Intel has officially confirmed what has long been rumored in the industry: its upcoming manufacturing node, Intel 14A, will cost more than the current Intel 18A, primarily due to the adoption of EUV High-NA (High Numerical Aperture) lithography developed by ASML. This cutting-edge technology promises a leap in performance and efficiency but also raises questions because of its high cost and the lack of external clients willing to pay the premium.
More efficiency, but at a higher price
According to Intel, the 14A node will deliver a 15% to 20% improvement in performance per watt, along with a 25% to 35% reduction in energy consumption compared to 18A. Regarding density, the company estimates nearly a 30% increase, which would theoretically allow for manufacturing smaller or more powerful chips within the same space.
However, as Intel CFO David Zinsner explained:
“The 14A node is more expensive than 18A. It’s not significantly more costly in terms of investment. However, the wafer cost is undoubtedly higher, partly because we anticipate using high-NA EUV tools in 14A—a technology that was not used in 18A.”
In other words, while the technological leap is assured, the production cost per wafer will be considerably higher, impacting the final price of CPUs and GPUs manufactured with this process.
Intel 14A innovations
The 14A node not only introduces EUV High-NA but also includes several key improvements:
- RibbonFET 2: an evolution of Intel’s GAA transistors, preceding the future Forksheet design.
- PowerDirect (formerly PowerVia): a backside power distribution network that enhances energy efficiency and reduces interference.
- Turbo Cells: double-height cells that optimize synchronization and enable higher frequencies in CPUs and GPUs without disproportionately increasing power consumption.
Combined, these technologies are expected to position Intel 14A at the forefront of the semiconductor industry.
The big challenge: profitability
The main obstacle is economic. The ASML Twinscan EXE:5200B EUV High-NA scanner costs around $380 million per unit, compared to $235 million for current EUV equipment. Additionally, the estimated cost of each factory built for this node is about $30 billion.
The risk is clear: if Intel cannot attract enough external customers for its foundry business (Intel Foundry Services, IFS), the viability of the 14A node will be in question. Zinsner himself acknowledged:
“If we don’t secure external customers for Intel 14A, it will be difficult to justify that node. Intel Products will be an important client, but we need more demand to ensure a reasonable return on investment.”
The role of ASML and pressure in Europe
This situation also impacts Europe, as ASML, the Dutch manufacturer of EUV scanners, heavily depends on the success of this node to promote High-NA adoption. If Intel and other major players like TSMC fail to attract clients, the billion-dollar investment in this technology could be compromised.
Indeed, TSMC has already shown caution: it has delayed its adoption of EUV High-NA and might continue to optimize its current EUV lithography, which is cheaper and sufficiently mature for its immediate processes.
What might happen?
The outlook presents several scenarios:
- Moderate success of Intel 14A: Intel convinces some large-volume clients (perhaps GPU manufacturers or cloud providers) and justifies the cost.
- TSMC dominance with conventional EUV: if Taiwanese company delays High-NA adoption and continues optimizing current EUV, it could keep competitive prices and maintain industry leadership.
- Impact on ASML and European industry: if adoption is slow, ASML might struggle to sell its higher-priced equipment, affecting its business model and European Union’s technological sovereignty ambitions.
Conclusion
Intel 14A promises revolutionary advances in transistors, efficiency, and performance, but its greatest challenge isn’t engineering—it’s profitability. The company faces the task of convincing the market that the premium cost of the node translates into clear benefits for customers.
Time will tell if Intel can turn 14A into the new standard or if, conversely, its strategy is hindered by a lack of confidence in EUV High-NA and pressure from competitors like TSMC.
FAQs about Intel 14A and EUV High-NA
Why will Intel 14A be more expensive than 18A?
Because it introduces EUV High-NA scanners from ASML, which are much more costly ($380 million compared to $235 million for current EUV).
What improvements does Intel 14A offer over 18A?
A performance per watt increase of 15-20%, energy savings of 25-35%, and a 30% increase in density.
Who will manufacture the EUV High-NA scanners?
The Dutch company ASML, which already dominates advanced lithography markets.
What if Intel fails to secure external clients for 14A?
Intel itself admits it would be difficult to justify the investment. In that case, the node could be delayed or even canceled, affecting the future of EUV High-NA technology.