How Far Has the AI Rally Progressed? CICC Examines Earnings, Valuations and Market Positioning

Since late March, global equity markets have been driven sharply higher by the AI theme. U.S. stocks, China’s ChiNext, and markets in Japan and South Korea have all strengthened, while the Philadelphia Semiconductor Index surged 65% in just one month—far ahead of the Nasdaq’s 26% and the S&P 500’s 17%. Although easing geopolitical tensions and improving sentiment contributed to the rebound, the decisive factor was the exceptional first‑quarter performance of technology companies. Without this earnings support, the market could easily have remained stuck at low levels, similar to the Hang Seng Tech Index, which has struggled due to a lack of industrial momentum.

Technology’s contribution to the U.S. market has been overwhelming. Information technology and communication services accounted for 77% of the S&P 500’s gains and 67% of its first‑quarter earnings growth, and contributed more than half of actual GDP expansion. In effect, AI has dominated market performance, corporate earnings, and economic growth.

This strong comeback surprised many investors. Only one or two quarters ago, concerns about an AI “bubble,” the impact on software and employment, and risks in private credit were widespread. These issues have not disappeared—April’s U.S. non‑farm payrolls show employment in information technology and finance continues to contract—but the market’s emotional swings have amplified both pessimism and optimism.

CICC notes that when bubble concerns peaked last November, its analysis concluded that the market resembled the 1996–1998 phase of the internet era rather than the 2000 bubble. Earlier this year, the firm also highlighted that China and the U.S. differ significantly in AI resource endowments and investment structures, with the infrastructure layer offering higher short‑term certainty and the technology and application layers offering greater long‑term potential.

With AI now at the center of market attention, CICC believes three questions matter most:
How strong is AI‑driven earnings growth? How far is the market from a bubble? And how much further can the AI rally extend?

AI Earnings: The Core Driver of Q1 Surprises
The first quarter delivered the strongest earnings growth since 2021. With 90% of S&P 500 companies reporting, earnings rose 28% year‑on‑year. Communication services and information technology led the gains, rising 51% and 48% respectively, while semiconductors and equipment surged 99%.

Tech companies also delivered the broadest earnings beats. The share of S&P 500 companies exceeding expectations rose to 85%, with information technology achieving a 96% beat rate. Communication services delivered the largest beat margin at 55%, supported by Google and Meta.

Technology contributed nearly 70% of total earnings growth. The Big 5 cloud companies—Microsoft, Amazon, Google, Oracle and Meta—alone accounted for 40% of overall earnings expansion. Revenue growth drove gains in semiconductors and hardware/software, reflecting strong upstream demand from AI infrastructure investment. Meanwhile, Google and Meta benefited from margin expansion driven by improved advertising algorithms and platform efficiency.

Capital expenditure was another highlight. Big 5 capex surged 91% year‑on‑year to USD 148 billion, lifting total S&P 500 capex to USD 381 billion, up 36%. Even excluding the Big 5, corporate capex accelerated from 9.6% to 15%.

CICC estimates that U.S. AI‑related capital expenditure could rise 70% this year to USD 730 billion. Data centers and supporting infrastructure remain the largest component at USD 570 billion, while model‑layer investment—compute procurement and training/inference costs—may rise from USD 45 billion to USD 130 billion, becoming the fastest‑growing segment.

AI investment has also become a major driver of U.S. GDP. Tech hardware and software investment reached an annualized USD 1.2 trillion in Q1, contributing 1.1 percentage points to the quarter’s 2% real growth. Even after adjusting for imports, the tech sector contributed 30% of real GDP expansion.

Where the Market Stands: A Comparison to the Internet Era
CICC evaluates the AI cycle across demand, investment intensity, and market valuation, concluding that the current phase resembles a blend of late‑1990s conditions.

Demand is comparable to 1998–1999.
U.S. labor productivity has risen 8.7% since 2023, faster than during the internet boom. AI adoption continues to rise, especially in enterprise paid scenarios, with AI Agents expanding from content generation to task execution. AI usage across U.S. industries has reached 20%, with IT and professional services leading.

Investment intensity is closer to 2000.
Big 5 capex now equals 94% of operating cash flow, approaching the threshold where external financing becomes necessary. However, leverage remains far below dot‑com levels, giving companies room to expand balance sheets. Differences across firms are notable: Meta, Microsoft and Google maintain healthy cash flow; Amazon’s free cash flow has turned negative; Oracle faces the greatest financing pressure but is shifting part of its infrastructure burden to customers through prepayment contracts.

Venture capital activity also resembles the late‑1990s. U.S. VC investment reached USD 267 billion in Q1, equivalent to 0.3% of total market capitalization—close to 1999–2000 levels. AI now accounts for 85% of VC funding, surpassing the 64% peak for information industries during the dot‑com era.

Valuations are near 1997–1998 levels.
The S&P 500 forward PE has rebounded to around 21x, while the Nasdaq trades at 29x—below the 33x peaks of 2023–2024. Big 5 cloud companies trade at 24x, also near late‑1990s levels. Historically, valuations above 30–31x have signaled short‑term correction risk.

How Much Further Can the AI Rally Go?
CICC believes the market has not entered a broad bubble, but investment has run ahead of demand, leading to periodic corrections. Since 2023, AI rallies have typically followed a pattern of two strong quarters followed by one weaker quarter as the market waits for new catalysts.

This year, leadership has shifted from cloud and chips to bottleneck segments such as memory and optical modules. Memory stocks have risen 197% and optical modules 103%, while chips gained 23.5% and cloud stocks only 3%, underperforming the S&P 500’s 8%.

This rotation reflects a shift in market focus from capital expenditure expansion to order visibility, earnings delivery, cash‑flow pressure and return on investment. High‑valuation segments such as semiconductor equipment, optical modules and power/cooling require strong earnings to justify current pricing. Cloud and chips, by contrast, still trade at relatively low valuation percentiles.

CICC expects the July Q2 earnings season to be the next major turning point for the AI trade.

Beyond AI fundamentals, macro factors—U.S. liquidity, Treasury yields, IPO activity, and geopolitical risks—will influence market direction. Under the baseline scenario, AI momentum continues and internal rotation within tech remains the main strategy. If geopolitical tensions ease, capital may rotate into cyclicals and broader AI‑related sectors. If tensions escalate or liquidity tightens, high‑valuation tech may face short‑term pullbacks.

CICC maintains a year‑end S&P 500 target of 7600–7800, upgraded after strong Q1 tech earnings offset the drag from earlier high oil prices.