I think in the case of supply-constrained GPUs, you can get the opposite of a volume discount. Google has the most capacity of anyone, the fact they’re paying so much per month to spacex is pretty remarkable
Which means more Grok degradation, more severe throttling, etc.
I can't understand why xAI charges 50% more per month for Grok over competitors when it doesn't even gracefully downgrade to a cheaper model when paid subscribers hit the limit.
I've used/use it. For a while it had one of the best lightweight coding LLM's, which actually lead to a #1 spot on openrouter usage ranking although they've fallen off the top 10 used now. It's also provided some good reasoning models, which perform better when dealing with non-PC topics.
Also, although I've never used it for this, I believe some of the paid models produce some of the best "adult" content, and I know there are even subreddits which do nothing but praise Grok and "content" produces who use it.
They already make it required to have a premium X account
Claude has tons of throttling already. Chat GPT is not as accurate at computational problems despite less throttling. Gemini has fewest restrictions but worse quality. Always a tradeoff.
Probably because it has all but zero filters on the input and output. It took widespread media outrage about "grok show me her in a bikini" to at least create a filter that bans such things.
Tangent alert: a couple of questions for folks who know far more than I do about compute capacity and Google these days...
Lately, like the past few months, I've noticed Google services (search, gmail, drive, maps) running very slowly to the point where, at the moment it happens, I always think it has to be my connection and not Google, but sure enough every time I check a couple of speed tests and they're... fine. And then I don't seem to be having the same latency from other sites/apps. Is there any chance that the commingling of the AI snippet and then directing users into the AI funnel through the text box is actually causing material performance impacts in other Google properties? Probably a dumb question because I can't imagine they would allow performance for broader properties to suffer for AI prompts/chats, but then again all this talk of compute starts making me think otherwise, like the prolific amount of prompting and chatting is causing massive across-the-board performance issues.
Somewhat related, but does anyone use Gemini and end up with the experience where you have a chat and it's obvious, to yourself and to Gemini, that you're trying to find a product to purchase, but Gemini doesn't even link you to what you would think would be the obvious place to purchase the product? This happens daily where I interact with it, it suggests some products, but won't even provide a link to that product or, if it does provide a link, it's to some no name site that wouldn't come up as a highly-ranked paid or organic result through regular Google search. Keeps making me think this is a Google performance problem where they have not figured out how to take the entire AI chat and engineer it back into a simple short keyword phrase to get an acceptable search result.
Btw, if anyone's thinking "why are you using Gemini because it's the worst?" I think that's fair and right. I have... reasons, but they're not super sensible ones.
Maybe agentic PRs made it to production and performance cratered.
On a serious note I feel the same. Google is slow these days and the slowest and least reliable service of them all is Gemini. Sometimes I don't even know if it hang already (no error messages) or if it's still "thinking".
I'm probably 50/50 with search in particular logged in vs. out and I do think I notice on both, but I'm not entirely sure. Just saying the search and maps algorithm is wading through so much of my history that it can't help but choke trying to deliver the "right" results?
No, but I was thinking it might be possible that your account is specifically afflicted by some storage problem. For example, it could be homed in the wrong part of the world, compared to where your browser is hitting frontend applications. Or a million other possibilities. When logged out those wouldn't be factors.
> SpaceX said in the filing that if it fails to “deliver access to the committed amount of GPUs by September 30, 2026,” Google can immediately end the agreement, or accept the number of GPUs provided at a reduced fee after a one-month grace period.
> After this year, the agreement can be terminated by either party provided they give 90 days’ notice.
Circular financing at its peak for the IPO. There has to be some regulatory body to not allow such shady things
Circular financing would require SpaceX to buy a similar quantity of stuff from Google. (Or invest in Google.) We have no evidence of that. Instead this looks like Google taking advantage of SpaceX’s desire to print revenue today versus a month from now.
(If the agreement is terminated with no exchange of goods, it might be market manipulation. But still not circular financing.)
