Patel: AI chip bottlenecks lie in interconnects and memory, not matrix multiplication
Dylan Patel · The State of Silicon and the GPU Poors - with Dylan Patel of SemiAnalysis · Dec 5, 2023 · at 24:06
Dylan Patel, Chief Analyst at SemiAnalysis, explains why AI hardware startups struggle against incumbents despite designing good compute units.
0:00 / 0:14exact quote · 14.3s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“Multiplying tensors is kind of, you know, anyone can, there's a lot of people who've made good matrix multiply units, right? But it's about, like, getting good utilization out of those, and interfacing with the memory, and interfacing with other chips really efficiently makes designing these chips very hard. And most of the startups obviously have not done that really well.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Dylan Patel
Dylan Patel: Political coup in Taiwan is more likely than full invasion
“What I think is most likely is that there is some sort of political coup or action. That destabilizes Taiwan in some way but doesn't actually, doesn't necessarily lead to an invasion, a full-scale invasion, and so sort of this is the best of the both worlds fo…”
Patel: China Has World's Most Advanced Standalone Chip Supply Chain
“China, China has by far the most advanced supply chain in semiconductors if you just look at China itself, right?”
Patel: Anthropic added $2B in monthly revenue at positive margins
“Anthropic doesn't just add two billion dollars of revenue in one month. You know, with, without having, you know, huge demand and they're doing it at positive margins, right?”
Patel: Google will have zero free cash flow in 2027 due to AI CapEx
“There's no reason why Google will have any profit in 27 at all, right, in terms of cash flow. They will just spend every dollar they make On, on AI infrastructure.”