Pinned Tweet
Thrilled to partner with @matrixvc for our next phase of growth
Today, we raised a $13M series A to solve the GPU shortage
Trillions are being spent on GPU CapEx while 80% of it sits idle
The capacity already exists
Virtualizing GPUs unlocks it
Backed by Matrix, YC, and CEAS
3
1
21
2,379
3/ @ThunderCompute
Thunder Compute raised a $13M series A to solve the GPU shortage.
Thunder Compute aims to improve GPU efficiency through virtualisation, and is backed by @ycombinator, Matrix Ventures and CEAS.
Today, we raised a $13M series A to solve the GPU shortage
Trillions are being spent on GPU CapEx while 80% of it sits idle
The capacity already exists
Virtualizing GPUs unlocks it
Backed by Matrix, YC, and CEAS
1
2
3
313
🚀 Just got through my first 24 hours of renting 8x Nvidia A100s(80gb each) on @ThunderCompute for my custom ~300gb quant of Kimi K3 by @Kimi_Moonshot ! Let me tell you how simple and affordable it is to spin up and spin down compute on this platform by connecting it to my agent via their MCP! Highly recommended if you can’t afford a $5000 DGX Spark like me 😆 (My last day with the two sparks I was working with was Friday 🥲) Try my referral link below if you want to give it a go!! 🙏 Extremely excited to share my Kimi K3 work with all of you!! My goal is to be right on par or just a bit below the PPL of the 1Bit @UnslothAI Kimi K3 model while being ~45% smaller! 🤓
console.thundercompute.com/s…
5
1
14
813
Thunder Compute retweeted
Furniture vendors: delivery in 6 weeks
Doordash: folding tables and chairs in 58 min
2
1
19
3,489
Thunder Compute retweeted
infra should bend to the workload
not the other way around
Thunder now lets you reshape an instance as you go
- gpu type
- storage
- vCPUs
without ever losing flow
2
7
555
Thunder Compute retweeted
~$2 an hour on-demand compute here. Just set your model to GLM Flash, create a snapshot, and then you can run your own private instance for all your agents, OpenClaw, and so on. Link below and you get $50 free with your $50 load!
thundercompute.com/?ref=memb…
1
1
3
769
We are rolling out a suite of infrastructure upgrades.
Expect some short-term hiccups for the sake of massive improvements👀
Please report errors to our devs on Discord, we're working hard to ensure a smooth transition.
2
2
7
2,132
We're proud to announce two security milestones:
GDPR Compliance - we have fully implemented controls for GDPR compliance.
SOC 2 Type II - Our three-month observation period with an independent auditor is underway. Controls in production; official certification soon.
2
14
1,699
@ThunderCompute is the easiest and cheapest way to try GPT oss. We wrote a 1-minute guide to show how:
1
2
10
1,195
Thunder Compute retweeted
@ThunderCompute has super cheap GPU instances, plus you can rent by a single GPU, instead of say 8xA100s for cloud providers.
1
3
536
Thunder Compute retweeted
We heard your feedback, we've cut RAM pricing in half.
Base configs now have double the memory (32GB instead of 16GB)
Additional vCPUs now include 8GB memory per vCPU (instead of 4GB)
This applies to all new instances on @ThunderCompute. Enjoy!
2
9
571
Thunder Compute retweeted
We're seeing a massive spike in demand at @ThunderCompute. New record highs every hour.
Please bear with us for any service interruptions, our team is working hard to keep everything online!!
1
3
15
857
If you need GPUs for your next deep learning project, check out the @ThunderCompute VSCode extension.
You can develop with a GPU with one click, in VSCode or Cursor, as if the GPU was local to your machine
With cheap prices
(Continued)
1
1
1
533
Try it for your next project with $20/month free: thundercompute.com/docs/
230