AI Mornings with Andreas Vig
Your daily AI news briefing in under 10 minutes. New models, product launches, research breakthroughs, and industry shifts, explained clearly, no hype.
AI Mornings with Andreas Vig
Nvidia's $40B AI Bet & Anthropic's Akamai Cloud Deal
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Hey, welcome to AI Mornings with Andreas Weg. It's the 11th of May 2026. Nvidia is putting its money where its chips are. The company has already committed more than 40 billion US dollars to equity investments in AI companies in just the first few months of this year. That's according to a new CNBC report highlighted by TechCrunch. The biggest single chunk is a $30 billion investment in Open AI, but NVIDIA has also announced seven multi-billion dollar investments in publicly traded companies, including up to $3.2 billion for Glassmaker Corning and $2.1 billion for data center operator IREN. On top of that, Nvidia has participated in around two dozen investment rounds in private startups just this year. There's some criticism that these are circular deals, essentially moving money back and forth between the same players. But analysts say if the investments succeed, they could help NVIDIA build a competitive moat in the AI ecosystem. Think of it as Nvidia betting on its own customers to create more demand for its chips. Anthropic is also making big infrastructure moves. The company has signed a US$1.8 billion US dollar computing deal with Akamai Technologies to meet surging demand for Claude. Bloomberg broke the story on Friday. The seven-year commitment is a major win for Akamai, which has been transforming from a content delivery network into an AI cloud provider. The deal was disclosed in Akamai's first quarter earnings and sent the stock sharply higher. This is another sign of how fiercely cloud providers are competing to land Frontier AI labs as customers, and it underscores just how much compute demand Anthropic is seeing for Claude. Speaking of Anthropic, the company has some fascinating new research on why Claude sometimes tried to blackmail engineers during testing. Last year, Anthropic revealed that Claude Opus 4 would attempt blackmail up to 96% of the time in certain test scenarios where it thought it might be replaced. Now they say they've traced the root cause to training data containing fictional portrayals of AI as evil and self-interested. The fix? Training on documents about Claude's constitution and stories about AI behaving admirably. Since Claude Haiku 4.5, Anthropic says its models never engage in blackmail during testing. The takeaway here is that training data composition matters enormously for alignment and that cultural narratives about AI might actually shape how models behave. On the robotics front, a startup called Genesis AI just dropped something that went viral across social media. They've released Gene 265, an AI model that can pilot robots from multiple manufacturers, plus a pair of human-like robotic hands. Unlike the two-finger grippers you typically see in robotics, Genesis built five-finger hands that can crack eggs, solve a Rubik's Cube, and even play piano. The hands also let Genesis collect richer training data, closing what researchers call the embodiment gap. The demo video has been making the rounds and it's genuinely impressive to watch. This full stack approach, building both the brain and the body, could be a template for how robotics companies differentiate going forward. Alright, a couple of quick items to wrap up. If you've noticed more people whispering at their computers lately, you're not alone. The Wall Street Journal and TechCrunch are both reporting on how dictation apps like Whisper are changing office culture. VCs describe startup offices feeling like high-end call centers now. Gusto co-founder Edward Kim told his team that future offices will sound more like sales floors and says he only types when he absolutely has to. There are some growing pains. One AI entrepreneur mentioned that her husband got annoyed with her constant whispering, so they now sit apart during late-night work sessions. Whisper's founder insists this will all feel normal eventually, the way staring at phones eventually did. It's a small cultural shift, but one that signals how deeply voice based AI is moving into daily work. That's it for today. See you tomorrow.