Intel's 14A chip process in jeopardy, while cutting jobs and pulling back global expansion

[Skye Jacobs]

Cutting corners: Intel's next leap in semiconductor manufacturing, known as the 14A process, now hangs in the balance. The company warned that it could slow down or even halt development of the advanced 1.4nm-class technology unless it secures a significant external customer and meets critical project milestones.

This admission, made public in a regulatory filing, marks a notable shift for the chipmaker, which has traditionally set the pace for pushing leading-edge technologies but now risks ceding ground to rivals such as TSMC and Samsung Foundry.

The future of the 14A node – a successor to Intel's 18A and 18A-P processes – was called into question as executives discussed the company's earnings and investment strategy. While Intel remains committed to its current technology roadmap, CEO Lip-Bu Tan emphasized a new approach of market-driven development.

"The 14A is the process node, but clearly I will make sure that I see the internal customer, external customer, and volume commitment before I put CapEx," Tan told analysts during the earnings call. "But that is something that has to meet my requirement in terms of performance and yield. It's a lot of responsibility to be serving our customers, make sure that we can deliver the result – consistent, reliable result – to them, so that their revenue can depend on us."

Developing technologies at the edge of what's possible isn't cheap. Intel spent over $16.5 billion on research and development in 2024, a sum that underscores the vast investment required to bring the 14A process to market. Its 14A node is expected to be its first to leverage high-numerical-aperture extreme ultraviolet (High-NA EUV) lithography for multiple critical layers, with each machine costing around $380 million. The cost of scaling 14A to high-volume manufacturing is so high that Intel insists these tools must serve both in-house and external customer needs.

The company's willingness to reconsider major technology projects reflects a hard pivot. If 14A does not achieve the commercial traction required, future high-performance Intel products could be outsourced to foundries like TSMC – a move that would have been unthinkable for Intel just a few years ago.

This tough assessment comes as Intel undertakes a dramatic restructuring of its global operations. According to Team Blue's latest earnings call and regulatory filings, the firm expects to end 2025 with roughly 75,000 "core employees" worldwide, down from nearly 100,000 at the end of the prior year. That reduction – amounting to approximately 24,000 jobs – reflects both layoffs and internal divestitures, shrinking the company by about one-quarter and marking the steepest workforce reduction in Intel's history.

Major international projects have now been shelved. Plans for large new chipmaking "mega-fabs" in Germany and Poland, once envisioned as bringing thousands of jobs to the region, are canceled. Assembly and test operations in Costa Rica will be consolidated or moved to larger facilities in Vietnam and Malaysia, though some engineering teams will remain in Costa Rica. In the United States, Intel is slowing construction on factory projects in Ohio and other sites to "align with market demand," according to its chief financial officer.

Addressing employees after the latest round of layoffs, Tan described the moment as a necessary inflection point. "We are making hard but necessary decisions to streamline the organization, drive greater efficiency and increase accountability at every level of the company," he said in a message to employees.

On the earnings call, he added, "I do not subscribe to the belief that if you build it, they will come. Under my leadership, we will build what customers need when they need it, and earn their trust."

Permalink to story:

Intel's 14A chip process in jeopardy, while cutting jobs and pulling back global expansion

 

[unoficialoficial]

If Intel wants to meet demand, they should offer a Threadripper-like processor with 3D cache at the back at the price of a Ryzen, produced on a 14+nm node. They should ditch the performance-weak cores completely; laptops are the only ones that need those. The demand will be driven by AI requirements, which means integrating an NPU and a lot of high-bandwidth DDR6 memory channels. It's not difficult to predict; nothing else is driving demand like AI. It's the elephant in the room. Create a consumer system capable of running the Deepseek R2 at 20 tk/second with full context. That's the target and they should aim to achieve it within the next 12 months. They are now a startup, made of seasoned veterans, but without the low expenses of a true startup. They need to move quickly.

 

[wiyosaya]

We've heard this story before. Wash, rinse, repeat. Intel hubris on display - STILL.

 

[Revolution 11]

We've heard this story before. Wash, rinse, repeat. Intel hubris on display - STILL.

I don't know wiyosaya, there is no evidence that Intel's hubris is gone but neither is there any evidence that it is still there.

These are pretty big workforce cuts by Intel and pretty much most of Pat's projects are canceled or reduced in scope.

What would make you say that the hubris is gone?

 

[wiyosaya]

I don't know wiyosaya, there is no evidence that Intel's hubris is gone but neither is there any evidence that it is still there.

These are pretty big workforce cuts by Intel and pretty much most of Pat's projects are canceled or reduced in scope.

What would make you say that the hubris is gone?

Time will tell. If they are half-admitting failure at this point when they stuck at it for so long, I rest my case.

 

[user556]

Man, not moving forward with those High-NA machines is going to sting down the track. How many did Intel buy in the end? It was rumored to be all the 2024 shipments from ASML.

 

[Kashim]

So tired of hearing about Intel's doom and gloom future. Who cares? Either do something or quietly go into the night, but shut up about it already. As far as I'm concerned Intel is getting exactly what they deserve. They spent decades abusing their market position, squeezing out competition, paying off big customers to use Intel parts exclusively, pushing proprietary garbage every chance they had, etc. You reap what you sow, as the saying goes. They didn't play nice when they had the chance and now they whine and moan about things not going their way. Bye Felicia.

 

[SRB]

I don't know wiyosaya, there is no evidence that Intel's hubris is gone but neither is there any evidence that it is still there.

These are pretty big workforce cuts by Intel and pretty much most of Pat's projects are canceled or reduced in scope.

What would make you say that the hubris is gone?

I think Pat gets the short end of this deal. The design side of Intel was screwed up long before Pat got back. Pat's idea was to shore up the company by getting other companies to use Intel foundries for their chips and bring in more revenue. The problem is the board is just as myopic as every other large U.S. company, making things costs money and requires investment. Service companies only have to pay employees (and Tan knows you can add or subtract those at will). Tan was hired to part out the company, starting with the foundries. He's just trying to do it without leaving his or the board's fingerprints on it.

So, no more foundry support or investment until they turn a profit. Good luck with that. Especially when you cancel a node you already started promoting and selling (al least as far as the outside customers are concerned).