Lambda

Your inbox. Our condolences. We'll be at booth 925 when you're ready to talk. Lambda at #HumanX 2026: https://lnkd.in/gS7jxTxf

Lambda’s MLCommons MLPerf Inference v6.0 results are out. Key findings: - NVIDIA Blackwell Ultra is ready: 29% higher throughput over Blackwell on MoE (GPT-OSS 120B) - NVIDIA Blackwell is mature: 9% more throughput from software progress alone - Our collaboration with Stevens Institute of Technology cuts time-to-first-token P99 by 31% on MoE Whatever your workload, we have you covered: Lambda is the only AI-native cloud with official MLPerf results for both inference and training on NVIDIA Blackwell Ultra. Read more: https://lnkd.in/gVmkDM2N

6,500 AI builders are heading to HumanX next week. If you're one of them, find us at booth 925. Agents are welcome, too. Join us for happy hour on Monday evening: https://luma.com/kn5mrex3 #HumanX

Forrester named Lambda among the large vendors in its Q1 2026 Public Cloud Platforms Landscape. AI-native clouds, the report says, are the top disruptors reshaping the market. Read the full report: https://lnkd.in/gV8ZJ8zA

Lambda is one of @FastCompany's Most Innovative Companies of 2026. We built the SuperIntelligence Cloud for the teams solving the hardest challenges in AI. Read more: https://lnkd.in/e-M4rAzY #FCMostInnovative

Day four at #NVIDIAGTC. The last full day on the floor, and one topic kept coming up: the flexibility the NVIDIA Vera Rubin platform unlocks. More possibilities to design the AI factory that fits your workload. Learn more: https://lnkd.in/etQq-27A

Day three at #NVIDIAGTC. The booth was packed, the cameras were rolling, and the energy held. The conversation that kept coming up: how to engineer AI factories ready for large-scale NVIDIA Vera Rubin deployments. Maxx Garrison and Doug P. break it down this morning: https://lnkd.in/gcN93W85

Tonight at #NVIDIAGTC. Last call for AI After Dark. Engineers, researchers, operators. A few spots left. Reserve yours: https://luma.com/0p9qv6we

Not a chat app. Not an image generator. AI has found its product-market fit in software engineering, and the infrastructure is here to match. On Thursday at #NVIDIAGTC, Maxx Garrison breaks down what deploying Bare Metal Instances on NVIDIA Vera Rubin NVL72 and NVIDIA GB300 NVL72 actually looks like: https://lnkd.in/gcN93W85

Open-weight models don’t train themselves. Behind OLMo is a pre-training run that needed real fault tolerance at scale. At #NVIDIAGTC, we’re showing visitors exactly how Ai2’s team kept that run on track: checkpoint recovery, hardware monitoring, and a cluster that didn’t flinch. Learn more: https://lnkd.in/gT4abNPf