The Big Question: Is Intel’s Turnaround Already Priced In?

Intel just reported its fastest revenue growth in almost 15 years, and the stock went up roughly 4% in after-hours trading.

Think about that for a second.

The company that was left for dead in 2025 — trading near $19 at its 52-week low — just posted $16.1 billion in Q2 2026 revenue, up 25% year-over-year, crushed EPS estimates by over 92%, and guided Q3 revenue between $15.8 billion and $16.8 billion, well above the Street’s $15.1 billion consensus. CEO Lip-Bu Tan called it the company’s strongest revenue growth in over 15 years. The stock closed Thursday at $100.23 and jumped toward $110 after hours.

That’s the reaction. A 9% to 10% pop. Restrained, all things considered.

Why Wall Street Is Fixated Right Now

The investment community isn’t debating whether Intel’s turnaround is real anymore. The numbers settled that. What professional investors are fighting over is something harder to model: whether the stock at roughly $100, up over 170% year-to-date and carrying a forward P/E near 94x, is still a buy or a monument to expectations already exceeded.

The stock had already run from $18.97 to a high of $142.35 before pulling back nearly 28% in July alone, partly on a Bank of America AI-valuation warning that hit the broader chip sector. Retail sentiment got loud. A viral post called the stock “trading at a level not seen even during the dot-com bubble.” Then earnings landed and the numbers made most of the bears look early.

This is the tension institutional investors are managing right now.

The Bull Case

The Data Center and AI group did the heavy lifting. DCAI delivered $6.3 billion in Q2 revenue, up 59% year-over-year and 24% sequentially. That’s an acceleration from Q1’s 22% year-over-year gain. Operating profit for the segment hit $2.5 billion at a 40% margin — a number that signals Intel’s server business isn’t just growing, it’s growing profitably.

Xeon 6, in the words of Tan on the earnings call, is “one of the fastest ramping products in Intel history.”

Then there’s the foundry story. Intel launched Xeon 6+, its first server processor built on 18A, during the quarter. Panther Lake client processors built on 18A shipped across more than 200 OEM designs. The 18A-P variant, which offers roughly a 9% performance improvement over base 18A, entered risk production. And 14A, the next node, is tracking ahead of where 18A was at the same point in its development cycle.

On the customer side, Google placed an order for 3 million custom Tensor Processing Units with Intel’s foundry. Fortinet announced it will manufacture its next-generation security chip on Intel 4. Reports link Microsoft, AWS, OpenAI, and Nvidia to 18A and 14A design projects. KeyBanc’s John Vinh wrote ahead of earnings that Intel is positioned to support 25% to 30% server CPU unit growth in 2026, driven by hyperscale AI deployments and what he described as agentic AI increasing the CPU intensity of AI clusters.

CFO David Zinsner raised the 2026 capital expenditure forecast from $18 billion to over $20 billion, with 2027 spending expected to rise “meaningfully.” That’s not a company hedging. That’s a company seeing real customer demand and committing capital against it.

HSBC’s Frank Lee, who had a $24 Sell target on Intel in late 2025, now has a $200 Buy target — roughly $100 above consensus. His reasoning: the foundry business, which he excluded from his model in April because external customers weren’t committed enough, now warrants inclusion. Susquehanna’s Christopher Rolland raised his target to $115 pre-earnings, expecting server demand to outpace supply well into 2028.

The Bear Case

Here’s where the closed-door conversation gets uncomfortable.

Intel Foundry posted another operating loss of roughly $2.1 billion in Q2. The full-year 2025 foundry operating loss was $10.3 billion. Q1 2026 foundry losses came in at $2.4 billion. The 18A ramp is real, but so is the cash burn. Intel’s own Q1 filing acknowledged that foundry losses were driven in part by the higher cost of early 18A wafers. Scale the volume up and the losses can expand before they contract. That’s the math no one loves discussing.

AMD passed Intel in quarterly data center revenue for the first time in July, a milestone that landed in the same week as the BofA valuation note and the early 18A yield concern reports. EPYC processors are winning data center contracts on performance-per-watt efficiency. ARM-based alternatives are growing. Intel is gaining momentum in AI server CPUs, but the competition isn’t standing still.

Slight tangent worth noting: the PC business is softening. Management warned PC consumption will be “subseasonal” in the second half of the year. Higher memory prices are compressing demand for entry-level and mid-range systems. The Client Computing Group, which generated $8.88 billion in Q2, could face real pressure in Q3 and Q4. That matters because CCG is still the single largest revenue segment.

