Asia’s Economic Crossroads: Echoes of 1997 and the Rising Threat of AI‑Driven Demand

When Frederick Neumann, HSBC’s chief economist, laid out his latest note on August 31, he couldn’t help but notice the uncanny resemblance between the current market backdrop and the years that culminated in the 1997 Asian Financial Crisis. He points to three big‑ticket items that are eerily familiar: soaring U.S. Treasury yields, a battered Japanese yen and a wave of tech‑centric optimism that feels almost feverish.

Back in the mid‑1990s, the U.S. 10‑year Treasury rate leapt from roughly 5 % in late‑1993 to almost 8 % by the end of 1994, hovering near 7 % as the crisis loomed. Fast‑forward to this year, yields have surged about 80 basis points since February, sitting just under 4.8 % – a steep climb from the historic lows of 0.5 % we saw in August 2020. It’s a jump that feels familiar, and it raises eyebrows among economists who remember how quickly cheap money can evaporate.

Currency movements tell a similarly stark story. The yen fell about 55 % against the dollar between April 1995 and April 1997, slipping from ¥80 to ¥130. Today the yen has endured an almost identical decline – from roughly ¥103 per dollar in early 2021 to a low of ¥163 in July – prompting an unusual joint intervention by Washington and Tokyo to hold it near ¥160. The parallels are hard to ignore.

What’s different, however, is what fuels the current optimism. In the mid‑1990s, investors were mesmerised by the nascent internet. Now it’s the glittering promise of artificial intelligence. U.S. tech giants are pouring money into AI‑ready data centres, and that spending chain feeds straight into Asia’s semiconductor factories in South Korea, Taiwan, Japan and Singapore.

Neumann stresses that the structural foundations have shifted dramatically. In the 1990s, many Asian economies were net importers of capital – they relied on foreign savings, ran large current‑account deficits and possessed fragile banking systems. When yields rose, capital fled, currencies collapsed and banks toppled, as we saw in Thailand, Indonesia and South Korea.

Today the picture is almost the opposite. Most of the same economies are now net exporters of capital, boasting sizable foreign‑exchange reserves that act as a cushion against sudden dollar‑funding squeezes. The banking sector, while not invulnerable, is far sturdier than it was three decades ago.

But there’s a new, subtler risk on the horizon – what Neumann calls a “demand vulnerability.” The region’s growth engine is now heavily tied to the appetite of U.S. hyperscalers for AI‑related chips and hardware. If higher U.S. Treasury yields force American tech firms to trim data‑centre spend, or if the yen’s volatility rattles broader funding markets, the ripple effect could be a swift, quiet contraction in orders for Asian semiconductor makers.

Unlike the bank‑run panic of 1997, this shock would hit through a sharp dip in export orders rather than a sudden loss of confidence on the trading floor. The danger is that the slowdown could be gradual enough to slip under headlines, yet severe enough to dent the region’s growth trajectory.

In short, while the old‑fashioned financial cracks may have been patched, Asia now finds itself balancing on a different edge – the relentless demand (or lack thereof) for AI hardware from across the Pacific. Policymakers and investors alike would do well to keep an eye on that demand gauge as the next potential fault line.