Google Blocks Meta From Gemini artwork

Google Blocks Meta From Gemini

The Daily AI Show

June 29, 2026

The hosts opened with Google limiting Meta’s access to Gemini capacity and what that says about AI compute constraints, Google Cloud demand, and internal model development.
Speakers: Brian, Beth, Andy, Karl, Gareth
**Brian** (0:00)
Hey, what's going on, everybody? It is June 29th, it's Monday, and this is The Daily AI Show Live. It's 10 AM on the East Coast. And with me today, we have Andy and Beth, and I'm Brian, episode 756, fast approaching that three-year mark, which should be coming up in just over a month. It's around the 8th of August, isn't it, guys? Somewhere in that range?

**Beth** (0:23)
Yeah.

**Brian** (0:24)
Is it the 8th?

**Beth** (0:25)
No, the 7th is our show birthday. The 8th is my personal birthday.

**Brian** (0:30)
8th is your first number. Must have been when I was thinking of that.
Now, wait, hold on. I do have a quick question on that. Is the 7th our pre-week of shows or that's our actual first live?

**Beth** (0:42)
I think it's our first live. Okay. Because we- I don't-
Yeah. Yeah.

**Brian** (0:48)
We talk about this a lot. We did a week of trial, which is, I'm sure, hilarious when you go back and listen to it. Jyunmi has said, threatened at times, maybe we'll release that. I said, hell, we will.
That will stay in a cold, dark cellar somewhere. But yeah, anyway, so that's what we have going on. It's Monday, which means all sorts of fun stuff has probably happened in AI over the weekend, because as we know by now, AI does not take the weekends off. Usually, by the time it comes around the Monday, there's well, two to three days worth of stuff that has been going on. And one of the ones, there's some bigger news out there, but one of the ones that caught my eye for sure was this, I saw it in Reuters, I saw it in the newsletters, Google limits Meta's use of its Gemini AI models. Basically, going back even a few months here, it looks like Google reached out to Meta in March, it said, this is Reuters saying this, and said it couldn't meet the full Gemini capacity the company had sought to purchase. What I find interesting about this is like Google is struggling to obviously meet the demands. We also know that they're working with, everybody's working with everybody. Meta's working with Google, Google, you know what I mean? Everybody's got hands in different pockets and stuff like that, but revenue from Google Cloud grew to 20 billion in the first quarter ending in March, but CEO, I need to say Sundar's last name, Pichai?
Anybody know?

**Andy** (2:17)
I'll go with that. Yeah.

**Brian** (2:19)
Pichai? Wow.
That's the first. Andy's like, yeah, you got it. He said, computer power constraints prevented even higher growth and contributed to Cloud units backlog, nearly doubling quarter on quarter. It's an interesting situation. Basically, you got Google Cloud can mint money as much as they want, if they can figure out the extreme demands on inference and everything else.
Other companies, I'm sure Meta can't be the only one that was affected, but of course, they're probably the biggest company that was affected. Meta has been said, but not verified Reuters is saying, but it's encouraged the staff to be more efficient with its AI tokens, which it says is even a measure of AI usage. What's interesting about that is we know at the beginning of the year, Meta was basically giving out bonuses and rewards for how many tokens people could use. So times have changed in six months because that only works if everybody is being super efficient with the tokens. I mean, on its face, it makes sense. If everybody is token maxing, what an awful term, token maxing as much as possible and doing as much of the efficient work, meaning that every token being spent is somehow positively impacting the business, then of course, sure, that makes sense or whatever, but we know that's not how tokens are used. There's a lot of trial and error. There's a lot of discovery. There's a ton of dead ends. There's a lot of oops, I shouldn't have done that and I paid myself into a quarter. Now I have to go back and all of that cost tokens. It's also the price of discovery and being in this, the edge of AI. We all are there. I've certainly blown through tokens and stuff. So now you have this same company who is basically saying he who uses the most he, she, right? Who uses the most tokens wins and gets bonuses for it too. Hey, maybe hold up a little bit on these set of tokens because Google can't meet what we're asking for. So kind of a big story. I think it just came out June 20th yesterday. This just came out yesterday. So I don't know. That was one of the ones I wanted to bring on a Monday.

54 more minutes of transcript below

Feed this to your agent

Try it now — copy, paste, done:

curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire.

Using your own key:

curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000774749770