line to read. Read the $5.8 billion and the answer is that the foundry business is large and growing. Read the $293 million and the answer is that the external business started from near zero, is growing extremely fast, and remains very small. The latter answers the question.

 

I also want to stress that this should not be read as dismissal. Going from $22 million to $293 million is 13.3 times, a rate rarely seen in semiconductor foundry. And the loss is narrowing fast, down $1.1 billion year over year with operating margin improving from negative 71.7% to negative 36.2%. The issue is not direction but timetable: 14A high-volume manufacturing arrives in 2028, process design kit version 0.9 was only delivered in October 2026, and customers typically need years between receiving a design kit and contributing meaningful revenue. So for a number of quarters ahead, the foundry segment will remain a business sustained mainly by internal demand while consuming profit.

 

On the profit structure, I calculated one figure: roughly 84% of the Data Center and AI segment's $2.5 billion operating income is offset by the foundry's $2.1 billion loss. Put differently, this company's actual source of profit right now is the $2.3 billion from client computing, while the fastest-growing, highest-margin data centre business has nearly all of its profit subsidising the manufacturing transition. That structure is the result of a strategic choice and is not necessarily a problem, but it means two things must hold simultaneously: data centre must sustain high growth, and the foundry loss must keep narrowing. Should either fail, consolidated profit deteriorates quickly, and that would not show up on the revenue line.

 

The margin figures are, I think, underrated in importance. Second-quarter non-GAAP gross margin rose 12.1 points year over year while third-quarter guidance implies just 2.0 points. The improvement shrinks to roughly one sixth. The reason is that last year's base was extremely low and the base has now risen. From this quarter onward, every point of margin improvement has to come from genuine cost structure optimisation or higher utilisation rather than from a low base. The 42.0% guidance looks modest, but it is the first quarterly target without help from an easy comparison.

 

Two traps readers are most likely to fall into. First, the second quarter's $11 billion GAAP net loss came chiefly from a mark-to-market adjustment on escrowed shares under the CHIPS Act Secure Enclave agreement, roughly $2.45 per share, while non-GAAP operating income over the same period was $2.8 billion against a $0.5 billion loss a year earlier. On an operating basis that was a significant turn to profit, and treating the $11 billion as evidence of operating deterioration misreads the basis. Second, the dispersion in analyst targets: a high of $200 against JPMorgan's post-upgrade target of $85, roughly 9.3% below the October 27 close of $93.74. An upgrade accompanied by a target below the current price is itself a statement that the market has no consensus on the foundry's terminal value. Under that kind of dispersion, a single quarter's results cannot settle the valuation question; they can only settle the timetable question.

 

 

 

After the US close on October 29, 2026, with the conference call at 5:00 p.m. Eastern (2:00 p.m. Pacific) the same day, corresponding to roughly 5:00 a.m. Singapore time on October 30. Note that a third-party source records the call at 2:00 p.m. Eastern, three hours apart from the time corresponding to the company's investor relations page, so anyone joining should rely on the company's page.

 

 

Company guidance calls for revenue of $15.8 billion to $16.8 billion with a $16.3 billion midpoint, non-GAAP gross margin of 42.0%, and non-GAAP EPS of $0.38. Consensus is $16.49 billion of revenue, up roughly 21%, and adjusted EPS of $0.40, both above company guidance. So merely reaching the guidance midpoint would register as a miss on market terms, and the company needs to land in the upper half of its own range.

 

 

Two figures need distinguishing. Second-quarter segment revenue was $5.8 billion, up 31%, but roughly 95% of that is the internal transfer price Intel's product groups pay the foundry. Revenue from external customers was $293 million, roughly 5% of segment revenue and roughly 1.8% of company revenue. That external revenue grew roughly 13.3 times from $22 million a year earlier, a high rate off a small base. These shares are derived from disclosed data.

 

 

Second-quarter operating loss was $2.1 billion, narrowing $1.1 billion year over year, with operating margin improving from negative 71.7% to negative 36.2%. That loss is roughly 7.2 times its $293 million of external revenue. Note this ratio is not a unit economics measure, since the segment simultaneously serves internal demand and carries the company's manufacturing assets and advanced node investment. The loss is narrowing faster than external revenue growth can explain, indicating the improvement comes mainly from higher utilisation and internal shipment volume.

 

 

The Data Center and AI segment earned $2.5 billion of operating income in the second quarter against the foundry segment's $2.1 billion loss, offsetting roughly 84% of it. The three segments' operating income sums to roughly $2.7 billion, broadly consistent with the $2.8 billion of non-GAAP operating income reported. That means the company's actual profitability currently rests on client computing's $2.3 billion of operating income. These ratios are calculated here from disclosed data.

 

 

The loss came chiefly from one non-cash item, a mark-to-market adjustment on escrowed shares tied to the CHIPS Act Secure Enclave agreement, worth roughly $2.45 per share. Over the same period non-GAAP net income was $2.2 billion with EPS of $0.42 and non-GAAP operating income of $2.8 billion, against a $0.5 billion loss a year earlier. Mark-to-market items move with the price of the relevant shares rather than with operating performance, so on an operating basis the second quarter showed significant improvement.

