Intel’s Turnaround Has a Real Deadline Now

Intel trades around $108.60 as I write this, 278% above its 52-week low of $29 and still 23.7% below the $142 high the stock touched earlier this year. A move that size in either direction usually means the market is betting on a story, not just pricing a quarter. So what’s the story here, and is it one you can actually check?

Intel’s own management has called 2026 the execution year, which is a useful phrase because it implies a deadline rather than an open-ended promise. Most turnaround stocks get sold as an ongoing process with no fixed checkpoint, the kind of framing that lets a company keep asking for patience indefinitely. I made this same point about a payments company mid-turnaround not long ago: a turnaround thesis is only as good as the specific, dated thing you can point to that would prove it wrong. Intel has given itself one. That’s worth taking seriously, and it’s why the stock deserves a real look rather than a reflexive pass.

A stock that already priced a comeback

Start with what’s already happened, because the market has not been shy about it. Intel’s quant rating moved from C a year ago to A now, about as large a swing as that scoring system produces, and the 278% climb off the 52-week low reflects a genuine re-rating, not just noise. Net income for the trailing year came in essentially at breakeven, $0.0 billion, against a loss of $-19.2 billion the year before. That’s one of the larger year-over-year net income swings I’ve written about on this site, and it’s the single number that explains most of the stock’s move better than any roadmap slide does.

Revenue tells a quieter, more mixed story than the net income line. Trailing revenue sits at $52.9 billion, down from $53.1 billion the year before and $63.1 billion three years ago, essentially flat to slightly down across a period when the rest of the semiconductor industry, particularly anything touching AI infrastructure, grew fast. The most recent quarter is the exception: revenue of $16.1 billion came in 25% above the year-ago quarter and 19% above the prior quarter, the kind of sequential acceleration that tends to catch a market’s attention even when the trailing twelve-month trend still looks flat.

Put those two lines together and you get the actual shape of this turnaround. Revenue isn’t growing the way a share-taking chipmaker’s revenue should be growing; it’s roughly stable after three years of decline. What changed is cost. Intel cut spending hard enough, and stabilized manufacturing well enough, that a business generating about the same top line as three years ago now clears breakeven instead of losing close to twenty billion dollars.

That’s a cost story wearing a growth story’s clothes. A pure cost turnaround can re-rate a stock once, from crisis pricing back to fair value. It can’t keep re-rating every quarter unless revenue eventually shows up too, and I think it’s worth being precise about which one you’re underwriting when you buy the stock here.

Panther Lake is the real test, not a slide

Intel’s 18A manufacturing process, the company’s most advanced node, is no longer a roadmap promise. Panther Lake, sold as the Core Ultra Series 3 line, is Intel’s first commercial chip built on 18A, and according to Intel’s own newsroom material it is now shipping and being produced at the company’s new Fab 52 plant in Chandler, Arizona alongside existing Oregon capacity. That matters because a chip either ships in volume or it doesn’t; there’s no partial credit, and no way to spin a delay as a rounding error the way a company can with a one-point miss on a margin target.

I don’t have an independently verified yield number for 18A, and I’d rather say that plainly than repeat a figure I can’t check. What I can say is that Intel chose to put a real product into real laptops on this node rather than keep it in qualification for another year, which is itself a decision with consequences if the yields underneath aren’t where the company needs them. Gross margin, 34.8% over the trailing period versus 32.7% the year before, has moved in the right direction, though a one-point gain across a year this consequential for the manufacturing side is a modest signal, not a verdict.

The foundry customer question is still open

The part of the turnaround that decides how big this gets, rather than whether it survives, is whether Intel can convince other chip designers to manufacture on its process instead of at TSMC. Public reporting this year has named companies like Apple and Amazon as parties Intel has courted for 18A capacity, alongside a later, more advanced 14A variant, though none of that reporting amounts to a signed, disclosed contract I can point to as fact. I’d treat every foundry-customer rumor as exactly that until Intel confirms one on an earnings call, the same discipline I’d apply to any company hoping the market will pre-price a deal that hasn’t closed.

