Intel is taking a cautious approach to its next-generation 14A semiconductor manufacturing process as it looks to take on TSMC more closely. Naga Chandrasekaran, head of Intel Foundry, said in a presentation to investors that its 14A process will be able to perform within 5 per cent of TSMC’s A14 technology. The two manufacturing platforms are scheduled for volume production in 2028 and set the stage for the next big battle between the world’s leading chip foundries.
The statement is notable since Intel has some technological advantages as it develops 14A. It is already working with ASML's High-NA EUV lithography equipment and plans its 14A process to implement backside power delivery technology. TSMC, however, has said it does not plan to incorporate High-NA EUV into its production nodes before 2029.
Despite these advantages, Intel is not making aggressive performance promises. It is because customers who are considering using Intel Foundry but not established manufacturing partners need predictable manufacturing yields, costs and delivery times.
Intel 14A And TSMC A14 Target 2028 Production
Chandrasekaran spoke about Intel's 14A expectations at an investor meeting at KeyBanc Capital Markets. He said Intel expects the process to be within 5 per cent of TSMC’s A14 in terms of performance.
The statement does not say whether Intel expects 14A to be faster or slower than A14. Instead, it sets a relatively narrow performance range between the two technologies.
Both 14A and A14 are part of the next generation of advanced semiconductor manufacturing processes and are scheduled to be in volume production in 2028. Intel plans to start risk production of its own 14A products in the second half of 2027 before moving to a more widespread production ramp.
The “1.4nm-class” terminology employed for these technologies is also not meant to be a literal measurement of every transistor feature. Modern semiconductor node names are mainly used to describe different generations of manufacturing technology as opposed to one physical dimension on a chip.
Intel And TSMC Have Different Performance Targets
Intel had previously forecast that 14A could deliver 15 per cent to 20 per cent higher performance than 18A at the same power level. It also said at the same performance level, 14A would reduce power consumption between 25 per cent and 35 per cent compared to 18A.
TSMC estimates that A14 could offer up to 15 per cent higher performance than its N2 process. At the same level of performance, TSMC expects A14 to provide up to 30 per cent lower power consumption.
However, comparing process technologies purely through published performance percentages has limitations. The final performance of a processor also depends on architecture, design, cooling, memory, software and other factors.
That means a few percentage points of difference between manufacturing nodes will not necessarily translate into the same difference between finished consumer products.
High-NA EUV Gives Intel An Early Manufacturing Advantage
One of Intel's most eagerly awaited advantages is its early adoption of High-NA EUV lithography. High-NA EUV, which is developed by Dutch semiconductor equipment manufacturer ASML, is designed for higher resolution than the state-of-the-art EUV lithography equipment.
Intel has been a major early adopter of the technology and installed its first commercial High-NA system at its Oregon research facility in 2024. The company has also worked with ASML on the factory-ready EXE:5200B High-NA platform.
The technology is expected to reduce the number of manufacturing steps required for certain critical chip layers. Intel has said that High-NA can reduce one layer's processing requirements from around 40 steps to less than 10 in some applications.
TSMC has chosen a different approach. The Taiwanese foundry intends to continue using its existing EUV equipment for several upcoming nodes and does not currently expect High-NA to enter production on its scheduled nodes through 2029.
TSMC has purchased High-NA equipment for research purposes, but the company has said that production adoption will depend largely on economics and technological maturity.
Backside Power Is Another Key Intel 14A Feature
Intel's other major technology advantage is backside power delivery. Its 14A process is expected to use PowerDirect, which moves power delivery connections to the backside of the semiconductor.
The traditional chip designs usually send the power and signals through structures above the transistor layer. Moving power delivery to the backside might free up space and may increase the efficiency of power delivery and signal processing.
But TSMC’s approach is different in terms of timing. The company has said that its backside power technology should come later in its roadmap and the A12 generation will be equipped with it in 2029.
This gives Intel another technology milestone to highlight as it looks to attract more external customers to its foundry business.
Why Intel Is Being Careful With Its 14A Claims
Intel's conservative performance guidance also reflects the company’s need to rebuild confidence among external foundry customers. Semiconductor customers do not evaluate a manufacturing process solely on peak performance claims.
Yield, production capacity, wafer costs, delivery schedules and long-term manufacturing consistency are equally important. Intel has faced delays with manufacturing technologies in the past like its 10nm generation and its 20A process was discontinued before reaching high-volume production.
Intel has said that 14A defect density is improving rapidly, with Chief Financial Officer David Zinsner describing the progress as the fastest improvement since the company's 22nm generation.
Accordingly, the company also warned that it would consider rethinking or pausing 14A investment if it was unable to find a major external customer. Now more recently, Intel has said potential customers have moved from evaluating the technology to discussions about capacity requirements.
High-NA Equipment Comes With A Major Cost
Intel’s early High-NA strategy also comes with a huge financial commitment. ASML’s High-NA EUV scanners cost hundreds of millions of dollars each; that is much higher than conventional EUV systems.
Intel has also increased its capital expenditure budget for 2026 to roughly US$20 billion, in line with the scale of investment required for expanding and modernising its manufacturing operations.
It is therefore not just an advanced process that Intel needs to build and develop. The company also has to have enough demand in its factories to justify the investment.
A cautious promise that 14A will remain within 5 per cent of A14 may be part of a broader strategy. Intel Foundry customers want confidence that the company will be able to deliver wafers at the expected yield and cost while maintaining production schedules.
What Intel 14A And TSMC A14 Could Mean For Future Chips
The competition between Intel 14A and TSMC A14 could impact processors used in laptops, smartphones, servers and artificial intelligence systems from 2028 onwards.
Intel has the advantage of producing many of its own processors internally and attempting to grow its external foundry business. TSMC, on the other hand, has a large established customer base that includes major semiconductor and technology companies.
Before 14A arrives, Intel is also developing new versions of its 18A manufacturing technology, such as 18A-P, which could give potential customers a chance to see Intel’s foundry capacity before committing to the more advanced 14A generation. The next major milestones will be customer commitments, risk production and yield improvements. Intel has indicated that customer decisions on 14A could begin taking shape during the second half of 2026.
For both Intel and TSMC, the coming battle will be more than a percentage of performance. The ability to produce advanced chips consistently, at competitive costs and in sufficient volumes might be just as important as the underlying technology.
