/AI5h ago

AI x Bio Labs Need 8 H200 GPUs, 300 CPU Cores, Petabyte Storage

212201.4K
Original post
Anshul Kundaje@anshulkundaje#1650inAI

Let's talk about compute hardware. To be seriously competitive in AI x Bio, we minimally need something in the range of 8 H200 GPUs, 300 CPU cores, a PetaByte of disk space (let's exclude back up costs for now).

Anshul Kundaje@anshulkundaje

A staff scientist ranges between 110K - 150K salary. Which comes out to at least 146K - 190K per year.

9:27 AM · Jun 9, 2026 · 697 Views
Sentiment
Sentiment building, check back later.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS692LIKES5RETWEETS1REPLIES1
Anshul Kundaje@anshulkundaje

Even if we went the cheapest route and try to buy hardware ourselves, this would cost close to at least 800K - 1M one time cost that would need hardware refresh once every 4 years at least. Would be much higher if it were a cloud option. This is excluding back up, power etc

Anshul Kundaje@anshulkundaje

Let's talk about compute hardware. To be seriously competitive in AI x Bio, we minimally need something in the range of 8 H200 GPUs, 300 CPU cores, a PetaByte of disk space (let's exclude back up costs for now).

5hViews 692Likes 5Bookmarks 0