It's circular since Google owns part of SpaceX. According to [1] they own 7% of SpaceX, so a $1.75T IPO would value their stake at $120B. The target IPO price is >90x revenue, so if Google increases SpaceX's revenue by $11B, SpaceX's valuation could increase by $990B to maintain the same multiple, which would increase the value of Google's stake by $69B.
Yeah, I just looked this up because my first thought was another circular financing deal (or not circular by definition but certainly backscratching). Looks like Google's SpaceX stake, diluted, based on a cursory search, at a $1.5T valuation is somewhere in the $80-$100B neighborhood (bought back in 2015 I think is what it said when SpaceX was valued in the low tens of billions if I'm remembering correctly). So you have Google sending $12B back to SpaceX annually in this deal, so maybe 12% or so of their equity stake at that valuation. I'm not sure how to feel about it other than a means of swaying people to buy into the IPO with the added benefit of actual compute value.
And seems silly to ignore that the Google founders and Elon are buddies, or were, based on which gossip rag you believe in, and there's zero chance these types of deals are being made independent of those guys talking (when are they ever, of course, but it's even more obvious in this case given the players and their histories).
I avoided adding too many details, made a base assumption that folks on this topic would already be aware of google's investment in spaceX, probably should have added that too
Google is getting in bed with some folks even more unsavory than themselves. The thing I noted most from I/O was how prominent and proud the are of partnering with Palantir. "Do no evil" has become "Let's do evil"
I thought it was notable that in Google's press release yesterday regarding their new facilities near Amarillo they seemed to go out of their way to point out that the applications are not AI, listing "Search, Gmail, Maps, Cloud, online banking, and 911 systems" instead. I wonder if they find it more convenient to rent an existing one rather than face public scrutiny for building another "AI data center".
Perhaps. It also seems to insulate Google from the risk that air quality regulators will be unexpectedly reinvigorated, while still providing to Google the benefits of xAI's lawlessness.
> How did Elon get so much NVIDIA hardware before everyone else?
He’s the richest man on the planet and doesn’t have a track record of not paying for shit he buys. If you want to reliably offload your chips, he’s safer than e.g. OpenAI who might or might not have the money when the bill comes due.
> He’s the richest man on the planet and doesn’t have a track record of not paying for shit he buys.
My impression was the other way around. The shenanigans he pulled around the Twitter acquisition were just farcical, and at Twitter he repeatedly refused to pay owed rent, etc. (I assume as a ploy to renegotiate terms).
Not an easy thing to gauge. X/Twitter stopped publishing MAUs after the acquisition.
External estimates vary, some point to growth, some to stagnation. We know that revenue suffered. LOTS of partisan and emotional opinions for either.
Google trends does paint a bleak picture for X but I am also questioning how much google itself can gauge that after LLMs exploded in popularity.
Anecdotally I did notice that references and embeds to X are way less prominent and common than before. My usual news used to be filled with them. My consumption of the platform also plummeted after not being able to read threads when not logged, its much harder for me to get drawn into it.
Still without data, I would be surprised if the changes to verification and logged out access did not massively hurt new adoption. With tiktoks prevalence amongst a new generation I would bet its a matter of time before X gets grouped to tumbler/facebook, not dead, but way past its peak and cultural relevance. If that has not already happened.
On the one hand, yes, this is really embarrassing for Grok. On the other hand, everyone except a couple Twitter randos already thinks Grok is worthless junk: there's no reputation left for it to to lose. And Elon is crying all the way to the bank with an extra $11bn in annual revenue.
Is this another circular investment to help pump up the stock valuation? Why would Google need to rent that much compute?