The valuation math also just doesn’t offer much margin for error. At 94x forward earnings, Intel needs to execute perfectly on a foundry turnaround that involves billions in losses, a manufacturing ramp with real yield uncertainty, and external customer commitments that, as analysts at RCR Tech noted, still look more like hedges than full migrations away from TSMC.

Of the 49 analysts covering the stock, 13 rate it Buy, 32 Hold, and 4 Sell. The 12-month average target sits around $108. HSBC at $200 is a clear outlier against a Hold-heavy field.

What the Evidence Actually Shows

The Q2 beat wasn’t marginal. Revenue came in at $16.1 billion against a $14.43 billion consensus estimate. EPS of $0.42 crushed the $0.21 expected. Q3 guidance at $15.8 to $16.8 billion topped consensus at the low end. The CFO said customer demand continues to outstrip growing supply. These aren’t inflection-quarter numbers. They’re acceleration-quarter numbers.

The stock’s muted response relative to the size of the beat is itself the signal. After a one-year gain of over 340%, the market had pre-loaded a lot of optimism into the price. Polymarket odds had an earnings beat at 99.95% probability before the release. When the entire world knows you’re going to beat and you beat big, a 9% pop is the outcome.

The foundry operating losses are real and worth watching closely, but the direction of travel is improving. Intel Foundry’s Q1 2026 loss narrowed meaningfully from Q4 2025’s $2.5 billion, according to Morgan Stanley estimates. The raised capex guidance — from $18 billion to over $20 billion — signals firming customer commitments, not speculative expansion.

What Investors Are Missing

Almost no one is talking about Intel’s ASIC business.

Intel’s purpose-built silicon segment posted revenue up nearly 3x year-over-year in Q2, approaching a $2 billion annual run rate. This is the business making custom chips for specific enterprise and AI applications — the Fortinet firewall ASIC is one example. Management described it as a potentially $100 billion total addressable market. That business is growing faster than any other Intel segment and hasn’t earned a line item in most analyst models.

The broader point: Intel is no longer just a CPU company trying to build a foundry. It’s becoming a full-stack semiconductor solutions provider at the exact moment AI infrastructure spending is forcing every major hyperscaler and enterprise buyer to think about custom silicon. The companies that can design, package, and manufacture custom chips at scale are acquiring a structural advantage. Intel, if 18A delivers, is one of the only companies in the Western world that can offer all three.

That opportunity still isn’t fully embedded in the $108 consensus target.

Stocks to Watch

Intel (INTC) — The central story. The turnaround is operationally confirmed. The debate is now entirely about valuation and foundry profitability timing. Watch Q3 for the first signs of whether the 18A ramp is compressing or widening foundry losses. That single data point will shape the stock for the next six months.

AMD (AMD) — The competitive pressure on Intel’s data center business is real. AMD passed Intel in quarterly data center revenue for the first time this year. EPYC’s momentum in the enterprise and hyperscale market remains the clearest near-term risk to Intel’s DCAI growth story. Worth monitoring how AMD’s own Q2 results shape the competitive framing.

TSMC (TSM) — The quiet loser in Intel’s foundry story. Every design win Intel Foundry captures with 18A is incremental share away from TSMC or Samsung. Reports that TSMC is struggling to keep up with demand from Nvidia, AMD, Apple, and others make Intel’s capacity expansion more strategically valuable for customers seeking supply diversification. TSMC’s pricing power is real, but the monopoly is under pressure.

Lam Research (LRCX) and Applied Materials (AMAT) — Semiconductor equipment stocks moved higher in after-hours trading on Intel’s capex guidance. Tooling spend is up 40% from 2025. A $20-plus billion capex plan from Intel, with 2027 spending expected to rise “meaningfully,” represents sustained demand for the equipment these companies supply. The overlooked beneficiaries of Intel’s manufacturing expansion.

The comeback quarter happened. The harder question is what comes next.

Intel’s foundry is still bleeding cash. The PC business is softening. AMD is not cooperating. The stock is trading at a valuation that requires near-perfect execution on a manufacturing ramp that has never been attempted at this scale in the Western world.

But the numbers keep moving in the right direction. And that $100 billion ASIC opportunity is one that Wall Street is barely looking at.

That gap between what the market is pricing and what might actually be building is where the most interesting investing conversations are happening right now.

Stay tuned.

— Wall St. Mavens Editorial Team