Anthropic's $20B round, the rise of agent swarms, and the AI acceleration gap

Anthropic finalises a massive raise, Chinese labs deliver competitive open models, and a new gap is opening between AI power users and everyone else.

Big-picture funding numbers, a real Chinese challenger in open-weight models, and the growing realization that AI proficiency is becoming the single biggest workplace skill gap.

Anthropic finalising a $20B+ funding round at a $350B valuation

My take: Leaked forecasts show Anthropic projecting $18B in revenue in 2026, potentially overtaking OpenAI by 2029. The scale of capital pouring into the top two labs is unlike anything we've seen. Practically, this means more aggressive pricing, faster model releases, and more pressure on every other player to consolidate or specialize.

The 'AI acceleration gap' is widening fast

My take: A small group of power users is achieving real breakthroughs with multi-agent systems. Most knowledge workers are still on basic prompts. The gap between the two will compound. If your team's AI usage is still mostly 'write this email for me,' you're already falling behind. The fix is structured experimentation, not a one-off training day.

Apple planning a Siri revamp powered by Google Gemini

My take: The new Siri will work more like ChatGPT, deeply integrated into the OS, with access to files and on-screen context. If you have any product or service that's consumed via voice, mobile, or in-car, the goalposts on what feels normal to users are about to shift dramatically. Plan your roadmap accordingly.

Studies show 40% of AI time savings lost to rework

My take: Translation: if your team isn't trained, AI makes them slower, not faster. There's an 81% to 27% training gap between C-suite and individual contributors. This is the single most under-discussed problem in enterprise AI. Solving it isn't expensive. Most companies just haven't bothered yet.

UK launches national AI upskilling program

My take: Free 20-minute online courses for every adult worker, targeting 10 million retrained by decade's end. Worth watching as a model. Singapore and others will likely follow with similar programs. If you have staff in any of these markets, the government may soon be doing some of your training for you. Take advantage of it.