Intel's Foundry Progress
Intel 18A-P Process Begins Risk Production, Signaling Foundry Progress and Strong AI Demand
Key Points
Intel said 18A-P is an enhancement to the existing 18A baseline rather than a ground-up redesign, built on the same RibbonFET gate-all-around transistors and PowerVia backside power-delivery architecture, with additional co-optimization across transistors, interconnect, and design-technology layers.
At the VLSI Symposium, Intel (via reporting corroborated by Tom’s Hardware) disclosed more specific manufacturing/thermal improvements than were highlighted in the summary—such as “20-40% thermal resistance improvement” and via-resistance gains that support the node’s power-efficiency profile.
Intel Foundry’s Naga Chandrasekaran framed the milestone as part of an ongoing effort, saying: “This is a journey, and while we have more work ahead, we appreciate the opportunity to share the progress we are making with Intel 18A-P and our longer-range R&D.”
Alongside the 18A-P risk-production announcement, Intel introduced its Core Ultra Series 3, which it described as its first AI PC platform built on the 18A node and “designed and manufactured in the United States,” reinforcing the domestic-fabrication pitch to external customers.
Intel linked demand momentum to AI-service clients, reporting that CPU demand from AI providers was so strong in the first quarter that inventory that had been “previously written off” was ultimately sold.
Intel has moved its 18A-P chip process into risk production, the company announced on June 16, marking a key step in its bid to become a serious foundry rival to TSMC. The new process delivers up to 9% better performance at the same power, or up to 18% lower power at the same performance, compared to the existing 18A baseline, according to Intel Tom's Hardware.
The news sent Intel shares up more than 4% in premarket trading on June 17, CoinCentral reported. Intel also issued Q2 revenue guidance of $13.8 billion to $14.8 billion, topping analyst expectations of roughly $13.5 billion.
What Risk Production Actually Means for Intel
Risk production is not full mass manufacturing. It is a small-batch phase where Intel makes real chips to prove its process works before committing to high volume. The step matters because Intel has a history of missing targets — its 10nm and 7nm nodes slipped for years, handing market share to TSMC. Getting to risk production on schedule is Intel saying: "We can execute now, not just promise."
Analyst Kevin Krewell of Tirias Research called risk production the industry's "proof of life" for a new node, according to reporting cited by Yahoo Finance. The critical next number Intel has not yet disclosed is yield — the share of working chips per wafer. Yield will determine whether this milestone turns into real foundry revenue.
18A-P Is a Refinement, Not a Rebuild
Intel built 18A-P on top of 18A rather than starting over. Both nodes use the same RibbonFET transistors — Intel's version of gate-all-around technology, which wraps the gate around the channel on all four sides for better control — and the same PowerVia backside power delivery, which moves power wiring to the back of the chip to cut interference.
At the VLSI Symposium on June 15, Intel engineers reported a 20–40% improvement in thermal resistance and gains in via resistance — the friction that slows electrical signals inside the chip, according to Tom's Hardware. Crucially, 18A-P uses the same design rules as 18A. That means outside chipmakers can reuse existing blueprints with minimal changes, lowering the cost and risk of switching from TSMC.
AI Demand Fuels Intel's Foundry Pitch to Outside Customers
Intel's timing is no accident. The company said demand from AI infrastructure providers was so strong in Q1 2026 that it sold CPU inventory that had been "previously written off" — stock once considered worthless. That surprise demand helped underpin the bullish Q2 revenue guidance, Yahoo Finance reported.
Intel is using that momentum to attract outside chip buyers. Alongside the 18A-P announcement, the company launched Core Ultra Series 3 — its first AI PC chip built on 18A and "designed and manufactured in the United States," per MacDailyNews. The domestic-fab angle is a direct pitch to customers who want supply chains outside Taiwan, especially amid ongoing geopolitical tension.
What Comes Next: Yields, Customers, and TSMC Competition
Intel Foundry head Naga Chandrasekaran kept expectations measured. "This is a journey, and while we have more work ahead, we appreciate the opportunity to share the progress we are making with Intel 18A-P and our longer-range R&D," he said in an official statement. The focus now shifts to ramping yields and integrating customer IP — the hard engineering work that follows a risk-production announcement.
Analysts at Bernstein say Intel appears to be capturing the "second wave" of AI spending, as buyers move beyond GPUs to high-performance AI CPUs. Some skeptics, however, note that design-rule compatibility is only part of the story — extracting the full 9% performance gain may still require significant chip-level optimization, Yahoo Finance reported. The real test will be whether a major outside customer like Apple or Nvidia signs on for high-volume production.
Sentiment
Mostly PositivePublishers
22
Articles
13
Reach
35
