Guangfa Liu Chenming: Is the Fed's interest rate hike what burst the tech bubble?

Since March this year, with changes in the US-Iran situation and the change of Fed Chairman, expectations of US interest rate hikes have been fluctuating, disturbing global markets.

Earlier, we discussed the impact of the Fed's interest rate hikes in several reports.

However, during recent roadshows, we still feel a lot of concern about the Fed's rate hikes, especially the potential impact on tech stocks.

From our observations, the lingering concern stems from a common first impression: the last tech bubble was burst by the Fed's rate hikes.

Is that really the case?

We believe it is necessary to dissect the details of 1999-2000 again.

This review is not about comparing today's AI to the 1999 tech experience, as the external environment is vastly different, and no historical situation can be fully replicated.

Rather, this article aims to use this classic case, from a first-principles perspective, to demonstratehow rate hikes, liquidity contraction, and rising interest rates actually affect tech stocks in the midst of an industry boom.****

I. The Rate Hike Path and Background Starting in 1999: Preliminary Guidance for Rate Hikes Began in February 1999

After the Asian financial crisis ended (the Fed cut rates three times for a total of 75bp between September and November 1998), the global deflation cycle reversed and commodity prices began to recover.

In addition, in early 1999, OPEC and non-OPEC jointly cut production, coupled with the outbreak of the Kosovo War (March-June 1999), which threatened transportation in the Balkans and the Mediterranean. The market feared supply disruptions, and oil prices rose from $10/barrel to over $30/barrel. At this time, US CPI inflation also picked up quickly. The Fed re-entered a rate hike cycle in June 1999, raising the federal funds rate from 4.75% to 6.5% by May 2000.

The rough timeline is as follows:

1. At the FOMC meetings in February-March 1999, the Fed did not give a clear signal of rate hikes, but initially expressed concerns about inflation risks and began to discuss the rationale for the current loose monetary policy.

2. At the FOMC meeting in May 1999, the Fed basically confirmed that it would begin raising rates in the future. The Nasdaq index adjusted about 8% in the short term, then continued its upward trend.

3. In June 1999, the Fed officially started its rate hike cycle. The Nasdaq saw no significant adjustment and continued its upward trend.

4. In August 1999, the Fed raised rates for the second time. The Nasdaq saw no significant adjustment and continued its upward trend.

5. In November 1999, the Fed raised rates for the third time. The Nasdaq saw no significant adjustment and continued its upward trend.

6. In February 2000, the Fed raised rates for the fourth time. The Nasdaq saw no significant adjustment and continued its upward trend.

7. In March 2000, the Fed raised rates for the fifth time. The Nasdaq had already peaked in the tech cycle before this rate hike. From the first Fed rate hike to the bubble peak, the Nasdaq cumulative gain exceeded 90%.

8. In May 2000, the Fed's last rate hike was 50bp in a single move. The total rate hike in this cycle was 175bp.

II. From the Moment of Confirming the Start of the Rate Hike Cycle in May 1999, the Dow Jones Index Was Suppressed First, Well Before the Tech Peak

The Dow Jones Industrial Index was suppressed in Q3 1999 by the dual pressures of high oil prices and rate hikes, briefly rallied in Q4, and peaked in January 2000.

In other words, the Dow entered a consolidation state from the May 1999 FOMC meeting when the Fed basically confirmed imminent rate hikes.

From a pricing logic perspective, this shows two things:

1. For sectors with weak demand, facing high interest rates (financing costs) and high oil prices, fundamentals may be more suppressed (compared to industries with surging demand).

2. For companies with relatively stable earnings cycles, pricing often considers long-duration cash flow discount models. When the central levels of interest rates and oil prices shift upward, valuations for such companies tend to be suppressed.

Therefore, for the relatively slow-growing Dow Jones index (EPS compound growth of about 3% in 98-99), even though valuations were not expensive, it was the first to be suppressed by rate hikes.

III. However, With the First Four Rate Hikes Starting in June 1999, the Nasdaq Soared Over 90%

From the Fed's first rate hike in June 1999 to the final tech bubble peak in March 2000, the Nasdaq cumulative gain was 91%, with the index peak lagging the rate hike by nine months.

This also illustrates several points:

1. When demand is surging, although rate hikes have some impact, companies do not care much about higher interest rates (financing costs), or at least it is not central to business decisions.

2. For companies with short-term earnings bursts, investors know that such bursts are likely short-term (2-3 years) rather than long-term (5-10 years). Therefore, these short-term EPS and cash flows cannot be discounted using discount rates, so pricing is not much related to the upward shift in interest rate levels.

3. When rate hikes lead to liquidity contraction, investors may prioritize selling stocks of slow-growing or non-growing companies and flock to a few booming sectors.

Thus, when the EPS growth of the Nasdaq 100 index surged from 15% in 98 to 60% in 99, even though the index's PE valuation was already 55x static and 45x dynamic in 98, it actually significantly increased its valuation during the Fed's consecutive rate hikes.

IV. What Factors Drove the Nasdaq's Earnings Explosion in 99?

Around 95, when the tech boom was just starting, the growth rate of the Nasdaq 100 index gradually increased, but generally remained in the 20-30% range.

Among them, in 1997-1998, the fundamentals of tech giants had already declined to some extent, but the expectation of order explosion from Y2K created a "temporary boom" for tech leaders in 1998-1999.

