Crowding, valuations, and ongoing rate hike expectations: Is it still worth sticking with tech?

Crowding, valuations, and ongoing rate hike expectations: Is it still worth sticking with tech?

```

Facing the triple pressures of renewed interest rate hike expectations, crowding warnings, and valuation disputes, the latest assessments by institutions such as GF Securities and Dongwu Securities point to the same core conclusion: The trend of EPS revisions and the AI industry cycle have not yet been disproven; various disturbances are more like temporary noise rather than sufficient conditions for a mid-term market peak. The technology sector remains worth holding, and the adjustment window may provide an opportunity for positioning.

The immediate event triggering this round of sharp market fluctuations was the US May non-farm payrolls data far exceeding expectations. Data released on June 5 showed that the US added 172,000 jobs in May, almost double the expected 88,000; March and April data were revised up by a combined 93,000. According to CME interest rate futures, the probability of a Fed rate hike within the year rose to 63%, with the probability of a hike before January next year nearing 100%. The Philadelphia Semiconductor Index (SOX) dropped over 10% in a single day, marking its biggest daily decline since March 2020; the Nasdaq fell 4.18%.

However, after analyzing interest rate expectations, technology valuations, crowding, and AI industry progress, the conclusions from multiple institutions are highly unified—before the continuation of the AI industry cycle and the slope of EPS revisions reverse, the core positioning of the technology style remains unshaken. As the mid-year earnings forecast season for A-shares and official earnings season for US stocks start at the end of June, the industry's ability to deliver results will undergo substantial testing.

Interest Rate Expectations: Historical Cases Show EPS Revisions Can Withstand Rate Shocks

GF Securities' strategy team pointed out in their June 7 weekly report, "There is no historical evidence that rising interest rates and liquidity contraction necessarily harm tech stocks or lead to growth stocks being devalued. This relationship mostly exists in the assumptions of discount rate models."

The team reviewed five classic cases: the US stock 1999 dot-com cycle, 2023 AI surge, and A-share cycles in 2013 mobile internet, 2017 supply-side reform, and 2021 new energy—reaching the same conclusion: When industry trends drive sustained EPS upward revisions for leading companies and robust performance is delivered, disturbances from interest rates and liquidity are only temporary, with no impact on medium-term stock price trends. Conversely, if price rises are mainly from valuation expansion rather than EPS, shocks from liquidity contraction tend to be more severe.

The historical reference from the 1999 dot-com cycle is the most convincing.

After the Fed's first rate hike in June that year, it raised rates six times in succession; the Dow Jones essentially traded sideways, while the Nasdaq surged about 91%, not peaking until March 2000. GF Securities pointed out, the root cause of this divergence was the EPS trend—Nasdaq 100 EPS grew 60% in 1999, while Dow Jones EPS grew just 19%; rising rates suppressed valuations of traditional low-growth sectors but did not block the performance explosion in growth sectors.

Returning to the present: Q1 2026 Nasdaq 100 EPS growth has risen to 36%, Dow Jones EPS growth has dropped to 4%, making the setup highly similar to 1999. GF Securities believes, given the latest industry trends, there is currently no evidence that EPS for core Nasdaq companies needs to be revised downward. Guosheng Securities (Xiong Yuan, Dai Kun) also commented on June 6, "This round of AI-driven tech rally has strong industry trends and performance support. We tend to think that after the adjustment, it should be another opportunity to position in tech stocks."

Technology Valuations: Investment Duration Determines Reference Framework

GF Securities pointed out that the starting point of valuation discussions lies in the set investment duration, and under different horizons, the effectiveness of valuation metrics is fundamentally different.

Historical data shows, over a one-year horizon, the PB level at the start of the year has no significant correlation to annual gains, but over three or five years, low PB effectiveness becomes apparent.

In sectors like electronics, communications, computers, power equipment, and defense manufacturing, over a one-year period, indicators like revenue growth and ROE change explain gains more than PE/PB valuation metrics, where correlation coefficients are "sometimes positive, sometimes negative," offering limited reference value.

In contrast, stable profit sectors like utilities, transportation, and home appliances—these sectors have limited profit elasticity, so the valuation when buying largely determines returns.

GF Securities also emphasized, The key to valuation judgment in growth investing lies in marginal changes in prosperity: during periods of accelerating growth, valuations often continue to rise, and immediate valuation levels are less meaningful; only when growth slows does valuation face downward pressure.

Historical data shows, during accelerating growth, median PB tends to rise; during decelerating growth, median PB tends to fall. Both rallies and sell-offs depend on the direction of prosperity, not absolute valuation levels.

Addressing the market's question of whether manufacturing companies' PBs over 40 signify a bubble, GF Securities cited 1990s cases: Dell's PB approached 50 in 1997, supported by extremely high ROE of around 80%; Cisco and Qualcomm reached PB peaks of 30–40 in 1999.

