On July 23, 2026, Intel revealed that it has fully committed to launching high-volume production of its next-generation process technology, Intel 14A, in 2028, following risk production for its own products starting in the second half of 2027. The 2028 timing itself is not new—CEO Lip-Bu Tan had already laid out the same schedule during the January earnings call. What's changed this time is not the date on the roadmap, but the management decision to commit capital to manufacturing equipment and design infrastructure. Intel has completed PDK 0.5 and will roll out PDK 0.9 in October. Demand for the company's own products has also grown, moving 14A into the phase of preparing for mass production.
2028 Mass Production Isn't a Pull-In—It's a Confirmed Investment Decision
The 14A manufacturing timeline has not changed from January 2026 through July 2026. During the Q4 2025 earnings call in January, Tan stated the outlook of moving to risk production in the second half of 2027 and mass production in 2028. In the July update, Intel specified that this applies to its own products while maintaining the same timing, and strengthened its language during the second quarter to say it had fully committed to launching mass production in 2028.
| Timing | What Intel Stated | Discernible Change |
|---|---|---|
| January 2026 | Risk production in H2 2027, mass production in 2028 | Presented the manufacturing timeline |
| Q1 2026 | Multiple customers evaluating; increase in internal tiles adopting 14A | Expanded internal and external design engagements |
| Q2 2026 | Completed PDK 0.5; fully committed to launching mass production in 2028 | Decided to invest in mass production readiness |
Therefore, interpreting this update as "pulling risk production forward from 2028 to 2027, and mass production from 2029 to 2028" does not match the record of events since at least January 2026. The real change is that management has now decided to commit to the equipment and development investments needed to make 2028 mass production happen.
Risk production and mass production also need to be read as distinct. The process planned for the second half of 2027 is a stage where Intel's own products are used to validate the process and product design ahead of mass production. It is not an announcement that external customers' products will ship in high volume that same year. The phrase "high volume ramps" in 2028 likewise refers to the start of ramping production, not a confirmed launch month or supply volume for any specific product.
PDK 0.9 Arrives in October, Moving Closer to a Design-Ready Process
Intel has completed PDK 0.5 for 14A and plans to roll out PDK 0.9 in October 2026. A PDK—process design kit—is a package of models, design rules, and verification environments used to design circuits that match a given manufacturing process. Even as process development advances, product designs cannot be finalized until a usable PDK and IP are available to customers. Reaching 0.9 marks a milestone that lets customers finalize their designs with mass production in view.
Intel also says development-side metrics are progressing. The company states that 14A's defect density and transistor performance now exceed those of Intel 18A at the same point in its development. As of Q1, Intel had already said that maturity, yield, and performance were ahead of where 18A stood at the same stage, but by July it had gotten more specific, naming the PDK version and its rollout month. Even so, these remain internal Intel comparisons—not figures that demonstrate proven mass-production yields or customer product performance.
14A employs the second-generation gate-all-around transistor, RibbonFET 2, along with PowerDirect, which delivers power from the wafer backside. According to Intel's internal analysis, compared to 18A, 14A can deliver 15-20% higher performance at the same power, or 25-35% lower power at the same performance, along with up to 30% higher chip density. Turbo Cells additionally allow only the circuits that need speed—such as those in CPUs and GPUs—to be boosted. Whether these stated figures translate into actual products will be determined by the designs finalized with PDK 0.9 and by real silicon results in 2027.
Committing to 14A's Mass Production Without Naming Customers
The grounds Intel cited for moving forward with mass production are progress in discussions with external customers and growing demand for its own products. As of Q1, multiple customers were said to be evaluating 14A, with initial design commitments expected to begin emerging from late 2026 into early 2027. For Intel's own products, the number of future tiles using 14A has also increased. Yet even in the July update, Intel disclosed neither the names of external companies adopting 14A nor their product names or order volumes.
Whether external demand materializes will determine 14A's economics. In its 2025 annual report, Intel explained that securing the economics of leading-edge manufacturing requires production volume beyond what its own products can generate. It also noted the possibility of pausing or discontinuing 14A and subsequent process technologies if it fails to secure major external customers. This full commitment represents a step forward from that cautious stance toward mass-production readiness—but it is not equivalent to announcing a major customer.
As it stands, Intel Foundry posted revenue of $5.8 billion in Q2, of which only $293 million came from external customers, alongside an operating loss of $2.1 billion. While manufacturing for internal products underpins the business's scale, external sales remain small. For 14A to change the revenue structure, Intel will need to stabilize mass production of its own products while converting external customer designs into sustained wafer demand.
Capital spending is moving ahead of that outcome. Intel raised its full-year 2026 capital expenditure outlook to over $20 billion and indicated plans to increase spending substantially further in 2027. Most of this will go toward the domestic U.S. manufacturing network, with orders for manufacturing equipment and cleanroom buildouts also being accelerated. However, Intel has not disclosed the amount specifically earmarked for 14A. Separate from the company-wide plan, Intel also said it increased investment in 14A during Q2 in preparation for risk production in 2027.
TSMC Also Targets 2028—The Contest Will Be Decided by Yield and Adoption
TSMC also plans to begin production of its next-generation A14 process in 2028. The company says that, compared to N2, A14 can deliver up to 15% higher speed at the same power, or up to 30% lower power at the same speed, along with more than 20% higher logic density. 2028 is shaping up to be a year in which both Intel and TSMC present closely comparable generations of process technology to customers.
However, similar naming—both companies use "14" in their process names—and the stated improvement percentages do not by themselves establish which is superior. Intel benchmarks against 18A, while TSMC benchmarks against N2, and the measurement conditions differ as well. The conclusion that Intel 14A is equal to or better than TSMC's A14 cannot be drawn from the publicly disclosed figures.
The evidence needed for a real evaluation will arrive in stages. First, the design environment will be locked down with PDK 0.9 in October 2026, followed by whichever customer names and adoption scale become clear. Then, in the second half of 2027, risk production of Intel's own products will test whether defect density and yield improvements hold up, setting the stage for whether Intel can ramp to mass-production scale in 2028. What Intel has confirmed this time is its decision to commit capital and move forward through this verification process. The ultimate verdict on this process technology will be decided at the stage where customers entrust their designs to it and reliably receive good-yield wafers as planned.