Nvidia’s dominant position, built on TSMC’s leading-edge capacity rather than Intel’s, is the scale Intel is chasing here, a point I made when writing about Nvidia’s moat as fundamentally a software and ecosystem advantage layered on top of manufacturing access it doesn’t own itself. Intel doesn’t need to beat TSMC outright to make this turnaround work. It needs enough external customers on 18A and its successor to prove the foundry business can be more than a subsidized cost center for Intel’s own chips, and that proof doesn’t exist in the numbers yet.

There’s a trust problem underneath the commercial one, and it’s specific to Intel rather than generic turnaround skepticism. A fabless chip designer handing its most advanced, most competitively sensitive design to a foundry is also handing that foundry a look at its roadmap years before launch. Doing that with TSMC, a pure-play manufacturer with no competing chip line, is a different risk calculation than doing it with Intel, which still designs and sells its own competing processors.

Intel has talked about operating its foundry business at arm’s length from its product group. That’s a governance claim, not a number. I’d want to see at least one large, named external customer stick around for more than one node generation before I called this problem solved.

A P/E that doesn’t exist yet

Valuing Intel on trailing earnings is close to a pointless exercise right now: the trailing P/E is negative, -52.0, on trailing earnings per share of $-2.09, a number distorted by the losses from before the turnaround argument had any evidence behind it. The market has moved on to forward numbers instead, pricing the stock at 89.1 times a forward EPS estimate of $1.22, an expensive multiple on a still-small profit base. Price to sales tells a similar story: 9.5 times trailing sales now against a five-year average closer to 3.0, a real premium the market is assigning to the recovery story rather than to results already in hand.

Metric	Value	Context
Price	$108.60	278% above 52-week low of $29
P/E (trailing)	-52.0	forward 89.1 on $1.22 fwd EPS
Revenue (TTM)	$52.9 billion	vs $53.1 billion prior year
Net income (TTM)	$0.0 billion	vs $-19.2 billion prior year
Analyst target	$121	range $80-$200, 23% buy-rated

That’s not automatically wrong, but it does mean the stock has very little room for a disappointing print. Sell-side coverage reflects the split: only 23% of the 26 analysts covering Intel rate it a buy, with an average target of $121, about 11% above the current price. The spread between the high target of $200 and the low of $80 is unusually wide for a company this size, which tells you the Street hasn’t converged on what Intel is worth so much as it’s hedging a real binary outcome. I’ve made a version of this same binary-outcome argument before when comparing AMD’s position as the second source in a market that mostly rewards the first, and Intel’s foundry bet has the same all-or-mostly-nothing shape to it.

What would actually break this call

Here’s the specific uncertainty I can’t resolve from where I sit: whether 18A yields hold up at real production volume, not just in early Panther Lake runs, over the next two quarters. If they don’t, and Intel has to quietly slow the ramp or delay the 14A follow-on, the stock’s forward multiple stops making sense and the $80 low-end target, already below today’s price, becomes the more credible number rather than the outlier. Average earnings-day moves for Intel have run 12.2%, among the largest of any stock I cover regularly, and the most recent report moved the stock -7.9%, so the market is not treating this quarter to quarter as settled business.

I’m not worried yet, mostly because Intel put a specific, checkable claim on the table instead of another open-ended promise, and because the net income swing from $-19.2 billion to roughly breakeven is real money, not sentiment. But I’d be lying if I said the foundry customer question and the yield question were anything other than open, and a stock priced at 89.1 times forward earnings has already spent most of the credit it gets for good intentions. Watch Intel’s next quarterly report, roughly three months out from the 2026-07-23 print, for whether gross margin keeps climbing past 34.8% and whether management names even one confirmed external 18A customer. Either one showing up would tell you more than another slide ever could.

Analysis and opinion only, not investment advice. Figures come from Intel’s filings on SEC EDGAR and its investor site; valuation multiples are approximate and were checked on 2026-09-18 20:02:20.