AI is the new gold rush, but the money isn’t moving how most people think. In 2025 alone, OpenAI (creator of ChatGPT) and its partners have signed over $1 trillion worth of deals. The twist here is that OpenAI only actually made $4.3 billion in the first six months of the year, so where on earth is the money coming from?
Well, it never actually leaves the same circle of mega-companies. Nvidia, OpenAI, Oracle, AMD, Microsoft, SoftBank, Meta and Google are all trading billions between each other in a suspicious, questionably legal financial merry-go-round. The result is soaring valuations, daily headlines, and stock prices that climb on promises rather than profits. Is it sustainable? Is it legal? And what happens when the music eventually stops?
Let’s untangle the web. On paper, OpenAI and its partners have signed over $1 trillion in deals this year – more than most countries spend on their entire economies.
With a revenue of just $4 billion from January-June this year, the money must come from somewhere:
It doesn’t stop there: who else is involved?
Bottom line: billions in, billions out, but it all just circulates within the same pool.
Money going in circles is known as round-tripping. In simple terms, it’s when companies send money out the door through investments, purchases or partnerships, and then quietly receive it back through another channel. It makes activity and revenue look bigger than it really is.
For example:
In accounting, this is strongly frowned upon. Regulators have cracked down on it in the past, but in big tech today, it’s happening in front of everyone – except it’s dressed up as “strategic partnerships” and “capacity commitments”.
When banks or individuals do this, it’s treated as money laundering and fraud. If you transferred a friend £10,000 just for them to send it straight back to you, so you could both claim higher income, you’d be breaking the law.
But when trillion-dollar corporations do it through investment agreements and supply contracts, it’s technically legal – because it all has a business justification. On paper, at least.
Every deal triggers a rush of market optimism. AMD’s stock jumped over 30% after its OpenAI announcement, adding billions to its market cap. Oracle’s shares hit all-time highs and made its CEO the richest man in the world for a short period, following its $300 billion infrastructure pledge. Nvidia’s valuation tops $4.5 trillion on continued optimism, growing with each announcement of GPU orders.
These stock moves are not based on delivered profits or increased performance – but rather anticipation alone. Essentially, press releases have become currency.
Analysts are nervous for a reason: OpenAI is now the single point of failure in this trillion-dollar chain. Every data centre, chip order and supply agreement depends on OpenAI continuing to grow at an impossible pace.
By 2030, it’s estimated that the industry needs to make $2 trillion in revenue per year to sustain its infrastructure commitments – but there’s already an $800 billion funding gap. If the capital markets tighten or investors lose patience, the whole system could collapse overnight.
And because the same few mega-companies are financially intertwined, a stumble by just one of them could bring them all down.
The AI boom is real, but the money behind it isn’t. Trillions are being pledged by companies that have no cash, and no real new capital is being generated. It’s the same pile of money cycling between a handful of tech giants, each transfer making everyone look richer, stronger and more dominant than they really are.
Is it sustainable? Will they all deliver on their promises? What happens if just one of them slips up?
Is this smart capitalism in action, or a trillion-dollar bubble ready to burst? Why can giant companies use round-tripping, but nobody else? Do you think they’ll pull it off, and what happens if they don’t? Let us know below.
