For most of this year, the market had made up its mind about Intel. The company that once defined American computing had become the cautionary tale of the semiconductor age, the chipmaker that missed mobile, missed the graphics boom, and missed artificial intelligence while Nvidia rewrote the rules. In July alone the stock fell roughly 28 percent, swept up in the chip bear market that dragged the hardware complex lower on the seventeenth and eighteenth. Then, on Thursday evening after the bell, the company everyone had left for dead reported its fastest revenue growth since 2011, and the shares jumped more than 12 percent after hours. As of this writing, early Friday morning in New York, the regular session has not yet opened, and Intel is about to walk into it carrying a comeback nobody priced in.
What makes the timing extraordinary is not just the number, but what it reveals about the machinery underneath this market. The same force that punished Alphabet and Tesla earlier in the week is the force that just resurrected Intel. To understand why, stop thinking about the AI trade as a single bet and start seeing it as two opposing sides of the same ledger.
The AI Boom Has Split the Market Into Payers and Receivers
This week drew the line cleanly. On the Friday before (the seventeenth), the chip stocks slid into a bear market, and the logic seemed obvious: the sellers of AI infrastructure were being punished, the trade was tired, the boom was rolling over. Then on Wednesday the payers bled. Alphabet and Tesla reported, and the market fixed on the same wound in both: the colossal capital expenditure required to build AI capacity had pushed free cash flow negative.
But capex does not vanish. Every dollar of negative cash flow on Alphabet’s or Meta’s books is a dollar that lands somewhere else as revenue, buying servers, accelerators, networking gear, advanced packaging, and the wafers those chips are printed on. That is the receiver side of the ledger, and on Thursday night Intel proved it is standing squarely on it. The spending that hollowed out the hyperscalers’ cash flow is exactly what drove Intel’s fastest top-line growth in almost fifteen years. Intel, of all companies, turned out to be the receiver comeback in the flesh.
The Numbers Behind the Headline
Start with the top line, the cleanest signal in the release. Intel reported revenue of 16.1 billion dollars, precisely 16.128 billion, up 25 percent year over year. That is the fastest revenue growth the company has posted since 2011, and it blew past a consensus that sat near 14.4 billion. A beat of roughly 1.7 billion dollars for a company this size is not a rounding error; it is a sign the analysts modeling the business had the wrong demand curve.
The margins tell the same story with more force. Non-GAAP gross margin came in at 41.8 percent, up roughly 12 percentage points from a year earlier and 280 basis points above the company’s own guidance; GAAP gross margin rose to 40.4 percent from 27.5 percent. Non-GAAP operating margin swung to 17.2 percent from minus 3.9 percent a year ago. Adjusted earnings landed at 42 cents per share against a consensus near 21 to 22 cents, essentially a double, an earnings beat on the order of 90 percent. Non-GAAP net income was 2.2 billion dollars, reversing a roughly 400 million dollar loss a year earlier, and operating cash flow for the quarter was a robust 7 billion dollars.
The segment detail is where the AI thread becomes undeniable. The Data Center and AI group generated 6.3 billion dollars, up 59 percent year over year, a record rate of growth, and the AI-driven businesses in aggregate expanded more than 70 percent. Client Computing, the PC-chip franchise, contributed 8.9 billion dollars, up 13 percent. Intel Foundry, the contract-manufacturing arm, brought in 5.8 billion dollars, up 31 percent and 6 percent sequentially. The 18A process node, on which Intel has bet its manufacturing future, ran roughly 25 percent above its output target and more than 50 percent higher than the prior quarter. On demand and execution, this was a blowout.
The Eleven Billion Dollar Loss and How to Read It
And yet the same release carried a GAAP net loss of 11 billion dollars, or minus 2.16 dollars per share. Read only the headline and you would conclude the quarter was a catastrophe. Read the footnotes and you learn something more interesting.
