Excited to share our work on envisioning Intelligent AI Delegation (arxiv.org/abs/2602.11865). Delegation in most existing AI systems is brittle, and relies on simplified hand-crafted control flows. As such, it fails to meet the requirements of what is needed to truly scale distributed reasoning and task completion in multi-agent systems and collectives.

Our Robotics Accelerator has launched with 15 startups helping shape the future of physical AI in Europe. 🤖 This three-month program will connect them with access to our AI stack, Gemini Robotics models and hands-on support from our teams. Meet the companies → goo.gle/4oeEk2K

When millions of AI agents interact with each other, new collective behaviors can emerge. 🌐 Together with @schmidtsciences, @coop_ai, @ARIA_research and supported by @GoogleOrg, we’re launching a $10M research fund to help understand how AI systems behave as a group. → goo.gle/3Si6rCl

I’m very excited that this is out. We’re looking forward to funding some great research

We’re teaming up @Palmeiras, the first football club to meaningfully build upon TacticAI: our AI system that can help simulate field scenarios and predict open play dynamics up to 8 seconds in advance. ⚽

Over the past few months I've been working on a very exciting project: a new $10m fund for research on multi-agent multi-principal AGI safety! Instead of focusing on single agent alignment and centralized control, we're looking to support research focusing on multi-agent settings, mechanism design, cooperative AI, and coordination problems. This is a joint initiative between @GoogleDeepMind, @Googleorg, @schmidtsciences, @coop_ai, and @ARIA_research. Huge thanks to @James_D_Fox, @weballergy, @FranklinMatija, @lrhammond, and @ObadiaAlex for their invaluable work! See: deepmind.google/blog/investi… Apply: schmidtsciences.smapply.io/p…

Congrats @sebkrier @weballergy @FranklinMatija @lrhammond! Oversight systems that help comps and govs oversee millions of AI agents from different labs are going to be hugely important and won't (and shouldn't be) be build by any of the big labs.

Check out my team mates' work on using AlphaEvolve to discover novel cognitive models of learning! This is such an important stepping stone towards building tools that help human scientists gain more insights from their data. We certainly did!

Building autonomous agents for scientific discovery? 🧬🤖 @GoogleDeepMind Science Skills is now available on GitHub. We've open-sourced this specialized toolkit to accelerate your agentic workflows with scientific grounding and higher token efficiency. Download now ↓ github.com/google-deepmind/s…

Computational models are a key part of science but discovering new ones is hard! DataDIVER discovers concise models from data, surfacing new mechanistic ideas and generating clear predictions for future experiments Preprint from @GoogleDeepMind Neuroscience Lab collaborators

We believe AI can be a dedicated research partner to help discover the next breakthrough. Enter Co-Scientist: our latest Gemini-based multi-agent system that can generate, debate and evolve novel hypotheses for complex scientific problems 🧵

RT @mweber_PU: Can we learn the curvature of a data manifold from a finite sample? We study continuum limits of Ollivier’s Ricci curvature…

AI agents are advancing research-level math. 🚀 I’m thrilled to share @GoogleDeepMind’s AlphaProof Nexus - an agentic framework for formal proof search powered by Gemini. When applied to a set of open formal math problems, our agent autonomously solved: ✅ 9 open Erdős problems (including two open for 56 years!) ✅ 44 Online Encyclopedia of Integer Sequences (OEIS) problems ✅ A 15-year-old open problem in algebraic geometry ✅ A 7-year-old open question in min-max optimization We are collaborating with mathematicians across disciplines - from combinatorics and graph theory to quantum optics. Ultimately, these results show the massive potential of even simple agentic loops powered by Gemini. Read the paper here: arxiv.org/abs/2605.22763v1

Your RL post-training may be sabotaging your LLM’s test-time scaling! Conventional RL pretends that you can collapse all reward signals *upfront* into a single *scalar reward*. We introduce Vector Policy Optimization (VPO), which natively maximizes *vector-valued* rewards, boosting test time search performance, even on the original scalar.

Some new results I found surprising that I’m tweeting for Chris (who isnt on here). With enough compute, the best data filter for LMs (on DCLM) might be no filter. Why? Large models can tolerate a surprising amount of nominally 'low quality' data, and can sometimes even benefit.

We want to help scientists discover their next breakthrough with AI. Gemini for Science is our new suite of experimental tools to help them explore more hypotheses, validate work at scale, unpack literature with ease, and more 🧵

Really happy to see @OpenAI adopt @GoogleDeepMind's SynthID for watermarking AI generated images. We need more such cross industry partnerships for enabling responsible use of AI systems.

Our paper “Accelerating scientific discovery with Co-Scientist” is published today in @Nature. Read it here: nature.com/articles/s41586-0… and learn more about our real-world partnerships on the DeepMind blog: deepmind.google/blog/co-scie…