AI digest: Giants raise billions as infrastructure shifts
Anthropic closes massive funding round whilst the industry rebuilds cloud infrastructure for AI agents.
Big money and bigger infrastructure changes this week, as AI companies secure war chests and the internet gets rewired for machines.
Anthropic nears trillion-dollar club with $65bn raise
Anthropic closed a $65 billion Series H at a $965 billion valuation, putting it within spitting distance of the trillion-dollar mark. The company’s annualised revenue has hit $47 billion, suggesting Claude is finding real enterprise traction. This could be their final private round before going public, which would be the AI IPO to watch.
Cloud giants rebuild internet for AI agents
The infrastructure layer is getting a complete overhaul as cloud providers redesign their systems for machine traffic rather than human users. AWS, Cloudflare, and others are betting that AI agents will dominate internet traffic as they move from experiments to production. This isn’t just about handling more requests, it’s about fundamentally different traffic patterns and requirements.
Claude Opus 4.8 ships with agent swarms
Anthropic released Claude Opus 4.8 with Dynamic Workflows, letting you coordinate up to 1,000 subagents for complex tasks. The model beats GPT-5.5 on most benchmarks and catches its own coding errors four times better than the previous version. The workflow orchestration could be more interesting than the raw performance gains.
Liquid AI’s efficient MoE runs locally
Liquid AI’s LFM2.5-8B-A1B activates just 1.5B of 8.3B total parameters whilst handling 128K context windows on consumer hardware. Mixture-of-experts models that actually run on your laptop could shift the on-device AI landscape, especially with reasoning and tool calling built in.