That loss was driven by a non-cash mark-to-market charge of 12.5 billion dollars on the shares held in trust in connection with the CHIPS Act arrangement, the same deal under which the United States government took a roughly 10 percent stake in Intel in 2025. Before that non-operating charge, Intel’s GAAP operating result was actually positive, around 1.8 billion dollars, against a GAAP operating loss of 3.176 billion a year earlier. The business made money at the operating line. The 11 billion dollar loss is an accounting revaluation of a government equity position, not cash that walked out the door.
Here is where the week offers a near-perfect mirror. Just days earlier, Alphabet’s headline earnings per share were inflated by an unrealized book gain of 99 billion dollars on its own equity holdings. Same mechanism, opposite direction: one headline was puffed up by a mark-to-market gain, the other buried under one. In both cases the honest number was the operating number. Alphabet’s operating EPS narrowly missed; Intel’s operating result was a genuine profit. The lesson is worth stating plainly: read the cash flow and the operating line, not the headline. The market that punished Alphabet for spending and is now rewarding Intel for receiving that spending is, if it reads only the top line of each report, being fooled twice by the same quirk of GAAP.
The Comeback That Is Not Finished
None of this makes the turnaround complete. The Foundry, the single most expensive and most strategically important piece of the whole plan, still lost 2.1 billion dollars in the quarter. That is an improvement, narrowed from a 2.4 billion dollar loss in the prior quarter, a reduction of 348 million, but it is still a loss. The part of the strategy meant to make Intel a peer to TSMC in contract manufacturing is not yet profitable, and it is consuming enormous capital to get there.
That capital is not shrinking either. Intel raised its 2026 capex guidance to more than 20 billion dollars and signaled still higher spending in 2027. So Intel is, in its own way, a payer as well as a receiver, plowing over 20 billion dollars a year into fabs on the bet that the AI demand it is now capturing will still be there when the 18A and future nodes mature. The GAAP balance sheet is red because of the government stake, the Foundry is still bleeding, and the stock had already run up roughly 170 percent for the year through Thursday’s close, on top of an 84 percent gain in 2025. A great deal of comeback is already in the price. Vindication, yes, but partial.
A Blowout Into a Risk-Off Tape
The cruelest detail is that Intel delivered its best quarter in years into one of the worst market days in months. Thursday’s regular session, before Intel even reported, was ugly. The S and P 500 fell 1.2 percent, its biggest one-day drop in a month, and the Nasdaq 100 sank 1.9 percent, with Alphabet down 7 percent and Tesla down 14 percent after their results. An index of the largest megacap stocks logged its worst session since the tariff-driven selloff of April 2025, though that superlative belongs to the megacap group, not to the broad market.
Underneath the equity weakness sat something more dangerous. Brent crude pushed above 100 dollars for the first time since late May, touching 100.66 dollars, capping a July rally of roughly 40 percent. The trigger was geopolitical: Houthi attacks on two Saudi oil tankers, the Encelia and the Layla, in the Red Sea, with a fire aboard the Encelia (the crew was unharmed) and five tankers turning back. The Houthis declared a naval blockade against Saudi ports, following the near-total halt of traffic through the Strait of Hormuz. Goldman Sachs warned that Brent could exceed 120 dollars in the fourth quarter and average 100 dollars in 2027 if Hormuz stays disrupted. An oil shock of that size raises the cost base of every industry, chipmakers included, and can reignite inflation just as central banks thought the fight was won. Intel’s after-hours pop, more than 12 percent to around 112 to 113 dollars and briefly above 13 percent, happened against exactly that backdrop.
The Stocks That Matter Around This Print
For US investors, the read-through runs in several directions at once. Nvidia and AMD are Intel’s most direct rivals, and Intel’s data-center strength complicates the tidy narrative that the incumbent is being carved up. It does not dethrone Nvidia, whose accelerators still define the training market, but a 59 percent jump in Intel’s data-center revenue suggests the pie is growing fast enough to feed more than one mouth. Micron, levered to the same server build-out through memory, and Broadcom, a custom-silicon and networking beneficiary, sit on the same receiver side of the ledger that just lifted Intel.
