And the return on that investment?

There isn’t one. Total industry revenue, according to the website isaiprofitable.com, is a mere US$613 billion. Chump change!

The website tracks the ongoing spend of all the major players in the market, from Amazon – the biggest spender (and loser), with an investment figure of US$313 billion – to smaller spenders such as Anthropic and xAI. Even China’s DeepSeek makes the cut.

And they’re all hemorrhaging money. NVIDIA is the only company in the black, which is hardly surprising, as it makes all the hardware the AI firms rely on. As it says on the site: “The AI economy is circular: Google funds Anthropic, Anthropic runs on Google Cloud, Amazon funds Anthropic, Microsoft co-invests with OpenAI. This means aggregate industry figures double-count some revenue flows. This site is one person’s best effort at an honest picture, not a financial audit.”

If you’re wondering, by now, the money spent while working on this article has hit more than US$51 million. While that number’s ticking up, we caught up with the site’s creator to ask him why it’s important to track these figures, even if – as the site says – they’re based on estimates and best guesses?

“Honestly, because no one else was doing it in one place, these companies are raising and spending at a historic scale, and the public deserves a transparent view of that,” the site’s creator – who wishes to remain anonymous – told Cyber Daily.

“Even though the numbers are estimates, they still convey the all-round idea of how the industry is doing and how close they are to being profitable. I’m not claiming these are audit-quality figures (I’m very upfront about that on the site), but the broad picture they paint is hard to argue with.”

The site is slickly designed, and the developer goes into a lot of detail about how they gather the data, but it’s watching that money spent figure climb that makes the most impact – and it’s based on real projections.

“Each company has a current annual burn rate. E.g. OpenAI is projected for US$14 billion in losses for 2026,” the developer said.

“I’ve divided that number by the number of seconds in a year (31,536,000) to get a $/second burned figure. Then the counter just ticks up in real time. People find the counter more visceral than a static bar chart, which was the reason I decided to go for it!”

The fact that almost no one is actually making money isn’t exactly a secret, and more than one observer has predicted that the bubble must eventually burst, but isaiprofitable’s developer thinks that the idea of AI either failing or succeeding is far too binary. As they point out, the dotcom bust is an obvious comparison, but it still left behind something of value – the internet as we know it and the infrastructure that it runs on.

“Even if several AI labs fail in the next few years, I think you’d still be left with massive datacentres, and fibre networks which still hold many use cases,” the developer said.

“I also think AI is already too embedded to disappear. It’s everywhere now: your search engine, emails, and company tools. Even if the bubble bursts, AI will survive and likely still be integrated into essentially everyone’s daily lives.”

However, when asked what they think the future holds, the developer is much more circumspect.

“Despite the site painting a pretty bad picture at the moment, I do think AI will be profitable for a handful of key players. For example, a week ago, WSJ reported Anthropic was about to have its first profitable quarter, although the jury is still out on whether or not this will hold,” the developer said.

“AI is still moving faster than anyone predicted, and I firmly believe we’ll have an answer as to whether the bubble will pop by 2030, as that is effectively the deadline the industry has set itself.”

Siri, remind me to check back in on isaiprofitable.com in 2030. In the meantime, as I’m about to hit publish on this article, the amount of cash splashed on AI spending while writing this article hit more than US$200 million.

Jensen Huang must be laughing all the way to the bank.