Insights archive
Every analysis we have published, newest first. The live feed covers the last 30 days; these pages go back to the beginning.
887 stories · page 22 of 45
Google Adds AI Voice Search to Gmail
Google has launched Gmail Live in beta, letting paying subscribers search their inbox by speaking instead of typing, with similar voice features coming to Google Docs and Keep later this summer.
· 2 min read
Sakana AI Launches Fugu After US Pulls Anthropic Access
A Tokyo AI company released a service that routes work across multiple AI models at once, pitching it as a way to avoid the disruption that hit businesses when the US government abruptly banned international access to Anthropic's most advanced models.
· 3 min read
Google Cloud Lists AI Models Built on Lab Data, Not Text
Google Cloud is adding SandboxAQ's science-trained AI models to its marketplace, giving pharma, materials, and semiconductor researchers access to tools that work from real laboratory data rather than internet text, arriving in Q3 2026.
· 2 min read
Albertsons Builds AI Tool to Guide Store Buying Decisions
Albertsons is rolling out a platform that lets store buyers ask plain-language questions about pricing, promotions, and shelf space, part of a $2 billion technology push as the grocery industry races to make data-driven decisions faster.
· 2 min read
Base44 Launches Its Own AI Model, a First for App Builders
Base44, the app-building platform owned by Wix, has launched its own custom AI model called Base1, becoming the first tool of its kind to stop relying entirely on outside AI providers, with direct implications for the growing cost pressure businesses face when using AI platforms.
· 3 min read
AI Spending Grows Headcount at Tech Firms, Not Others
A new report from Ramp and Revelio Labs finds companies spending heavily on AI are actually growing headcount faster, including entry-level roles, but the gains are concentrated in tech-forward firms with the capital and people to act on AI properly, while everyone else stays stuck in pilot mode.
· 3 min read
California Signs Claude Deal at Half the Standard Price
California has signed a statewide deal giving all its government agencies and local cities access to Anthropic's Claude AI at a 50% discount, at the same time the federal government has effectively banned Anthropic from Pentagon work, creating an unusual split between state and federal AI policy that other public and private sector operators should watch closely.
· 2 min read
Huawei Takes China's AI Chip Market From Nvidia
Nvidia has gone from holding 95% of China's AI chip market to effectively zero, with Huawei stepping in as the dominant supplier, and the two countries are now building entirely separate AI technology stacks that will affect how businesses globally source and use AI tools.
· 3 min read
Claude Now Runs on Azure With Faster Hardware
Microsoft and Anthropic have made Claude available in Microsoft's Azure cloud platform, running on NVIDIA's latest hardware, which makes it easier for businesses already using Microsoft tools to add Claude to their workflows without a separate contract.
· 2 min read
Arena AI Hits $100M Revenue in Eight Months
Arena, the free leaderboard that millions use to compare AI models, has quietly become a $100 million business by selling detailed performance analysis to the AI labs whose products it ranks, raising a real question about whether the world's most-watched AI report card can stay independent.
· 3 min read
US Bill Would Ban AI Chatbots From Selling Your Health Data
US lawmakers are preparing a new law to stop AI companies from selling health and location data people share with chatbots like ChatGPT or Claude, exposing a gap most people do not know exists.
· 2 min read
xAI Launches Grok 4.5 in Private Beta at Tesla and SpaceX
Elon Musk's AI company xAI has put its most powerful model yet, Grok 4.5, into internal testing at Tesla and SpaceX, with plans to release brand-new AI models every month for the rest of 2026, as the company races to catch up with Anthropic and OpenAI in a market where it currently trails.
· 2 min read
Samsung and SK Hynix Commit $518B to New Chip Factories
South Korea's two largest chipmakers are building four new factories to keep pace with AI-driven demand for memory chips, a shortage already pushing up prices for laptops, smartphones, and electronics across every industry.
· 2 min read
OpenAI Launches GPT-5.6, US Government Controls Access
OpenAI launched three new AI models on June 26: Sol, Terra, and Luna, but three days later general access still has no confirmed date, as the US government is approving customers one by one under a new pattern that now applies to both OpenAI and Anthropic.
· 3 min read
AI vendor lock-in is now a board-level cost problem
A wave of new data and policy changes from major software vendors shows that companies which adopted AI tools quickly are now finding it expensive, complicated, and sometimes technically impossible to change course, while employees on the ground remain skeptical and underequipped.
· 3 min read
OpenAI Maps Which EU Jobs AI Will Change First
OpenAI released a country-by-country framework mapping which European jobs face automation, which will grow, and which will simply change, giving employers and workers a clearer picture of where pressure is coming first.
· 2 min read
Palantir Brings Private AI to US Government Agencies
Palantir and NVIDIA have launched a system that lets US government agencies run powerful AI on their own isolated computers, completely cut off from the internet, which is a model that regulated businesses in finance, healthcare, and manufacturing are now watching closely for their own use.
· 2 min read
LinkedIn Automates Job Applications for Paying Users
LinkedIn has launched a tool that writes and pre-fills job applications on behalf of paying subscribers, hiding the AI involvement from recruiters, at a moment when hiring is already drowning in AI-generated noise.
· 2 min read
HP Deploys OpenAI Agents Across Its Entire Business
HP Inc. has formalized a full-scale partnership with OpenAI's enterprise platform called Frontier, moving from a four-month pilot into company-wide deployment across customer support, security, software, and internal operations, making it one of the clearest public examples so far of a major traditional company treating AI as an operating layer rather than a side experiment.
· 2 min read
Ford Rehires 350 Engineers After AI Quality Failure
Ford admitted its automated quality systems failed to replace experienced engineers, spent three years quietly rehiring 350 veterans to fix the damage, and has now climbed from near the bottom of the US vehicle quality rankings to the top mainstream spot for the first time in 16 years.
· 2 min read