Crusoe
@CrusoeAI
Followers
9K
Following
907
Media
215
Statuses
796
The AI factory company. We are on a mission to accelerate the abundance of energy and intelligence.
Denver, CO
Joined January 2019
The future of LLM inference is up to 9.9x faster. 🚀 Introducing Crusoe Managed Inference, now available in the new Crusoe Intelligence Foundry! Our inference engine is powered by Crusoe MemoryAlloy — a cluster-native KV cache and fabric that targets the core bottlenecks of
0
4
36
Is LLM inference latency killing your app's UX? The issue is memory fragmentation. Our latest AI Lab video breaks down Paged Attention; the key optimization built into vLLM and SGLang that solves it. You'll learn: 🧠 Paged Attention in 60 seconds (it's how we eliminate KV
0
0
3
🚨Breaking: Crusoe is integrating @NVIDIA’s full-stack, co-designed AI platform into our infrastructure! GB200 NVL72s — deployed on Crusoe Cloud — power complex #MoE models and enable up to 9.9x inference performance gains. ➡️ Learn more: https://t.co/qemAq4L0Pz
#GB200 #AI
blogs.nvidia.com
Kimi K2 Thinking, DeepSeek-R1, Mistral Large 3 and others run 10x faster on NVIDIA GB200 NVL72.
2
0
10
Huge thanks to everyone who joined our exclusive AI Tech Talk in London last week — and an extra shoutout to our incredible speakers from @StackOneHQ, @inephany, @reimaginerobots, and @DecartAI. We're proud to support innovative AI native startups across the UK and EMEA. 🇬🇧
1
2
8
The NVIDIA GB200 NVL72 performance puzzle is solved. 💡 We used @PyTorch on our Crusoe NVIDIA GB200 NVL72 clusters to crack the code. By leveraging SLURM block scheduling to respect the rack's full 72-GPU NVIDIA NVLink™ domain, we unlocked a 13% throughput gain in Llama 4 Scout
crusoe.ai
Unlock a 10%+ throughput increase on NVIDIA GB200 NVL72. This guide details how to leverage SLURM's topology-aware block scheduling to align large-scale AI training (Llama 4 Scout 17B) with the...
1
1
11
Our VP of Public Affairs @AxelrodSara delivered a keynote at @NVIDIAGTC in D.C. last month on how Crusoe is helping America win the AI race. 🇺🇸 Watch it here ▶️ https://t.co/1ax6aieMcy We’ll be at the next #NVIDIAGTC from March 16–19, 2026 in San Jose! See you there.
nvidia.com
The rise of AI has made gigawatt-scale data centers a critical necessity for winning the AI race
0
0
11
Doubling down on thanks! 🙏 We’re grateful for our Crusoe teammates and customers whose resilience and ambition fuel the future of abundant AI. Here’s to another year of momentum and innovation. 🧡
0
2
18
AI infra founders, researchers, and investors — you’re invited to Compute Poker Night. ♠️ ⌚️When: Wednesday, December 10 at 6:30 p.m. PST 📍Where: San Francisco, CA Ready to try your hand at winning $30K in Crusoe Cloud credits? Reserve your spot today ➡️
0
0
8
GPU users: Stop wasting cycles. We certified NVIDIA Run:ai on Crusoe Managed Kubernetes (CMK). 🔥 AI lifecycle ready: Configure a Run:ai cluster to accelerate both distributed training and elastic, serverless inference workloads. 👉 Result: Dynamic GPU allocation + full MLOps
0
0
11
Our CEO and co-founder @ChaseLochmiller joined leaders from @EngineNo_1, @Vertiv, @VantageDC, and @Oklo at @NVIDIAGTC in Washington, D.C. last month for a session on building and right-sizing AI factories. 🎥 Catch the full conversation:
nvidia.com
Learn how to design and right-size AI factories that power the age of intelligence while optimizing performance, energy use, scalability, and sustainabilit
1
1
15
Crusoe will be at #TheAISummit in NYC on December 10 + 11. Be sure to visit us at Stand #717 — and don’t miss our SVP of Product Management, Erwan Menard, taking the Headliners Stage on Wednesday at 11:55 a.m.! 🎤 Book time with our team: https://t.co/pXIlgu0XuB
0
1
8
Stop waiting. Start winning. 🧠 Crusoe’s inference engine with MemoryAlloy technology delivers: • Up to 9.9x faster TTFT • Up to 5x more tokens/sec processed Seamless scaling. Serious results. Sign up today: https://t.co/nPNyZrvDdt
2
1
9
🇬🇧 Join Crusoe in London on Wednesday, November 26, at 5 p.m. for an AI founders' panel with @StackOneHQ’s CTO @GuillaumeLebed1, @inephany’s CEO @johnptorr & @reimaginerobots’s CEO Jon Scholz, moderated by @DecartAI’s VP of R&D. Don't miss it! Register ➡️ https://t.co/A5f3kL1pjz
2
2
8
Running into bottlenecks and capacity crunches on your traditional cloud? Crusoe has a better way to build. ☁️ Download our guide to see if the neocloud path is right for you: https://t.co/mCwRPrbmf7
#AI #cloudcomputing #neocloud #MLOps #aiinfrastructure
crusoe.ai
Get the ultimate guide to AI cloud, featuring a 30-day roadmap, real-world case studies, and expert insights. Download now to jumpstart your AI journey.
0
2
9
In AI, your model is only as good as the data you can feed it. Crusoe’s new AI Lab video series unpacks how holistic, InfiniBand architecture keeps your models fed and running at peak performance. ▶️ If you’re tired of running into performance bottlenecks, we encourage you to
0
0
5
Hello, Bellevue! 👋 Crusoe has officially opened our newest office right here in the Pacific Northwest. It’s a stunning new space with room to grow as Crusoe and the AI sector close out a year of tremendous momentum. Onward to 2026! #SeattleTech #CrusoeCulture
1
0
11
🎤 “We rely on clouds like Crusoe to give us the best price per performance.” — @FireworksAI_HQ co-founder Chenyu Zhao speaking with Crusoe’s SVP of Product at @AMD’s #AIDevDay 2025. Listen here: 🔗 https://t.co/Bw9VPNubAf
#CrusoeCloud #AMD
0
0
5
🇬🇧 London! There’s still time to sign up for Crusoe’s AI happy hour & tech talk on Wednesday, November 26 in Shoreditch. On the agenda: 🧠 Founder panel w/ @StackOneHQ, @inephany & @reimaginerobots 🎤 Keynote by @DecartAI’s VP of R&D Secure your spot: ➡️ https://t.co/A5f3kL1pjz
0
0
7
Geomagnetic storms lit up the night sky in Denver last week. 💫 📸 The Crusoe team snapped these incredible photos of the Northern Lights! #CrusoeCulture #DenverTech
0
0
10