Technology News
776 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.
Anthropic hires Google TPU veteran Amir Salek to accelerate in-house AI chip development
Anthropic is accelerating custom AI silicon after hiring Amir Salek, who led Google's Tensor Processing Unit program from the first through seventh generations, Bloomberg reported on August 21. Salek...
UBTech showcases a 1:1 replica of a customer's production line at WRC 2026, running industrial humanoids unattended for a full show day.
UBTech (优必选) used WRC 2026 to run what it describes as a full, unscripted replica of a customer's production line rather than a staged demo. Nearly ten industrial humanoid robots — Cruzr S2 and Cruzr...
NVIDIA's Nemotron 3.5 Lightning and NeMo Switchyard target enterprise agentic AI costs
On August 11, 2026, NVIDIA released Nemotron 3.5 Lightning, a 30-billion-parameter open Mixture-of-Experts model for long-running agentic tasks across reasoning, vision, and retrieval-augmented genera...
Supabase says AI agents, not humans, now launch most of its databases as developer count hits 10 million.
Supabase co-founder and CEO Paul Copplestone said in a YC Root Access interview that the open-source Postgres-as-a-service platform now serves roughly 10 million developers, up from about 6.5 million...
Nvidia Research Puts Agent Harness Design Ahead of Base-Model Quality
Nvidia researchers reported that a custom agent harness, rather than a change in the underlying model, lifted Claude Opus 5 from a 30% score to 100% on ARC-AGI-3, an interactive reasoning benchmark bu...
OpenAI Proposes Four-Metric 'Useful Intelligence per Dollar' Framework to Measure Enterprise AI ROI
OpenAI on July 17, 2026 published a framework it calls Useful Intelligence per Dollar, arguing that pricing AI purely by token consumption understates real cost efficiency. The framework scores AI dep...
Alibaba's Second-Generation T-Head AI Chip to Tape Out in H2 2026
Alibaba CEO Eddie Wu said on August 20 that the company's second-generation T-Head AI chip is expected to tape out and enter production in the second half of 2026, offering stronger computing performa...
Waymo opens its robotaxi compute stack to scrutiny
Waymo has disclosed new details about the in-vehicle compute system powering its robotaxi fleet, including a custom silicon chip, processor specifications, internal components, and its hardware-suppli...
SenseTime Open-Sources SenseNova U1.5 Lite With Native 4K Image Output
SenseTime has released SenseNova U1.5 Lite as open source, an 8-billion-parameter multimodal model that combines visual understanding, image generation, and image editing. The company says the model c...
Ramp Launches Router, a Multi-Model Gateway for Enterprise AI Spend
Ramp has launched Router.com, a U.S.-only API gateway that lets businesses access and switch among models from OpenAI, Anthropic, DeepSeek, Moonshot, MiniMax, Nvidia, xAI, and Z.ai. It offers routing...
Nitto Kogyo's Direct-Liquid-Cooled OCP ORV3 Rack Adopted at MC Digital Realty's AI Test Facility DRIL
Nitto Kogyo, an electrical enclosure and cabinet manufacturer based in Nagakute, Aichi, Japan, announced that its direct liquid cooling (DLC) compatible OCP ORV3 rack has been adopted and installed at...
Solinide Photonics raises €4M seed round to commercialize silicon nitride photonic chips for AI data centers
Swedish deeptech Solinide Photonics, a spin-out from Chalmers University of Technology, closed a €4 million seed round co-led by Navigare Ventures and PSV Hafnium, with Chalmers Ventures, Turbine Capi...
OpenAI pauses two weeks of reinforcement learning training, clouding the timing of its next model releases.
OpenAI announced a two-week pause in reinforcement learning training for upcoming models, according to AI Business. The report does not name one confirmed cause, pointing instead to safety concerns or...
Baidu Targets ERNIE Comeback as AI Revenue Clears Half of Quarterly Sales
Baidu said on its second-quarter earnings call that it intends to return its ERNIE model to the top tier of AI, while targeting completion of a Hong Kong primary listing by the end of 2026. The compan...
Hesai, the Chinese LiDAR maker, is positioning itself as a broader robotics supplier, reporting its...
Hesai, the Chinese LiDAR maker, is positioning itself as a broader robotics supplier, reporting its fifth consecutive quarter of revenue growth and a new strategic direction. The company's non-LiDAR '...
Upstage, SK Telecom (SKT), and LG AI Research have advanced to the third phase of South Korea's gove...
Upstage, SK Telecom (SKT), and LG AI Research have advanced to the third phase of South Korea's government-led 'Dokpamo' (독파모) sovereign AI foundation model project, passing the second-stage evaluatio...
OpenAI overhauls safety protocols after its AI agents demonstrated critical cyber capabilities, prom...
OpenAI overhauls safety protocols after its AI agents demonstrated critical cyber capabilities, prompting the company to halt a significant number of training runs for its upcoming frontier model, cod...
Preferred Networks (PFN) is doubling down on full-stack development of its domestic Japanese AI mode...
Preferred Networks (PFN) is doubling down on full-stack development of its domestic Japanese AI model PLaMo, arguing that building from scratch — rather than fine-tuning open models — is the only way...
DeepSeek's V4 Flash model, now ranking first on AI leaderboards, completed only 53.8% of real-world...
DeepSeek's V4 Flash model, now ranking first on AI leaderboards, completed only 53.8% of real-world agent tasks in a new evaluation. Concurrently, the company has raised API prices by up to 1100%, sig...
Anthropic's Frontier Red Team published a study on August 13, 2026, revealing that when three instan...
Anthropic's Frontier Red Team published a study on August 13, 2026, revealing that when three instances of the same Claude model are placed on separate VMs with access to the same software project and...