The Superintelligence Cloud
San Francisco, CA
- Tweets 1,878
- Following 240
- Followers 20,326
- Likes 1,879
Pinned Tweet
AI infrastructure is the largest industrial buildout of our lifetime. Lambda is assembling the leadership team to match the opportunity ahead.
Today, Lambda welcomes global infrastructure operator Michel Combes as CEO and former AT&T Communications CEO John Donovan as Chairman of the Board.
Co-founder Stephen Balaban takes on the CTO role full-time, shaping the technology that will define the next decade of AI compute.
Read the exclusive from @business: bloomberg.com/news/articles/…
6
6
42
6,722
Lambda keynote at ALVR, co-located with @aclmeeting (#ACL 2026) in San Diego.
Our research team will cover published multimodal AI papers and walk through the infrastructure patterns behind them: how data moves, how training scales, and where the bottlenecks hide.
July 3: lambda.ai/blog/lambdas-keyno…
1
419
Lambda retweeted
As promised, more model cards for you:
Kimi K2.7-Code and GLM-5.2!
Both of which need zero introduction, as GLM-5.2 especially has taken the world by storm, but there's a new addition to our cards going in now: @NoumenaAI `code` usage directions!
Kimi: lambda.ai/inference-models/m…
GLM: lambda.ai/inference-models/z…
1
2
13
1,277
Lambda retweeted
New model card up, @MiniMax_AI M3! (Working through the Colorado backlog)
At 400B+ parameters, using the unquantized weights ends up needing a full HGX B200 (and I don't believe we can run the MXFP4 on hopper)
Nice addition (on top of performance) is the multi-modality 😍
2
1
22
10,839
.@raais 2026, 10th edition. DeepMind on RL and robotics, Odyssey's live GoldenEye world model, ElevenLabs on serving at scale. 400+ researchers and builders in London.
One of the few rooms where infrastructure and research collide. Thanks, @nathanbenaich.
2
8
3,660
Lambda's @chuanli11 and Amir Zadeh co-authored this one. 3D-DLP learns object-centric 3D representations from raw RGB-D and voxel inputs, with no supervision. Wins 9 of 10 RLBench tasks against matched baselines. Accepted at ICML 2026.
🧩 #ICML2026 💥 How can a model discover the 3D objects in a scene—their shape, color, and position—without any labels? Introducing 3D-DLP, a self-supervised object-centric model that decomposes colored 3D scenes (RGB-D and voxels) into a set of 3D latent particles.
1
5
15
2,950
.@claudeai Code + spare GPUs + an experiment tracker = a Gemma 4 model that learned Tetris in 2.5 days.
3
2
12
898
Autonomous by default. Human on escalation. the_lab.api is open source: lambda.ai/blog/what-happens-…
336
.@KodiakRobotics runs 28 driverless trucks. Training the model behind them required GPU memory, inter-node bandwidth, and data throughput at scale.
On Lambda's 1-Click Clusters: zero to pre-training in one week and 2x iteration speed.
2
2
10
829
Robots act in partially observed worlds. 3D-Belief lets them reason about what they can't see: completing unseen scenes in 3D using diffusion and Gaussian splatting, then acting on that belief.
Joint work: @JohnsHopkins, @Cambridge_Uni, Lambda. Tested on a @HelloRobotInc Stretch platform.
Paper: 3d-belief.github.io
How can robots make decisions when they can't see the whole picture? Researchers introduced 3D-Belief, a generative world model that helps robots reason about unseen parts of their environment. A fascinating look at the future of robot autonomy.
🤖arxiv.org/pdf/2605.11367
1
4
23
2,086
Lambda is a platinum sponsor of the Agentic AI Summit, Aug 1-2 at @UCBerkeley.
CSO @chuanli11 opens the summit on the Infrastructure and Platform panel alongside Amazon, Google Cloud, and NVIDIA. Plus a Lambda-led workshop on open-source agent security and optimization.
Schedule: rdi.berkeley.edu/events/agen…
🚨 The full program for Agentic AI Summit 2026 is now live.
📍 Aug 1–2 @ UC Berkeley
🔥 The largest Agentic AI event ever held
Last year: 2,000+ in person, 40,000+ online
This year: 5,000+ in person, hundreds of thousands on livestream
Want to understand where Agentic AI is headed next?
Join us to get the most comprehensive view of the frontier of Agentic AI, from cutting-edge research to production deployments, covering every layer of the stack:
⚡ Infrastructure
⚡ Foundation models & capabilities
⚡ Agent frameworks & platforms
⚡ Evaluation & benchmarks
⚡ Enterprise & consumer applications; agentic AI for Science, Math, Finance, Legal, Healthcare
⚡ Safety, security & governance
📣 Also excited to announce the Startup Spotlight:
Building something exciting in Agentic AI? Apply to pitch directly to 5,000+ decision-makers, investors, practitioners in the room and hundreds of thousands watching worldwide.
Application form in thread🧵
Deadline: July 6, 11:59pm PT
The future of AI won't just be discussed here—it will be built here.
#AgenticAI #AIAgents #ArtificialIntelligence
10
2,072
Happy Father's Day to @stephenbalaban and every dad out there. The original 24/7 uptime. Always on call, no maintenance window. Thanks for the years of supervised learning.
1
12
1,165
Compute was never a commodity.
Lambda co-founder and CTO @stephenbalaban joins @mattturck on the MAD Podcast on the state of AI compute in 2026: what the H100 price indices get wrong, the real cost stack from photons to tokens, and why we're still underbuilding.
Full conversation: piped.video/watch?v=0NttU4Cb…
1
2
13
2,149
CVPR 2026. Two papers, two workshops, a @KodiakRobotics autonomous truck demo, and one theme underneath all of it: physical AI is scaling hard.
We wrote up what we learned.
1
11
1,687
Training data is running out. Human annotation doesn't scale. Sim2Reason generates synthetic physics data that outperforms real-world corpora.
+17.9% on JEEBench. Zero human-annotated data.
Accepted at @icmlconf 2026. Built by @CarnegieMellon and Lambda.
lambda.ai/blog/the-age-of-la…
3
1
8
891
Lambda is on the 2026 NatSec100 for the fourth consecutive year. Ranked 20 among the top 100 venture-backed companies shaping U.S. national security.
natsec100.org
6
431
.@MLCommons MLPerf Training v6.0: Lambda posts the fastest pre-training for LLM and MoE models.
3
3
12
1,384
GPT-OSS-20B on @nvidia HGX B200: 96.46 minutes. MLPerf's first-ever MoE benchmark.
Fastest among single-node HGX B200 submitters.
1
613
Read the full breakdown: lambda.ai/blog/mlperf-traini…
2
307