ARM's debut as a chip manufacturer threatens its own licensees just as TSMC confirms an AI chip shortage running through 2027 — reshaping the entire semiconductor supply chain. Plus: HBM4 supply locked for Vera Rubin, and Google's $30B SpaceX GPU deal signals just how strained Nvidia allocations have become.
Audio is available on Spreaker — see link below.
ARM just made a move that changes how you think about it. In March twenty twenty-six, the company announced its AGI CPU, its first foray into actual chip manufacturing.
Underpinning all of this is a supply environment that gives ARM timing leverage it couldn't have planned better. TSMC's CEO confirmed the AI chip shortage runs through twenty twenty-seven.
On the memory side, Jensen Huang confirmed what the market needed to hear. Samsung, SK hynix, and Micron are all in active HBM4 production, with customer shipments targeted for Q3 twenty twenty-six to support Vera Rubin.
The clearest evidence of that near-term pressure came from Google. Alphabet agreed to pay SpaceX nine hundred twenty million dollars monthly for a hundred and ten thousand H200 GPUs through June twenty twenty-nine.
Meanwhile, TSMC is quietly reinforcing something important. Two-nanometer production stays in Taiwan through the end of the decade.
Coming back to ARM. The question isn't whether the AGI CPU is technically sound.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.