> Google parent Alphabet has made a windfall from backing SpaceX, which was worth $12 billion at the time of its 2015 investment, and is looking to go public at a valuation of over $1.75 trillion
So space data centers are absolutely possible then. I heard a lot of skepticism about the feasibility but it looks like Google and Anthropic looked at SpaceX and trusted them to deliver on the promise and even signed deals worth billions.
google: we are commited to carbon free data centers by 2030 (https://sustainability.google/reports/247-carbon-free-energy...)
also google: renting capacity from a data center powered by 27 methane gas turbines on trailers
https://www.epw.senate.gov/public/index.cfm/2026/4/whitehous...
The deal ends in June 2029, which is before 2030.
So you mean to say Google knows AI definitely won't be needed after June 2029... Understood...
With ~$2.1B/month of GPU rental revenue, is xAI now profitable? Are all divisions of SpaceX now profitable?
Yes, but as Elon himself remarked, these are short term leases (90 day exit window), not guaranteed long term contracts.
Do people think those numbers are correct? 920 million a month for 110,000 GPUs is $11.61 per GPU hour. That seems very high to me.
I think in the case of supply-constrained GPUs, you can get the opposite of a volume discount. Google has the most capacity of anyone, the fact they’re paying so much per month to spacex is pretty remarkable
Which means more Grok degradation, more severe throttling, etc.
I can't understand why xAI charges 50% more per month for Grok over competitors when it doesn't even gracefully downgrade to a cheaper model when paid subscribers hit the limit.
Who uses Grok? Not even SpaceX engineers that is known...
Tesla drivers, at least for a few minutes each day before hitting the limit.
I've used/use it. For a while it had one of the best lightweight coding LLM's, which actually lead to a #1 spot on openrouter usage ranking although they've fallen off the top 10 used now. It's also provided some good reasoning models, which perform better when dealing with non-PC topics.
Also, although I've never used it for this, I believe some of the paid models produce some of the best "adult" content, and I know there are even subreddits which do nothing but praise Grok and "content" produces who use it.
> which actually lead to a #1 spot on openrouter usage
that was only because it was free
+1. The coding model was fine and it was fast but the fact that it was free was a massive boost.
> reasoning ... with non-PC topics
do you have examples? Do you ask "reason why women belong in the kitchen" or what?
They already make it required to have a premium X account
Claude has tons of throttling already. Chat GPT is not as accurate at computational problems despite less throttling. Gemini has fewest restrictions but worse quality. Always a tradeoff.
Why would you even want to use that dumpster fire of a model in the first place..?
Probably because it has all but zero filters on the input and output. It took widespread media outrage about "grok show me her in a bikini" to at least create a filter that bans such things.
[1] https://counterhate.com/research/grok-floods-x-with-sexualiz...
Does the thing where it wanted to move every conversation onto the topic of "white genocide in South Africa" not count as a filter on the output?
Tangent alert: a couple of questions for folks who know far more than I do about compute capacity and Google these days...
Lately, like the past few months, I've noticed Google services (search, gmail, drive, maps) running very slowly to the point where, at the moment it happens, I always think it has to be my connection and not Google, but sure enough every time I check a couple of speed tests and they're... fine. And then I don't seem to be having the same latency from other sites/apps. Is there any chance that the commingling of the AI snippet and then directing users into the AI funnel through the text box is actually causing material performance impacts in other Google properties? Probably a dumb question because I can't imagine they would allow performance for broader properties to suffer for AI prompts/chats, but then again all this talk of compute starts making me think otherwise, like the prolific amount of prompting and chatting is causing massive across-the-board performance issues.
Somewhat related, but does anyone use Gemini and end up with the experience where you have a chat and it's obvious, to yourself and to Gemini, that you're trying to find a product to purchase, but Gemini doesn't even link you to what you would think would be the obvious place to purchase the product? This happens daily where I interact with it, it suggests some products, but won't even provide a link to that product or, if it does provide a link, it's to some no name site that wouldn't come up as a highly-ranked paid or organic result through regular Google search. Keeps making me think this is a Google performance problem where they have not figured out how to take the entire AI chat and engineer it back into a simple short keyword phrase to get an acceptable search result.
