TECH NEWS – According to one market analyst, it’s four times larger than the risky mortgage bubble that caused the 2008 financial crisis.

The Associated Press recently reported that OpenAI’s valuation reached $500 billion, making this unprofitable company the most valuable startup in history. One market analyst believes this madness has gone too far, warning that companies and their investors will soon face diminishing returns. He advises clients to stay away from them. So far, investors have bought into the hype, and even governments are rushing to join the AI revolution. Huge sums of money are at stake in AI, delivering transformative results in the near future.

So much money, in fact, that if the bubble bursts, the economic consequences could be brutal and define entire eras. Even financial professionals are beginning to suspect that something is wrong. A new client letter by Julien Garran of the independent research firm Macrostrategy Partnership echoes this concern — and the firm has long taken a conservative stance on AI.

Garran argues that the AI bubble is 17 times larger than the dot-com bubble and four times the size of the subprime mortgage bubble that triggered the 2008 global crisis. Artificially low interest rates have led to misallocation — in economic terms, money and labor are being directed to the wrong places, destabilizing the situation as production, products, or promises fail to materialize. Garran used Wicksell’s differential model to calculate the GDP gap, which includes AI, real estate, venture capital, and NFTs. Based on this metric, resource misallocation before the 2008 crash accounted for around 18% of GDP. Garran estimates that figure could now be an astounding 65%.

Garran illustrated AI’s productivity performance with real-world examples. He cited a study showing that at a software company, AI’s task completion rate ranged from 1.5% to 34%. Even in tasks where AI outperformed humans, it failed to replicate this success consistently over time. Another economist, using data from the Department of Commerce, created a table showing that AI adoption among large companies is declining.

“We don’t know exactly when LLMs will reach the point of diminishing returns because we lack tools to measure the statistical complexity of language. To find out, we need to watch LLM developers closely. If they release a model that costs ten times more, likely requires twenty times the computing power, and isn’t much better than existing ones, then we’ve hit the wall,” says Garran.

Garran also noted that the heaviest LLM users consume more computing resources than their monthly subscriptions cover. He believes that once the bubble bursts, the economy could enter a Zone 4 deflationary crisis during our investment cycles.

Source: PCGamer, AP, MarketWatch, MarketWatch Live, Wikipedia