Western countries often underestimate China's potential, as it leads globally in both the quantity and quality of engineers. Despite underlying fears, major U.S. tech companies avoid publicly acknowledging that AI hardware could become obsolete before recovering their investments, according to industry sources cited by elEconomista.es. Underestimating China is a strategic error: the Asian giant could surprise the world with new generations of microprocessors, as seen with DeepSeek, whose AI model surpassed industry leaders in efficiency, causing a stock market drop that affected Nvidia, Nasdaq, and global markets. Almost a year after that episode, multi-billion dollar investments in AI microprocessors are at risk due to Asian competition and cross-investment dynamics. A setback from any major player could trigger a domino effect in the sector. Nobel laureate in Economics Paul Krugman has warned about the similarities between the current AI investment frenzy and the dot-com bubble. In his newsletter, Krugman highlights the excessive growth of AI investment and its economic impact, recalling his track record in economic predictions. Nvidia is the barometer of the sector: if the AI industry maintains its expectations, the company will continue to lead. However, any doubt shakes its foundations, as recently occurred when Meta chose to acquire $72 billion in Tensor Processing Units from Alphabet instead of Nvidia. In the past year, Nvidia's stock value has increased by 36%, reaching $4.58 trillion. Palantir Technologies, on the other hand, has risen by 157% to $461 billion, and since 2023, it has accumulated a revaluation of over 2900%. However, if the AI bubble bursts, Palantir would be the first to be affected. Currently, the combined value of Nvidia and Palantir exceeds $5 trillion. Looking ahead, the AI industry must face the challenge of profitability. Huge investments have yet to translate into significant revenues, and monetization remains a challenge for many Nasdaq companies specializing in AI. Analysts warn that if AI cannot be monetized, the sector could collapse quickly. Investors are beginning to question whether major players like Amazon, Google, and Microsoft will be able to profit from their massive investments. For now, revenues from user subscriptions to AI services are minimal, and the profitability of enterprise licenses is limited. There are alternative models, such as pay-per-use through tokens, but these revenues remain insufficient. The key application of AI could be in cost savings and improvements in business productivity, although recent studies, like the one from McKinsey cited by Michael Chui, show that only 1% of companies have integrated AI maturely into their processes, indicating that real transformation is still far off. Some experts, like Will Nutting, compare AI investment to building a cathedral: costly and with long-term returns. The market, which is priced as if growth is immediate, may be forced to readjust expectations if technology stops challenging the logic of accelerated growth. As Nutting points out, in 2026, AI does not need to collapse to disappoint; it just needs to stop growing at a rapid pace. Source: elEconomista.es