Technology News
886 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.

Exina demonstrates MX1 CXL memory expansion for AI inference with Intel
Exina demonstrated its MX1 memory platform with Intel's Xeon 6 server CPU at FMS 2026, showing how generative AI services can expand KV-cache storage into CXL memory instead of GPU memory. The fabless...

LangGrant proposes an open standard for reusable enterprise AI reasoning
LangGrant, formerly Windocks, announced the Enterprise Reasoning Initiative on September 29, supported by 10 innovators across the AI ecosystem. The open-source project proposes an interoperability st...

Ivo releases open-weight Sage model for multi-step contract review
San Francisco-based legal AI company Ivo has released Ivo Sage, a free, MIT-licensed contract model built on DeepSeek V4 Flash. According to LawFuel, Sage was post-trained with reinforcement learning...
Reflection prepares open-weight model and localized AI ecosystem offering
Reflection is preparing to release an open-weight AI model, according to Axios. The Nvidia-backed startup's model is expected to compete with leading Chinese open-weight models while initially trailin...
Meshy reports $100M ARR as AI 3D assets enter production workflows
Meshy says its annual recurring revenue grew from $1 million to $100 million in under two years, according to QbitAI. The report describes revenue channels spanning individual subscriptions, enterpris...

Anthropic reportedly weighs new model as enterprise competition with OpenAI grows
Anthropic is reportedly considering a new AI model in response to OpenAI's expansion among enterprise customers. AI Life Economy, citing Analytics Insight's account of Reuters reporting, says a releas...
Appier introduces SMITH to train small AI agents to build and reuse tools
Taiwan-based Appier unveiled SMITH, a research method that jointly trains AI agents to create and use software tools. Developed with National Taiwan University researchers, the paper was accepted at N...
Utopai Studios reports second-place video benchmark result for Utopai X
Utopai Studios' customized video-generation model, Utopai X, ranked second globally and first among U.S. entrants on Artificial Analysis' text-to-video leaderboard as of September 30, 2026, according...
openJiuwen introduces X-Router for adaptive model selection in AI agents
openJiuwen, an open-source AI agent platform jointly built by Huawei teams and external developers, has introduced X-Router, a model-routing algorithm within a modular routing framework. The announcem...

TakeMe2Space launches MOI-1A satellite for AI inference on Earth observation data
India's TakeMe2Space launched its MOI-1A satellite aboard SpaceX's Transporter-18 rideshare mission, according to Inc42. The 14-kg satellite carries 117 trillion operations per second of onboard compu...

OpenAI patches ChatGPT Mac flaw that could expose chats and connected applications
OpenAI acknowledged a security flaw and fix for its ChatGPT macOS app in its September 25 change log, according to WIRED. Objective-See Foundation researchers found that a trusted script interpreter c...
ELYZA releases two commercially usable Japanese models based on LLM-jp-4
ELYZA, a Tokyo-based AI developer in the KDDI group, released ELYZA-Thinking-1.0-llm-jp-4-33b and ELYZA-Thinking-1.0-llm-jp-4-32b-a3b on October 2. Both are available free on Hugging Face under Apache...

Google launches Gemini 4 Argon with restricted access for defensive cybersecurity
Google has launched Gemini 4 Argon, a model for coding, research, writing, and long-horizon reasoning, with initial access limited to selected cybersecurity partners through its Fairwind Program. Goog...

Psyonic plans to turn everyday prosthetic-hand use into robot training data
San Diego startup Psyonic is developing a plan to deploy thousands of bionic hands globally and pay wearers to generate training data for robotic manipulation. The effort is in its early stages, with...
DeepSeek Adds Native Huawei Ascend Support to TileLang and Five Core Libraries
DeepSeek has released native Huawei Ascend support for TileLang, alongside Ascend versions of DeepGEMM, DeepEP, FlashMLA, TileKernels and DeepSelect, according to Leiphone. The package covers kernel d...

Snyk reaches $309.2 million in revenue as losses widen during AI security push
Snyk reported 2025 revenue of $309.2 million, up 11%, according to UK financial disclosures reported by CTech. Operating losses increased 8% to $202.7 million, while net losses rose 19% to $197.9 mill...

Lablup introduces model routing and centralized token management
Lablup presented Continuum Router and Continuum Hub at its sixth annual technical conference, held September 29 at COEX in Seoul, according to Platum's September 30 report. Router directs requests to...

OpenAI releases MentalHealthBench for evaluating AI mental-health conversations
OpenAI has released MentalHealthBench, an open benchmark developed with more than 80 licensed mental-health professionals from 22 countries and regions. It covers 19 languages and synthetic conversati...
Fermi Cosmos launches FermiQLLM 1.0 with quantum-inspired model techniques
Fermi Cosmos has introduced FermiQLLM 1.0, which it describes as a model enhanced across data representation, architecture, training, reinforcement learning and evaluation. The approach applies method...
Naive AI Releases N0.5-Flash Model for Automated AI Research
Naive AI has released Naive-N0.5-Flash, an open-weight model designed to read research papers, reproduce experiments and support model development. According to Leiphone, the company began with an exi...