AI News
The latest developments, breakthroughs, and analysis from the world of artificial intelligence.
Microsoft, Google, xAI Will Give US Early Access to AI Models
Three of America's largest AI developers will share unreleased models with the Commerce Department for national security reviews.
Norway Joins US-Led Pax Silica AI Supply Chain Coalition
Norway becomes the 15th member of Pax Silica, bringing its sovereign wealth fund and critical minerals to the AI supply chain alliance.
How OpenAI Built Voice AI That Feels Instant
OpenAI reveals the WebRTC architecture powering real-time voice for 900M+ users with minimal latency.
OpenAI Reveals How It Delivers Voice AI to 900 Million Users
OpenAI details the WebRTC infrastructure overhaul behind ChatGPT voice mode, serving 900M weekly active users with low-latency audio.
SAP Acquires Prior Labs and Dremio in €1B Enterprise AI Push
SAP bets big on tabular AI with Prior Labs acquisition and €1B investment to crack enterprise structured data.
Saperly Launches as the First Phone Carrier Built for AI Agents
Saperly gives AI agents their own phone numbers for calling and texting, with MCP compatibility and built-in compliance.
Sierra Raises $950M at $15.8B Valuation as AI Agents Scale
Bret Taylor's Sierra hit $150M ARR in eight quarters and has 40% of the Fortune 50 deploying its AI customer agents.
xAI Launches Grok Connectors for GitHub, Notion, and More
Grok now integrates directly with GitHub, Notion, Linear, Google, and Microsoft apps — plus custom MCP connectors.
Mistral Medium 3.5 Hits 77.6% on SWE-Bench, Adds Cloud Coding Agents
Mistral's new 128B flagship folds chat, reasoning, and code into one model and ships remote coding agents in Vibe plus Work mode in Le Chat.
Parag Agrawal's Parallel Hits $2B Valuation, Builds Web Layer for AI
Sequoia leads a $100M Series B in Parallel Web Systems — the AI-agent web infrastructure startup founded by ex-Twitter CEO Parag Agrawal.
Reflection AI Lands Pentagon Deal, Plans Open Frontier Model in 2026
Reflection AI — the only newcomer on the Pentagon's classified-AI list — is building a frontier model on tens of trillions of tokens to challenge DeepSeek.
Salesforce Launches Agentforce Operations to Automate Back-Office Work
Salesforce's new Agentforce Operations targets manual workflows like data validation, compliance checks, and approvals — generally available now.
NSA Tests Anthropic's Mythos AI to Find Microsoft Security Flaws
The NSA is testing Anthropic's Claude Mythos to find security flaws in Microsoft software — and officials say its bug-hunting speed is impressive.
Pentagon Clears 8 AI Firms for Classified Networks, Excludes Anthropic
The DoD signed deals with OpenAI, Google, Microsoft, AWS, Nvidia, SpaceX, Reflection, and Oracle to deploy AI on IL6/IL7 networks. Anthropic is locked out.
Half of Google and Amazon's Q1 AI Profits Came From Anthropic
Alphabet booked a $36.9B equity gain on its Anthropic stake. Amazon booked $16.8B. Together it's nearly half of their Q1 profits — not from AI products.
OpenAI Building AI Agent Smartphone With Qualcomm, MediaTek for 2028
OpenAI is reportedly building an AI agent smartphone with Qualcomm and MediaTek that replaces apps with task-driven assistants. Mass production targeted for 2028.
EU Expands Digital Markets Act to Target Cloud and AI
European regulators will extend the Digital Markets Act to cloud services and AI, targeting Amazon, Microsoft, and major virtual assistant platforms.
Meta Reserves 1 GW of Space Solar to Power Its AI Data Centers
Meta signed deals with Overview Energy and Noon Energy for up to 1 GW of orbital solar and 100 GWh of long-duration storage to power AI infrastructure.
Musk Takes Stand vs Altman in $134B OpenAI Trial
Elon Musk testified for two days in a federal trial seeking $134B from OpenAI and the removal of Sam Altman and Greg Brockman from its board.
Google Signs Classified Pentagon AI Deal as 950 Employees Revolt
Google grants the DoD access to Gemini for 'any lawful' classified use after Anthropic refused — and 950 of its own engineers are publicly objecting.
UPDATE: Musk Takes Stand in OpenAI Trial, Predicts AGI Next Year
Day two of Musk v. Altman: Elon Musk testifies he founded OpenAI as a charity to counter Google and predicts AI smarter than any human within a year.
OpenAI Misses Revenue and User Targets, WSJ Report Rattles AI Stocks
WSJ says OpenAI fell short of internal revenue and ChatGPT user goals as it races toward an IPO. Oracle and AI chip stocks tumbled on the report.
Google DeepMind's Vision Banana Beats SAM 3 by Generating Pixels
DeepMind's new Vision Banana model proves image generators can outperform specialized vision systems on segmentation, depth, and surface normals.
Ineffable Raises $1.1B Seed: AlphaGo Creator Bets Against LLMs
AlphaGo creator David Silver emerges from stealth with $1.1B at $5.1B — a record European seed for an AI built without LLMs or human data.