Chen Li, Global Chief Economist at Soochow Securities and Chief Executive of Soochow Hong Kong, argues that attributing the recent sharp decline in US tech stocks solely to strong non-farm payrolls data is an oversimplification. The real causes are more complex and rooted in market structure and liquidity.
Last Friday, US stocks experienced a significant sell-off, with the Nasdaq Composite plunging over 4%, its largest single-day drop in over a year. Companies like Nvidia saw massive market value erosion, losing over 2.2 trillion RMB in a single session. The immediate narrative linked the sell-off to May's non-farm payrolls data, which significantly exceeded expectations, leading to heightened expectations for Federal Reserve rate hikes and subsequent selling in tech stocks.
Examining the Non-Farm Data Argument
While the May non-farm payrolls added approximately 172,000 jobs, beating forecasts and leading to a slight uptick in Treasury yields and the US dollar, the magnitude of the Nasdaq's decline seems disproportionate. For instance, the 10-year US Treasury yield rose by about 6 basis points to 4.54%, and the dollar index strengthened modestly. However, the Nasdaq fell by 4%.
This reaction is puzzling. In late April and early May, the 10-year yield rose 24 basis points in a single week due to inflation concerns, yet tech stocks quickly rebounded and continued their ascent. Furthermore, the Russell 2000 small-cap index fell 3.47% on Friday, a smaller decline than the Nasdaq's. If the logic of "strong jobs data leads to rate hike fears leads to stock market decline" held universally, more economically sensitive small-cap stocks should have fallen more sharply, not less.
Additionally, the quality of the May jobs data is debated. Some economists suggest a significant portion of the gains may be attributed to one-time hiring for the upcoming FIFA World Cup, particularly in leisure, hospitality, and local government security roles. If these are temporary positions, the Fed is unlikely to base monetary policy tightening on them.
Broader Market Weakness Preceded the Tech Sell-off
The sell-off was not isolated to tech. Broader market pressures have been building for months due to unresolved geopolitical tensions in the Middle East affecting oil shipping lanes, persistent inflation expectations, and consistently high bond yields.
Over the past two to three months, a clear divergence emerged in global asset performance: AI-related tech stocks, like Nvidia, rallied to new highs, while virtually all other sectors—non-tech US stocks, Hong Kong stocks, and non-AI A-shares—had already been weakening.
In this macro environment, the AI sector alone was propped up by consistently strong earnings expectations. Other sectors have been under pressure from high rates for over a month. The Hang Seng Tech Index, lacking key AI hardware components, has consistently underperformed neighboring markets like Japan and South Korea.
In the A-share market, even when the Shanghai Composite Index surpassed 4000 points, about one-third of all stocks (roughly 1700) were trading below their levels from late September 2024. Mutual fund products without exposure to semiconductors or optical communications have significantly lagged the broader market this year. Investors without AI tech exposure have been quietly bearing the brunt of inflation expectations for two months.
The Core Issue: Liquidity and Positioning
The primary driver of last Friday's sharp adjustment is a liquidity and positioning issue, not a fundamental deterioration in the AI industry.
First, valuation disparities have grown extreme. After nearly a year of gains, the dynamic price-to-earnings ratio for the tech sector is at a relatively high level, leaving little room for positive earnings surprises. Meanwhile, non-AI sectors have been declining for over a month, widening the valuation gap further.
Second, leveraged trading has created extreme crowding. AI-related stocks have been the most concentrated long position for global institutions over the past six months. This crowded trade involves significant leverage; any market tremor can force leveraged positions to unwind, exacerbating the sell-off through a cascade effect.
These factors are liquidity-driven and unrelated to the long-term demand outlook for AI.
Specific Concerns Within Tech
Beyond macro factors, specific company news added pressure. While the market is focused on SpaceX's upcoming mega-IPO, Meta Platforms Inc's announcement of a planned multi-billion dollar stock sale on Friday, which caused its shares to drop 7%, raised more immediate concerns. This follows Alphabet Inc's recent $80 billion fundraising. While framed as supporting AI capital expenditure, these large equity issuances also signal significant funding pressure, potentially diluting existing shareholders.
Furthermore, Broadcom Inc reported earnings after the market closed on Wednesday. While AI chip revenue surged 143% quarter-over-quarter and total revenue grew 48% year-over-year, hitting record highs, its stock plunged over 12% in after-hours trading. The market reacted negatively because, first, the guidance for the AI chip business merely met high expectations without being raised further, and second, CEO Hock Tan indicated that Google might diversify its chip suppliers away from Broadcom's exclusivity and that the company would pivot from selling full AI systems to selling chips only.
This highlights a critical market dynamic: for AI stocks, even stellar results can be viewed negatively if future guidance fails to exceed already lofty expectations. Against this backdrop, the strong non-farm payrolls data simply added another blow to an already wounded market.
AI Industry Fundamentals Remain Intact
Examining Nvidia's latest quarterly report reveals robust fundamentals: revenue grew 85% year-over-year, data center revenue surpassed $75 billion for the quarter, marking the 14th consecutive quarter of sequential growth, and next-quarter guidance points to a median of $91 billion, indicating accelerating growth. These are not the results of a company facing peak demand.
More importantly, the accelerating deployment of AI Agents is shifting compute demand from an "infrastructure build-out phase" to a "scale application phase," suggesting hardware demand is expanding, not contracting.
Market concerns are not entirely baseless. Fluctuations in free cash flow among major cloud providers, high customer concentration for data center chips, and debates over the profitability of AI investments are variables worth monitoring. However, they do not yet constitute evidence of a fundamental downturn. In summary, the current state of the AI industry can be described as having no major flaws except for being expensive.
Investment Outlook Post-Adjustment
Since the adjustment's root cause is liquidity, not industry fundamentals, the investment framework should adapt accordingly.
The logic for selling requires a premise: either AI industry demand falls short of expectations, or systemic macro liquidity tightens. Currently, there is insufficient data to support the former, while the latter depends on the evolving path of inflation and interest rates.
The logic for buying on dips rests on a different premise: the industry trend remains unchanged, and the current adjustment is merely a rebalancing of positions and capital, not a trend reversal.
The inclination is that, barring a major setback in AI industry fundamentals, this round of tech stock adjustment likely does not mark the cycle's peak. There may be opportunities for bargain hunting in AI tech stocks.
However, a more difficult question remains for every investor: can we truly identify a "major fundamental setback" before it occurs?
Reflecting on the 1999 dot-com bubble, warning signs were evident in hindsight—most companies were unprofitable, business models relied on continuous financing, and valuations were completely detached from reality. Yet, at the time, these signals were absorbed by a new narrative: "This is a new economy; old frameworks don't apply." The data wasn't necessarily strong, but the market actively changed its framework for interpreting it.
Today's AI sector differs substantially from the dot-com era. Nvidia's valuation is supported not by narrative alone but by quarterly revenue growth of 85% and real, tangible profits—a fundamentally different asset class from the likes of Pets.com.
Nevertheless, the logic of narrative frameworks remains relevant. The true warning signal for an inflection point is not a single piece of weak data but the moment the market's narrative framework for explaining negative signals begins to crack. The day the view that "AI capital expenditure is unsustainable" shifts from a minority opinion to a mainstream narrative will be more significant than any single quarterly report.
Currently, there are no signs of this framework loosening. Cloud providers' procurement is accelerating, and AI Agent applications are expanding. Volatility will persist, but the overall direction will depend more on the industry's ability to continue meeting expectations than on the next non-farm payrolls report.
The real challenge is whether investors can identify that inflection point. If not, the focus should return to protecting investment portfolios through appropriate strategies.
Comments