The Debate Over Anthropic’s New Product: Price or Existential Dread?

**Nathaniel Whittemore** (0:00)
Today on the AI Daily Brief, a big dust up around Anthropix new product. How much of it is about price and cost versus some larger existential ennui? Before that in the headlines, the open qualification of the world continues. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.
We have been tracking closely the Claw-fication of the world, and last week, no less than Jensen Wong had some very positive words for the project, calling it maybe the most important release of software ever. It felt perhaps like a bit of hyperbole, but with a new report from Wired, it makes a little bit more sense. Wired reports that Nvidia is planning to launch their own AI agent platform that is not dissimilar to OpenClaw. They write Nvidia is planning to launch an open source platform for AI agents, the chip maker has been pitching the product, referred to as NemoClaw, to enterprise software companies. The platform will allow these companies to dispatch AI agents to perform tasks for their own workforces. Companies will be able to access the platform regardless of whether their products run on Nvidia chips. Now, the timeline for this seems to be around Nvidia's annual developer conference, which happens next week. Nvidia apparently has been reaching out to very premier partners like Salesforce, Cisco, Google, Adobe and CrowdStrike for partnerships around the platform. Now, Wired writes, For Nvidia, Nemoclaw appears to be part of an effort to court enterprise software companies by offering additional layers of security for AI agents. It's also another step in the company's embrace of open-source AI models, part of a broader strategy to maintain its dominance in AI infrastructure at a time when leading AI labs are building their own custom chips. Nvidia's software strategy until now has been heavily reliant on its CUDA platform, a famously proprietary system that locks developers into building software for Nvidia's GPUs, and has created a crucial moat for the company. What's interesting is that last year, there was a lot of discourse about this idea of Nvidia moving up the stack and diversifying away from just pure chips, sort of as a hedge against how the world might change in positioning themselves for those potential different outcomes where people are less reliant on Nvidia chips, whether it's because they've got their own custom silicon or because the nature of the field has changed. I feel like there's less of a sense of that being a likely outcome right now than there was last year. You've basically seen a lot of the big players like Meta seemingly back off of their custom silicon projects. Not I don't think because they're not interested anymore but because the simple reality is that right now they just need to compute at basically any cost and don't have time to wait around to figure out their own systems. Now I don't think that Nvidia, as smart as they are, is going to stop hedging against future changes but it will be interesting to see if and where any of these various experiments that they have outside of chips themselves start to actually become a more significant business line for the company in the future. Next up, Microsoft gets in the Coworking game. On Monday, Microsoft's CEO Xiaotian Nadella tweeted, Announcing Copilot Cowork, a new way to complete tasks and get work done in M365. When you hand off a task to Cowork, it turns your request into a plan and executes it across your apps and files, grounded in your work data and operating within M365 security and governance boundaries. Axio sums up the move this way. Microsoft launched Copilot Cowork on Monday, an enterprise AI agent built on Anthropic's technology and named after the Anthropic product that wiped hundreds of billions off of Microsoft's market cap. In other words, if you can't beat them, join them. Indeed, this is not just a copycat version of Cowork, this is actually a collaboration with Anthropic. Working closely with Anthropic, they write, we have brought the technology that powers Claude Cowork into Microsoft 365 Copilot. Microsoft is also increasingly pushing the idea of being able to select between different models. In that same blog post, they write, your work is not limited by one brand of models. Copilot hosts the best innovation from across the industry and chooses the right model for the job regardless of who built it. There is of course a lot of memeing going around about Microsoft being behind or just copying others, but in this case, I think their speed to response on this is actually pretty good. There are lots and lots of people who by virtue of their work environments are stuck in the Copilot ecosystem and for there to be less than a two-month gap between when Anthropic drops Cowork and when Microsoft offers their version in Copilot, that's a lot better than Copilot users have expected in the past. I also think there's a certain humility and intelligence in not trying to do a janky version of it but just partnering with Anthropic to actually get the thing close to the same level of capability that the Claude version has. Sean Wang writes, Wait, did Microsoft really clone Claude Cowork? That's kind of based. Still, Ethan Molek brings up the big question that will or won't probably dictate success for this that will have a big impact on whether this thing is seen as successful. Molek writes, Microsoft seems to be launching its own branded version of Cowork. A big question is whether it will continue to use lower-end models without telling you. Also, whether it will keep pace as the space evolves, or is it a one-off? To me, the question of whether Microsoft will give access to the most recent and best models is big. Given that GPT-5 beat or tied humans in expert tasks less than 38% of the time, while months later GPT-54 beat or tied human experts 82% of the time, this really matters. Another big question, he writes, is this limited to producing materials that use Microsoft apps, how does it handle the fact that so much of what makes Claude Cowork interesting is the fact that it can improvise all sorts of output using code? Adding a little bit of trajectory context to this, Brett Winton from Arc shared the revenue projections from Anthropic and OpenAI as compared to Windows and Office's top line revenue, showing that if indeed Anthropic and OpenAI are correct, they will exceed Windows and Office revenue sometime in 2028 As Brett wrote, what Microsoft built in around 40 years, they will have surpassed in around five.