AI's gravity: Why global money is drifting from Dalal Street to North Asia

Foreign portfolio investors (FPIs) have pulled Rs 44,166 crore from Indian equities so far in October, taking their 2026 withdrawals to Rs 3.04 lakh crore. While elevated crude prices, a firmer dollar and high US bond yields have hurt sentiment, a quieter force is reshaping the flows: global capital is chasing the artificial intelligence boom in North Asia, where semiconductor-heavy markets offer more direct exposure to the world’s fastest-growing technology theme.

The result is an uncomfortable contrast for Indian markets. The Nifty has delivered negative returns of 13.87 per cent year-to-date in 2026, while Taiwan and South Korea -- home to chipmakers at the centre of the AI supply chain -- have become the preferred destinations for funds seeking immediate AI-linked earnings.

The numbers behind the exit

FPIs had already withdrawn Rs 35,861 crore in September after investing Rs 20,200 crore in July and Rs 29,631 crore in August. The October selling has pushed 2026 outflows to Rs 3.04 lakh crore, far above the Rs 1.66 lakh crore recorded in the whole of 2025, according to NSDL data cited in market reports.

The scale of the shift is not merely an India-specific phenomenon. Foreign investors sold Asian equities at the fastest pace in at least 16 years in the first half of 2026, pulling a net $137.36 billion from South Korea, Taiwan, India, Indonesia, Thailand, Vietnam and the Philippines.

South Korea and Taiwan bore the largest outflows -- $70.8 billion and $29.6 billion respectively -- as funds trimmed positions in markets that had rallied sharply on AI optimism.

That detail matters: the AI trade has become so concentrated and expensive that investors are even taking profits in the biggest winners. But India’s problem is different -- it lacks the large-cap semiconductor, memory-chip and AI-hardware companies that have powered the regional rally.

Why Taiwan and Korea win the AI trade

Taiwan’s market has been propelled by Taiwan Semiconductor Manufacturing Company (TSMC), which accounts for roughly 42 per cent of its benchmark index and rallied 49 per cent in 2026, according to market data cited by The Economic Times.

Taiwan’s market capitalisation rose to $4.95 trillion, overtaking India’s $4.92 trillion and making it the world’s fifth-largest equity market.

South Korea presents a similar story through Samsung Electronics and SK Hynix, both central to memory chips and high-bandwidth memory used in AI data centres.

Nomura raised its 2026 KOSPI target to 10,000–11,000, citing a “commodity memory and high-bandwidth memory supercycle,” and projected 200 per cent year-on-year earnings growth for Korean corporates in 2026.

By comparison, India’s listed universe is dominated by financials, consumer companies, energy firms, infrastructure players and IT-services exporters.

Indian IT firms may benefit from AI-led outsourcing demand over time, but they do not offer the direct, high-beta exposure to chips, servers, advanced packaging and AI infrastructure that global funds can find in Taipei or Seoul.

India’s structural disadvantage

Analysts argue that India’s limited exposure to the AI-hardware cycle is not a short-term sentiment issue but a relative-earnings problem. Kotak Securities’ Sanjeev Prasad has described India’s exposure to the AI and semiconductor cycle as “negative,” warning that the mismatch could persist for one to three years.

The numbers illustrate the gap. Jefferies estimates cited by NDTV Profit put India’s 2026 earnings growth at about 12 per cent on an April–March fiscal-year basis, against 58 per cent for Taiwan and more than 300 per cent for Korea.

For a global fund allocating capital across Asia, the choice between a broad consumption-and-financials market and markets with explosive AI-linked profit growth becomes easier.

India’s weight in the MSCI Emerging Markets index has also fallen to about 12 per cent from 19 per cent a year earlier, reflecting both underperformance and the reallocation of global portfolios towards AI-linked economies.

Not a verdict on India -- but a competition problem

Market participants caution against reading the outflows as a rejection of India’s long-term prospects. Vedant Gupte, co-founder and CEO of Trackk, said the selling should be viewed as a “global repositioning of capital” rather than a judgment on India’s investment case, pointing to crude prices, dollar strength, US yields and the AI rally in North Asia.

That distinction is important. India’s economy remains large, diversified and domestically funded; its market has not collapsed despite sustained foreign selling. But in the short term, global investors are not asking only whether India will grow—they are asking which market gives them the cleanest exposure to the current earnings cycle. On that question, India is currently losing to North Asia.

VK Vijayakumar, chief investment strategist at Geojit Investments, added that FPI selling was rational while the 10-year US government bond yield remained above 5.2 per cent. As long as safe dollar returns remain high and Indian valuations do not become compelling, foreign funds have little immediate incentive to return.

The domestic cushion

The counterweight to FPI selling is India’s domestic investor base. Systematic investment plans alone contributed over Rs 31,961 crore in July, more than ten times the level a decade ago, helping prevent a sharper market decline even as overseas funds exited.

This domestic support has become the central stabiliser of Indian equities. It explains why the market has corrected rather than crashed despite foreign ownership declining—from around 23 per cent to roughly 15 per cent, according to independent market analyst Ambareesh Baliga.

But the cushion has limits. If FPI outflows continue, the rupee weakens further, crude stays elevated and corporate earnings disappoint, domestic institutions may eventually demand lower entry valuations before adding aggressively.

Can India join the AI rally?

India’s opportunity may lie less in building a TSMC-like chip giant overnight and more in becoming a large-scale deployer and services layer for AI. The country’s digital public infrastructure, large multilingual user base, data-centre build-out and IndiaAI Mission could support AI applications in financial inclusion, healthcare, language translation and government services.

That could eventually create an India-specific AI investment theme – centred on data centres, power transmission, cloud services, digital platforms and enterprise software – rather than one dependent on semiconductor manufacturing. But global funds are unlikely to wait for that theme to mature while Taiwan and Korea are delivering immediate earnings momentum.

For now, the market message is stark: India remains a compelling long-term growth story, but it is not yet the most efficient vehicle for the AI trade. Until Indian equities offer either cheaper valuations, stronger near-term earnings or a credible AI-infrastructure theme, foreign capital will continue to look east—to Taipei, Seoul and the wider North Asian technology complex.