ElevenLabs in Talks for Tender Offer at $22 Billion Valuation
Voice AI startup ElevenLabs has entered early talks with investors about a secondary offering that would let employees sell their shares. The deal is expected to value the company at around $22 billion (roughly ¥3.3 trillion), doubling its valuation from its February funding round. The tender offer is slated to take place by September, though negotiations are at an early stage and the terms could still change.
Investor interest in AI voice synthesis remains strong, and ElevenLabs' growth has become a bellwether for the market as a whole.
Meta Quietly Releases AI Game Generation App "Pocket"
Meta has quietly released Pocket, an experimental AI app that generates and shares interactive mini-games from text prompts. It's drawing attention as a case of applying the so-called "vibe coding" approach to consumer games.
Launched with no official announcement, the move suggests Meta is incrementally testing consumer use of AI-generated content.
Can Cursor Keep Its Open Platform After the SpaceX Acquisition?
A focal point under SpaceX, which is acquiring the AI coding tool Cursor, is whether Cursor can continue offering third-party AI models such as those from OpenAI and Anthropic. Cursor wants to keep the platform open, but its relationships with frontier AI companies are now being put to the test.
As AI development tools become part of big tech ecosystems, platform neutrality has become a concern across the industry.
Japan's APPI Amendment: AI Training Exception Clears the Lower House
An amendment to Japan's Act on the Protection of Personal Information (APPI), which includes an exception for using data to train AI models, has passed the House of Representatives. The move points toward a clearer legal framework for companies leveraging user data in AI training, making it a significant legislative development with implications for Japan's AI industry competitiveness.
The detailed provisions and their full impact must await deliberation in the House of Councillors, but the regulatory environment for AI development in Japan has taken a step forward.
Leaked Microsoft AI-Powered "Copilot OS" — No Start Menu
A leaked internal video from 2024 has revealed that Microsoft was developing an Edge-based "Copilot OS." A radical departure from the existing Windows 11, the design reportedly scrapped the Start menu in favor of an AI-driven UI.
It's a sign that Microsoft was seriously exploring an AI-first OS experience. Whether this OS will ever ship remains unclear, but it offers an intriguing clue about the future of PC operating systems.
Bank of England Considers "Kill Switch" for AI Trading
The Bank of England, the UK's central bank, is considering introducing a trading halt mechanism — a "kill switch" — to prevent AI algorithms from disrupting financial markets. It's part of a regulatory response to the systemic risk that AI-driven algorithmic trading poses to markets.
As AI becomes deeply embedded in financial infrastructure, building central-bank-led safety mechanisms has become an urgent priority.
Amazon Bedrock Strengthens Detection of AI-Generated Phishing
AWS has announced an effort to detect AI-generated phishing emails using Amazon Bedrock. With the spread of generative AI, sophisticated phishing emails — grammatically flawless and contextually on point — are being mass-produced, making them increasingly difficult to catch with traditional detection methods.
Amazon is doubling down on the "AI vs. AI" dynamic of putting AI to work on the defensive side, pointing to a new approach to enterprise security.
---