Tech • IA • Crypto
Musk vs OpenAI ends; Google Anti-Gravity 2.0 sparks revolt
Briefing
Musk loses OpenAI lawsuit
Elon Musk has lost his legal case against OpenAI, closing a high-profile dispute between former collaborators. The ruling removes a major uncertainty hanging over the company’s governance and strategy. With litigation resolved, attention shifts back to competition and commercialization. The outcome symbolically resets a rivalry that has shaped the AI narrative.
OpenAI Anthropic SpaceX eye IPOs
OpenAI, Anthropic, and SpaceX are reportedly preparing near-simultaneous U.S. IPOs later this year. Early filings suggest a tightly clustered timeline that could intensify competition for capital. The listings would mark one of the largest waves of tech offerings in recent years. Investor demand is expected to hinge on AI growth prospects and infrastructure scale.
Trillion-dollar AI valuations emerge
Projected valuations reach extraordinary levels, with OpenAI near $1 trillion and SpaceX up to $1.7 trillion. Anthropic could command even higher multiples due to its profitability profile. These figures reflect expectations that AI will underpin multiple industries. They also raise questions about sustainability and market concentration.
Google Anti-Gravity 2.0 backlash
On May 19, Google pushed an automatic update replacing traditional coding tools with Anti-Gravity 2.0. Developers reported missing terminals, file explorers, and direct editing, disrupting active projects. Complaints spread quickly across forums, with many calling workflows unusable. The rollout highlights tensions between rapid AI adoption and developer control.
From IDEs to agent platforms
Anti-Gravity 2.0 transforms the coding environment into an agent orchestration system. It introduces a desktop control hub, CLI, SDK, and managed agents integrated with Google Cloud. Multiple agents can work asynchronously across project components. This signals a shift from writing code to supervising AI-driven development processes.
Alibaba unveils Qwen 3.7 Max
Alibaba launched Qwen 3.7 Max, targeting long-running, multi-agent workflows. The model is positioned as up to four times cheaper than rivals from OpenAI and Anthropic. It emphasizes endurance and coordination rather than peak benchmark dominance. The release strengthens China’s push in enterprise AI systems.
Persistent multi-agent systems advance
Qwen 3.7 Max can reportedly operate for 35 hours while coordinating dozens of agents. It supports up to 1,000 MCP connections, enabling deep tool integration. Systems can maintain continuity even when individual agents fail and are replaced. This reflects a broader move toward durable, autonomous AI workflows.
HBM shortage drives hardware prices
Demand for High Bandwidth Memory (HBM) has surged, becoming a key bottleneck for AI infrastructure. Prices for GPUs, RAM, and SSDs have risen 2–4× as production shifts to enterprise clients. AI memory output has grown from 1.5% to nearly 100% of supply in three years. Shortages are expected to persist until at least 2028, reshaping the hardware market.
Videos covered
IPO SpaceX et OpenAI : choc
- •End of a major legal battle
- •Toward a wave of IPOs
- •Colossal valuations ahead
Google Bans Coding With Its Biggest Attack On Devs Ever: AntiGravity 2.0
- •Sudden rollout disrupts developers
- •From code editor to agent platform
- •Multi-agent workflows redefine development
Cette nouvelle puce fait exploser le prix des PC
- •AI Boom Reshapes Hardware Markets
- •HBM Memory Becomes Critical Bottleneck
- •Production Shift Away from Consumers
IA Alibaba: 35h sans s’arrêter
- •Launch of Qwen 3.7 Max
- •Focus on multi-agent orchestration
- •High MCP connectivity capacity
Demis Hassabis on AGI by 2030, Curing Every Disease, Life After AGI, and More
- •AGI Timeline Narrowing
- •Rapid Product Integration
- •Rise of Multimodal and Agent Systems
Ferrari EV, Enhanced Games, Pope's Encyclical, Alcohol & Podcasting
- •Ferrari enters EV era with Luce
- •Unusual design and collaboration
- •Performance strong but not class-leading