The shortage prevented Google from offering the model at scale.
Google reportedly limited Meta’s use of Gemini due to a shortage of compute resources. — FT
Google is in a position where it can’t sell Gemini to Meta as freely as it might want to.
Compute remains power, and the scarcest resource in AI.
Negative users criticize Gemini's poor quality and express frustration at Google's compute limits on Meta, while positive users praise Nvidia efficiency and local models from Google and Apple.
No Digg Deeper questions have been answered for this story yet.
wild.
Google reportedly limited Meta’s use of Gemini due to a shortage of compute resources. — FT
Google is in a position where it can’t sell Gemini to Meta as freely as it might want to.
Compute remains power, and the scarcest resource in AI.
I tried out Gemini for automatic PR reviews on one of our repos. We used it for a few months alongside other LLM reviewers to compare, but its reviews were so bad and just...dumb, that i recently asked to turn Gemini PR reviewer off again for that repo.
Over the whole duration, it made ONE genuinely useful comment; ironically on the exact day that I asked to turn it off lol
@jukan05 Google has a huge cloud business. How’s the fuck is this possible?
@jukan05 Why are they even using Gemini its the worst model ever
@jukan05 @zephyr_z9 Why the hell meta using so much Gemini tho
@HenrySidebiz_2 @jukan05 Demand exceeds supply.
Pichai in Q1 2026 earnings call: “We are compute constrained in the near term. Our cloud revenue would have been higher if we were able to meet the demand.”
They are investing heavily to increase supply, but they expect demand to grow even faster.
This is the whole AI trade in one headline. The scarcest layer wins, and it isn't the models. The twist: it isn't even the famous chip names, half of them are lagging. The real leaders in our rankings sit one layer down, in the gear and materials feeding the build. $AMAT making the machines, $GLW making the glass that wires the data centers, both breaking out while the household names lag. Own the bottleneck, not the brand.
@esper1971 @jukan05 스엑스아이는 이미 거의 다 팔아버렸죠…
@jukan05 Compute is the new oil and Google just turned off Meta’s tap.
@jukan05 My strategy analysis !
🔽as follow
@jukan05 아직 유휴 Computing 자원이 있는 SpaceX에 호재가 될까요?
@jukan05 This is why $TSLA is a sleeping giant.
Every car.
Every Optimus…
Shared compute. And they have the network too. And direct to consumers.
This is why all the big companies are railing against Elon. •
@jukan05 The scarcest resource is no longer talent or models—it's compute.
@jukan05 $NBIS still has an incoming 1.2gw data center.. announced during May's ER and currently waiting for more details on whether they will sign on any hyperscaler deals or will $meta upsize on their existing needs
@HenrySidebiz_2 @jukan05 Underinvestment, like with electricity. If Meta had built electricity stations instead of the metaverse they'd be worth more than nvidia lol
@jukan05 算力短缺说明 AI 竞争已经从模型口号变成资源战争。谁能稳定拿到芯片、电力和数据中心，谁才有资格把产品规模化卖出去。
@jukan05 It’s compute v the end use battle for now
@jukan05 AI's biggest bottleneck isn't models it's compute.
@jukan05 Details of my analysis.
👇Next steps
@jukan05 Power bottlenecked.
