A growing sell-off in US mega-cap technology stocks is sending ripples across Wall Street and Asian markets, rattling investors in Seoul, Tokyo and Taipei who have heavily bet on AI plays. At the heart of the anxiety is a critical question: can US tech giants’ massive investments in AI infrastructure deliver the returns investors expect?
The question will be put to the test this week as US tech giants, including Microsoft, Amazon and Alphabet, with a combined market capitalisation of $10 trillion, report quarterly earnings. The outcome could have a bearing on Asian chipmakers and India, an underperformer that could emerge as a ‘reverse AI’ trade.
"Our recent investor feedback notes concerns about returns on AI investment given the rising capex relative to cash flows and uncertain monetization," UBS said last week in a note on the Asia tech supply chain.
Concerns over AI-related capex amid rising Fed rates are also evident in a post on X by Elon Musk, the world's richest man. "As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers. SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue. That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many," he warned.
The tech-heavy US index Nasdaq Composite is down 6 per cent since October 6, with Meta and NVIDIA falling 12-13 per cent and Alphabet slipping 8 per cent over the same period. Broadcom, Advanced Micro Devices and Micron Technology have seen deeper corrections of 14-16 per cent. In Asia, Samsung Electronics is down 21 per cent, while SK Hynix has plunged 27 per cent in the past one month.
Manish Sonthalia, director and chief investment officer at Emkay Investment Managers, said the AI trade is starting to unravel, attributing it as a major reason global 10-year bond yields have begun to correct.
Data showed that 10-year US Treasury yields had slipped below the 5 per cent mark as investors flocked to safe-haven assets amid a retreat from riskier technology stocks.
The Philadelphia Semiconductor Index, a useful proxy for the global AI trade, has corrected and is now showing weakening market breadth, measured by the percentage of index constituents trading above their 200-day moving average (DMA).
"Historically, such deterioration in breadth has often preceded prolonged downtrends or sideways phases, particularly when accompanied by a break below the index's own 200DMA," Elara Securities said.
Post-Covid, the first such episode occurred in January 2022, when the index broke below its 200-DMA alongside a sharp collapse in breadth, triggering a steep correction that lasted almost a year. A similar pattern emerged in July 2024, when the index breached its 200DMA and breadth deteriorated sharply, preceding an eight-month correction. In contrast, the September 2023 deterioration in breadthwas not accompanied by a break below the index's 200DMA, and the weakness proved shortlived, with breadth recovering quickly.
Sonthalia said the question of how much capex is going into AI creation vis-à-vis the return on investment has been debated for a long time."But currently, this far, no further in terms of productive use of AI and the revenue model around these LLMs could be found. The whole safeguard initiative taken by the Frontier Lab in terms of safeguarding against the use of AI, the likes of OpenAI and Anthropic, whatever caution they have been sounding...," Sonthalia said in an exclusive interview to Business Today TV (BTTV).
Late-cycle characteristics Nuvama, in a recent strategy note, said high rates are likely to weigh on the AI trade, which is exhibiting late-cycle characteristics, akin to the dot-com era. This, along with fading domestic monetary easing, further bolsters its call for a contrarian shift towards the Nifty over emerging markets.
"Rising rates amid weakening earnings tend to weigh on the pockets of mania - à la NBFCs in 2018, IT in 2022. Today that pocket is AI. The AI mania is already exhibiting late cycle characteristics, à la dot-com era (link) - high valuations, leveraged retail participation, consolidation in chip stocks post melt-up."
Does AI pain equal India's gains? Not an easy answer. On face of it, yes.
"The weakening of the AI trade is positive for India. In September and October, so far, the capital flight from South Korea and Taiwan from FPI selling is more than the capital outflow from India. This is an important trend," said V K Vijayakumar, Chief Investment Strategist, Geojit Investments Limited
But India too has its own challenges. Nuvama last week noted that India's cyclical earnings recovery faces headwinds. These include moderating margins as inventory gains fade. RBI rate hikes amid weak incomes, El Niño and the fading tailwind from GST cuts pose further risks to domestic growth.
These concerns are visible on market returns. Relative to emerging markets, India has almost given up its entire post-Covid outperformance. In fact, India's 20-year returns and those of emerging markets have nearly converged, a sharp contrast with the scenario in September 2024.
Even as mutual funds are investing, "almost entire retail flows in recent years have been used to fund promoter exits through IPOs, OFS as well as blocks rather than fund companies’ expansion plans or propel secondary markets. This certainly does not warrant a re-rating," Nuvama said.
On the earnings front, Bernstein said earnings are not seeing upward revisions for FY27, with only a modest 0.2 per cent increase over the past three months and several sectors continuing to see cuts.
Bernstein said a lower long-term capital gains (LTCG) tax is desirable for a capital-starved economy, but an LTCG cut alone is unlikely to drive sustained FII inflows without a recovery in India's growth and earnings cycle, which remains elusive.
"A sharp 1-2 day market rally that a cut would trigger may actually be used to offload positions," it said.
The foreign brokerage said India is unlikely to benefit materially from a near-term pause in the AI trade, as it offers neither the return profile nor the investment opportunities attracting global capital into AI today. It may be some time before portfolios turn black.