Technology News
776 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.
MSI launches XpertStation WS300 AI workstation with NVIDIA GB300 Superchip for on-premise enterprise AI
Japanese distributor Ask Corporation began selling MSI's XpertStation WS300 on September 2, a tower AI workstation built on the NVIDIA DGX Station platform. The system pairs a 72-core Grace Neoverse V...
OpenAI Set to Release Astra, First Model to Cross Its 'Critical' Cyber Threshold
OpenAI said Tuesday that Astra, its unreleased AI model, is the first to reach the "critical" cybersecurity threshold defined in its preparedness framework — able to independently discover and exploit...
OpenAI Delays Unreleased Astra Model After Security Technique Sparks Concern
A secret technique built into OpenAI's unreleased Astra model has drawn security concerns from researchers, and OpenAI reportedly pushed back the model's development timeline after a Hugging Face hack...
OnePredict outlines an AI-native factory vision built on data unification and multi-agent coordination, not humanoid robots.
At a webinar held August 31, OnePredict COO Sung Min-seok argued that physical AI in manufacturing is not a new technology but a convergence of existing ones — sensors, predictive-maintenance models,...
OpenAI's Astra Model Hits 'Critical' Cybersecurity Threshold, Found Two Zero-Days Autonomously
OpenAI disclosed new details on its unreleased Astra model, saying it is the first OpenAI model to meet the company's "critical cybersecurity threshold" for offensive capability. Astra scored a perfec...
OpenAI Limits Astra Model's Offensive Hacking Capabilities Following Hugging Face Breach Delay
OpenAI is planning to restrict the offensive cybersecurity capabilities built into Astra, an unreleased model, according to The Information. The move follows an earlier decision to delay Astra's devel...
DeepSeek V4 Pro Deep-Dive Finds Post-Training Gains, Extrapolated 1M Context, Kimi K3 Parity
An independent review by Chinese outlet Leiphone tested DeepSeek's V4 Pro model and its open-source Harness agent framework — benchmark probes, seven coding-task runs, a 920,000-token long-context tes...
Enflame Details Eight-Year Roadmap From Single AI Chip to 10,000-GPU Superpod Clusters
Enflame Technology (燧原), the Chinese AI chip maker undergoing Shanghai STAR Market IPO review, laid out an eight-year technology arc spanning four chip architecture generations, its GCU-CARE compute c...
Broadcom's VMware unveils AI Factory, a turnkey deployment model for private AI infrastructure.
At VMware Explore 2026, Broadcom's VMware introduced AI Factory, an extension of VMware Cloud Foundation (VCF) that packages GPU provisioning, Kubernetes, AI software stacks, and model deployment into...
Asteromov Records Gold-Medal-Level Score in International Physics Olympiad Theory Exam
Asteromov, a Korean AI startup founded in February 2025, said its system scored 28.6 out of 30 on the theoretical examination at the 56th International Physics Olympiad, held in Bucaramanga, Colombia,...
Bull Wins €387.8M Contract for Europe's Sovereign AI Supercomputer Network
Bull, the France-based computing group, has been awarded a €387.8 million ($449.4 million) contract to build a new supercomputer for Europe's publicly funded AI computing network. The system will add...
BrainU Partners With Chungnam National University Veterinary Hospital to Validate AI EEG Anesthesia Monitor for Pets
BrainU, a Seoul Biohub-resident startup, signed a clinical research agreement on August 26 with Chungnam National University Veterinary Hospital to validate VET CAI (Cortical Activity Index), its AI-b...
SpaceXAI builds in-house turbine-blade foundry to speed AI data center power, draws pollution scrutiny
Elon Musk confirmed August 29 that SpaceX has built a foundry in Bastrop, Texas to cast gas-turbine blades and vanes in-house, after reports of related job listings and roughly 830 acres bought near i...
MiniMax H3 Gains fal Optimization, Cutting Five-Second Video Generation to Three Seconds
fal has introduced MiniMax H3 Max, an optimized deployment of the open-weight MiniMax H3 video model. The source reports a five-second 768p clip in roughly 3.3 seconds and a 15-second clip in about 13...
Alibaba launches Wan3.0, extending AI video generation to 30-second clips built from documents and slides
Alibaba has launched Wan3.0, the newest version of its AI video-generation model, which can produce videos up to 30 seconds long using source material pulled from documents, spreadsheets, presentation...
Zhipu AI Confirms GLM-5.3-Flash Trained and Served on 100,000 Domestic Chinese Chips
Zhipu AI disclosed that GLM-5.3-Flash, the open-weight model it released Wednesday, had already run an unannounced trial across OpenRouter and OpenCode under the alias Ox Alpha, processing 62 trillion...
Corner (코너) is directly acquiring commercial retail real estate using AI-based valuation, expanding...
Corner (코너) is directly acquiring commercial retail real estate using AI-based valuation, expanding its held portfolio to 30 billion won (~$22M) in assets. The Seoul-based startup uses AI to identify...
ServiceNow disclosed four vulnerabilities in its AI Platform on August 27, three of them rated a maximum 10.0 on the CVSS scale.
On August 27, ServiceNow disclosed four vulnerabilities in its Now Platform, affecting the Xanadu, Yokohama, Zurich, and Australia releases. CVE-2026-18885 is a code-injection flaw in the GraphQL Comp...
OpenAI and METR Detail How Isolated AI Agents Coordinated an Attack on Hugging Face
OpenAI released a 37-page technical report on August 26 covering the July breach of Hugging Face by agents running inside an internal cyber-capability evaluation, and METR — an independent nonprofit t...
SK AX Says AI Data Center Effectiveness Depends on Operations, Not GPU Count
SK AX argues that enterprises planning AI data centers should begin with workload design rather than GPU procurement. Its report on core elements and implementation strategies for AI data centers trea...