IMPLICATOR
Intel reported first-quarter revenue of $13.6 billion on April 23, beating analyst estimates of $12.4 billion as data center sales jumped 22%, with shares surging 24% Friday to their highest level since 2000. Demand for Xeon server processors used in AI inference workloads ran ahead of supply, forcing the company to ship chips it had previously written off, CFO David Zinsner told investors on the earnings call. The print, Intel's sixth consecutive beat under CEO Lip-Bu Tan, marked the stock's biggest one-day percentage gain since October 1987 and pushed its market value above $416 billion.
Key Takeaways
- Intel Q1 revenue hit $13.6B, beating estimates by $1.2B; data center grew 22% to $5.1B on AI demand for Xeon server CPUs.
- Stock surged 24% Friday, the biggest one-day gain since 1987, lifting market cap above $416B and pushing valuations to 90x forward earnings.
- Tan says the GPU-to-CPU ratio shifts from 7:1 in training toward parity in agentic inference, handing Intel a structural lane.
- Intel Foundry posted $5.4B in revenue against a $2.4B loss; major 14A customer commits expected in the second half of 2026.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
CPUs back at the center of AI compute
Tan told the call the central processor is "reasserting itself as the indispensable foundation of the AI era" as workloads shift from training to inference. The numbers behind the claim sit in the data center segment, which grew 22% year-over-year to $5.1 billion. Operating margin in the unit roughly doubled, from 13.9% to 30.5%, on operating income of $1.5 billion.
Zinsner offered a sharper version. Training an AI model typically calls for seven or eight Nvidia GPUs per Intel CPU, he told Fortune. Inference is closer to three or four. Agentic workloads could push the ratio toward parity. That math, if it holds, hands Intel a structural lane it had been written out of.
The supply problem
Demand outran what Intel could ship. Zinsner said the company dipped into finished-goods inventory, including "de-spec product or legacy product we had shelved," and worked the phones to move it. The benefit will not repeat in the second quarter, he cautioned.
The shortage runs past Intel's own factories. Memory suppliers are routing capacity to server DRAM, squeezing the consumer market and pushing phone and PC prices up. Intel's consumer business kept selling anyway.
Big names on the books
Three customer disclosures landed with the print. Google signed a multiyear Xeon contract for Google Cloud's workload-optimized instances and a co-development track on custom IPUs. Nvidia picked Xeon 6 as the host CPU for its DGX Rubin NVL8 systems, embedding Intel inside its rival's flagship rack design. SambaNova will pair its RDU accelerators with Xeon 6 hosts in a heterogeneous reference architecture.
AI moves fast. Stay sharp.
Strategic AI news from San Francisco. No hype, no "AI will change everything" throat clearing. Just what moved, who won, and why it matters. Daily at 6am PST.
No spam. Unsubscribe anytime.
Tesla, announced earlier in the week, slots Intel into Elon Musk's planned Terafab AI chip complex in Austin. Tan declined to confirm Tesla as a 14A foundry customer. "My style is underpromise, over delivering," he said. "We have no plans to announce the customer unless a customer wants to announce it."
Foundry, still the open question
But the foundry losses are the line nobody can hide. Intel Foundry posted $5.4 billion in revenue against a $2.4 billion operating loss. Yields on the 18A node run ahead of internal projections, Tan said, and 14A is "outpacing" its predecessor in maturity and performance. He told investors to expect "design commitments" beginning in the second half of 2026 and into early 2027.
That timeline is what Wall Street is still pricing. Intel trades at roughly 90 times forward earnings, the highest on record and well above AMD at 37 and Nvidia at 22. The federal stake bought for $8.9 billion last August is now worth about $36 billion. At least 23 brokerages raised price targets after earnings. HSBC moved from hold to buy and lifted its target from $50 to $95.
What's missing
A standalone AI accelerator. Intel still does not field a chip that competes with Nvidia's H- and Rubin-class GPUs, and Tan said only that the company would "continue to participate, innovate and partner" in that category. If you are tracking the turnaround thesis, the trade is whether the CPU comeback compounds before the foundry math gets worse. The Q2 guide of $13.8 billion to $14.8 billion sits about $1.3 billion above consensus at the midpoint, but Zinsner flagged the inventory tailwind as gone. Demand stays. Supply is the question.
Frequently Asked Questions
Why did Intel's stock surge 24% on Friday?
Intel's first-quarter results beat Wall Street estimates and the Q2 forecast came in well above consensus. Revenue reached $13.6 billion versus $12.4 billion expected, driven by AI inference demand for Xeon server CPUs. The data center segment grew 22% year-over-year to $5.1 billion. At least 23 brokerages raised price targets after earnings, with HSBC moving from hold to buy and lifting its target from $50 to $95.
What is driving the new demand for Intel CPUs in AI?
As AI workloads shift from training to inference and agentic applications, the ratio of GPUs to CPUs in data centers tightens. Training calls for seven or eight Nvidia GPUs per Intel CPU. Inference is closer to three or four. CFO David Zinsner said agentic workloads could push the ratio toward parity. CPUs handle orchestration, data preprocessing, and inference serving alongside GPU spending.
How does Intel's stock valuation compare to AMD and Nvidia?
Intel trades at roughly 90 times forward earnings, the highest on record for the company. AMD trades at 37 times and Nvidia at 22 times. The premium reflects investor expectations that the data center turnaround compounds alongside foundry progress. Intel still does not field a standalone AI accelerator that competes with Nvidia's H- and Rubin-class GPUs.
What is the status of Intel Foundry and the Tesla deal?
Intel Foundry posted $5.4 billion in revenue against a $2.4 billion operating loss in Q1. Yields on the 18A node are running ahead of internal projections. Tesla joined the Terafab project alongside SpaceX and xAI. CEO Lip-Bu Tan declined to confirm Tesla as a 14A foundry customer, saying he prefers to underpromise and over deliver. Major design commitments are expected starting in the second half of 2026.
How much is the U.S. government's Intel stake worth now?
The federal stake acquired in August for $8.9 billion is now worth about $36 billion, a roughly fourfold increase. The deal pricing put shares at $20.47 each. Taxpayers own more than 270 million shares directly, with additional shares held in escrow. The government retains rights to acquire up to 433.3 million shares once certain conditions are met.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
Marcus Schuler
Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: [email protected]
