Chamath says Nvidia controls most AI memory supply, leaving Google short
Microsoft announced its third‑generation Maia 300 AI chip on Oct. 12 and said it will scale the chips to multi‑gigawatt deployments starting in 2027. The chip is designed for higher memory bandwidth, but Chamath Palihapitiya warned on the Oct. 16 All‑In podcast that Nvidia now controls most HBM packaging and memory capacity, leaving even Google short on memory.
Key points
- Microsoft unveiled the Maia 300 AI chip on Oct. 12, promising multi‑gigawatt deployments starting in 2027.
- Chamath Palihapitiya says Nvidia controls most HBM packaging and memory, leaving Google short on capacity.
- Micron completed agreements for most of its calendar 2027 HBM supply and nearly doubled capital spending to $30.7 billion in fiscal 2026.
Micron Technology reported that it has completed agreements for the vast majority of its calendar 2027 HBM supply and that its capital spending has nearly doubled to $30.7 billion in fiscal 2026. The company added that most of the new spending will fund clean‑room space arriving in late 2028 and beyond, when rival chips could begin to scale faster. Bloomberg Intelligence’s Vasu Kasibhotla noted that chip, optical and memory makers have sold out capacity in 2027 and likely much of 2028 as well.
The story so far
6 episodes →- Chamath says Nvidia controls most AI memory supply, leaving Google shortthis story
Chamath: Even Google “Tin Cups for Memory” Thanks to Nvidia. Here’s What Happens Next
finance.yahoo.com · 19 October 2026
Loading the full article…
This text was published by finance.yahoo.com and written by David Beren. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Hardware & Compute
All →- Tenstorrent raises $1.5B Series E at $6.2B valuation · 2 src
- Nvidia says Vera CPU delivers 3.4 TB/s bandwidth and higher single-thread performance · 6 src
- DeepSeek plans to use Huawei Ascend 950DT for pre‑training by end of 2026 · 2 src
- Ijah launches Metal‑based inference engine for Qwen 3.8‑Flash‑Next on Apple M5 · 1 src
- PNY offers NVIDIA RTX PRO 6000 Blackwell Workstations with up to 14.2 petaflops · 1 src
Comments
via GitHub Discussions