Btw, if anyone's thinking "why are you using Gemini because it's the worst?" I think that's fair and right. I have... reasons, but they're not super sensible ones.
Maybe agentic PRs made it to production and performance cratered.
On a serious note I feel the same. Google is slow these days and the slowest and least reliable service of them all is Gemini. Sometimes I don't even know if it hang already (no error messages) or if it's still "thinking".
Do you still perceive search and maps to be slow even when logged out?
I'm probably 50/50 with search in particular logged in vs. out and I do think I notice on both, but I'm not entirely sure. Just saying the search and maps algorithm is wading through so much of my history that it can't help but choke trying to deliver the "right" results?
No, but I was thinking it might be possible that your account is specifically afflicted by some storage problem. For example, it could be homed in the wrong part of the world, compared to where your browser is hitting frontend applications. Or a million other possibilities. When logged out those wouldn't be factors.
> SpaceX said in the filing that if it fails to “deliver access to the committed amount of GPUs by September 30, 2026,” Google can immediately end the agreement, or accept the number of GPUs provided at a reduced fee after a one-month grace period.
> After this year, the agreement can be terminated by either party provided they give 90 days’ notice.
Circular financing at its peak for the IPO. There has to be some regulatory body to not allow such shady things
> Circular financing
Circular financing would require SpaceX to buy a similar quantity of stuff from Google. (Or invest in Google.) We have no evidence of that. Instead this looks like Google taking advantage of SpaceX’s desire to print revenue today versus a month from now.
(If the agreement is terminated with no exchange of goods, it might be market manipulation. But still not circular financing.)
It's circular since Google owns part of SpaceX. According to [1] they own 7% of SpaceX, so a $1.75T IPO would value their stake at $120B. The target IPO price is >90x revenue, so if Google increases SpaceX's revenue by $11B, SpaceX's valuation could increase by $990B to maintain the same multiple, which would increase the value of Google's stake by $69B.
1: https://finance.yahoo.com/markets/stocks/articles/alphabet-s...
Their exposure is more than just their ownership of SpaceX.
If the SpaceX IPO bombs (or even merely underperforms), the expectations for the Anthropic/OpenAI IPOs collapse, and with that, everything else AI.
AI companies can't afford to let any AI company go down.
> Alphabet has made a windfall from backing SpaceX. Musk’s company was worth $12 billion at the time of Google’s 2015 investment
How come its not a circular deal where google is investing little bit more money to make a whole lot more money
that's not circular though
> Circular financing
Keep in mind, Google has a 6% stake in SpaceX, so this is more like exchanging millions to gain billions.
Yeah, I just looked this up because my first thought was another circular financing deal (or not circular by definition but certainly backscratching). Looks like Google's SpaceX stake, diluted, based on a cursory search, at a $1.5T valuation is somewhere in the $80-$100B neighborhood (bought back in 2015 I think is what it said when SpaceX was valued in the low tens of billions if I'm remembering correctly). So you have Google sending $12B back to SpaceX annually in this deal, so maybe 12% or so of their equity stake at that valuation. I'm not sure how to feel about it other than a means of swaying people to buy into the IPO with the added benefit of actual compute value.
And seems silly to ignore that the Google founders and Elon are buddies, or were, based on which gossip rag you believe in, and there's zero chance these types of deals are being made independent of those guys talking (when are they ever, of course, but it's even more obvious in this case given the players and their histories).
There is nothing shady or circular in the text you quoted
I avoided adding too many details, made a base assumption that folks on this topic would already be aware of google's investment in spaceX, probably should have added that too
Google is getting in bed with some folks even more unsavory than themselves. The thing I noted most from I/O was how prominent and proud the are of partnering with Palantir. "Do no evil" has become "Let's do evil"
Space-X is an AI/datacenter company that also makes rockets
[delayed]
I thought it was notable that in Google's press release yesterday regarding their new facilities near Amarillo they seemed to go out of their way to point out that the applications are not AI, listing "Search, Gmail, Maps, Cloud, online banking, and 911 systems" instead. I wonder if they find it more convenient to rent an existing one rather than face public scrutiny for building another "AI data center".
