Intel stock surges on AI-driven turnaround

For most of the last three years, the AI trade has been a GPU trade — and Intel was the company the AI era was supposed to leave behind. Lip-Bu Tan's Q1 2026 message inverts that framing. 'In recent months we have seen clear signs that the CPU is reasserting itself as the indispensable foundation of the AI era,' Tan told investors, pointing to inference and agentic workloads where orchestration, memory, and general-purpose compute live on the host CPU rather than the accelerator. The numbers back the rhetoric: Data Center and AI revenue grew 22% year-over-year to $5.1 billion, AI-driven businesses now make up roughly 60% of total revenue and grew about 40% year-over-year, and Intel sold chips it had previously written off because demand exceeded supply.

The NVIDIA deal anchors the thesis architecturally. Jensen Huang described the partnership as 'tightly couples NVIDIA's AI and accelerated computing stack with Intel's CPUs,' and Intel Xeon 6 was named as the host CPU for NVIDIA's DGX Rubin NVL8 systems — meaning every Rubin-class rack NVIDIA ships pulls Xeon along with it. CFO David Zinsner reinforced the demand picture by noting Q1 results would have been even stronger if supply could keep up. The bullish read is that AI inference is not a GPU monopoly; it is a system problem, and Intel sits at the orchestration layer of that system.

The most underappreciated character in the Intel rally is the U.S. government. In August 2025, the Trump administration converted $5.7 billion of unpaid CHIPS Act funds and $3.2 billion of Secure Enclave funds into 433 million Intel shares at $20.47 — an $8.9 billion cost basis for what is now a 9.9% stake. After Thursday's close, that position is worth roughly $36 billion, a 300% return and an unrealized gain of about $26.5 billion. By any normal accounting, U.S. taxpayers are now Intel's biggest cornerstone investor and the single largest beneficiary of the AI turnaround.

That changes the political economy of the story. The CHIPS-to-equity conversion was controversial when it happened: critics framed it as state capitalism, supporters as a smart use of leverage to align incentives. The April 2026 print is a powerful piece of evidence for the supporters' side — the same administration that inherited a struggling national champion now holds a paper position larger than the entire market cap of most semiconductor peers. It also tightens the political alignment around Intel's foundry strategy. Washington has $36 billion of skin in the game on 14A working, on Tesla's Terafab being delivered, and on Intel keeping its U.S. manufacturing footprint — meaning policy support is structurally biased toward Intel for as long as that stake is outstanding.

The single-day move was historic in the most literal sense. Intel's roughly 24% gain on April 24, 2026, was its best trading day since October 1987 — a 39-year record that the chipmaker held through the dot-com bubble, the 2008 financial crisis, the 2020 COVID crash, and the 2022 chip-cycle reset. The rally also pushed Intel above its August 2000 dot-com era high, an all-time peak that had stood for roughly 26 years. Year-to-date 2026 gains now exceed 100%, the 12-month gain is approximately 224%, and market capitalization is above $416 billion.

The ripple effects are nearly as striking. AMD rallied around 14% and Arm Holdings around 15% in the same session — both companies that Morgan Stanley flagged as Intel's structural competitive risk — on the broader signal that AI inference revives CPU demand industry-wide. Q1 specifics: revenue of $13.58 billion versus a $12.42 billion estimate, non-GAAP EPS of $0.29 versus a $0.01 consensus (a roughly 29x beat), and a sixth consecutive quarter above Intel's own guidance. On the foundry side, Tesla's announced 14A commitment is the first major external customer; broader 14A commitments from prospective customers are not expected until the second half of 2026 or first half of 2027, once PDK 1.0 ships. The chart below frames how the day-of move sits against Intel's 12-month run.

The bull case is loud, but the bear case is structured. Wedbush flagged Intel's valuation as 'stretched' even as it acknowledged improving fundamentals. Morgan Stanley raised its price target to $73 from $56 but kept a Neutral rating, explicitly citing 'market share to Advanced Micro Devices (AMD) and Arm Holdings' as the reason not to upgrade. The implication: even if Intel is winning the next eighteen months, the longer-run server CPU socket is increasingly contested — AMD has been taking share in cloud and enterprise for years, and ARM-based custom hosts inside the hyperscalers are quietly displacing x86 sockets in workloads that scale to the tens of millions.

Execution risk on 14A is the other open question. Tan has previously warned that Intel would exit manufacturing without an outside 14A customer; Tesla's Terafab announcement clears the existential bar but does not yet translate into firm committed volume from Google, Apple, AMD, or NVIDIA, all of whom are expected to make decisions in late 2026 or 2027. Supply constraints are a near-term cap on the upside — CFO David Zinsner said Q1 results would have been even stronger if Intel could produce more chips, which is bullish on demand but bearish on how much of that demand actually converts to revenue this year. Community sentiment captures the split: Wall Street Bets is processing the move through nostalgia and meme — one viral post counts the years since the 'Nana' inheritance bet became a winner — while cable-news analysts frame the move as 'cautious optimism' on Tan's turnaround rather than a clean victory lap.