Key Points:

1. Under the impact of Trump-era tariffs in 2025, the Mag-7 led the market decline due to excessive valuation premiums and skepticism

over the sustainability of AI investments.

2. Resilience in AI-related capital expenditures and outperformance in cloud businesses may now serve as the core driver for the Mag-

7’s rebound.

3. However, long-term risks such as tariffs and intensifying industry competition cannot be overlooked.

In the first half of 2025, the U.S. stock market experienced a valuation reset triggered by a combination of factors, particularly Trump’s tariff

policy. The “Magnificent 7” (Apple, Microsoft, Google, Meta, Amazon, Nvidia, and Tesla), which had driven the bull market over the past

decade and soared on the back of the AI boom, were also the hardest hit amid concerns about excessive valuations and doubts over the

sustainability of aggressive AI investments.

However, market sentiment appears to be shifting between the volatility of Trump’s trade policy and a round of earnings surprises. Since the

rebound after April 8, the Mag-7 have rallied 14.8%, outperforming the S&P 500’s 10.9% gain. This suggests that investors are not ready to

give up on the group entirely. The resilience in AI-related capital expenditures and better-than-expected cloud earnings are emerging as

central to this market turnaround.

Apple stock data. Source: Futu NiuNiu. Data as of May 6, 2025

The Value Reset: Shaken AI Belief

In early 2025, the AI frenzy appeared to cool. Enthusiasm for infrastructure-focused AI capital expenditures rapidly declined, especially after

the release of China’s open-source DeepSeek model on January 24. With capabilities comparable to GPT-4, DeepSeek shattered the myth

of U.S. dominance in AI technology and sparked concerns over the sustainability of U.S. tech giants’ unchecked AI spending.

Meanwhile, Trump’s announcement on April 2 of retaliatory tariffs directly disrupted the cost structures and markets of Mag-7 firms. Apple’s

Q1 2025 earnings revealed an additional $900 million in costs due to tariffs. Amazon, meanwhile, cut its full-year guidance, citing tariffs

and weakened consumer demand.

Expectations of sharply rising supply chain costs further clouded the investment case for AI hardware and data center expansion. Key

components such as GPUs and power infrastructure faced cost spikes, exacerbating concerns about the return cycle of AI investment. The

combination of policy shocks and rising tech competition triggered the collapse of “AI faith.”

From a valuation perspective, the Mag-7’s forward P/E ratio fell from 30x at the beginning of the year to 21.7x by early April. Though still

above the S&P 500’s overall 19.2x, the compression is notable. The price-to-sales (P/S) ratio also declined from 7.7x to 5.6x. The Roundhill

equal-weighted Mag-7 ETF (MAGS) plunged nearly 27% between February and early April, far worse than the S&P 500 ETF (SPY), which

lost 19% over the same period.

Roundhill Mag-7 ETF data. Source: Futu NiuNiu. Data as of May 6, 2025

AI Returns Highlight Earnings Resilience

Despite tariff pressures, the Mag-7’s earnings reports have sent a key message: AI investments are translating into actual revenue growth,

rather than merely being a cost sink.

Recent earnings from tech giants show that AI-related businesses are maintaining strong momentum. Google Cloud posted Q1 revenue of

$12.26 billion, with operating margins rising from 9.4% to 17.8%. Gemini, Google’s LLM, reached 350 million monthly active users in March.

Alphabet also announced a 43% YoY increase in 2025 capital expenditures to $75 billion.

While Apple was hurt by higher hardware costs, Microsoft’s Azure cloud business offset tariff-related impacts thanks to robust AI demand.

Azure revenue grew 33% YoY, with AI services contributing 16% of that growth. Microsoft also guided for higher capital spending in Q4,

with CEO Satya Nadella noting that “software is the best weapon against inflation.” Meta and Amazon remained resilient in advertising and

cloud respectively, and both raised their 2025 CapEx guidance.

Nvidia, despite growing competition from Google’s TPUs and Apple’s in-house chips, continued to see strong data center GPU demand.

Major clients like Meta and Microsoft have yet to fully utilize their compute capacity and continue to scale up, with AI infrastructure

accounting for over 50% of 2025 CapEx plans.

According to Bloomberg, the Mag-7 are expected to post 21.6% earnings growth and 9.7% revenue growth in 2025—both revised upward.

Importantly, their forward profit margin has climbed to 26.2%, more than double the 11.9% average for the rest of the S&P 500, highlighting

tangible returns on AI investments.

Valuations Cooling, But Growth Momentum Persists

Notably, while the Mag-7’s weight in the S&P 500 has dropped from 32% to 28.4%, their contribution to index earnings has risen to 22.6%,

with a revenue share of 11.8%. This suggests that although valuation multiples have compressed, their structural advantages in earnings and

revenue have strengthened.

Meanwhile, the rest of the S&P 500 (excluding the Mag-7) trades at a mere 2.1x forward P/S, reflecting a lack of growth premium in non-

tech sectors. In other words, the Mag-7 remain rare and relatively certain growth assets in the current market.

In addition, recent signs from the White House of possible tariff moderation have helped fuel a two-week V-shaped market rebound. The

S&P 500 posted a rare nine-day winning streak, pushing investor sentiment higher and sending the VIX fear index down to 22.68. While

concerns about a tariff-induced recession remain, investor confidence is gradually returning, with the Mag-7 leading the charge.

S&P 500 Index data. Source: Futu NiuNiu. Data as of May 6, 2025

Conclusion

The recent rally in the Mag-7 likely reflects a sentiment reset following the absence of worst-case outcomes. Earnings resilience from AI

capital spending has stabilized market expectations around the sustainability and return potential of AI investments. At the same time,

continued increases in AI spending by tech giants have injected fresh momentum into the AI cycle.

However, long-term risks remain. First, if the White House escalates tariffs further, hardware and e-commerce players will bear the brunt.

On May 4, Trump announced a 100% tariff on all foreign-made films entering the U.S. and hinted at future levies on pharmaceuticals. This

may mark the start of a new phase in Trump’s trade war.

Second, the rise of low-cost models like DeepSeek and Hong Kong-listed tech firms such as Alibaba and Tencent is beginning to attract

global capital. If AI dividends begin to materialize more rapidly in China, markets may reassess the valuation foundation of U.S. AI plays. The

AI race has now shifted from a “narrative gap” to a “delivery gap.” Companies that fail to turn AI into real profits and efficiency will face

continued valuation compression.