I wonder if they are more than happy to let someone else take on the burden of the massive writedowns that are bound to hit in a couple of years.
Perhaps. It also seems to insulate Google from the risk that air quality regulators will be unexpectedly reinvigorated, while still providing to Google the benefits of xAI's lawlessness.
How did Elon get so much NVIDIA hardware before everyone else?
> How did Elon get so much NVIDIA hardware before everyone else?
He’s the richest man on the planet and doesn’t have a track record of not paying for shit he buys. If you want to reliably offload your chips, he’s safer than e.g. OpenAI who might or might not have the money when the bill comes due.
tesla not paying bills: https://www.cnn.com/2025/07/31/us/elon-musk-company-unpaid-l...
x not paying bills: https://www.cnbc.com/2023/02/24/musks-twitter-has-been-sued-...
spacex not paying bills: https://www.fastcompany.com/91124157/spacex-contractors-texa...
It has been leaked the have a huge open bill with AWS....
> He’s the richest man on the planet and doesn’t have a track record of not paying for shit he buys.
My impression was the other way around. The shenanigans he pulled around the Twitter acquisition were just farcical, and at Twitter he repeatedly refused to pay owed rent, etc. (I assume as a ploy to renegotiate terms).
I've heard the harder part is to have data centers to put the nvidia hardware than getting the hardware currently.
And the secret ingridient to building those faster than others is... crime.
https://www.theguardian.com/technology/2026/jan/15/elon-musk...
Because he went all in with money before a lot of other people.
He moved as fast as he could with known financing
Related / same story from different source (Bloomberg), submitted earlier: https://news.ycombinator.com/item?id=48416941
Another way to read this is, "Twitter's AI dept. has been doing so badly they don't have a use for billions of dollars of hardware they bought."
The downfall of Twitter is one of the most spectacular self inflicted business failures I have ever witnessed.
Amazing to think it was once so ingrained in mainstream society.
What metrics are you using to classify it as a downfall?
I don't care either way but I got curious.
Not an easy thing to gauge. X/Twitter stopped publishing MAUs after the acquisition.
External estimates vary, some point to growth, some to stagnation. We know that revenue suffered. LOTS of partisan and emotional opinions for either.
Google trends does paint a bleak picture for X but I am also questioning how much google itself can gauge that after LLMs exploded in popularity.
Anecdotally I did notice that references and embeds to X are way less prominent and common than before. My usual news used to be filled with them. My consumption of the platform also plummeted after not being able to read threads when not logged, its much harder for me to get drawn into it.
Still without data, I would be surprised if the changes to verification and logged out access did not massively hurt new adoption. With tiktoks prevalence amongst a new generation I would bet its a matter of time before X gets grouped to tumbler/facebook, not dead, but way past its peak and cultural relevance. If that has not already happened.
On the one hand, yes, this is really embarrassing for Grok. On the other hand, everyone except a couple Twitter randos already thinks Grok is worthless junk: there's no reputation left for it to to lose. And Elon is crying all the way to the bank with an extra $11bn in annual revenue.
Is this another circular investment to help pump up the stock valuation? Why would Google need to rent that much compute?
> Google parent Alphabet has made a windfall from backing SpaceX, which was worth $12 billion at the time of its 2015 investment, and is looking to go public at a valuation of over $1.75 trillion
antigravity is a token hog
So space data centers are absolutely possible then. I heard a lot of skepticism about the feasibility but it looks like Google and Anthropic looked at SpaceX and trusted them to deliver on the promise and even signed deals worth billions.
They are paying for the existing earth-bound data centers SpaceXai owns.
Oh wow did not realize xai has data centers. So are they completely abandoning xai or do they just have that much capacity left over?