Navigate Select ESC Close

Dylan Patel — The Single Biggest Bottleneck to Scaling AI Compute

2026-03-13 Science & Technology
58.6k
1.4k
165
Dwarkesh Patel
Dwarkesh Patel
1.3m subscribers

Unlock all features

FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.

Description

Dylan Patel, founder of SemiAnalysis, provides a deep dive into the 3 big bottlenecks to scaling AI compute: logic, memory, and power. And walks through the economics of labs, hyperscalers, foundries, and fab equipment manufacturers. Learned a ton about every single level of the stack. Enjoy! 𝐄𝐏𝐈𝐒𝐎𝐃𝐄 𝐋𝐈𝐍𝐊𝐒 * Transcript: https://www.dwarkesh.com/p/dylan-patel * Apple Podcasts: https://podcasts.apple.com/us/podcast/dylan-patel-deep-dive-on-the-3-big-bottlenecks-to/id1516093381?i=1000755126873 * Spotify: https://open.spotify.com/episode/5qiibwoBWY5rXyflK7WJzH?si=SX4ajSKXT-KeNtaHsiTNzw 𝐒𝐏𝐎𝐍𝐒𝐎𝐑𝐒 - Mercury has already saved me a bunch of time this tax season. Last year, I used Mercury to request W-9s from all the contractors I worked with. Then, when it came time to issue 1099s this year, I literally just clicked a button and Mercury sent them out. Learn more at https://mercury.com - Labelbox noticed that even when voice models appear to take interruptions in stride, their performance degrades. To figure out why, they built a new evaluation pipeline called EchoChain. EchoChain diagnoses voice models’ specific failure modes, letting you understand what your model needs to truly handle interruptions. Check it out at https://labelbox.com/dwarkesh - Jane Street is basically a research lab with a trading desk attached – and their infrastructure backs this up. They’ve got tens of thousands of GPUs, hundreds of thousands of CPU cores, and exabytes of storage. This is what it takes to find subtle signals hidden deep within noisy market data. If this sounds interesting, you can explore open positions at https://janestreet.com/dwarkesh To sponsor a future episode, visit https://dwarkesh.com/advertise. 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 00:00:00 – Why an H100 is worth more today than 3 years ago 00:24:52 – Nvidia secured TSMC allocation early; Google is getting squeezed 00:34:34 – ASML will be the #1 constraint for AI compute scaling by 2030 00:55:47 – Can't we just use TSMC's older fabs? 01:05:37 – When will China outscale the West in semis? 01:16:01 – The enormous incoming memory crunch 01:42:34 – Scaling power in the US will not be a problem 01:54:44 – Space GPUs aren't happening this decade 02:14:07 – Why aren't more hedge funds making the AGI trade? 02:18:30 – Will TSMC kick Apple out from N2? 02:24:16 – Robots and Taiwan risk

Top Comments (10)

@Jobox05 2026-03-13

Not even a minute in and Dwarkesh asks exact, detailed questions about the future of the industry. Other podcasts even with the same guest dont have this density of conversation

223 10 replies
@bulltheknicksfan3140 2026-03-13

Leopold next please

200 10 replies
@TheDoubleMvp 2026-03-13

I'm always in awe at just how much Dylan knows about the semi industry. Human ChatGPT for semis.

130 1 replies
@baerstudio 2026-03-13

"hey bae wanna hear 2 and a half hours on supply chain constraint issues related to highly technical euv 2-3nm chip production?"

77 4 replies
@RichardFarland-yt7kf 2026-03-13

I believe the “Carl Zeiss” Dylan and Dwarkesh are talking about (the ASML optics bottleneck) is actually **Carl Zeiss SMT** — a private subsidiary of the unlisted Carl Zeiss AG. The $2.5B public market cap they quoted belongs to the completely unrelated **Carl Zeiss Meditec AG**, which makes eye surgery equipment and has zero connection to EUV lithography or AI chips.

60 3 replies
@edgarsfishchannel 2026-03-14

So glad YouTube randomly suggested a short about ww2 history with Sarah Paine on DK podcast 1 year ago and now I binge watch all of these ai/ semiconductor ceo interviews😅😂

19 2 replies
@imdanielpiva 2026-03-14

It's shocking how a 2h30 video podcast of pure and dense conversation at 2x speed about scaling AI compute can fix me.

8
@mikecounsell 2026-03-14

I expect to listen to this at least 3 times

4
@dB2-gg 2026-03-15

First time commenter, long time listener. Conversations with you and your guests are consistently the most consciousness-expanding part of my routine. Thanks for broadcasting to those of us out here in the long tail.

1
@ClipMindset0 2026-03-14

20 minutes in and these insights are incredible. Way more depth and knowledge here than any other podcast or platform. Good work dwarkesh!

1

Unlock the Data Inside
Turn Videos into Knowledge

  • Get FREE 10/day: transcripts, summaries, chats
  • Chat with videos, export text & PDF
  • $1 free API credit for RAG, chatbots & research

Free forever plan • All features unlocked

App screenshot