Google plans to invest up to $40B in Anthropic • Hacker News Board

Re: Google plans to invest up to $40B in Anthropic

#531

Post by darkwater » Sat, Apr 25, 2026, 7:15 AM UTC

fc417fc802 wrote: post #476

hattmall wrote: post #456

Earlier quoted context omitted.

That seems really paradoxical and I think it would just burn up compute. The AI really doesn't have any way to know it's getting better without humans telling. As soon as the AI begins to recursively improve based on its own definition of improvement model collapse seems unavoidable.

If humans are able to judge, and if the AI is more capable than a human in every respect, then why can't the AI be the judge of its own performance? Humans judge their own output all the time.

The difference IMO is that every single human is a slightly different model, not the same one with a different prompt, or weights.

Re: Google plans to invest up to $40B in Anthropic

#532

Post by GiselW » Sat, Apr 25, 2026, 7:17 AM UTC

[dead]

Re: Google plans to invest up to $40B in Anthropic

#533

Post by throwaway2037 » Sat, Apr 25, 2026, 7:18 AM UTC

zymhan wrote: post #307

skybrian wrote: post #276

Context: a few weeks ago, Anthropic signed a deal to buy "multiple gigawatts of next-generation TPU capacity" from Google and Broadcom [1]. There have been several previous deals, too. Some people call this sort of thing a "circular deal", but perhaps a better way to think of it is as a very large-scale version of vendor financing? The simple version of vendor financing is when a vendor gives a retailer time to pay f…

To be honest, I think "vendor financing" is still a very risky premise. Vendors may be positioned to know how a customer is doing, but they're also incentivized to overestimate how well a customer is going to perform. GE Capital (edit: and GMCA) is a great example of how seemingly reasonable vendor financing can cause the lender serious problems.

    > To be honest, I think "vendor financing" is still a very risky premise.

Are you aware that all heavy industry in all highly developed nations make extensive use of vendor financing to sell their products? Siemens is a perfect example of a well-run, stable, industrial giant. They offer vendor financing for large purchases. Same for the "heavies" (Mitsubishi, Kawasaki, IHI, Hyundai, Doosan, Hanjin) in Japan and Korea.

If anyone is interested to learn about the damage that the financialisation of General Electric (USA) brought upon itself, you can ask ChatGPT to tell you the story. It is too long to repeat here.

Here is a sample prompt that I used to remind myself:

    > I am interested in the history of General Electric and the trouble that their financing units brought in the early to mid 2000s. Can you tell me more?

Re: Google plans to invest up to $40B in Anthropic

#534

Post by netcan » Sat, Apr 25, 2026, 7:23 AM UTC

skybrian wrote: post #276

Context: a few weeks ago, Anthropic signed a deal to buy "multiple gigawatts of next-generation TPU capacity" from Google and Broadcom [1]. There have been several previous deals, too. Some people call this sort of thing a "circular deal", but perhaps a better way to think of it is as a very large-scale version of vendor financing? The simple version of vendor financing is when a vendor gives a retailer time to pay f…

So yes, but that doesn't negate the circular investment aspect, for most intents and purposes.

The risk is from this structure is mostly to do with how this affects market cap. Companies using the value of their shares to fund demand for their services.

That's a risk.

Re: Google plans to invest up to $40B in Anthropic

#535

Post by throwaway2037 » Sat, Apr 25, 2026, 7:24 AM UTC

Spooky23 wrote: post #459

zymhan wrote: post #307

Earlier quoted context omitted.

To be honest, I think "vendor financing" is still a very risky premise. Vendors may be positioned to know how a customer is doing, but they're also incentivized to overestimate how well a customer is going to perform. GE Capital (edit: and GMCA) is a great example of how seemingly reasonable vendor financing can cause the lender serious problems.

GE Capital was a different creature, riding the line of fraud in some ways. They misapplied accounting rules and had to write down or capitalize over $20B for long term care insurance.

