@vast_ai

Vast is the infrastructure layer where AI agents autonomously design, procure, and optimize their own compute.

 California, USA
vast.ai/
Joined August 2018
  • Tweets785
  • Following71
  • Followers14.9K
94Media
Vast.ai is hiring. We're growing. Fast. Agent-ready AI infrastructure. Profitable, venture-backed. New roles across engineering, research + DevRel, mostly deep systems + GPU work. On-site in SF or LA. vast.ai/jobs
53
18
147
45,927
Elon Musk
@elonmuskOct 25
Replying to @Yuchenj_UW
I think we’re going to see a lot of stranded compute in partially complete data centers
2
7
501
Proud to be among @brexHQ’s 25 fastest-growing software vendors this fall. Thanks to the builders choosing Vast and the GPU hosts making it possible. 🚀
Brex
@brexHQOct 21
Spring was an abstraction. Summer was the stack. Fall? The “chips” fall where they may. The 25 fastest-growing software vendors on Brex this fall👇
1
3
10
1,409
650,000 compute transactions ran through Vast in the last 30 days. Today we filed a comment with the @CFTC on compute derivatives. We’re excited about what futures could mean for compute buyers and providers. Here’s why:
1
3
1
4
1,260
more replies
CME Group and ICE, the NYSE’s parent, have both announced GPU compute futures, pending CFTC review. These proposals would give the industry more tools to manage price risk as demand for compute grows.
1
1
154
Vast delivers compute for people and the agents they build. We filed our CFTC comment to support the next stage of compute markets with properly licensed, auditable data. Read more about our comment: vast.ai/article/compute-futu…

Compute futures mark the next step in GPU markets

Vast.ai supports well-regulated compute futures and auditable, lawfully sourced settlement benchmarks to help GPU buyers and providers plan for growth.

vast.ai
1
137
vast.ai retweeted
Nate
@NatemarmorsteinOct 19
Check us out @vast_ai, we currently have 40 B200s from $6.25/hr, 20+ H100s from $1.79/hr, and 100+ A100s from $0.44/hr
Elon Musk
@elonmuskOct 18
What it’s like trying to rent AI/SI compute
1
3
275
okay, but how much VRAM?
Tesla
@TeslaOct 16
Real life, baby
2
2
28
2,820
Our COO Travis Cannell joined @WCMC_Podcast to talk about how the GPU market works. youtu.be/_XneivTDCqA

GPUs Are Snowflakes: Why a Compute Marketplace Doesn’t Race to the...

Travis Cannell, COO of Vast AI, joins Where Capital Meets Compute t...

youtube.com
2
5
2,157

GPUs Are Snowflakes: Why a Compute Marketplace Doesn’t Race to the Bottom

Article · @vast_ai
1
1,369
vast.ai retweeted
Nate
@NatemarmorsteinOct 15
Excited to share that I've joined @vast_ai as a Developer Relations Engineer! Looking forward to working with Vast users and helping grow the community. If you're building something on Vast (or got stuck trying), tag/dm me!
1
3
346
vast.ai retweeted
vast.ai
@vast_aiOct 12
We’re hosting LA Tech Week. Thursday 10/15, 6-8pm at our Westwood office. Food, open bar, and a conversation about what AI compute actually costs in 2026. Limited spots, free to attend. RSVP here: partiful.com/e/cNUPXCGv3BjOy…
3
3
6
2,042
We’re hosting LA Tech Week. Thursday 10/15, 6-8pm at our Westwood office. Food, open bar, and a conversation about what AI compute actually costs in 2026. Limited spots, free to attend. RSVP here: partiful.com/e/cNUPXCGv3BjOy…
3
3
6
2,042
84 GB is enough to hold some 27B-class MoE models at full context that would otherwise need two RTX 5090s. That’s the case for the RTX PRO 5500. Memory bandwidth is 1,398 GB/s against the 5090’s 1,792, so it’s built for capacity, not peak speed. Full comparison with the PRO 6000: vast.ai/article/everything-w…

Everything We Know (and Don't) About the NVIDIA RTX PRO 5500

Explore NVIDIA RTX PRO 5500 specs, 84 GB memory, RTX 5090 and RTX PRO 6000 comparisons, and pricing and availability uncertainties.

vast.ai
6
6
1
27
2,105
New MiniMax M3 template now available in the Vast.ai model library @MiniMax_AI M3 is a native multimodal model from MiniMax with a 1M-token context window. It has about 428B total parameters with about 23B activated per token, and it takes text, image, and video input. MiniMax built it for long-horizon coding and agentic work, and reports frontier-level results across long-horizon agentic benchmarks in both coding and cowork tasks. Learn more here: vast.ai/model/minimax-m3
3
4
20
2,831
Come find us at The AI Conference in San Francisco @ Pier 48. Booth 202. Last day! We've got a Nintendo Switch 2 raffle going on 👀
13
5
31
3,084
Setting up an AI agent and not sure where to start? Here's a walkthrough on picking a framework and model, sizing your GPU, and getting it running on Vast.ai. vast.ai/article/how-to-deplo…

Rent GPUs | Vast.ai

Rent high-performance cloud GPUs at low cost with Vast.ai. Instantly deploy GPU rentals for AI, machine learning, deep learning, and rendering. Flexible pricing, fast setup, and global availability

vast.ai
8
4
13
2,530
Vast.ai is at The AI Conference 2026. September 29 to October 1. Pier 48, San Francisco. Booth 202. If you are planning GPU capacity for the next two quarters, book time with us. Bring the workload and the timeline. You leave with real pricing and availability across our network. Booking link below. @AIconference
4
4
1
16
3,801
Book with us here: vast.ai/the-ai-conference-20…

Vast.ai at The AI Conference 2026

Meet Vast.ai at The AI Conference 2026 in San Francisco, September 29–October 1. Visit Booth 202 or book time with the team.

vast.ai
7
2,392
Load more