Microsoft Research has developed an end-to-end machine learning pipeline that uses solar wind observations at L1, forecasts of the AE and Dst indices, and local geological conductivity data to estimate local space weather risks 30 to 60 minutes ahead for 66,935 substations in the contiguous United States.
CoreWeave announced that NVIDIA Vera Rubin NVL72 systems with Spectrum-X 102.4T Ethernet were available on CoreWeave Cloud, with Cognition becoming the first customer to run them in production.
Chipmaker AMD is acquiring World Labs, a startup focused on spatial intelligence. AI pioneer Fei-Fei Li will join AMD's leadership team as Executive Vice President and Chief Scientist. The all-stock deal is valued at about $8.2 billion, according to AMD. The transaction is expected to close by the end of 2026, pending regulatory approval.
AMD is acquiring World Labs, one of the leading developers of deep learning models intended to understand physical reality, in an $8.2 billion deal, the two companies said today. World Labs justified the deal in a statement saying that AI development required “close collaboration across model research, systems and compute.”
NVIDIA released Kumo Tabular, an open-source tabular foundation model that predicts labels for new rows in a single forward pass given labelled rows, without training, tuning or feature engineering. It supports classification and regression, offers three sizes from 28M to 215M, and was pretrained solely on artificially generated tables. It uses the commercially usable OpenMDW-1.1 licence and ranks first on TabArena, BeyondArena, TALENT and ScoringBench.
Microsoft Research released Quine, an AI research system for biological complexity, comprising a biological world model trained on multimodal data spanning sequences, structures, functions, cell states and imaging, and an interactive harness connecting the model, scientific tools, literature and experimental researchers.
AWS released a solution for automating Amazon Textract adapter lifecycle management across accounts. By externalising adapter IDs to AWS Systems Manager Parameter Store, production adapter references can be updated without downtime or application redeployment, reducing coordination that previously took hours or days to seconds.
According to The Information, Tesla had factory workers in Texas and California wear specialised motion-recording suits to collect imitation-learning data for Optimus humanoid robots. Some complained that the robots would eventually replace them, after which Tesla moved collection to dedicated teams and established training centres.
AWS presents an architecture combining EFA and DeepEP on Amazon EKS to scale MoE reinforcement learning training, increasing throughput by 40%.
Using the open-source SkyRL framework on Amazon SageMaker HyperPod, multimodal reinforcement learning post-training with GRPO increased the Qwen3-VL-8B vision-language model’s maze-solving success rate from 43.75% to over 95% on a fixed evaluation set of 64 mazes.
Amazon SageMaker HyperPod with Cloud Native Qumulo (CNQ) and Cloud Data Fabric (CDF) lets training clusters read remote datasets directly across AWS Regions without copying data or changing code.
Drawing on work at Endura Therapeutics, Adrian Sanborn divides AI's impact on science into two categories: Foundries use next-generation sequencing, high-throughput microscopy and physical automation to reduce experimental measurement costs by an order of magnitude, while Navigators apply cheaper thinking to decisions and processes.
NVIDIA joined a global research consortium including Google DeepMind and EMBL-EBI to release predicted 3D protein complex structures for over 2,800 viruses through the AlphaFold Database, free for any scientist to use.
NVIDIA validation engineer Sakeena Fiza validates new hardware before mass production in its data centre systems lab. One memorable moment was the first successful system-level enumeration of the Rubin GPU. She compares validation to solving crimes, finding and reproducing issues before customers do, across trays, racks, clusters and customers' AI factories.
Radical Numerics co-founder and CEO Eric Nguyen argues that models enhancing biological capabilities can also support defence and advocates more aggressive frontier research. His team's Evo and Evo 2 genomic language models were used by Arc/Stanford researchers to generate complete phage genomes and synthesise functional viruses.
Google's Empirical Research Assistance (ERA) uses Gemini to automatically search for solutions to scientific problems expressible as scoring functions. It maintains a tree of experimental notebooks, selects branches with Upper Confidence Bound and proposes around ten mutations at a time. Between Gemini 2.0 and 2.5, it went from unusable to highly effective.
Xiaomi released the MiMo-V2.6 family, including omnimodal models MiMo-V2.6-Pro and MiMo-V2.6-Flash, plus MiMo-V2.6-Pro-UltraSpeed with up to 20 times faster output.
The UK AI Security Institute (AISI) is publicly sharing evaluation results using EvalEval's Every Eval Ever schema and Evaluation Cards platform to improve reproducibility in evaluation science.
Microsoft published RetroChimera, a retrosynthesis prediction framework, in Nature and open-sourced its implementation and weights. It combines the Transformer model R-SMILES 2 with GNN-based NeuralLoc, using a learned ensemble to rerank predictions. Chemists preferred its single-step reaction predictions in blind testing.
