scripod.com

20VC: Cerebras CEO on Why Raise $1BN and Delay the IPO | NVIDIA Showing Signs They Are Worried About Growth | Concentration of Value in Mag7: Will the AI Train Come to a Halt | Can the US Supply the Energy for AI with Andrew Feldman

Shownote

Andrew Feldman is Co-Founder & CEO of Cerebras, building the world's fastest AI inference and training. Cerebras recently closed a $1.1BN Series G round at an $8.1 billion valuation, backed by top names including Fidelity, Atreides, Tiger Global, Valor Equ...

Highlights

In this episode, Andrew Feldman, CEO and Co-Founder of Cerebras, dives into the evolving landscape of AI infrastructure, discussing the company's recent $1.1 billion funding round and its strategic path toward an IPO. The conversation explores the technological, economic, and geopolitical forces shaping the future of AI compute.
05:03
Incremental improvements in AI won't lead to significant progress
12:29
NVIDIA's investment in OpenAI is designed to be unclear and lock up demand
13:59
Older AI chips like H100s and A100s remain valuable, challenging the 18-month depreciation rule.
24:58
Using AI in fundamentally new ways leads to massive productivity gains.
29:26
The Mag7 stock concentration creates a false sense of security in the S&P 500.
31:57
Demand for AI infrastructure is five times current supply capacity
32:59
Top AI talent commands unprecedented compensation due to extreme demand
34:25
No company goes bankrupt paying extraordinary technical talent well.
40:56
Major leaders missed the largest compute market in the 21st century, while ARM won it
45:18
China's aggressive AI policy and drone robotics lead contrast with US structural hurdles
47:31
75-80% of Cerebras' 2024 revenue comes from UAE, signaling strong regional innovation adoption.

Chapters

AI Infrastructure and Cerebras' Growth
00:00
Why We Did Not IPO and Raised $1BN From Fidelity
02:43
Analysis of Chip and Compute Landscape Today
05:03
NVIDIA Showing Signs They Are Running Out of Ideas
07:14
The Real Questions to Ask on Chip Depreciation
13:57
Energy Requirements for AI: Is it Feasible?
24:54
Mag7 Value Concentration: Feature or a Bug
29:25
Talent is the Bottleneck and Trump Makes it Worse
31:57
The War for Talent: Secrets No One Sees
32:55
Evaluating the Data Centre Economy: Many Will Lose Money
34:22
Three Changes the US Could Make to Beat China in AI
38:01
Why 80% of our Revenues are in the UAE
42:30
Quick Fire Questions
47:26
Why Work Life Balance is Total BS
58:59

Transcript

Andrew Feldman: Things are moving at a rate that 6, 8, 12 months out, everybody's unsure. It's so fast. It's so big. There is unbelievable demand, and nobody knows where it will go in the future. The question of depreciation is, how much faster are future ...