Now Playing
High-NA EUV Race: ASML, Terafab 2028 & Japan's Yttrium Crisis
Audio is available on Spreaker — see link below.
ASML's High-NA EUV acceleration is reshaping every advanced AI chip roadmap — and a critical yttrium shortage in Japan could stop it cold. Today's briefing covers Nvidia's Vera Rubin revenue gap, Broadcom's hyperscaler shift, Samsung's memory push, and the lithography bottleneck tightening across TSMC, Intel, and Samsung.
Audio is available on Spreaker — see link below.
ASML just accelerated deployment of its High-NA EUV machines across TSMC, Intel, and Samsung, and that single move is reshaping the timeline for every advanced AI chip on the roadmap. Here's the key thing to understand.
That timing connects directly to Nvidia. The company's Vera Rubin platform generates roughly forty billion dollars in revenue per gigawatt of compute deployed, compared to twenty-five billion for its Grace Blackwell generation.
The pressure on ASML's allocation schedule gets more acute when you factor in Terafab. The SpaceX-Tesla joint venture has now publicly committed to volume production in Q3 twenty twenty-eight, with a test SRAM chip milestone targeted for Q2 twenty twenty-five.
Samsung's announcement of its Core Team R&D unit is a more contained signal, but the direction is clear. The unit focuses on next-generation DRAM and NAND for AI infrastructure demand, which is a direct response to competitive pressure from Micron's HBM3e progress and SK Hynix's current dominance in high-bandwidth memory.
The story that carries the most underappreciated risk is Japan's yttrium shortage. China restricted yttrium exports to Japan starting around twenty ten, tightened those restrictions in early twenty twenty-six, and stopped key shipments by July twenty twenty-six.
One more signal worth tracking. Broadcom's AI semiconductor revenue grew two hundred twenty-one percent to sixteen point seven billion dollars.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.