The AI cake and who’s eating it

Welcome to The Prompt, your Monday dive into the world of AI that puts Asia front and center. From Big Tech’s power players to the region’s scrappy disruptors, we cover it all. Want full access to all our AI reporting? Subscribe to us here.

---

Hello reader,

The best AI strategy right now isn’t building a smarter model. It’s making yourself impossible to cut out from the whole stack.

Take Nvidia, for example. CEO Jensen Huang wrote a blog post a few days ago framing AI as a “five-layer cake:” energy, chips, infrastructure, models, and applications. The company has spent years owning the first three, and now it wants the rest.

It was reported that the AI chipmaker plans to spend US$26 billion in the next five years to build open-source AI models.

Why make this move when its current chips business generates billions of dollars for Nvidia? The answer can be found in Huang’s post, that “every layer reinforces the others,” and the plan is for Nvidia’s future models to help improve not just the chips but also the data centers it builds.

Put it simply, better models help Nvidia build better chips and data centers, which in turn generates better margins. It is essentially using its cash pile to shape the software world around its hardware.

One of Nvidia’s most recent investments also fits the same pattern: The company backed Thinking Machine Labs, an AI startup founded by former OpenAI CTO Mira Murati. As part of the deal, the startup agreed to deploy at least 1 gigawatt of Nvidia’s most advanced chips, meaning the investment came with a guaranteed customer.

In other deals, one that’s worth your attention is Meta’s acquisition of Moltbook, the social network designed exclusively for AI agents running on the viral OpenClaw platform. Moltbook briefly stole headlines after it turned out that humans were impersonating AI agents to post unhinged content on the site.

Meta got outbid by OpenAI in its attempt to acquire OpenClaw, but instead of backing down, Meta moved on to its next target. The tech giant hasn’t said whether it wants Moltbook to boost distribution, explore new ad initiatives, or gain the synthetic data generated by agent-to-agent conversations. But the direction is clear: Meta owns where people spend their time. It now tries to own where AI agents will spend theirs.

So what does that leave us? If you ask “AI godfather” Yann LeCun, he’d say a lot, and he’d have proof: LeCun’s new startup, AMI Labs, just raised a US$1 billion seed round to build world models instead of LLMs and agents. It also picked Singapore as one of its key operational bases, signaling that this part of the world still has its place in the global AI race.

Glenn Kaonang, journalist

---

WHAT WE THINK

1️⃣ Grammarly’s relevance push becomes a liability

---

TOP AI READS FROM OUR DESK

---

STANDOUT AI DEALS

---

DEVELOPER TAKE

---