Amazon in Talks To Invest Up To $50 Billion in OpenAI

An anonymous reader shares a report: Amazon is in talks to invest up to $50 billion in OpenAI, according to people familiar with the matter, in what would be a giant bet on the hot AI startup. The ChatGPT maker is seeking up to $100 billion in new capital from investors, a round that could value it at as much as $830 billion, The Wall Street Journal previously reported.

Andy Jassy, Amazon's chief executive, is leading the negotiations with OpenAI CEO Sam Altman, according to some of the people. The exact shape of a deal, should one be reached, could still change, the people said. Investing tens of billions of dollars in OpenAI could make Amazon the biggest contributor in the AI company's ongoing fundraising round. SoftBank is in talks to invest up to $30 billion more in OpenAI as part of the round, adding to the Japanese conglomerate's already large stake in the startup.

Keep inflating that bubble

[SumDog]

Well I guess everyone is going to keep on inflating that bubble so it doesn't pop. I hate this shit. Let it burn so we can buy reasonable priced RAM again.

Re: Keep inflating that bubble

[liqu1d]

Doubt people will be buying much for a while if this pops. It's so intermingled with so many huge companies now there's no way to be safe from it popping.

Re:

[geekmux]

Doubt people will be buying much for a while if this pops. It's so intermingled with so many huge companies now there's no way to be safe from it popping.

Well, at least we'll know how the hell we managed to speedrun right past the Recession and straight into a full-blown Depression.

When a bubble that big pops, even hot-swapping to shitcoin mining won't save society.

Re: Keep inflating that bubble

[liqu1d]

Well at least the prices might come a bit then! Not that it'll do many any good reducing prices.

Re:

[gweihir]

That is "when" it pops. The profits they somehow would need to find to make this still work are completely outside of what is possible.

Re:

[gweihir]

Indeed. And the longer the bubble gets inflated, the harder it pops. "Smart" has left the building regarding AI quite a while ago. These people apparently really believe they can keep inflating things forever. That is not possible.

Rufus and ChatGPT ads

[mhocker]

Has anyone tried Amazon's pathetic chatbot, Rufus, which only exists because they block other chatbots from accessing Amazon's catalog? It's pathetic and frequently factually wrong on the only dataset which it should be an expert at - Amazon itself.

Given that ChatGPT is going to start advertising, this seems like a logical move. It gets rid of the Rufus problem and is in line with OpenAI's strategy.

My only question is whether this is another circular deal where no money is actually changing hands.

Re:

[timeOday]

Oh I hate the Rufus-suggested questions on product reviews so much.

"Is the item easy to use and reliable?"

Re:

[Tony Isaac]

Yep, and good luck asking it to narrow criteria for you because the regular search sucks so badly. Try prompts like "Exclude any products shipped from China" or "Include only hard-wired versions of this product." Their AI is just as bad as the regular search.

Re:

[gweihir]

Hate? No. Ignore because they are worthless? Yes.

Amazon review averages are meaningless. You need to look at a few really bad ones to find out whether the writers had a point or were too stupid to use the product right. That works with pretty good reliability.

Re:

[Wolfrider]

Yah, try asking Rufus just " TBW " on an SSD. About half the time there is no data.

I hope OpenAI retains some independence

[timeOday]

OpenAI has a piece to the puzzle but Oracle, Microsoft, Amazon etc have the apps so it makes sense partner. But lack of business pull is what makes ChatGPT good. Amazon made Alexa into a joke because it was driven solely by Amazon's business interests. Right now I pay OpenAI directly and it connects me straight to answers without having to go into the mess the Web has become (unless I want to check by clicking the links it gives as source), and I hope this remains an option.

Re:

[Tony Isaac]

Here's the thing about large financial investments.
- They are *never* without strings attached.
- The recipient will always cave, if the investment is big enough.

It's kind of like how universities put the names of contributors on buildings...if they don't, the benefactors wouldn't contribute. And the name is just the tip of the iceberg. With those large contributions, the benefactors stipulate all kinds of things that the university must do with the money. OpenAI is not immune to this kind of influence. He w

Re:

[allo]

Their main problem is Google. Google has money to burn, Google has users, and since some time Google has (one of) the best AI models. Without enough investments, Google may just squash them by burning more money than OpenAI can effort, refinancing it in the years after. On the other hand its a risky bet for Google, because Microsoft could take over the remains and have the users in its own ecosystem (remember how they force people to use Microsoft accounts now) learning to use MS AI.

Wonder

[liqu1d]

If this is going to be one of those cyclic deals where we invest $50billion in you and you buy $50billion in our compute platform. "Yeah we got a $50billion investment. Obviously we're valuable" "we just made another $50billion in sales" whilst both hoping no one looks too closely and just pumps the stock.

Re:

[gweihir]

Since that is how they keep inflating the bubble, very likely.

Also note that this practice is illegal in many countries, because it destroys the economy.

Electric bill's come due

[Pollux]

Daddy Bezos has to cut a check to keep the lights on.

I feel like my mind has just blown itself to pieces just now. How on God's Green Earth did humanity decide to value a company that lost $12 billion dollars in just Q3 of 2025 [wsj.com] at $830 billion dollars?

Re:

[allo]

Electricity is negligible at the moment while there are still huge hardware investments. Also all compute arguments play in favor to the AI companies, as models keep getting smarter at smaller sizes and newer hardware keeps getting more efficient (compute/watt or equivalently compute/dollar). Remember the huge GPT-3 model? A model of equivalent intelligence runs on a middle class smartphone by now. For AI, Moore's law is back into effect at the moment.

If you're interested, have a look at this semi-logarithm

Re:

[gweihir]

How on God's Green Earth did humanity decide to value a company that lost $12 billion dollars in just Q3 of 2025 [wsj.com] at $830 billion dollars?

Collective hallucination, belief that "AI" is the second coming and that finally we will be in a golden age and complete ignorance of tech history. The same as in the numerous other AI hypes we have had. This is by far the largest, but in stupidity and sheer ignorance concerning the facts, it is about as deranged and stupid as the other ones.

I really would not mind if these morons were only destroying their own wealth. But from the excessive size of things, this may well cause a long-lasting, global recessi

But they're NOT replacing workers with AI

[goldspider]

We're supposed to just take them at their word that this is unrelated to the white collar job cuts.

Re:

[radarskiy]

"We're supposed to just take them at their word that this is unrelated to the white collar job cuts."

The truth is even worse. They're cutting the jobs just to pay for the GPUs... they don't even have the AI to replace the workers yet.

OpenAI now mainly serves Microsoft's interests ?

[Mirnotoriety]

‘Musk co-founded OpenAI in 2015 as a nonprofit, but says the company now mainly serves Microsoft's interests [dw.com]. OpenAI has accused Musk of leading a campaign of "harrassment."’

Re:

[gweihir]

Assholes accuse assholes of being assholes. What else is new?

Re:

[Mirnotoriety]

What is a thought-terminating cliché?

its a circle jerk of money

[Growlley]

being passed around.

Crap!

[gweihir]

That means the deranged hype will go longer and crash harder.