MIT Technology Review and Times of San Diego spent 15 months creating the first comprehensive map and analysis of deaths near US border surveillance towers, examining migrant deaths since 2015. They requested records from 17 Texas county sheriff's offices and obtained over 4,000 pages from 14 counties, used Anthropic's Claude API to extract coordinates where remains were found, then manually checked samples.
NVIDIA highlighted five AI clean-energy companies at New York Climate Week. ThinkLabs AI's NVIDIA CUDA-based digital twins and agents help Southern California Edison cut grid connection application assessments from 30–45 days to two minutes.
Hex's data agent uses GPT-6 Astra to turn analytical answers into interactive visual reports that employees want to share.
Salesforce unveiled its first CRM reasoning model, Koa, at Dreamforce. Post-trained on NVIDIA Nemotron 3 Super, it uses a proprietary synthetic dataset derived from nearly three decades of enterprise CRM deployments across more than 14 industries, including manufacturing, financial services, healthcare and travel.
Good Start Labs turned railway board game 1830 into a training environment and tested a 30B model. Both single-turn Q&A and multi-turn terminal-agent training improved game performance, but only the terminal-agent design improved Finance-Agent benchmark results.
Google Research proposed Retrieve-for-Train, using offline reinforcement learning to find reward-aligned query fan-out and compile it into supervision, then distilling it into a 53.9M-parameter diffusion retriever. At inference, it performs one non-autoregressive query fan-out without generating CoT reasoning tokens.
NVIDIA announced several advances for Vera Rubin and the DSX platform at AI Infra Summit, focusing on energy-efficiency improvements in token throughput per megawatt.
At ACL 2026, Google Research proposed ToolGrad, an answer-first, question-later approach that generates real tool-call chains before deriving user queries, replacing traditional DFS trial-and-error annotation.
OpenAI announced Data agent in ChatGPT Work is available to everyone, connecting enterprise data, discovering insights and building interactive dashboards with AI through natural language.
TRL v1.14's AsyncGRPOTrainer now supports training and synchronising only LoRA adapters to vLLM. A rank-1 adapter is just a few MB and can transfer through a Storage Bucket mounted to each job, without NCCL.
Google DeepMind launched AlphaGenome Atlas, a platform containing predicted effects for 9 billion single-nucleotide variants across the human genome. At 1 PB, it is over 30 times the size of the AlphaFold Database.
Google Research evaluated cross-population transfer of polygenic risk scores (PRS) on eight clinical measures using European UK Biobank data and nearly 200,000 Japanese participants from Biobank Japan. At target-population samples of 15,000 or more, target-specific models outperformed mixed European-data training, while highly genetically correlated traits continued to benefit from European data until target samples reached 25,000–40,000 or more.
Google Research partnered with HHMI Janelia, the University of Cambridge and others to publish in Cell a complete connectome of a male fruit fly’s brain and central nervous system. It contains over 166,000 neurons and 125 million synaptic connections, making it the largest brain map to date by neuron count.
A public low-cost approach fine-tunes LFM2.5-350M with TRL GRPO using around 500 samples and 100 steps, raising IFStruct from 22.6% to 29.7%. Training fits free Colab or Kaggle GPUs, with local llama.cpp evaluation on a MacBook and code open on GitHub.
Google Research proposed MAPL-EMIT, a Swin-S Transformer-based deep-learning framework automatically detecting methane plumes, predicting enhancements and locating them in NASA EMIT hyperspectral satellite data, with 84% recall on expert-labelled plumes.
Google Research released TimesFM-3, its first time-series foundation model pretrained natively for multivariate forecasting, with 330 million parameters and over 1 trillion pretraining time points.
Microsoft Research introduced GigaPath-Flash and GigaTIME-Flash, greatly reducing pathology compute requirements with a 22M-parameter ViT-S backbone distilled from the billion-parameter GigaPath encoder, open-source under Apache 2.0.
Google Research introduced the experimental Planetary Prediction Engine (PPE) under Google Earth AI. From a natural language query, it autonomously discovers geospatial data, engineers features, trains and evaluates models and produces reports, compressing weeks of manual data engineering into minutes.
Google released self-supervised GlucoFM, separating slow glucose trends from short-term fluctuations. Pretrained on 109,066 hours of unlabelled CGM data, it averaged 5.8 percentage points higher PR-AUC than the best GluFormer variant across 14 evaluations on seven clinical tasks in four cohorts, and achieved the lowest post-meal response prediction MAE.
Sentence Transformers v6.0 adds a fourth model type, MultiVectorEncoder, for ColBERT-style late-interaction retrieval, with a complete training approach.
Multiverse Computing published a paper proposing Quantization-Aware Healing (QAH). After compressing GPT-OSS 120B to 60B parameters and quantising it to MXFP4, the method distils directly from the original uncompressed model rather than a reconstructed bfloat16 checkpoint.