2028: The Deadline? Speculations on the Endgame for US Tech Stocks

```

This round of US tech stock rally this year has not automatically fizzled out due to the word "bubble." Tech stock profits and capital expenditure continue to strengthen, driving US stocks and the global semiconductor supply chain to new highs. Meanwhile, since June, US stocks have fluctuated at high levels, Asia-Pacific markets have pulled back sharply, and the market has started to ask the same question: How will the US AI bull market ultimately end?

Qian Wei and the overseas & major asset team at CITIC Construction Investment Securities point out in their latest research: “In the short term, risk of pullback is highlighted, and the Nasdaq may fall more than 10% from its previous high. In the medium term, 2028 could be a critical ‘limit’ for US tech stocks.” CITIC Construction Investment believes that overheated leverage, China’s AI catch-up, style shift brought by traditional economic recovery, political backlash, as well as industry evidence or Fed tightening, are important contradictions that could end the US tech bull market.

The subtler point is that this does not mean an immediate shift to bearishness. Most of the "market top" logics are still at the narrative level, but the driving force for valuation expansion has already weakened. In other words, the short term may feature an upward trend with a 10% pullback mixed in, rather than a one-sided collapse; annual gains have essentially priced in profit expectations for 2026, and the next leg up will have to wait until the second half when the market shifts its focus to 2027.

The real medium-term contradiction is: AI either relies on explosive applications to turn the bubble into a new cycle, or is resolved early by low-cost Chinese models, CAPEX pullback, rate hikes, or regulatory shocks. If the K-shaped divergence persists for another year or two, political risk may become harder to price than earnings risk, and the 2028 election will be the key time point.

Leverage signals are already in the red, but it's not a precise countdown

Margin balances rising with the stock market is not surprising. In a bull market, investors are more willing to add positions, and leverage naturally increases. The danger arises when margin growth obviously outpaces the index gains, usually meaning incremental funds have been excessively used in advance.

Since 1990, when US stock margin balance growth outpaced index gains by over 20%, it happened five times. The latest three correspond to the bursting of the dot-com bubble in 2000, the financial crisis in 2008, and the US stock crash after Fed rate hikes in 2022. After 1992, US stocks didn’t fall sharply, but moved sideways for more than half a year; the impact wasn’t obvious in 1994.

This indicator is now at 25%. Linear extrapolation suggests the chance of a market top in US stocks within the next year cannot be ignored. However, this is not a “stocks will fall tomorrow” signal. The indicator may not have peaked yet, and could still rise further in the short term, corresponding to continued US stock gains. It's more like a red light: the market is still running, but the speed is high.

China’s AI catch-up shakes faith in US CAPEX

The hardest support for US tech stocks now is capital expenditure. As long as the market believes AI computing investment can bring future profits, high CAPEX is not just a cost, but part of EPS expectation.

The problem is, once CAPEX expectations sharply retreat, US tech stocks will face revaluation pressure. In early July, Meta announced compute sales, triggering debate on whether compute and CAPEX are already in surplus. What could truly break this logic is Chinese AI models catching up with the US using fewer resources.

Industry evaluation frameworks show the gap between top US and Chinese models is narrowing. After Kimi K3 was released, US advantages shrank from nearly 300% to 5%. Meanwhile, Chinese models’ average costs are significantly lower than US, including equipment and energy expenses.

This directly raises a question: If US top models can’t keep delivering breakthroughs, why should the market keep paying for huge capital expenditure?

In early 2025, DeepSeek triggered a sharp Nasdaq drop. In the short term, whether Kimi K3’s performance evaluation will cause similar disruption is a key variable for the market.

Traditional economy can't fully take over the market’s main line

Ending a tech bull market doesn’t necessarily mean tech collapses itself. Another path is the traditional economy gets stronger again, funds find better destinations, and the market completes a style shift.

The year 2000 is an example. The Nasdaq began a downward trend from March, falling over 50% within a year; but the Dow didn’t crash during the same period, but moved sideways near highs for almost a year, showing clear relative and even some absolute returns. The US economy was overheated then, GDP growth reached 5%, and traditional sectors like consumption were stronger.

A similar switch occurred from October 2025 to early 2026. After the Nasdaq peaked and corrected, the Dow rose and made new highs. The market at the time traded the recovery and inflation from rate cuts and tax cuts, improving earnings expectations for sectors like industry and consumer, while tech expectations plateaued. When the recovery was disproved, tech EPS expectations resumed upward, and the K-shaped divergence deepened again.

Back to now, under high interest rates, it’s hard for real estate, industry, consumption and other traditional sectors to internally improve comprehensively. Relying on the traditional economy to take over the tech main line isn’t likely unless AI applications truly explode, bringing all sectors into a new growth cycle. Otherwise, the main market line will probably still stay with tech, at most temporarily revert to a balance.

Political risk in 2028 may be harder to digest early than valuation

The farther the AI rally goes, the sharper the problem of K-shaped divergence. Resources, income, and market pricing keep concentrating in giant companies, non-AI sectors can hardly benefit and may even be harmed. If this structure lasts another 1–2 years, anti-AI social sentiment could emerge.

Historically, policy shocks have been more fatal at the peak of tech or financial bubbles. In April 2000, Microsoft was found guilty of monopoly and ordered to be split up, seen as a key negative factor for the dot-com bubble’s burst.

In the long cycle, profit and power concentration from tech revolutions often enters election narratives. In the 1896 election, Bryan criticized industrialization, railroad monopolies, and financial capital exploiting Midwest farmers via new tech; in 2008, Obama supported broad financial regulation.

By 2028, it’s uncertain if both parties will target AI giants. But if campaign platforms feature “strengthen AI giant regulation,” “fight AI giant monopolies,” “tax AI giants,” etc., and the tech bull is already high, the market won’t treat it as ordinary noise.

Valuation expansion has dulled, further gains rely more on EPS

Industry narrative collapse or Fed tightening are more familiar paths for the market. The 2022 US bear market was a typical case of the Fed suppressing valuations.

Now these two long-term trends aren’t decisive, but worries are reflected in valuations. Forward valuations for tech challenged 30x twice in mid-2024 and end-2025 but didn’t succeed; valuations compressed in Q1, rebounded to around 25x in Q2, but have fallen again.

In this period, CAPEX and earnings still grew fast, but valuations didn’t keep expanding. That is, industry and Fed narratives can hardly lift valuations any further, so future price gains depend more on EPS delivery.

2028 is not a precise prediction, but a year when risk gets crowded

In the short term, a Nasdaq pullback of over 10% from previous highs is not surprising, but the narratives ending the tech bull haven’t yet become actual shocks. The market is more likely to experience high volatility and upward oscillation, rather than enter a long-term bear right away.

Structurally, continuous recovery of the traditional economy is still unlikely, so the market main line will probably remain in tech, at most reverting to a balance between tech and non-tech. Persistent shift to cyclical sectors requires stronger fundamental evidence from traditional sectors.

In the medium term, 2028’s importance isn’t about the calendar itself, but that several contradictions may converge before then: whether AI applications will explode, whether Chinese models continue catch-up, whether US CAPEX can be sustained, whether the Fed tightens again, whether K-shaped divergence gets amplified politically in the elections. For the tech bull to last past then, it needs not just continued profits, but proof that AI can help more sectors improve together.

Risk warning and disclaimerThe market is risky; investment needs caution. This article does not constitute personal investment advice, nor does it take into account the particular investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions herein are suitable for their specific situation. Invest accordingly at your own risk. ```