Sent to 16 readers
Race to AGI Daily Digest - Tuesday, February 10, 2026
Share:
TLDR
The EU is challenging how Meta controls access to its WhatsApp AI assistant, extending the policy focus from data centers to messaging platforms.
India’s Bhashini platform is moving to a Nvidia H100‑based sovereign cloud, turning national language AI into critical infrastructure.
OpenAI is adding ads to ChatGPT Free and Go in the US, turning a core AI agent into an ad‑driven discovery surface.
ByteDance’s Seedance 2.0 video model signals that high‑end AI video is becoming a key front in the model race.
Legacy data and chip players like Oracle, Palantir, Snowflake, AMD, and Broadcom are rallying as markets price in an AI infrastructure boom.
The Full Story
Building on Monday’s New York story, today felt like the same fight playing out at different layers of the stack.
On the surface, the EU is now warning Meta that locking its WhatsApp AI assistant to its own services may cross competition lines. So yesterday it was a city telling data centers to slow down; today it’s Brussels asking who gets to plug agents into the world’s default messengers. You can dig into the case here: EU warning to Meta over WhatsApp AI assistant ->.
At the same time, governments are quietly building their own AI backbones. India’s Bhashini project is shifting onto Yotta’s sovereign AI cloud, running on Nvidia H100s. That’s not just a hosting change. It’s a state saying: we want our own stack, for our own languages, on hardware we can point to. More on that move here: Bhashini moves to Yotta’s sovereign AI cloud -> and on Nvidia’s broader strategy with specialized hardware here: Nvidia’s Groq talent and technology deal ->.
Now look at the agent side. OpenAI is turning ChatGPT into an ad‑supported assistant for free and Go users in the US. That means the “universal helper” many people lean on will start recommending things with economic incentives baked in. Details are here: OpenAI starts testing ads in ChatGPT ->. At the same time, ByteDance’s Seedance 2.0 is being hailed as a top‑tier video model, exactly the kind of engine that powers those real‑time translators and creative tools we’re seeing on Reddit: Seedance 2.0 video model coverage ->.
Underneath all this, the “old guard” of data and infra—Oracle, Palantir, Snowflake, AMD, Broadcom—rallied again. So while workers swap stories about AI making their jobs more intense, the market is clearly betting on a future where agents sit on top of thick, regulated, and very profitable infrastructure. If you want to see who is building that base layer, the full map is here: Browse all AI companies ->.
Get This Delivered Daily
Join thousands of AI professionals who start their day with Race to AGI.