$110 Billion Is Simply Not Enough For OpenAI
The cash-eating machine can't be satiated that easily.
OpenAI recently announced it is raising $110 billion, including $30bn from SoftBank, $30bn from Nvidia, and $50bn from Amazon. That is more than double the $40 billion they raised last year in the largest private tech deal on record. It also valued the company at an astonishing $840 billion! But, OpenAI has also seemingly reduced its spending targets, with investors being told that it is targeting around $600 billion in total compute spend by 2030. That is significantly less than half its previous $1.4 trillion commitment. So, with more money coming in and less going out, is this the turning point for OpenAI? Can this turn OpenAI from a cash black hole into something even remotely sustainable? In short, no, not at all. $110 billion isn’t even going to touch the sides. Let me explain why.
Let’s start with the fact that, by OpenAI’s own estimates, this isn’t really enough. Back in September 2025, OpenAI dramatically increased its projected cash burn, stating that it expects to rack up losses of $115 billion between now and 2029. That is $80 billion, or 44% higher than their previous projections! It also means this record-breaking investment will only keep the fires going for a few years.
But, there is something telling here. OpenAI posted a loss of $5 billion in 2024 and $8 billion in 2025, meaning its losses are growing 60% each year. If you assume losses keep growing at the same rate of 60% per year, then between 2026 and 2029, OpenAI would rack up $118 billion of losses, which is almost identical to OpenAI’s own estimates.
Why is that telling? Well, the entire AI industry operates on the idea that scaling AI will eventually lead to sustainability and profitability. In other words, the industry assumes AI revenue will scale faster than the cost to develop and run AI models, so eventually, at a certain scale, they will break even. This argument of scale to profitability is foundational to the current frenzied AI investment drive.
But OpenAI’s assumption that their losses will keep growing at the same rate, rather than at a slower rate as before, strongly implies they know the scale-to-profit idea simply isn’t real. They know that scaling this way will actually push them further away from break-even.
In other words, once this $110 billion is used up, they will need even more to keep the ash bonfire burning. But, more on that in a second.
Does reducing spending to $600 billion by 2030 make any of this better? At the very least, does it mean this $110 billion will last longer?
Take a wild guess!
OpenAI’s $1.4 trillion in commitments were for roughly 26 GW of compute power, suggesting this $600 billion is for roughly 11 GW of compute power by 2030. How much would that cost?
Well, as I discussed in my previous article, IBM CEO Arvind Krishna has stated that a single GW of AI-capable data centres costs $80 billion to build. However, we also know that a GW of AI data centres costs around $1.3 billion in energy costs each year and has a realistic operational lifespan of three years. The annual cost of a single GW of AI compute power (including the annual spread build cost and energy cost) is $27.97 billion.
OpenAI ended 2025 with 1.9 GW of compute, which will cost around $53 billion a year to operate! Let’s assume OpenAI adds 2.75 GW of compute power each year between now and 2030, for a total of 11 GW. That would add another $76 billion a year to their operational costs! As such, by the end of 2027, OpenAI would have blown through all of this $110 billion investment.
But it would also mean that by the end of 2028, OpenAI would have a $205 billion compute bill to pay. But, OpenAI itself only predicts that it will have $100 billion in revenue in 2028, and even that is highly optimistic. In other words, in just two years, OpenAI will have to raise this colossal amount of cash yet again just to avoid bankruptcy.
Now, these are very rough estimates. For example, they don’t consider how data centre operators soak up some of the data centre losses, or how some of OpenAI’s partners give it discounted compute. So this isn’t 100% accurate, but it gives a sense of the size of the hole OpenAI has dug itself, and how little this $110 billion will help it.
But, speaking of OpenAI’s “partners”, this $110 billion “investment” round has a huge cannibalistic problem to it.
You see, Amazon didn’t invest $50 billion into OpenAI. It invested $15 billion, with another $35 billion “in the coming months when certain conditions are met.” But this deal also locks OpenAI to buying 2 GW of Amazon-owned compute, which, as we have seen, will likely cost more than $50 billion. That sounds like a bit of a sour deal for OpenAI. It’s not so much of an investment as an extraction. Likewise, Nvidia’s investment in OpenAI is explicitly intended to lease chips from Nvidia.
So, OpenAI is selling off its share, its equity, to lease the very cloud computing it needs to operate. That is like chopping off and selling your body parts to buy food. Sure, it might work for a little while, but eventually, you won’t have any body parts left to chop off. OpenAI didn’t want to do it this way. It wanted to build and operate its own data centres, but struggled to raise the funds needed to do so. The only way it could raise enough was to enter these arguably predatory agreements.
And that is the huge issue here. The fact that OpenAI has to sell off such a large portion of itself to lease cloud computing from these very “investors” is terrifying. What kind of circular payday loan economics is that? It is not only a huge red flag that this is all a giant bubble, but it also leaves OpenAI high and dry. At some point, OpenAI won’t have enough equity to keep its cash bonfire going, and it will burn to the ground.
So, no. $110 billion is nowhere near enough, and lowering their projected expenditure to $600 billion by 2030 won’t make a difference. What’s more, how OpenAI raised this cash is a sign of just how desperate they are for funds, and just how much they have backed themselves into a corner. In my opinion, you would have to be blind to think any of this is “good news” for OpenAI.
Thanks for reading! Everything expressed in this article is my opinion, and should not be taken as financial advice or accusations. Don’t forget to check out my YouTubechannel for more from me, or Subscribe. Oh, and don’t forget to hit the share button below to get the word out!
This is so typical Sam Altman; he’s as big a con artist as Trump or Musk. This deal will keep Open AI going while he finalizes how he’ll guarantee his personal financial future. He knows the bubble is going to burst and his wealth will not all be in Open AI equities.