AI Monetization Struggles? US Tech Giants Face Cost Crisis with Closed Models — Guo Zhengliang Highlights China's Surging Token Usage

The seven major US tech companies—Microsoft, Apple, Nvidia, Alphabet, Amazon, Meta, and Tesla—have recently released their Q2 2024 financial reports. Among them, Alphabet and Tesla reported negative free cash flow, with capital expenditures exceeding expectations. Meta's free cash flow also plummeted by over 90%, sparking market concerns over the massive capital spending on artificial intelligence (AI).

In response, former legislator Guo Zhengliang, on the program 'Dada Pingpingli,' questioned the future of US closed AI models: 'Where will the future customers come from?' Regarding allegations that the seven tech giants are 'hiding liabilities,' Guo noted that industry insiders were already aware of Alphabet's negative free cash flow. He questioned whether the company's profitability model remains sustainable after its massive investments in AI.

Guo compared the development trajectories: the US has built 59 large models since starting in 2001, while China, beginning in 2014, has already developed 35 large models. He added that many US companies are now adopting cheaper Chinese AI models, further undermining confidence in US closed models.

When comparing total AI 'Token' usage volume, Chinese models are already four times higher than US models and have maintained this lead. Although China currently has fewer data centers than the US, it is constructing them at scale, meaning China's future computing power and total model capacity will inevitably grow.

Guo concluded that China's total Token usage will continue to lead, potentially widening the gap to over 10 times. He acknowledged that the US still holds technological deterrence, with Taiwan, Japan, South Korea, and the EU currently relying on US models. However, he warned of potential cracks in this dominance. Even if the US develops open-source models, the costs would be enormous, making it difficult to compete with China.