The team believes, high PB in growth companies is driven by extremely high ROE and hidden assets like R&D, patents, and customer barriers. For leading overseas computing companies in A-shares now, if high turnover and gross margins can be maintained, high PB plus reasonable dynamic PE does not necessarily mean the risk of a valuation bubble.

Guotou Securities noted that its constructed A-share "high-cut-low index" reached 60% as of June 5, touching the top range of three years of split market. Reviewing eight typical high-cut-low rounds since 2016, high sectors averaged a 12.8% decline; when tech chains led, declines expanded to 16.1%.

The team also noted that since May, this round of AI hardware rally has displayed differences from previous rounds: the overall market decline has not changed the clear core trend, and instead, group concentration is tighter; tech gains have not led other sectors upward, but have continuously siphoned from dividend and consumption styles, with small-micro cap indices down over 11%—seen as a typical sign that grouping has entered the second phase (concentration of funds).

However, Guotou Securities emphasizes, current AI tech is experiencing "left-side temporary high-cut-low", not the formation of a clear first top.

The key distinction: If the industry trend is intact, macro "grey rhinos" are unclear, and industry competition hasn't collapsed, then even short-term adjustments will be followed by a market return to the core trend. The team believes AI capital expenditure slowdowns (reassessment needed for 2027) are hard to disprove within half a year; observing all features of the first top, "the first top of AI tech pricing waves has probably not yet arrived".

AI Industry Progress Continues, Style Rotation Needs "Perfect Combination"

The core logic supporting the tech style—AI demand expansion—is still being sustained in the latest industry data.

According to GF Securities, OpenRouter platform token calls continue to break records; the Google I/O conference disclosed that by May 2026, Google's monthly token consumption reached 32 quadrillion, up more than sevenfold year-on-year. Anthropic's ARR was revised up to $44 billion in early May, then further to $47 billion, and major tech companies' capital expenditure guidance continues to be revised up. The Wall Street Journal previously reported that Anthropic may achieve operating profitability for the first time in Q2.

Dongwu Securities pointed out, because model giants like Anthropic keep revising up ARR, capital now finds it hard to question AI’s commercial potential; once positive signals pile up, a rebound can happen quickly, making it hard to give up core positions.

GF Securities’ market pricing framework gives two core judgments:

First, whether the tech rally ends depends on the AI industry cycle; monetary liquidity is just “the icing on the cake”.

Second, for style rotation (to consumption, pharma, finance), a "perfect combination" is needed—strong economy, accelerating inflation, entering a rate hike cycle, and AI growth peaking; all four must be present, and currently are not.

Dongwu Securities, comparing with the 2021 new energy rally, pointed out that peak was reached only after preliminary signs of weakening industry competition (intensifying price game in the PV chain, multi-supplier commercialization in battery) appeared, indicating that industry change itself is a more core end signal than liquidity. For now, no similar signs have appeared in AI infrastructure chain manufacturing.

Short-term Adjustments May Be Entry Window; Four Key Events Determine Next Phase

On the operational side, GF Securities suggests, starting from late June, as A-share interim forecasts, US stock official earnings, and A-share official earnings windows open in succession, the adjustment in overseas computing (optical modules, chips, PCB, fiber, etc.) in June may be another positioning window. Returning to "first principles," EPS and industry trend remain core drivers for stock prices.

Guosheng Securities highlighted four key upcoming observations:

First, developments in the US-Iran situation and oil prices, which will determine inflation trend and timing for rate cuts;

Second, US May CPI data (released June 10), watching for energy price transmission to core goods and services prices;

Third, major central bank meetings in mid-June; OIS market pricing indicates high probabilities of rate hikes by ECB and BOJ at their June meetings;

Fourth, Waller’s debut statement on June 18, whose judgement on inflation rebound and rate paths will be a key reference for market reassessment of policy prospects. If the Fed’s communication mechanism changes, market logic for pricing future rate paths will undergo restructuring.

Guotou Securities suggests, with unclear liquidity conditions, the tech segments lacking EPS support only offer contingent opportunities and higher risks; but AI infrastructure leaders with high yields, strong client stickiness, and supply chain control will show greater resilience in liquidity headwinds. Dongwu Securities also advises, after the market adjusts into a low volatility phase, leading stocks can be replenished. Key metrics to watch include Anthropic and other model giants’ ARR and token consumption growth. Absent clear industry termination signals, total retreat is not advised.

Risk Warning and DisclaimerThe market carries risks; investment requires caution. This article does not constitute personal investment advice, nor does it take into account individual users’ unique investment goals, financial situation, or needs. Users should consider whether any opinions, views, or conclusions herein fit their specific circumstances. Investments made accordingly are at one’s own risk. ```