DeepMind AI News, Models and Product Updates
Track latest Deepmind AI news, launches, research, and ecosystem moves.
Deepmind news, model releases, product launches, research updates, and major announcements in one place.
Latest Articles
- Predicting protein structures at any folding stage from sequence — FoldScape generates protein conformations at any specified folding stage from sequence, covering roughly ten times more of the folding coordinate than seed-varied AlphaFold 3 ensembles across 20 human proteins.
- Agentic cheminformatics with Google Antigravity — Google DeepMind added Cheminformatics Skills to Google Antigravity, letting an agent search compounds, run matched molecular pair analysis and propose new macrocycles with synthetic routes in about an hour.
- Auto-Research Agent on MSA Organization to Improve Protein-Protein Complex Structure Prediction — An autonomous agent that reorganizes MSAs per target improves AlphaFold 3 protein-complex predictions on 1,015 interfaces while using 60% fewer sequences on average.
- BridgeDock: Taming Generative Co-Folding Prior for Molecular Docking with Diffusion Bridge — BridgeDock, accepted at NeurIPS 2026, adapts pretrained co-folding models like AlphaFold 3 to flexible docking by modeling the apo-to-holo transition as a diffusion bridge, beating co-folding baselines even with few denoising steps.
- Trajectory raises $45M to make AI model training accessible to nontechnical experts — Trajectory, a one-year-old continual-learning startup, raised a $45M Series A led by Sequoia at a $300M valuation and opened its post-training platform to the public, aiming to let domain experts improve open models without research teams.
- Recycle only when you have to! — Researchers show AlphaFold 3-style confidence metrics can be predicted from the trunk at early recycling steps, cutting compute for protein binder screening and design.
- Discovering Temporal Structure: An Overview of Hierarchical Reinforcement Learning — A comprehensive HRL survey by Doina Precup, Marlos Machado and collaborators — updated September 2026 — maps the landscape from classical options theory through LLM-guided temporal abstraction, framing hierarchical structure discovery as a path toward long-horizon general agents.
- Tapes Together Strong: The Co-evolution of Computation and Cooperation — Researchers from Google DeepMind and MIT introduce Autopoietic Game Theory: a computational model where cooperation, replication, and their computational costs co-evolve in programs running Z80 machine code — finding that coupling computation capacity to energy budget can make defection self-limiting even in well-mixed populations.
- Ex-Deepmind VP Vinyals says AI self-improvement is coming but won't trigger an intelligence explosion — Former Google DeepMind VP of Research Oriol Vinyals argues at Agentic AI Summit 2026 that AI self-improvement is real but won't cause an intelligence explosion — the hard limits are generating novel ideas ('research taste') and reliably evaluating results. He's launching Discovery Loop with Jeff Dean, Sanjay Ghemawat, and Quoc Le to automate end-to-end scientific research.
- NVIDIA Details BioNeMo Inference Runtime (BioIR): 2.90x Higher Boltz-2 Folding Throughput and 58.5K Residues per GPU-Hour on 8xH100 — NVIDIA's BioNeMo Inference Runtime (BioIR) hits 2.90x higher Boltz-2 folding throughput at 58.5K residues per GPU-hour on 8xH100s — already deployed at production scale to generate 31 million protein-complex predictions for the AlphaFold Database expansion.