Intel Has Beaten Six Quarters Straight. None of Them Answered the Real Question.

The market has paid for the turnaround. The one foundry number on July 23 is what proves it, or exposes it.

Intel has cleared Wall Street’s bar for six quarters in a row. It also lost more than ten billion dollars in a single year inside the one business its whole future depends on. Both are true. Only the first is in the stock.

On July 23, Intel reports the second quarter. It is another chance to confirm the comeback, or to show that the part of the comeback the bulls are paying for is still missing. So before the print, we separate what Intel has proven from what it is asking you to assume.

The assumption needs proof, by the Firewall’s own grade

Our Stock Story Firewall run on Intel, dated 28 June 2026, names the bet and grades it Needs Proof. Investors are wagering that Intel’s 18A node and foundry platform will attract enough external customer volume to make the foundry profitable and restore Intel’s place in advanced manufacturing.

Read the grade. Needs Proof is not a knock on the story. It means the financial evidence has not yet confirmed the story is true.

Turnaround Analysis asks one thing: is the reset real and durable

The Firewall prescribes Turnaround Analysis. The question is not whether the narrative is compelling. It clearly is. The question is whether the operational and financial evidence confirms that the structural reset is real and that it lasts.

For Intel, that comes down to the foundry. The products business is healthier. The foundry is the bet, and the foundry is where the proof has to show up.

The case the bulls are buying

The momentum is real. First-quarter revenue of $13.6 billion beat the guidance midpoint, the sixth straight quarter of exceeding expectations. Intel guided the second quarter to $13.8 to $14.8 billion and, for the first time in a while, a GAAP profit at the midpoint. Xeon 6 won a design slot in NVIDIA’s DGX platform, a concrete sign Intel’s data-center CPUs still matter in AI builds. And management says 18A output is running ahead of internal projections, with 14A maturing faster at the same stage.

That is a company executing. None of it is hype.

The hole the bulls are stepping over

Now the foundry, by the numbers. In all of 2025, Intel Foundry generated $17.8 billion in revenue, but only $307 million came from outside customers, and the unit lost $10.3 billion. In the first quarter of 2026, the foundry booked $5.4 billion in revenue and still lost $2.4 billion, with most of that revenue coming from Intel building chips for itself.

So the business that is supposed to sell capacity to the world is, so far, mostly selling to Intel, at a loss measured in billions. Add roughly $50 billion of debt, and the margin for error is thin.

The crux is external revenue against a ten-billion-dollar loss

Here is the number the print turns on. External foundry revenue, the part outside customers actually pay, measured against that $10.3 billion annual loss and the break-even management hopes to reach by the end of 2027.

The card frames the test cleanly. Does external foundry revenue meaningfully exceed the $307 million full-year 2025 base, and is it accelerating? That is the difference between a foundry that is becoming a business and one that is still a subsidized internal workshop. Anything else in the release is secondary to that one line.

What the print will not give you is just as important. Yields at high volume are not disclosed in dollars. The economics of any signed customer are not laid out. So the print can move the thesis forward, but it cannot finish it.

A fair counterpoint

The bull case deserves a real hearing. Six straight beats is a pattern, not luck. The Xeon 6 win shows the product side is competitive in the exact market that matters. 18A running ahead of plan, and 14A maturing faster, is how a process comeback is supposed to look early. The first GAAP-positive guide suggests the worst of the bleeding may be passing. And if 18A lands a publicly named, high-volume external customer, the re-rating has room, because almost no one is modeling that yet.

That case is coherent. It also rests on a single proof that has not arrived: a real outside customer base, produced at competitive yields, growing fast off a tiny start.

The insiders are selling into it

One more fact belongs in the file. As the stock ran, two senior executives sold. The chief legal officer sold 40,256 shares on May 1 at about $99.53, and an executive vice president sold 21,024 shares on May 29 at about $118.28, both in the open market. No insider made an open-market purchase, and neither filing noted a preset trading plan.

A sale or two proves nothing on its own. But a run this strong, met with selling and no buying, is a contradiction worth holding in view rather than explaining away.

The scorecard, written before the print

The Firewall’s three questions, pointed at the filings, are the test.

First, the foundry line. In the Q2 release and 10-Q, what is external foundry revenue as a standalone figure, and is it growing fast enough to make break-even by end 2027 credible against a $10.3 billion annual loss?

Second, the proxy. In the 2026 proxy, what share of stock do officers and directors hold, and does that concentration suggest their interests track long-term holders rather than near-term compensation?

Third, the customer. Does 18A have a publicly named external customer committed to volume production, not a pilot or a collaboration? If not, at what point does that absence become the answer?

Confirm looks like external revenue clearly above the $307 million base and accelerating, the foundry loss narrowing, and ideally a named volume customer. Break looks like external revenue still tiny, the loss not improving, and no customer beyond pilots.

A pass moves it one notch, not to the finish

Say July 23 shows real external acceleration. It does not make Intel a buy, and it does not end the work. It moves the company from Needs Proof toward Clear for Deeper Research, and into In Review.

But this is where the distance shows. One quarter of external growth is a data point, not the durability the thesis needs, because break-even is years and many quarters away. Business quality is unproven while the unit loses billions. And the insider selling sits unresolved. So even a strong print most likely routes Intel to Watchlist re-entry, with the next report as the trigger, where a second and third quarter can turn a data point into a trend. Proof gets Intel further into the gauntlet. It does not carry it through.

The reusable takeaway

A turnaround is a verb, not a noun. The market keeps pricing Intel as a comeback that already happened, while the foundry numbers say it is still happening. Beating the quarter measures execution. Proving the thesis takes a customer base that does not yet exist at scale. Until the external line moves, the most important thing about Intel is the one number it has not yet printed.

Intel reports July 23. We grade this exact test the next morning and roll it into the Tracker. Subscribe to get the score in your inbox.

Not investment advice. The subscriber decides.