Intel's 2026 Profit Call: AI Demand Is Real, but the Turnaround Must Still Fund Itself

Q2 results made Intel's recovery harder to dismiss

Intel is no longer being judged on hope alone. The latest quarter gave investors actual numbers to anchor the story. The company delivered Q2 revenue of $16.1 billion, adjusted profit of 42 cents a share, and adjusted gross margin of 41.8%. That moves the discussion away from the old survival narrative.

The next test is whether the quarter was a one-off

What matters now is the forward call. Intel guided Q3 revenue of $15.8 billion to $16.8 billion, above $15.10 billion in estimates, and Q3 adjusted profit of 38 cents, above 27 cents expected. Bulls see a company starting to produce the earnings stream investors have chased for months. Bears can still point to weaker demand from the PC market and margins that remain below historical highs.

Why investor patience now depends on execution

The setup changed because management is asking investors to fund the next leg with a backlog that still needs to prove its durability. Intel lifted this year's capex to $20 billion and said spending will be "up meaningfully next year." That only works if AI-related CPU demand stays strong enough to absorb the investment. The stock's drop of more than 25% from its June peak suggests the market no longer rewards the story by itself; it wants repeated proof.

AI CPU demand looks real, but capex is what turns it into profit

The key question is no longer whether AI is helping Intel at all. It is whether AI CPU demand is converting into durable profit rather than just a strong quarter.

How AI demand could improve operating leverage

The mechanism is fairly direct. Intel is benefiting from a shift in AI spending from training toward deployment, where CPUs still play a major role. Reuters says cloud providers are shifting from training models to deploying them, and Intel executives have said demand for data center CPUs has outstripped the company's ability to manufacture the CPU chips. When demand runs ahead of supply, revenue can accelerate, utilization can improve, and fixed costs can be spread across more shipped units.

That also helps explain why the narrative has become easier to defend. Earlier this year, Intel said it was unable to meet demand for server chips. Now management is raising this year's capex to $20 billion from $18 billion. In practical terms, customers appear to want more product, and Intel is choosing to add capacity rather than lean too heavily on the story.

The bull case: more spending implies more throughput

Bulls see the capex increase as evidence that demand has enough visibility to justify building ahead. The same Reuters report says AI data-center buildout is increasing demand for Intel's central processing units, while Melius said the raise signals confidence in cash flow upside and demand visibility from long-term agreements. If that holds, current spending should support next year's profit base.

The important part is that the demand is showing up in CPUs, not just in broad AI enthusiasm. Intel has said it is starting to be a meaningful beneficiary of AI investment in data centers, which makes the upside case more concrete than the old "maybe AI helps later" debate.

The risk case: capex can come before cash flow

Bears still have a credible argument. Capex is certain; profit is not. If demand cools, or if supply constraints ease and pricing softens, the extra spend could weigh on returns instead of lifting them. That risk is why higher spending only helps if investors believe the revenue follow-through will be durable.

What to watch next

• Revenue above guidance plus rising capex: suggests demand is pulling capacity expansion.
• Margin resilience: the clearest sign that AI CPU demand is turning into profit, not just volume.
• Another capex increase without demand follow-through: the cleanest warning sign that Intel may be spending to create growth rather than fund it.

External customer wins now matter as much as management commentary

The market has already shown what can move Intel. After reports that Google ordered over three million tensor processing units for 2028 production, the stock rose as much as 12%. That was not just a quarter-by-quarter reaction; it was a reaction to a dated external commitment that could validate demand years out.

After the latest earnings and the decision to raise this year's capital expenditure, investors are no longer paying for management commentary alone. They are looking for outside confirmation. On foundry, Intel already has a foot in the door with a Tesla tie-up. The more market-sensitive watchpoint is the potential deal to manufacture processors for Apple that neither company has confirmed.

Catalysts that could keep the rerating alive

• Another dated customer commit tied to future production, like the Google TPU order for 2028, would suggest the demand story extends across customers and timeframes.
• Any movement on the potential Apple deal could matter more than most headline quarters because it would strengthen Intel's foundry reputation.
• Evidence that Intel is still riding growing adoption of its data center CPUs would show the AI demand story is broadening beyond a single product or customer.

What could break the rerating

• If foundry progress stays narrow while the Apple deal remains unconfirmed, the stock could lose momentum quickly.
• If customer wins remain abstract while spending keeps rising, the market may stop paying for future proof and start questioning execution.
• If external proof fades, earlier enthusiasm can unwind fast; the market has already shown it punishes gaps in evidence after the shares declined more than 25% from a record close.

That is the setup now: Intel is less a story trade than a proof trade. Dated external commitments matter more than hopeful commentary, because the next rerating is more likely to come from customer evidence than from another narrative update.