Y2K refers to the "millennium bug" replacement cycle. At that time, US scientific circles, media, and even the government were promoting the possibility of a huge cyber risk: if everyone did not upgrade servers, systems, and PCs to the latest version before the year 2000, the millennium would cause code and system crashes.

Under policy directives from agencies like OCC, FDA, and the Department of Defense requiring priority fixes for system bugs, governments and businesses globally began a "panic buying spree" of old servers, mainframes, personal computers, and software operating systems in 1998-1999, fully igniting the capital expenditure boom for device replacement.

Consequently, many software and hardware companies' earnings surged again, pushing the EPS growth of the Nasdaq 100 index in 99 to the fastest level of the 1990s: 60%.

V. If Not Rate Hikes, What Actually Burst the Bubble?

The capital expenditure boom brought by Y2K not only allowed the Nasdaq to withstand the consecutive rate hikes starting in 99, but also created the largest tech bubble in history.

However, what made it also broke it.

The Y2K crisis of the year 2000 did not occur. The pre-ordered hardware purchases borrowed future demand, making hardware capital expenditure unsustainable in the next phase, and the industry faced pressure from high inventories.

From 1998 to 1999, to prevent the date coding problem of the "millennium bug" that could cause widespread system paralysis, global enterprises massively advanced IT department capital expenditures. But when 2000 actually arrived, people found Y2K was not that severe. Even without replacing hardware and software, simply modifying code could avoid cyber risks.

This signaled that demand for tech equipment procurement would soon decline, leading to shrinking orders and increasing inventory pressure for leading companies like Dell, HP, and IBM. The global semiconductor industry's inventory level also reached a historical high by the end of 2000.

In addition, in Q1 2000, negative signals from the tech industry and core company fundamentals came one after another:

In February-March 2000, US media widely reported that the Department of Justice's antitrust lawsuit against Microsoft was about to be adjudicated, with extensive evidence supporting the monopoly allegations, negatively affecting market sentiment.

Earnings data released starting in March 2000 showed poor sales of tech products during the 1999 Christmas holiday.

Microsoft's earnings report in April showed revenue below expectations and anticipated slowing PC market growth. IBM's first-quarter earnings report in April also showed revenue below expectations. Several once-prominent tech startups went bankrupt (e.g., luxury fashion e-commerce Boo.com, online news platform APB online, online retail company Value America, etc.).

Ultimately, in 2000, the Nasdaq 100 index's EPS growth fell from 60% in 99 to only 12%.

VI. Back to the Present: If Rate Hike Expectations Trigger Concerns about a US Recession, AI Will Also Be Vulnerable

Returning to the current AI situation, besides changes in AI demand itself, if concerns about a US recession arise due to rate hike expectations, people will further worry about the impact of a recession on the traditional businesses of US CSP cloud giants, thereby affecting their cash flow and future capital expenditure capabilities, and ultimately impacting overall computing power demand.

Around April 2025, during the global market's trading of US recession expectations, sectors like GPU, optical modules, and PCB experienced one of the biggest declines in recent years.

But do rate hike expectations necessarily trigger recession concerns?

A typical case is 2022-2023. Facing high oil prices (Russia-Ukraine war) and high inflation, the Fed enacted the fastest rate hikes in history.

At that time, most investors believed the US would experience a recession in 2022-2024. However, ultimately, expansionary fiscal policies like the three major acts offset the rise in interest rates.

In the current environment, many global economies are similar: monetary policy is showing signs of tightening or has already tightened, but major large economies are still adopting very active fiscal policies, ensuring relative stability in traditional demand.

Historically, during war crises, an official recession is triggered only when war causes a sustained and sharp surge in oil prices, combined with the US economy being in a fragile state at the end of an expansion cycle with high inflation and entering a rate hike cycle. If the war's impact on oil prices is short-lived and there are multiple fundamental drivers, a recession may not be triggered even during a rate hike cycle.

At the same time, the US fundamentals since the pandemic have been characterized by "no clear recession on the downside, no strong recovery on the upside." The support items have gone through several rounds of driver changes, but there has been no full-blown overheating.

No surge, no crash. If the US economy is at the peak of a cycle and all indicators are overheating, then facing high oil prices, high inflation, and rate hike expectations could be dangerous.

Currently, US fundamentals are supported by resilient consumption, high growth in AI investment, and possible fiscal stimulus under the political cycle. Looking ahead, growth momentum may slow, but recession concerns are premature.

VII. Otherwise, Rate Hikes Are Not a Concern; the Core Returns to the AI Industry Itself

In fact, the best way to resist rate hikes, rising interest rates, and liquidity contraction is rapid demand-side explosion.

The following chart shows several historical instances in A-shares and US stocks where industries boomed despite liquidity contraction, proving this point.

Finally, there is a sharp question: Will rate hikes increase financing costs and affect the capital expenditure of CSP cloud giants?

This also seems to depend on the demand of the AI industry itself.

If capital expenditure is driven solely by FOMO, then higher financing costs will make giants think twice.

But if there are continuously surging cloud service orders and demand, will giants slow down capital expenditure because of a 25bp increase in financing rates?

In the first two decades of the 21st century, would Chinese people stop buying houses because of high interest rates?

Source: Chenming's Strategic Deep Thinking

Risk Warning and Disclaimer

        Market has risks, investment needs caution. This article does not constitute personal investment advice, nor does it consider the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Investment based on this information is at your own risk.