The sharpest contrast, though, is TSMC. On the sixteenth of July the Taiwanese foundry reported a record quarter, and the stock fell anyway, a classic sell-the-news reaction after a long run. A week later Intel reports a comeback off a beaten-down base, and the stock rises. The difference is expectations: TSMC was priced for perfection and delivered it, while Intel was priced for decline and defied it. The same earnings quality can produce opposite reactions depending on what the market had already assumed.
There is also the unusual matter of the shareholder register. The US government owns roughly 10 percent of Intel, so the American taxpayer holds a direct stake in this turnaround, and the same 12.5 billion dollar charge that buried the headline is a revaluation of that public position. For any US investor booking gains on Intel’s run, the tax treatment matters: positions held a year or less are taxed as short-term capital gains at ordinary income rates, while those held longer qualify for the lower long-term rate, a distinction worth weighing after a stock has climbed 170 percent in a year.
The Risks and the Counterarguments
The bull case is straightforward and strong. The 59 percent gain in data center is a genuine demand signal, not an accounting artifact; the 18A node is running ahead of plan; gross margin jumped double digits; and 7 billion dollars of operating cash flow is real money. Management also raised guidance clearly: third-quarter revenue is now guided to a midpoint of 16.3 billion dollars, in a range of 15.8 to 16.8 billion, against a prior expectation nearer 15.1 billion, with non-GAAP EPS of 38 cents (versus an expected 27 cents) and gross margin around 42 percent. The operating turnaround is credible.
The bear case is equally honest. The Foundry still loses 2.1 billion dollars a quarter, the most expensive part of the strategy remains unprofitable, and the GAAP books show an 11 billion dollar loss even if the cause is non-cash. The stock has run hard and was already down 28 percent in July, so both euphoria and despair have been extreme within a single month. And the macro backdrop is hostile: oil above 100 dollars lifts everyone’s costs, a sustained shock could force central banks back into a tightening posture, and growth stocks suffer most when discount rates rise. As CEO Lip-Bu Tan framed the opportunity, “AI is driving unprecedented demand for compute, and as we continue to execute, Intel is well-positioned to capture sustainable growth across our CPU franchise, ASICs, advanced packaging and vast wafer foundry network.” The demand is real. The execution, node by node and quarter by quarter, is the part that still has to be proven.
What to Watch From Here
Intel has handed the market a clean thesis: the AI capex that hurts the payers lands as revenue at the receivers, and the operating line, not the GAAP headline, is where the truth lives. But a thesis is only as good as the confirmation that follows, and the calendar is dense. Microsoft and Meta report on the twenty-ninth of July, and their capex commentary will tell us whether the spending that feeds companies like Intel is still accelerating. Apple and Amazon follow on the thirtieth, and Nvidia, the keystone of the entire trade, reports at the end of August. Each of those prints is another test of whether the receiver side of the ledger can keep growing while the payers absorb the cost.
Layered on top is a macro wildcard that has nothing to do with silicon. The Section 122 tariff, the 10 percent duty imposed on the twenty-fourth of February, expires today, 150 days later, at one minute past midnight Eastern time, though the separate Section 232 and 301 tariffs are unaffected. And the oil shock keeps grinding, with the European Central Bank holding its deposit rate at 2.25 percent on Thursday, still wary after its June hike. The lesson of this week is that even a blowout does not guarantee a rally when it arrives inside a risk-off tape. Intel walks into today’s regular session with its fastest growth in almost fifteen years, a government on its shareholder register, a foundry still in the red, and an oil market threatening the cost base of the whole economy. The comeback is real. Whether the market is in any mood to reward it is a question only the trading day, which has not yet begun, can answer.
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.
