Latest publications
OpenAI to nearly double headcount by 2026 to boost ChatGPT sales for businesses
OpenAI is preparing to nearly double its team, expand its San Francisco offices, and bet on ChatGPT sales to enterprises in response to pressure from Anthropic and Google.
Huawei unveils Atlas 350 and claims it outperforms Nvidia H20 in AI inference
Huawei announced the Atlas 350 accelerator based on the Ascend 950PR chip and says it is more powerful than Nvidia H20 for the Chinese market, reinforcing its bet on its own AI semiconductor platform.
OpenAI slowed data center construction to reassure investors before going public
OpenAI decided to slow the pace of deals to expand its compute infrastructure: investors were concerned about the high costs and the complex financing structure ahead of the company's possible stock market debut.
Capcom will not add generative AI to Resident Evil Requiem and new games
Capcom told investors it will not insert AI-generated materials in its games, but will continue testing generative AI as an internal tool for graphics, audio, and code.
Claude from Anthropic learned to manage Mac computers on behalf of users
Anthropic updated Claude: now the AI opens files, manages browsers and launches development tools on Mac — tasks can be assigned directly from smartphones.
OpenAI Cites Microsoft and Azure Dependence as Risk Factor for Investors
OpenAI warned investors during fundraising that its long-standing reliance on Microsoft funding and Azure cloud creates a risk: the company's growth is too heavily dependent on a single partner.
SoftBank prepares another $30 billion in debt for OpenAI after asset sale
SoftBank, which already sold assets for the previous OpenAI funding round, now plans to raise another $30 billion in loans and intends to relax its internal debt guidelines.
Google Photos adds virtual try-on for your own wardrobe
Google has built an AI wardrobe into Google Photos: the app now virtually tries on your own clothes and shows how any outfit will look on you.
Meta Develops AI Agents for Individuals and Business Based on Muse Spark Model
Meta, through its Superintelligence Labs division, is preparing personal and business agents based on Muse Spark to simplify task automation for regular users and companies.
Nvidia B300 servers in China surge to $1 million amid supply shortage and grey market imports
Nvidia B300 servers in China have nearly doubled in price, now reaching $1 million on the grey market: US export restrictions and crackdowns on smuggling have intensified AI computing capacity shortages.
OpenAI Shuts Down Sora and Winds Down Separate Video App After Viral Launch
OpenAI has halted its Sora app for iPhone just months after launch: viral interest failed to translate into sustained demand, prompting the company to shift focus to a broader product.
Arm launches its own AI processors and targets $15 billion in annual revenue
For the first time in 35 years, Arm is exiting its licensing model to directly sell server chips, expecting this new business to generate $15 billion annually by 2031.
OpenAI to raise another $10B from investors for data centers and training new models
OpenAI seeks to raise an additional $10B in its current funding round: funds will go toward compute infrastructure, data centers, and training next-generation language models.
Microsoft Adds AI Restyle to OneDrive for Artistic Photo Redesign Without Distortion
Microsoft launched AI Restyle in OneDrive — a tool that transforms ordinary photos into anime, movie posters and other styles, preserving the scene, faces and key details without noticeable distortion.
Walt Disney Abandons Nearly $1 Billion OpenAI Deal Following Sora Closure
Following Sora's closure, Walt Disney exited a roughly $1 billion deal with OpenAI, according to The Hollywood Reporter, underscoring how fragile partnerships in AI video remain.
OpenAI secures 10 GW computing capacity ahead of schedule
OpenAI has secured contracts for 10 GW of computing power ahead of its original timeline: the company initially planned to reach this milestone by 2029 but has now completed it.
Microsoft claims 20 million paid Copilot users in enterprise Microsoft 365
Satya Nadella announced that Microsoft 365 Copilot already has 20 million paid enterprise users, which should dispel doubts that the service is purchased but not actually used.
Anthropic Prepares May Funding Round: Valuation Could Rise to $900B and Surpass OpenAI
Anthropic could reach a $900B valuation as soon as May after raising some $30B since February, which would position the company above OpenAI in private valuation.
Apple Will Allow Connecting Gemini and Claude to Siri in iOS 27 Through Extensions System
In iOS 27, Apple may allow iPhone owners to choose their own AI model for Siri responses, connecting not only ChatGPT but also Gemini, Claude, and other AI assistants through an Extensions system.
Anthropic Prepares October IPO and Seeks Over $60 Billion for AI Development
According to Bloomberg and The Information, Anthropic is discussing an IPO in October 2026 and expects to raise over $60 billion to invest in AI infrastructure and data centers in the US.
Google adds memory and chat import from ChatGPT and other AI assistants to Gemini
Google has enabled transferring chats and saved memory from ChatGPT and other AI services to Gemini, lowering barriers to switching and intensifying competition in the assistant market.
Anthropic wins sanctions relief: court orders Pentagon to remove company from blacklist
U.S. federal court ruled unlawful the inclusion of Anthropic in the Pentagon's list of unreliable suppliers and ordered the department to lift restrictions within seven days.
Anthropic Accelerates Claude Rate Limit Depletion During Peak Hours for Free, Pro, and Max Users
Anthropic has modified how it counts rate limits in Claude: during high-load periods, five-hour sessions for Free, Pro, and Max users now deplete faster, though weekly limits remain unchanged.
xAI loses all cofounders: Ross Nordeen departs Elon Musk's company
Startup xAI has lost its last cofounder besides Elon Musk: Ross Nordeen departed amid massive company restructuring ahead of SpaceX's potential IPO and hiring restart.
Anthropic Wins Pentagon Case, But Final Victory Remains Uncertain
A federal court in California ordered the Pentagon to remove Anthropic from its list of unreliable suppliers, but experts warn that the final outcome will be decided by a higher court.
OpenAI shut down Sora due to resource shortage, breaking major Disney deal
WSJ revealed that OpenAI halted Sora support not due to new strategy, but due to computing power deficit: the service cost up to $1 million per day and competed with other priority projects.
Microsoft taught Copilot Researcher to cross-check GPT and Claude answers in a single process
Microsoft added a Critique mode to Copilot Researcher, where GPT prepares the answer and Claude verifies it for accuracy, while simultaneously beginning rollout of the Copilot Cowork agent.
Google and Meta Lose Leading AI Researchers as New Startups Rapidly Raise Billions
Former researchers from Google DeepMind, Meta, OpenAI, and Anthropic are launching AI startups and securing hundreds of millions or even billions of dollars from investors nearly immediately.
OpenAI Claims ChatGPT Learned to Count Letters, But Still Fails on Simple Words
OpenAI announced that ChatGPT now correctly answers questions about the letters in strawberry, but users quickly demonstrated: behind this narrow fix lies the same old problem of confident errors.
Russian companies sued over AI-generated work by employees, but courts rarely rule in their favor
Russia is experiencing a growing number of disputes over work completed with neural networks, but courts typically focus not on AI itself, but on the contract, work acceptance, and proven damages.