Comments for SemiWiki

Comment on Breker Hosts an Energetic Panel on Spec-Driven Verification by Daniel Nenni

Daniel Nenni — Thu, 19 Mar 2026 01:01:55 +0000

“The future of DV isn’t fewer engineers- it’s higher-leverage verification architects.”

I agree 100%. The future is much more complex designs that meet spec the first time, every time.

Comment on Breker Hosts an Energetic Panel on Spec-Driven Verification by shellyhenry

shellyhenry — Wed, 18 Mar 2026 15:49:37 +0000

Great insights! This nails the real shift happening.

AI in verification isn’t about replacing engineers, it’s about restructuring the workflow around spec-driven, human-verifiable stages. The real challenge (and opportunity) is the last 5%—where trust, alignment, and debugging matter more than raw generation.

The future of DV isn’t fewer engineers- it’s higher-leverage verification architects.

Comment on Tesla and Samsung Relationship Update by benb

benb — Tue, 17 Mar 2026 16:51:57 +0000

Sam Altman has been quiet about spending big on fabs. He was the first to raise the AI megafab idea.

It is the right place to spend big, I think–you gain more control. Vertical integration is a benefit in a big new market.

I don’t think Intel, Samsung or TSMC are for sale. But GF may be. And GF has done technology transfer from Samsung before.

Comment on Intel Foundry: How They Got Here and Scenarios for Improvement by Jim351

Jim351 — Tue, 17 Mar 2026 00:26:38 +0000

Starting to run into some speculation that Elon might want help from Intel in setting up his TeraFab in TX for 2nm node products. As an industry vet in the materials space, I can’t feature how Elon could do this alone without help from one of the big three. Could be a good catalyst for Intel’s future stability if it comes to pass.

Comment on Unraveling Dose Reduction in Metal Oxide Resists via Post-Exposure Bake Environment by Fred Chen

Fred Chen — Mon, 16 Mar 2026 06:41:51 +0000

Pollentier and Dorney’s team at IMEC have published multiple papers in the area of metal oxide resist (MOR) characerization, using their BEFORCE platform. What this latest result is saying is that without oxygen exposure at bake after EUV, this resist will not work. The threshold dose is sensitive to the amount of oxygen and humidity in the ambient. Lowering the threshold dose will of course worsen the stochastics (<10 photons absorbed/nm2). While maintaining the same dose, the lower threshold also means more thinning, which also worsens stochastics (from the other side).

Come to think of it, there have been no printing of features. The oxygen diffusion can be an extra source of blur, as with acids for chemically amplified resists.