After the AI frenzy, when will the productivity revolution take the baton?

Last week, U.S. stocks once again experienced a rollercoaster ride—the $S&P 500(.SPX)$ fell 2% from its record high in late October, while AI-related stocks saw a sharp pullback (-8%), marking their largest correction since Q1 this year. Investors were on a rollercoaster ride, fixated on $NVIDIA(NVDA)$ (earnings on November 19) and a host of consumer giants ($Home Depot (HD)$, LOW, TJX, $Target (TGT)$, $Wal-Mart(WMT)$ , ROST, WSM), torn between worries over AI investment trajectories and concerns about the health of American consumers' wallets.

This earnings season, discussions about AI among S&P 500 companies have quietly gained traction and become mainstream.

AI Discussions Heat Up: 47% of Companies Focus on Productivity, Communications and Finance Emerge as "Hardest-Hit Sectors"

On the broader trend: In Q3, 47% of S&P 500 companies specifically mentioned AI for "productivity and efficiency" during earnings calls—a significant jump from Q2. Why the surge? Goldman Sachs research shows that 37% of investment banking clients have already implemented AI, while Census Bureau data (though stringent, counting only "deep adoption") still reached 13%.

By industry, communication services (74%) and finance (66%) lead the pack—consider how $Meta Platforms, Inc.(META)$ and $Alphabet(GOOGL)$ leverage AI to optimize ad algorithms, while banks harness it for risk management. In contrast, utilities, energy, and materials saw mention rates barely exceeding 20%, with AI still feeling like a distant prospect for these sectors.

The Usual Three Use Cases + Quantified "Main Course": Coding and Customer Service AI-Powered, Profit Margins Quietly Rising

When companies discuss AI, which areas do they favor most? It's still the same old topics: coding/engineering (31%) and call centers/customer support (29%)—up 7pp and 5pp respectively from Q2. Marketing (23%), supply chain (19%), and HR/expense management (18%) also made the list, signaling AI's shift from back-end operations to front-line applications.

But what truly stings is the quantitative impact. Goldman Sachs singled out two "model students":

• $Snowflake(SNOW)$ (ServiceNow): AI-driven operational efficiency has prompted an upward revision of the annual operating margin target by 50 basis points, from 30.5% to 31%, equivalent to squeezing out hundreds of millions of dollars in additional profit.

• $C.H. Robinson Logistics (CHRW)$: Its Lean AI strategy propelled 2026 adjusted operating income projections from $220 million to $336 million, with the additional $116 million solely attributable to AI-driven margin expansion.

8% of companies have quantified "soft metrics": AI-generated code volume, AI-powered customer service agent share, and doubled document processing volume. These firms report average profit margins below the S&P median, yet their profit growth has outpaced the median over the past two years—AI acts like a "patch," helping inefficient businesses catch up. The issue is that most haven't dared to report specific profit contributions, leaving the market rife with doubts.

The Matthew Effect in Infrastructure Investment: Infrastructure Feasts While Productivity Scrapes by—When Will Phase 4 Turn the Tide?

Despite AI's widespread adoption, returns remain highly concentrated: Phase 2 infrastructure basket (chips, data centers) is up 35% YTD, outperforming the equal-weighted S&P 500 by 27 percentage points; Phase 3 (direct AI revenue beneficiaries, e.g., platforms) is essentially flat with the market; Phase 4 (productivity gains) lags by 12 percentage points—uncertainty remains high, with unclear distribution, magnitude, and timing.

EPS revisions confirm this: Starting in Q4, NVDA +6%, Phase 2 tech hardware +4%; Phase 4 is nearly zero. Investors favor "imminent" catalysts, and infrastructure is that "shot in the arm."

The return curve offers a clearer picture: Since June 2023, infrastructure investment has steadily climbed while productivity stagnated at rock bottom—yet the EPS curve reversed course, with Phase 4 surging 24% from late 2023 and outperforming the market by 9 percentage points. This reminds me of Alibaba Cloud's CapEx strategy: initial heavy spending yields slow returns, but later stages gain tremendous momentum.

"AI Productivity Dark Horse," HRB/CTSH Shine Bright

Given infrastructure risks (chip bans, geopolitics), Goldman Sachs screens AI productivity beneficiaries: Companies in the Russell 1000 where labor costs account for the top 25% of revenue, wage exposure to AI automation ranks in the top 25%, and AI efficiency was highlighted in Q2/Q3. Excludes Phase 2/3 companies, focusing on banking/IT services, etc.

Top 5 Average Rankings: HRB (H&R Block, Tax Services) RHI (Robert Half, Staffing) CTSH (Cognizant, IT Services) EPAM (EPAM, Software) IQV (IQVIA, Life Sciences) IQV (IQVIA, Life Sciences). This group is up 17% YTD (slightly underperforming the equal-weighted S&P 500's 23% gain), yet consensus EPS is projected to rise 24% (exceeding market expectations by 9 percentage points) — indicating significant room for valuation recovery.

Industry exposure charts reveal standout sectors: Software services (32% workforce, 42% exposure) and banking (30%, 38%) stand to benefit most—regulatory moats enable banking AI to "achieve twice the result with half the effort." High uncertainty? Indeed, software may become a "risk factor," but execution-driven players will dominate.

AI Platforms Take a Breather: SNOW, MDB, and Others See Revenue Momentum Build

Don't overlook Phase 3's "AI platform stocks"—databases/development tools that bridge infrastructure and applications. Goldman Sachs' picks like SNOW, MDB, and DDOG have rebounded strongly in recent weeks, with their revenue potential rising alongside the adoption of enterprise AI.

These stocks' EPS curves have turned upward, with a median YTD gain of +8%—consider the ecosystem impact of Alibaba's Qwen open-source model, where the platform serves as the central hub for "self-building chains."