 

 

Because the tailwind from a low base has ended. Second-quarter non-GAAP gross margin was 41.8%, up 12.1 points year over year, while third-quarter guidance is 42.0%, up just 2.0 points, shrinking the improvement to roughly one sixth. From this quarter onward every point of margin improvement must come from genuine cost structure optimisation or higher utilisation. Per third-party summaries, the margin trajectory depends on product mix, factory utilisation and the cost of advanced node investment.

 

 

The 2026 capital expenditure plan was raised to above $20 billion, roughly $3 billion above earlier expectations. The company completed a share offering in August 2026, upsized from an initial $15 billion to $20 billion, pricing roughly 210.5 million shares at $95 each for expected net proceeds of roughly $19.7 billion, implying roughly 4.2% dilution against roughly 5.044 billion shares outstanding, with proceeds for general corporate purposes including capital expenditure and working capital. Roughly $19.7 billion of net proceeds equates to roughly one year of capital expenditure.

 

 

The company disclosed in the second quarter that 18A has successfully ramped with improving yields and cycle times, and that 18A-P risk production has progressed. For 14A, risk production is planned for the second half of 2027, high-volume manufacturing for 2028, and process design kit version 0.9 was delivered in October 2026. That timetable means external customer revenue on 14A cannot reach scale until 2028, and a design kit delivery is a precondition for customers to begin designing rather than an order. Any external customer news tonight should be sorted between design collaborations, letters of intent and signed capacity commitments.

 

 

The information above is provided for general market and product information only and does not constitute investment advice, financial advice, legal advice, tax advice or a recommendation to buy or sell securities, nor does it constitute an endorsement of any issuer or trading product. This article is a preview written before the earnings release, verified before 16:30 Singapore time on October 28, 2026, at which point Intel's third-quarter 2026 results had not been published; all figures herein are previously published historical financial data, company guidance or third-party consensus estimates, and none constitutes a forecast of actual results. Consensus estimates, price targets and ratings come from third parties and may change at any time, different providers use different bases and samples that are not comparable, and the dispersion in targets referenced comes from different institutions at different times. Figures identified as this article's calculations or derivations, including the foundry segment's intersegment amount, the various external revenue shares, the ratio of foundry loss to external revenue, and the share of data centre profit offset, are not figures disclosed by the company; they are calculated from published data and actual results may differ on a different basis. Foundry segment revenue and external customer revenue measure different things, segment revenue includes internal transfer pricing, and the two cannot substitute for one another; the ratio of loss to external revenue illustrates relative scale only and does not constitute a unit economics measure, since the segment simultaneously serves internal demand and carries the company's manufacturing assets and advanced node investment. Non-GAAP gross margin, operating income, net income and earnings per share are non-GAAP measures defined by the company, are not comparable in absolute level with the GAAP basis, and are not directly comparable with other companies; the GAAP net loss includes non-cash items such as mark-to-market adjustments whose movement depends on the price of the relevant shares rather than operating performance. The conference call time carries a source conflict noted in the text, and anyone joining should rely on the company's investor relations page. Some third-party material contains internally conflicting definitions, and this article has excluded and not cited the relevant free cash flow figure. Process node timetables and manufacturing plans are set by the company and subject to adjustment, and a design kit delivery does not constitute an order or capacity commitment. Company guidance may be raised, maintained or cut. The stock perpetual contract referenced is a derivative rather than a share, conferring no ownership, voting rights or dividends, and leveraged products can produce losses exceeding the initial outlay, with order books typically thinning and volatility rising around earnings. Readers should consult the company's original filings before making any decision, judge against their own financial circumstances, investment objectives and risk tolerance, and consult a qualified professional where appropriate. The MEXC Crypto Pulse team accepts no liability for any direct or indirect loss arising from the use of this information.

 

 

James Mitchell specializes in technical analysis, market trends, and trading strategies for both Bitcoin and altcoins. Based in London, he has over 10 years of experience in financial markets. Before joining MEXC Learn, James worked as a senior analyst at a leading European investment firm, where he developed expertise in risk management and quantitative trading. His transition to cryptocurrency markets began in 2017, and he has since become recognized for his data-driven approach. He holds a Master's degree in Financial Economics from the London School of Economics. His analytical approach combines traditional technical analysis with on-chain metrics to provide readers with actionable insights.

 

Areas of Expertise: Technical Analysis, Market Trends and Cycles, Trading Strategies, Bitcoin and Altcoin Analysis, Risk Management.

 

 

 

Want the fastest access to MEXC's latest updates? Join our official Telegram group now!

MEXC Platform: Download App | Zero Fee Trading | Buy USDT Now

Join MEXC Community: X (Twitter) | Telegram | Discord

Get Insights: MEXC Learn | Terms of Service | MEXC Blog | MEXC News

Account Verification: Understand KYC | How to Complete KYC

External Content Platforms: Substack | Medium | Paragraph | LinkedIn | X(News)