Intel’s comeback bet: Apple, Nvidia and Washington are in, and Thursday’s earnings decide if it worked

Filing Intel under “legacy PC chipmaker that missed the AI boom” would have been a fair reading of the company. Not anymore. In the space of about twelve months, Intel has taken on a roughly 10 per cent equity stake from the US government, a $5 billion investment from Nvidia, $2 billion from SoftBank, and landed a preliminary manufacturing deal with Apple, the one customer its foundry business had chased unsuccessfully for years. That’s why we built Winvesta Crisps, to decode what is actually driving the companies you own, in plain language, before the consensus catches up. 60,000+ investors from all over India are already in. What about you?

🔔 Don’t miss out! Add winvestacrisps@substack.com to your email list so our updates never land in spam.

Most investors who have glanced at Intel’s chart this year have seen a stock up well over 150 per cent and assumed it is a squeeze in an unloved chipmaker riding the general AI wave. That is not quite what happened. Apple has a preliminary agreement to manufacture chips on Intel’s newest process node, ending a roughly six-year exile to TSMC exclusivity. Microsoft, Amazon and Google have all signed on as data centre or foundry customers. Nvidia, a company that was busy trying to make Intel irrelevant two years ago, now owns a $5 billion stake in it, and the US government owns roughly a tenth of the company outright. Thursday’s second-quarter results, due after market close on 23 July 2026, are where all of that turns from a narrative into a set of numbers.

---

🧩 Not one chipmaker, but four

Most coverage of Intel still leads with the Client Computing Group, the unit that makes the processors sitting inside laptops and desktops, because that is the Intel most people grew up buying. It generated $7.7 billion in the first quarter of fiscal 2026, up a modest 1 per cent year on year, per the company’s earnings release. It is the legacy anchor: large, cash generative, and barely growing.

The more interesting segment is Data Centre and AI, or DCAI, which posted $5.1 billion in Q1, up 22 per cent year on year, with an operating profit of $1.5 billion, a 31 per cent margin, per Intel’s earnings transcript. This is the business that CFO David Zinsner told analysts now represents around 60 per cent of Intel’s total revenue when combined with other AI-linked lines, and it grew about 40 per cent year on year. Within DCAI, Intel’s custom ASIC business, purpose-built silicon designed for specific customer workloads, more than doubled year on year and is already running at an annualised rate above $1 billion.

Then there is Intel Foundry, the division at the centre of the turnaround story: $5.4 billion in Q1 revenue, up 16 per cent year on year and 20 per cent sequentially, but still posting an operating loss of $2.4 billion, narrowed by $72 million from the prior quarter as yields improved across the Intel 3, Intel 4 and 18A nodes, per the company’s Q1 transcript. External foundry revenue, meaning chips Intel makes for other companies rather than itself, was just $174 million for the quarter. That is a small number today. It is also the number every investor betting on this turnaround is watching most closely.

The fourth bucket, All Other, houses what remains of Altera and Mobileye after Intel sold a majority stake in Altera to Silver Lake for $3.3 billion in 2025 and separately monetised part of its Mobileye holding. Revenue there fell 33 per cent year on year to $628 million, though it still turned a small operating profit of $102 million.

---

🏭 The comeback nobody priced in

Intel’s foundry ambitions have been a running joke on Wall Street for the better part of a decade: a company promising to catch up to TSMC’s manufacturing lead while losing billions of dollars a year trying. What changed in 2025 and 2026 is not the ambition. It is who showed up to fund it and buy from it.

Start with the balance sheet. In August 2025, the Trump administration converted roughly $8.9 billion in CHIPS Act grants into a 10 per cent equity stake in Intel, receiving about 433.3 million shares at a cost of roughly $20.50 apiece, per CNBC’s original reporting. At Intel’s current price of around $93, that stake is worth well over $40 billion, more than four times what it cost. SoftBank followed with a $2 billion investment at $23 a share the same month. Then, in September 2025, Nvidia announced a $5 billion investment of its own, alongside a joint plan to co-develop data centre and PC chips: Intel building custom x86 processors that Nvidia’s GPU chiplets plug into. Analyst Dan Ives at the time called it a turning point that brought Intel back to the centre of the AI conversation rather than watching from the sidelines. Nvidia CEO Jensen Huang has pointed to a market opportunity he pegs at roughly $50 billion a year for the combined product line. The Nvidia deal closed in December 2025 after the FTC signed off on it.

That capital did two things. It gave Intel cash to pay down debt (about $4.3 billion in Q3 2025 alone) and it gave the market a reason to believe Washington, Tokyo and Silicon Valley all had a financial stake in Intel’s survival, not just its recovery.

Then came the customers. Microsoft signed on for custom 18A-based chip design in February 2024. Amazon Web Services followed in September 2024. Google signed a long-term DCAI agreement disclosed on the Q1 2026 call. And in May 2026, the Wall Street Journal reported that Intel and Apple had reached a preliminary agreement for Intel to manufacture some Apple silicon, a deal that would end Apple’s near-total reliance on TSMC. President Trump announced the arrangement on Truth Social on 18 June 2026, and Intel’s stock jumped roughly 11 per cent that day to close near $134, per RTTNews. Reports since then, including from TechPowerUp, suggest Apple is looking at Intel’s 18A-P node for a future MacBook chip and the newer 14A node for an iPhone chip, with production not expected to begin before around 2027.

The technical milestone behind all of this is 18A-P, a refined version of Intel’s most advanced process, which entered what the industry calls risk production in mid-June 2026, per an announcement at the VLSI Symposium in Honolulu. Intel’s foundry chief, Naga Chandrasekaran, framed it as an early step in a long process rather than a finished product, per CNBC’s coverage. The node reportedly delivers about 9 per cent higher performance or 18 per cent lower power draw than standard 18A. Some reporting in early July suggested profitable yields on 18A-P may not arrive until late 2026 or 2027, a concern that fed directly into the broader semiconductor sell-off that has weighed on Intel’s stock this month. That gap, between a real customer list and a foundry business that still loses billions of dollars a quarter, is the entire investment debate in one sentence.

---

Want to add Intel to your portfolio? Trade INTC directly from India on the Winvesta app. No US bank account needed!

🚀 Join 60,000+ investors, become a paying subscriber or download the Winvesta app and fund your account to get insights like this for free!