←back to thread

Ask HN: Why hasn’t AMD made a viable CUDA alternative?

183 points spacebanana7 | 6 comments | 01 Apr 25 14:37 UTC | HN request time: 0s | source | bottom

I appreciate developing ROCm into something competitive with CUDA would require a lot of work, both internally within AMD and with external contributions to the relevant open source libraries.

However the amount of resources at stake is incredible. The delta between NVIDIA's value and AMD's is bigger than the annual GDP of Spain. Even if they needed to hire a few thousand engineers at a few million in comp each, it'd still be a good investment.

Show context
danielmarkbruce ◴[01 Apr 25 14:49 UTC] No.43547462[source]▶
They likely haven't put even close to enough money behind it. This isn't a unique situation - you'll see in corporate america a lot of CEOs who say "we are investing in X" and they really believe they are. But the required size is billions (like, hundreds of really insanely talented engineers being paid 500k-1m, lead by a few being paid $3-10m), and they are instead investing low 10's of millions.

They can't bring themselves to put so much money into it that it would be an obvious fail if it didn't work.

replies(3): >>43547563 #>>43547805 #>>43547934 #
1. spacebanana7 ◴[01 Apr 25 14:57 UTC] No.43547563[source]▶
Given how the big tech companies are buying hundreds of thousands of GPUs at huge prices, most of which is pure margin, I wonder whether it'd make sense for Microsoft to donate a couple billion to make the market competitive.

https://www.datacenterdynamics.com/en/news/microsoft-bought-...

replies(1): >>43547645 #
2. danielmarkbruce ◴[01 Apr 25 15:03 UTC] No.43547645[source]▶
The big players are all investing in building chips themselves.

And probably not putting enough money behind it... it takes enormous courage as a CEO to walk into a boardroom and say "I'm going to spend $50 billion, I think it will probably work, I'm... 60% certain".

replies(1): >>43547778 #
3. spacebanana7 ◴[01 Apr 25 15:13 UTC] No.43547778[source]▶
You're probably correct, but I feel like I have to raise the issue of Zuckerberg spending a comparable amount on VR which was much more speculative.
replies(2): >>43547864 #>>43548018 #
4. hnlmorg ◴[01 Apr 25 15:20 UTC] No.43547864{3}[source]▶
Zuckerberg owns Facebook though. It’s a lot easier to make bold decisions when you’re the majority shareholder.

Edit: though emphasis should be put on “easIER” because it’s still far from easy.

replies(1): >>43547935 #
5. danielmarkbruce ◴[01 Apr 25 15:26 UTC] No.43547935{4}[source]▶
This. Without knowing the guy, he seems to be a) very comfortable taking a lot of risk and b) it's actually not that risky for him to blow $20 billion.

There aren't many cases like this. Larry/Sergey were more than comfortable risking $10 billion here and there.

6. wavemode ◴[01 Apr 25 15:33 UTC] No.43548018{3}[source]▶
Zuck is founder and owner. So is Huang (Nvidia CEO). They call all the shots.

Whereas AMD's CEO was appointed, and can be fired. Huge difference in their risk appetite.

I'm reminded of pg's article "founder mode": https://paulgraham.com/foundermode.html

I think some companies simply aren't capable of taking big risks and innovating in big ways, for this reason.