메타, AMD와 1,000억 달러 규모 AI 계약 체결 — 주가 15% 급등 🚀

I struggle to understand why Meta needs this level of AI spend. Their platforms are largely slop and the AI integrations are so basic and just add clutter to the UI. I’d rather see some actual vision for what they want to accomplish before spaffing 100bn on chips

OP spent an hour writing a sentence - it's down to 9% now.

So AMD is willing to sell 10% of their company for a foot in the door with Meta?  Ooof

That level of spending suggests they are planning for way more than just chatbots.

Meta has now struck deals with Nvidia, AMD and Google for TPUs. Are they doing this to prevent their competitors from getting the chips thereby slowing them down?

I was wondering about that big green candle. Thanks for the info.

**Posted a few days ago by me. Very relevant to this deal. In before the inevitable “circular economy” posts. It’s clear why AMD makes these deals. It’s because they’re so far behind Nvidia that they need their customers to own a piece of AMD so they can get into the AI tech stack. They can’t get large players to seriously consider their much less mature systems without these deals.** https://www.reddit.com/r/stocks/s/e66p44SOji

Saying "AI circle jerk" sounds smart and gets free upvotes but is devoid of actual facts and analysis.

The "circle jerk" money is relatively small compared to the actual money in infrastructure buildout. Big tech alone is going to spend $700b in 2026 for data centers. These deals pale in comparison. The buildout is mostly paid by real customer demand.

In case people wonder why companies are making these deals:

* Some big tech like Amazon and Microsoft aren't able to compete in frontier model training. So Amazon invested in Anthropic. Microsoft backs OpenAI mostly. They want access to frontier models. Without these deals, Microsoft and Amazon are on the outside looking in.
* OpenAI and Anthropic need as much compute as they can get. They're constrained and can't serve all their demand. The need to make deals with companies who have chips, energy.
* Nvidia makes this deal because they don't want OpenAI to use non-Nvidia chips too much
* Another reason Nvidia backs OpenAI is because Anthropic is mostly backed by Google and Amazon, who are trying to compete against Nvidia chips with their own custom AI chips. Also, Anthropic CEO keeps on telling congress to block Nvidia chip sales to China.
* AMD made a deal with OpenAI because they want OpenAI to use AMD AI chips so AMD can eventually become a viable alternative to Nvidia
* Oracle's deal with OpenAI does not involve equity. It's just Oracle building data centers for AI, most of which will be sold to OpenAI.
* Nvidia's deal with xAI is same as OpenAI's. They're trying to make sure xAI stays on Nvidia's ecosystem. To Nvidia, the AI winners need to be running on Nvidia hardware and not Google or Amazon hardware. Backing xAI also means getting on the good side of Elon, who like it or not, has big influence in government and social.

* Meta buys Google and AMD chips to keep Nvidia in check.


If you actually look at the underlying reasons for these deals, it's all about companies jockeying for position in what is likely one of the most transformative inventions ever. For OpenAI and Anthropic, it's all about securing compute and distribution. AI is an existential threat for many companies.