AI chip founders

Founders designing energy efficient AI chips, discovering new chip materials and moving compute into space, on how they compete with the incumbents.

Y Combinator batch P26
20 Years of R&D. Matched in 3 Months by 2 Friends.Advaith Sridhar / Discovered MaterialsOct 22, 2026 · 1h 11m
Y Combinator batch S24
Is NVIDIA’s Moat a Psyop? 28x More Efficient AI ChipsElias Almqvist / ZettascaleMar 10, 2026 · 1h 10m
Y Combinator batch S24
Data Centers in Space: The Future of AI ComputePhilip Johnston / StarcloudOct 3, 2025 · 51m

Moments

3:53
Claude writing all his code sparked the materials idea
Advaith Sridhar, Discovered Materials
20:18
Seeding the chips with frontier researchers first
Elias Almqvist, Zettascale
2:04
What a data center in space actually looks like
Philip Johnston, Starcloud
5:21
Agents did in three weeks what took Akash a year
Advaith Sridhar, Discovered Materials
25:12
Monolith, the in-house cluster researchers can SSH into
Elias Almqvist, Zettascale
8:17
The top investor objections and a slide for each one
Philip Johnston, Starcloud
26:20
Why heat is the main problem in chips today
Advaith Sridhar, Discovered Materials
26:36
Switching from GPUs means installing a driver, not rewriting code
Elias Almqvist, Zettascale
19:32
Why Starcloud launches every year instead of one big bet
Philip Johnston, Starcloud
31:21
Thermal pastes matching Dow, Honeywell and Henkel in three months
Advaith Sridhar, Discovered Materials
30:13
The trade-off: higher latency in exchange for efficiency
Elias Almqvist, Zettascale
20:29
Building and launching the first satellite on a $2.5M pre-seed
Philip Johnston, Starcloud
32:22
Why the robotic arm helped less than changing equipment
Advaith Sridhar, Discovered Materials
33:38
A team with no roles, from AI research down to silicon
Elias Almqvist, Zettascale
24:52
Processing Earth observation imagery in orbit instead of downlinking it
Philip Johnston, Starcloud
43:18
Pausing thermal interface materials for bigger markets
Advaith Sridhar, Discovered Materials
45:40
Why spatial reconfigurability was dropped
Elias Almqvist, Zettascale
27:27
Why waste heat will force compute off Earth
Philip Johnston, Starcloud
49:17
Pitching more than 100 times in two weeks
Advaith Sridhar, Discovered Materials
47:07
Meeting co-founder Prithvi at a week-long event in London
Elias Almqvist, Zettascale
43:27
The worst take: build in space because space is cold
Philip Johnston, Starcloud
1:07:51
Contrarian takes: CUDA is not a moat, LLMs are plateauing
Elias Almqvist, Zettascale
48:04
What a 40 megawatt module costs to build and launch
Philip Johnston, Starcloud

AI chip founders

Moments

Claude writing all his code sparked the materials idea

Seeding the chips with frontier researchers first

What a data center in space actually looks like

Agents did in three weeks what took Akash a year

Monolith, the in-house cluster researchers can SSH into

The top investor objections and a slide for each one

Why heat is the main problem in chips today

Switching from GPUs means installing a driver, not rewriting code

Why Starcloud launches every year instead of one big bet

Thermal pastes matching Dow, Honeywell and Henkel in three months

The trade-off: higher latency in exchange for efficiency

Building and launching the first satellite on a $2.5M pre-seed

Why the robotic arm helped less than changing equipment

A team with no roles, from AI research down to silicon

Processing Earth observation imagery in orbit instead of downlinking it

Pausing thermal interface materials for bigger markets

Why spatial reconfigurability was dropped

Why waste heat will force compute off Earth

Pitching more than 100 times in two weeks

Meeting co-founder Prithvi at a week-long event in London

The worst take: build in space because space is cold

Contrarian takes: CUDA is not a moat, LLMs are plateauing

What a 40 megawatt module costs to build and launch

Pull up a chair.