David Roche predicts a controlled global market correction within 6 months

In an interview with ET Now, he said that non-financial credit in the U.S. has grown to nearly four times the size of the country’s economy and has doubled over the past decade. On the asset side, he noted that while AI has real practical value, the rapid inflow of capital risks creating a “bubble.”

Roche added that a correction could be “measured” rather than catastrophic, suggesting that markets may adjust downward in an orderly way instead of crashing.

According to Roche, AI does increase productivity, but too much money is now flowing into the sector in the U.S., creating bubble-like conditions. He warned about large amounts of unprofitable AI spending – reaching a scale he compared to four times U.S. GDP – and contrasted this with China’s more capital-efficient approach to AI. He also expects gold prices to rise due to central bank buying and ongoing de-dollarization.

Because the AI sector is overpriced, Roche believes it could face disappointments ahead, as companies struggle to justify their current valuations.

Concerns about an AI investment bubble have grown throughout 2025, especially in recent months. Analysts, investors, and media outlets are debating whether the trillions being poured into chips, data centers, and infrastructure represent a repeat of the dot-com bubble or a justified boom driven by real economic value.

They point to huge tech spending, mostly unprofitable AI ventures, debt-heavy financing, and risks such as energy shortages or future overcapacity. In the first half of 2025, OpenAI reported a loss of $13.5 billion despite $4.3 billion in revenue. Analysts estimate that Big Tech collectively invests around $1.5 trillion into AI infrastructure, chips, data centers, and research. This scale is considered a “hybrid market,” creating risks for startups that can’t keep up with those costs.

Recently, Nvidia CEO Jensen Huang addressed the growing fears of an AI bubble following Nvidia’s earnings call and a 62% profit jump. Huang argued that AI demand is real and likely to continue growing.