Live data from Hacker News

Google to pay SpaceX $920M a month for compute capacity at xAI data centers

cnbc.com

991–1000 of 1001 posts

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#991

Earlier quoted context omitted.

You said Nazism is being accepted like other political movements. But nobody (in this thread) ever said something like that.

From my and seemingly others’ reading, you have equated Nazi and woke very directly in your comments. Do you feel you have you not?

I haven't equated them, I've used them as examples of opposite extremes.

I think Nazism is extreme right, and wokeism is extreme left. It's both out of balance. And Google and xAI both acknowledged that, and addressed the problems.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#992

Earlier quoted context omitted.

> Why do you think "Semi is DoA"? They started producing and selling the Semi in 2022 (after its unveiling in 2017, when they started taking pre-orders) and from everything I've dug up with a bit of Googling it seems they have shipped fewer than 200 trucks by 2025. We'll see if this new 50k per year factory will actually have customers to ship to, but I wouldn't hold my breath given the current track record. > The ec…

No, the early units from 2022 were essentially beta testing for both Tesla and their early customers (Pepsi, etc.). Wiki says: https://en.wikipedia.org/wiki/Tesla_Semi > Volume production of the Semi started on April 29, 2026. Note volume in that statement. You wrote: "If the battery is too heavy". The 2026 version of Tesla Semi is 450kg lighter than 2022 model because they switched the internal voltage from 12W to 4…

Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. As ever, the future that Tesla paints is extremely rosy, and suggests we should disregard what has happened so far. There is nothing whatsoever to indicate that Semi will actually work to the extent advertised and actually be desired by anyone - especially in the current anti-green climate in the USA, with no subsidies for electrification of the kind Pepsi used to buy the tiny pilot program.

Note that they never announced that the original run of the Semi would be just tiny. When they unveiled it in 2022, they explicitly said that this was the production version, as opposed to the 2017 concept. They even had a few more (still small 100-200 count) contracts where they kept delaying because they couldn't deliver enough - again suggesting that they were having problems, not intentionally running a pilot program.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#993

Earlier quoted context omitted.

No, the early units from 2022 were essentially beta testing for both Tesla and their early customers (Pepsi, etc.). Wiki says: https://en.wikipedia.org/wiki/Tesla_Semi > Volume production of the Semi started on April 29, 2026. Note volume in that statement. You wrote: "If the battery is too heavy". The 2026 version of Tesla Semi is 450kg lighter than 2022 model because they switched the internal voltage from 12W to 4…

Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. As ever, the future that Tesla paints is extremely rosy, and suggests we should disregard what has happened so far. There is nothing whatsoever to indicate that Semi will actually work to the extent advertised and actually be desired by anyone - especially in the current anti-green climate in the…

     > Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular.
I hate asking this question: "Sources?" If this was true, why does PepsiCo/Frito Lay continue to use Tesla Semi heavy haul electic trucks?

    > no subsidies for electrification
This is factually incorrect. California has a massive subsidy programme for electric trucks -- as I understand, the highest/largest for any state in the United States.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#994

Earlier quoted context omitted.

Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. As ever, the future that Tesla paints is extremely rosy, and suggests we should disregard what has happened so far. There is nothing whatsoever to indicate that Semi will actually work to the extent advertised and actually be desired by anyone - especially in the current anti-green climate in the…

> Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. I hate asking this question: "Sources?" If this was true, why does PepsiCo/Frito Lay continue to use Tesla Semi heavy haul electic trucks? > no subsidies for electrification This is factually incorrect. California has a massive subsidy programme for electric trucks -- as I understand, the highes…

It's likely not true.

We're certainly not as far along with the electrification of heavy duty trucks as we are with light duty cars, but the Semi seems fairly popular where it's suitable.

https://x.com/Tesla/status/1598490490613432321

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#995
post #939

Earlier quoted context omitted.

And if you look at the composition of the indices and ETFs you realize that you aren't participating in the innovative China, but get typical developing country stocks and very limited exposure to innovation. Also -10% over the last 5 years vs. +103% for the S&P500

> the innovative China Can you give some example listed stocks that you consider (1) innovative and (2) not a member of CSI 300?

Well, I mean Alibaba, Tencent, Xiaomi, PDD/Temu, Baidu, Bilibili, NIO, XPeng...

These aren't in the CSI 300 and the legal constructs of listing them via HK or other stock exchanges is unfortunately often questionable.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#996

Earlier quoted context omitted.

Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. As ever, the future that Tesla paints is extremely rosy, and suggests we should disregard what has happened so far. There is nothing whatsoever to indicate that Semi will actually work to the extent advertised and actually be desired by anyone - especially in the current anti-green climate in the…

> Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. I hate asking this question: "Sources?" If this was true, why does PepsiCo/Frito Lay continue to use Tesla Semi heavy haul electic trucks? > no subsidies for electrification This is factually incorrect. California has a massive subsidy programme for electric trucks -- as I understand, the highes…

They are operating 100 trucks that they bought with subsidy money - out of probably 10k trucks or more that they use in the USA. I'm not claiming the Semi is completely non-functional, it's obviously a real working vehicle. But this doesn't prove in any way that the Semi is actually as cheap and reliable as it was advertised as - if it were, why didn't Pepsi order far, far more?

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#997

Earlier quoted context omitted.

> Everything that has actually happened so far with the Semi is that it didn't work as advertised and was deeply unpopular. I hate asking this question: "Sources?" If this was true, why does PepsiCo/Frito Lay continue to use Tesla Semi heavy haul electic trucks? > no subsidies for electrification This is factually incorrect. California has a massive subsidy programme for electric trucks -- as I understand, the highes…

It's likely not true. We're certainly not as far along with the electrification of heavy duty trucks as we are with light duty cars, but the Semi seems fairly popular where it's suitable. https://x.com/Tesla/status/1598490490613432321

> but the Semi seems fairly popular where it's suitable.

What are you basing this on? Again, they have sold 200 trucks in 4 years. There are some hundreds of thousands of trucks being used in the USA alone. Tesla themselves are claiming they are going to produce (and presumably try to sell) 50k trucks per year. So, by any possible measure so far, Semi has basically 0 adoption. Maybe this is strictly based on production issues and there is huge un serviced demand - I admit this is a possibility, theoretically. But I don't see any reason to actually believe it, and certainly neither the current sales, nor the link you provided, in any way show that this demand will materialize. We'll see soon enough, I guess.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#998

Earlier quoted context omitted.

It's likely not true. We're certainly not as far along with the electrification of heavy duty trucks as we are with light duty cars, but the Semi seems fairly popular where it's suitable. https://x.com/Tesla/status/1598490490613432321

> but the Semi seems fairly popular where it's suitable. What are you basing this on? Again, they have sold 200 trucks in 4 years. There are some hundreds of thousands of trucks being used in the USA alone. Tesla themselves are claiming they are going to produce (and presumably try to sell) 50k trucks per year. So, by any possible measure so far, Semi has basically 0 adoption. Maybe this is strictly based on producti…

One company ordered 370 trucks last month.

https://www.teslarati.com/tesla-massive-order-semi-370-units...

2026 production is estimated to be 5k-15k.

https://sherwood.news/tech/what-we-know-about-teslas-semi-tr...

IIRC, 50k is full production, but it'll take a year or few to ramp up to that.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#999

Earlier quoted context omitted.

> but the Semi seems fairly popular where it's suitable. What are you basing this on? Again, they have sold 200 trucks in 4 years. There are some hundreds of thousands of trucks being used in the USA alone. Tesla themselves are claiming they are going to produce (and presumably try to sell) 50k trucks per year. So, by any possible measure so far, Semi has basically 0 adoption. Maybe this is strictly based on producti…

One company ordered 370 trucks last month. https://www.teslarati.com/tesla-massive-order-semi-370-units... 2026 production is estimated to be 5k-15k. https://sherwood.news/tech/what-we-know-about-teslas-semi-tr... IIRC, 50k is full production, but it'll take a year or few to ramp up to that.

They claim 5-15k trucks produced in 2026, but the company who apparently ordered those 370 will only get 50 in 2026, per the article you shared. The numbers are not adding up.

I think all signs keep pointing to another huge flop by the end of the year, and that Semi will remain a small niche product that various companies order a handful of and never go on to get more.

370 trucks is also still a tiny order, for a truck that supposedly is cheaper to operate than any other on the market, in the midst of an oil crisis no less.

Re: Google to pay SpaceX $920M a month for compute capacity at xAI data centers

#1000
post #995

Earlier quoted context omitted.

> the innovative China Can you give some example listed stocks that you consider (1) innovative and (2) not a member of CSI 300?

Well, I mean Alibaba, Tencent, Xiaomi, PDD/Temu, Baidu, Bilibili, NIO, XPeng... These aren't in the CSI 300 and the legal constructs of listing them via HK or other stock exchanges is unfortunately often questionable.

All good points. So why not just buy Hang Seng Index? Most of the growth in HSI comes from these innovative Mainland Chinese companies. You don't need to worry about CSI 300 and all of the silliness around CNH current!
Post reply on HN