The Zhitong Finance App learned that Société Générale strategist Albert Edwards, who has always been known for his extremely pessimistic market views and even calls himself a “giant bear,” compared this round of artificial intelligence (AI) investment boom to the economic bubble before the 1997 Asian financial crisis. He warned that large-scale capital expenditure may mask the current situation of weak productivity growth, raising questions about whether the AI boom can deliver on investors' high expectations.
In a research report released on October 8, Edwards indicated that total factor productivity in the US is slowing down. Total factor productivity measures how efficiently an economy uses labor and capital. The weakening of this indicator is in stark contrast to the surge in AI infrastructure investment and market optimism driving the S&P 500 index to a record high.
Edwards said that this divergence reminded him of the East Asian economic expansion period in the 1990s — when rapid investment and strong growth inspired general optimism, then the region experienced a financial collapse.
The strategist quoted Apollo Global Management's chief economist Thorsten Slocke's research as pointing out that despite the AI investment boom in full swing, the growth rate of total factor productivity in the US is declining. The purpose of this indicator is to separate out the growth brought about by simply increasing manpower and equipment, and to measure the improvement in economic efficiency separately.
Lessons from the Asian Financial Crisis
Edwards recalled how economists and policy makers praised the rapid development of the Asian economy before the 1997 crisis began. The World Bank published a report entitled “The Miracle of East Asia” in 1993, followed by a study on Thailand's economic success in August 1996 — and less than a year later, Thailand fell into financial collapse.
Economist Paul Krugman challenged mainstream optimism at the time in his 1994 article “The Myth of the Asian Miracle.” He believes that much of the region's economic growth reflects increases in investment and labor rather than substantial increases in productivity.
Edwards said he also came to a similar conclusion at the time, warning that cheap and abundant capital was being allocated inefficiently. The financial crisis that followed revealed vulnerabilities that were overshadowed by rapid economic expansion.
The strategist also reviewed his experience comparing the US tech bubble in the late 90s to Thailand's previous investment boom. He believes that today's confidence in AI may be ignoring similar warning signs.
AI spending fails to boost net investment
Edwards emphasized the differentiation between total investment and net investment in US companies.
Total investment has increased as companies spend heavily on AI-related equipment and infrastructure. However, net corporate investment as a share of GDP after deducting depreciation has remained basically the same.
This difference is critical because increased input does not necessarily translate into an equivalent increase in productive capital. Businesses must constantly replace aging equipment, particularly in the tech sector, where assets can quickly become obsolete.
Citing economist Rob Palento's observations, Edwards pointed out that the increase in total investment did not bring about a corresponding improvement in net investment. The report points out that in recent years, the total investment of actual enterprises as a share of GDP has risen, while net investment has basically been flat.
The strategist also questioned the composition of corporate spending. Investment in equipment has accelerated in nominal terms, but after adjusting for inflation, the increase was not significant.
At the same time, investment in non-residential buildings has shrunk despite data center expansion. Edwards pointed out that the contribution of corporate fixed investment to annual GDP growth is still within the normal range of history. Productivity dividends remain uncertain.
Edwards acknowledged that AI could eventually deliver the productivity gains investors expected. The technology is still being deployed in various industries, and its economic benefits may take time to show in official statistics.
Slock also said that it may be too early to expect AI investments to produce measurable improvements in total factor productivity.
Despite this, Edwards believes that the reality of weak productivity growth is enough to make one doubt the scale of the investment boom.
The analogy with the Asian financial crisis does not mean that AI investments will eventually end in a similar collapse. But Edwards' analysis suggests the possibility that spending driven by expectations of future technological breakthroughs may exceed the economic benefits ultimately realized.
The bond market poses another risk
Edwards also warned that the global bond bear market is still a major concern for investors, and government borrowing costs may face a new round of pressure. He pointed out that bond investors have recently focused on France, but as they assess the financial weaknesses of major economies, their attention may turn to Japan, the United Kingdom, or the US.
The strategist believes that as investors test the ability of governments to maintain borrowing, the pressure on the bond market may shift rapidly between countries.
For stock investors, the broader concern is that while evidence of a continued productivity boom is still limited, enthusiasm for AI has supported record share prices.
Edwards' analysis shows that whether AI-driven market gains can continue depends ultimately on whether this technology can bring measurable improvements in economic efficiency rather than simply setting off another wave of capital expenditure.