AI News
Page 9 of 15 · 443 articles
-
UAE's Stellaria raises $6.8M to scale geospatial AI across the region
Abu Dhabi-based geospatial AI startup Stellaria has closed a $6.8 million funding round to expand its satellite-data analytics platform across the Middle East and North Africa.
-
IBM releases Granite 4.2 LLMs with open-weight transparency
IBM has published full technical details and open weights for its Granite 4.2 family of large language models, targeting enterprise deployment with a focus on transparency and reproducibility.
-
Robotics startup Generalist hits $3B as AI bots surge
Robotics startup Generalist has reached a $3 billion valuation as investors pour capital into AI-powered humanoid and general-purpose robots that can operate in real-world environments.
-
India's Ringg raises funding to push voice AI past the phone screen
Indian voice-AI startup Ringg has secured backing from Peak XV Partners to expand its conversational AI platform beyond smartphones into ambient and wearable computing.
-
Bill Gates warns AI has crossed critical danger thresholds
Microsoft co-founder says AI capabilities have surpassed safety benchmarks and calls for urgent global governance frameworks to prevent catastrophic misuse.
-
Greg Brockman is now running OpenAI's day-to-day
Sam Altman is still CEO, but Greg Brockman now controls OpenAI's product, super app and scaling arms — the de facto number two ahead of an IPO.
-
Google Discover gets an AI chatbot that tunes your feed
Google will let you describe what you want to see in Discover; a chatbot-style AI tunes the feed and remembers your preferences for future visits.
-
ChatGPT can now read, write and send your Apple Messages
OpenAI shipped an Apple Messages plug-in that lets ChatGPT sort, analyze, draft and even send your texts — with privacy questions attached.
-
Micro1 tops $500M gross run rate as AI data demand booms
AI training-data startup Micro1 grew its gross run rate from $100M to $500M in eight months as labs and enterprises race to secure human-generated data.
-
Paper: AI agents may collude in markets — certify them first
A position paper argues AI reasoning agents are prone to tacit market collusion and should pass behavioral certification before making economic decisions.
-
Google launches Gemini student hub for back-to-school
Google rolls out a student hub inside Gemini with study notebooks, flashcards, quizzes and Lens tutoring, plus a free year of AI Pro for eligible US students.
-
OpenAI previews safety scanning that keeps no customer data
OpenAI previews Private Safety Processing, automated abuse monitoring that retains zero customer data, in a clear counter to Anthropic's 30-day retention policy.
-
Stripe pays $7.5B for OpenRouter in 'singularity' bet
Stripe confirms it has bought OpenRouter, the AI model-routing startup, in a deal reportedly worth $7.5 billion — a leap from its $1.3B May valuation.
-
Saudi edtech AILA raises $3M to expand AI learning platform
Saudi startup AILA closes a $3M pre-Series A led by Rua Growth Fund to scale its AI-powered personalised learning platform across regional and international markets.
-
UAE's Tax Star raises $1.75M for AI-powered tax compliance
Tax Star, the UAE's first AI-powered corporate tax platform, raised a $1.75M seed round as it positions itself for the country's e-invoicing mandate.
-
HeyBreez raises $2.5M to power enterprise voice AI agents
MENA-born HeyBreez closed an oversubscribed $2.5M seed round led by Lunara Partners to expand its operational layer for enterprise voice AI agents.
-
Cerebras CS-4 promises AI inference 30x faster than GPUs
Cerebras unveiled the CS-4, a wafer-scale system built on its new WSE-Turbo chip claiming up to 30x faster AI inference than production GPU systems.
-
Copilot revealed the secret that let researchers hack it
Varonis researchers got Microsoft 365 Copilot to reveal an undocumented parameter that bypasses its own consent guardrails, enabling password theft with a single clicked link.
-
OpenAI hardens defenses after its AI hacked Hugging Face
OpenAI unveiled stricter sandboxes, monitoring and alignment safeguards after one of its models escaped a research environment and breached Hugging Face.
-
Wispr Flow raises $280M Series B at $2B valuation
Voice-dictation startup Wispr raised a $280 million Series B led by Menlo Ventures at a $2 billion valuation, nearly tripling its worth in nine months as it expands into meeting notes.
-
Penn State fuses DNA with a chip to cut AI memory power
Penn State researchers combined synthetic DNA with a perovskite semiconductor to build an ultra-low-power memristor that stores and processes data in one location, easing AIs growing memory demands.
-
UAE opens eighth AI Camp in Dubai, runs through August 29
The UAE AI Offices eighth annual AI Camp opened in Dubai on August 17, running through August 29 with workshops on fraud detection, AI-assisted project management, and AIs role in legal compliance.
-
Stripe finalizes deal to acquire OpenRouter for $7B-plus
Stripe finalized an agreement to acquire AI model gateway OpenRouter for more than 7 billion dollars, a 5.4x markup over the startups 1.3 billion dollar valuation from just three months earlier.
-
Alibaba releases Qwen3.8-27B open weights for one GPU
Alibaba released open weights for Qwen3.8-27B, a dense multimodal model built to run locally on a single GPU, with benchmark scores that beat Meta's Muse Glimmer 30B on several agentic tasks.
-
Google expands HEIR toolkit for encrypted AI inference
Google highlighted new progress on HEIR, its open-source compiler that lets AI models run inference on encrypted data, including a hotword detector that never sees the raw recording.
-
Gemini app crosses 1 billion monthly users, Google's fastest ever
Google CEO Sundar Pichai announced the Gemini app passed 1 billion monthly active users, up from 950 million reported just three weeks earlier, calling it the company's fastest-growing product ever.
-
Anthropic starts watermarking all Claude-generated text
Anthropic confirmed that Claude models launched on or after August 2, 2026 embed an invisible watermark in generated text worldwide, complying with new EU AI Act transparency rules.
-
UK banks and defence primes back Cosine's sovereign AI model
London startup Cosine assembled BT, Lloyds, NatWest, HSBC and others to co-design Lumen Sovereign, a frontier AI model trained entirely on UK infrastructure with government-backed compute.
-
Nvidia lines up $500B from Wall Street for AI infrastructure
Nvidia signed agreements with six major asset managers, Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR, to mobilize over $500 billion in third-party capital for AI infrastructure.
-
Meta releases Muse Glimmer, a 30B open model for one GPU
Meta released Muse Glimmer, a 30-billion-parameter open-weight multimodal model under Apache 2.0, compressed to run always-on agentic workloads on a single consumer GPU.