I don't know the full the history of this story, but I honestly wonder if type of scandal is still possible in the United States. After Enron and Worldcomm, the US introduced Sarbanes-Oxley reporting regulations. Additionally, after the Global Financial Crisis of 2008/2009, there was a dramatic increase in regulations for banks (of all kinds) and insurance companies.

Re: Google plans to invest up to $40B in Anthropic

#536

Post by aurareturn » Sat, Apr 25, 2026, 7:26 AM UTC

jayd16 wrote: post #400

Petersipoi wrote: post #318

Earlier quoted context omitted.

I think the opposite. AI will get cheaper as models become more efficient and we solve the datacenter/energy problem. I bet 10 years from now AI, that is way better than what we have today, will be close to free.

Just like how cloud costs got cheaper and we solved the datacenter/energy problem over the past 10 years.

You can practically host a website that serves millions of users a day for nearly free using Cloudflare. Imagine doing that in the year 2000.

Re: Google plans to invest up to $40B in Anthropic

#537

Post by throwaway2037 » Sat, Apr 25, 2026, 7:27 AM UTC

everly wrote: post #511

zymhan wrote: post #307

Earlier quoted context omitted.

To be honest, I think "vendor financing" is still a very risky premise. Vendors may be positioned to know how a customer is doing, but they're also incentivized to overestimate how well a customer is going to perform. GE Capital (edit: and GMCA) is a great example of how seemingly reasonable vendor financing can cause the lender serious problems.

$40 billion is about a quarter’s worth of profits for Google. They make that much every 3 months, what’s the risk

Hat tip. Great point. To quote J Paul Getty: "If you owe the bank $100, that's your problem. If you owe the bank $100 million, that's the bank's problem." In this case, yes, the investment is large, but not bankrupting for Google if it goes wrong.

Re: Google plans to invest up to $40B in Anthropic

#538

Post by fc417fc802 » Sat, Apr 25, 2026, 7:27 AM UTC

Aeolun wrote: post #514

fc417fc802 wrote: post #387

Earlier quoted context omitted.

I mean as in living through the industrial revolution would have been wild. So whether we have an AI revolution or an AI bubble it's bound to be a roller coaster. And that's without accounting for the various wars (and resultant economic impacts) that are already in progress. A large part of what drove the meat grinder of WWI was (very approximately) the various actors repeatedly misjudging the overall situation and…

> minimize loss of life this time around is (ironically enough) the rise of autonomous weapons systems I think that just means the relative civilian loss of life will increase once again.

What strategic merit is there in targeting civilians or life critical infrastructure in a fully automated battlebot scenario? Perhaps it's naive but I would expect stockpiles, datacenters, and any key infrastructure on which the local semiconductor fabrication depends to be the primary targets.

Re: Google plans to invest up to $40B in Anthropic

#539

Post by aurareturn » Sat, Apr 25, 2026, 7:27 AM UTC

sethops1 wrote: post #301

blueblisters wrote: post #218

Earlier quoted context omitted.

I feel the same until I’m reminded I’m paying Anthropic $100 every month for something that’s indispensable to me now and would probably pay a lot more. Very inelastic demand as long as competition is low at the frontier.

I pay TMobile $100 a month but they aren't worth a trillion dollars.

Yes but Tmobile enterprise customers don't pay much more for plans. In fact, they may pay less because of volume.

However, Anthropic can and will charge much more for enterprise customers.

Re: Google plans to invest up to $40B in Anthropic

#540

Post by TeMPOraL » Sat, Apr 25, 2026, 7:28 AM UTC

fc417fc802 wrote: post #476

hattmall wrote: post #456

Earlier quoted context omitted.

That seems really paradoxical and I think it would just burn up compute. The AI really doesn't have any way to know it's getting better without humans telling. As soon as the AI begins to recursively improve based on its own definition of improvement model collapse seems unavoidable.

If humans are able to judge, and if the AI is more capable than a human in every respect, then why can't the AI be the judge of its own performance? Humans judge their own output all the time.

Humans ultimately judge their output by comparison and competition. When we get to the point an AI is capable of participating on the market directly, it'll no longer make sense to proxy judgement through humans anymore.