US AI Software Stocks Surge 24% in One Month—Highest Gain in 15 Years as Commercialization Accelerates

Explosive Surge in U.S. AI Software Stocks: A Commercialization Inflection Point Behind the Strongest Monthly Performance in Fifteen Years

In May, the U.S. AI software sector achieved a historic breakthrough: the S&P Global AI Software Index surged 24% month-on-month—its strongest single-month performance in fifteen years, since the index’s inception in 2009. Snowflake soared 87.3%, Datadog jumped 87.6%, New Relic rose over 72%, and Palantir extended its momentum with another 21% gain. By comparison, the S&P 500 edged up just 1.8%, while the Nasdaq Composite rose 5.4%. This pronounced outperformance was not driven by short-term sentiment—but rather signals a pivotal shift: enterprise AI applications are moving decisively from labs to production lines, and from proof-of-concept (POC) pilots to scalable, revenue-generating deployments. The market is voting with real capital: the “software layer” of AI has already pierced through the fog of commercialization—and entered its profit-realization phase.

Weighted Leaders Rally Collectively: Index Composition Reflects Authentic Demand Shifts

The core catalyst behind this rally lies in the sustained strength of high-weight AI software leaders. Microsoft’s Azure AI services posted triple-digit year-on-year revenue growth for three consecutive quarters; its Copilot for Microsoft 365 now serves over 100 million commercial users, with average revenue per user (ARPU) for the commercial subscription reaching $22 in Q3—significantly above traditional Office 365. Oracle, meanwhile, leveraged AI-powered assistants embedded in Oracle Fusion Cloud ERP (e.g., Adaptive Intelligence) to deliver 31% cloud application revenue growth in FY24 Q4—with AI-driven cross-selling accounting for over 40% of new orders. Palantir’s Gotham platform completed operational deployment within the U.S. military’s Joint All-Domain Command and Control (JADC2) program, while simultaneously rolling out its AI decision-making core to European energy giants and Japanese manufacturing clients—propelling its commercial segment revenue up 52% year-on-year in FY24 Q1. These figures collectively confirm a fundamental shift: enterprises are no longer purchasing AI software for “technology tasting,” but to solve concrete pain points—including supply-chain forecast inaccuracies, low ERP process automation rates, and soaring compliance audit costs. In earnings calls, phrases like “customer renewal rate improved to 95%+” and “AI module paid adoption rate reached 68%” have become ubiquitous—underscoring that demand has evolved from “Should we adopt AI?” to “How deeply can we integrate it?”

High-Margin SaaS Model Regains Favor: A Structural Shift in Valuation Logic

The AI software surge reflects, at its core, a fundamental reassessment by capital markets of tech companies’ earnings quality. While hardware and compute-layer players retain strategic importance, they face structural constraints—including rigid capex requirements, heavy depreciation burdens, and margin pressure (GPU cluster operations account for over 35% of total costs). In contrast, AI-native SaaS firms—leveraging asset-light models, sticky subscriptions, and near-zero marginal costs—demonstrate superior earnings elasticity. Snowflake maintained a stable 76% gross margin in FY24 Q4; Datadog’s operating leverage became evident as its Non-GAAP operating margin turned positive for the first time—at 12.3%. Capital is rapidly reallocating toward such “cash-cow” models: in May, ARKQ increased its Datadog position by 18%; BlackRock’s iShares U.S. Technology ETF raised its AI software weight to 23.7%, a record high. This portfolio shift not only bolsters overall tech valuation resilience (the Nasdaq-100’s price-to-sales ratio holds at 8.2x—well above the 2022 low of 5.1x) but also pressures upstream suppliers to re-evaluate value distribution: chipmakers must now prove their AI accelerators directly lift customer lifetime value (LTV), not merely stack raw compute specs.

Dual Implications for China’s A-Share Market: Pressure and Benchmarking in Parallel

The U.S. AI software surge delivers a clear and urgent mirror to China’s A-share market. On one hand, it exerts valuation pressure: with Snowflake trading at 35x price-to-sales (PS), domestic peers such as Kingsoft Office (WPS AI) and Wanxing Technology (EdrawMax AI) trade at only 8–12x PS—prompting investors to ask, “Can Chinese AI applications achieve comparable commercial efficiency?” On the other, it offers benchmarking reference: Seres’ collaboration with Volcano Engine to co-develop a new smart vehicle brand essentially replicates Palantir’s “vertical-industry AI operating system” playbook—embedding foundation models across the full intelligent-driving stack and using data flywheels to accelerate model iteration, ultimately exporting solutions to overseas automakers. MiniMax’s initiation of A-share IPO tutoring further signals regulatory clarity and explicit support for listing “AI-native application companies.” Notably, the recent rollout of the Guangdong-Hong Kong-Macao Greater Bay Area’s yacht free-entry policy—superficially about cross-border transport convenience—actually provides a live-sandbox environment for AI-driven maritime intelligence (e.g., vessel trajectory prediction, real-time risk alerts), closely mirroring Datadog’s deployment of AI anomaly detection systems in the shipping industry.

Underlying Drivers of Accelerated Commercialization: Policy Coordination Meets Technical Maturity

This surge is no coincidence. U.S. Defense Secretary Hegseth, speaking at the Shangri-La Dialogue, emphasized that “mutual respect and communication between the U.S. and China are vital to world peace,” sending a positive signal on great-power friction management and technology supply-chain stability—easing investor concerns over geopolitical fragmentation of AI development. Meanwhile, Cuba’s agreement with the U.S. military to maintain command-level communication further underscores global critical infrastructure sectors’ acute need for AI-powered situational awareness and risk-prediction systems. Technically, the maturation of RAG (Retrieval-Augmented Generation) architecture has lifted private-knowledge-base query accuracy beyond 92%; MoE (Mixture of Experts) models have cut inference costs by up to 60%; and cloud providers—including Microsoft and Oracle—have deeply integrated AI toolchains into existing ERP/CRM systems, dramatically lowering enterprise AI adoption barriers. When technical readiness, policy certainty, and demonstrable ROI converge in a self-reinforcing triangle, AI software commercialization becomes irreversible.

The U.S. AI software sector’s strongest monthly performance in fifteen years functions as a prism—refracting both the global enterprise digital transformation’s deepening entry into the “AI-native” era, and the stark commercialization gap China’s tech industry still needs to bridge. As Snowflake’s code reshapes global data governance paradigms, A-share investors must look beyond valuation differentials—and instead focus on identifying domestic AI application companies with genuine “scenario penetration power” and “commercial闭环 capability” (i.e., end-to-end monetization capacity). After all, in the AI era, the future will be defined not by the flashiest algorithms—but by the most solid, sustainable revenue.