Articles

The Stock Market’s AI Epoch

The Stock Market’s AI Epoch: Recent Performance, What’s Driving It, and What May Come Next

Executive summary
  • Market leadership remains highly concentrated in AI beneficiaries. Record highs in late 2025 mask a “split market” where outsized gains are driven by a narrow group of AI-exposed mega caps and adjacent suppliers. 
  • Earnings and capex are the engines. AI leaders are contributing disproportionately to S&P 500 earnings growth as hyperscaler capex soars and is expected to remain elevated into 2026. 
  • Valuation and breadth have shifted. The “Magnificent Seven” (and a new cohort of AI winners) still represent an unusually large share of index value, though leadership within the group has evolved. 
  • Near-term (≈6 months): Base case favors continued resilience for core AI leaders, but sensitivity to earnings, supply chains (compute), and policy is high.
  • 2–3 years: If productivity gains materialize, AI leaders likely outgrow the market; broader S&P 500 outcomes hinge on diffusion of AI beyond the current cluster and the trajectory of rates and capex.
1) Where markets stand: 2025 in context

The S&P 500 has printed new highs with leadership concentrated in AI-linked technology and suppliers, while many non-AI cohorts have lagged. Several pieces of evidence point to this concentration:

  • Analysts describe a market “split” in which the headline index is strong but advances are driven by AI-powered names. 
  • The Magnificent Seven’s footprint in the S&P 500 has swelled from ~12% in 2015 to ~37% by Oct-2025, underscoring index concentration. 
  • Leadership within that group has rotated: NVIDIA achieved world-leading market capitalization in 2025 as the key supplier to AI build-outs, while others have seen more mixed performance. 

At the single-name level, NVIDIA has symbolized the era, reflecting investor conviction that AI compute remains the critical bottleneck input. 

2) What’s powering the leaders: earnings + capex flywheel

Two reinforcing forces explain why AI leaders have outrun the market:

  1. Earnings leadership. For Q2-2025, FactSet highlights that several AI leaders (NVIDIA, Microsoft, Alphabet, Broadcom) were among the top contributors to S&P 500 earnings growth, not just price action. 
  2. Capex supercycle. Goldman Sachs and others expect hyperscaler AI capex to remain a dominant use of cash, with investment measured in the hundreds of billions through the mid-2020s and potentially >$1T by 2027, supporting component suppliers and cloud platforms. 

Strategically, this aligns with independent research (McKinsey) that continues to rank GenAI and related compute among the most impactful enterprise technology trends and a major potential contributor to productivity.

3) Leaders vs. early adopters: how performance has differed

AI leaders (illustrative cohort: NVIDIA, Microsoft, Alphabet, Amazon, Meta, Broadcom, AMD, Oracle, Palantir) have benefited from either:

  • Direct monetization of AI (chips/accelerators, cloud AI services, model-served advertising/engagement, AI software platforms), or
  • Indirect operating leverage (workflow automation, content personalization, developer productivity).

Coverage through 2025 increasingly notes a broadening beyond the original Magnificent Seven, with Broadcom, Oracle, Palantir among the new large-cap contributors to AI-driven gains—while some prior leaders have cooled.

By contrast, early-stage or low-exposure adopters across defensives or structurally challenged industries have not captured the same re-rating, and several high-profile non-AI or slower-to-monetize tech names have underperformed during 2025. (Bankrate’s running lists and similar trackers show many notable laggards year-to-date.) 

Why the gap? Markets are rewarding visible AI revenue/earnings today (in silicon, cloud, and model-driven platforms) over potential long-dated AI benefits in companies still retooling.

4) Six-month outlook (through ~Q2 2026): what’s reasonable to expect

Base case (most likely):

  • AI leaders remain supported by hyperscaler capex, product cycles (new accelerators/servers), and expanding software monetization. Earnings beats remain possible but tougher comps and any deceleration in capex plans could inject volatility. 
  • Early adopters / lower exposure likely continue to trail unless they show hard ROI cases (clear AI productivity gains in margins or growth).
  • Index level: The S&P 500’s path is still highly sensitive to a small number of AI heavyweights, given their ~one-third weighting; any stumble by the core group can sway the index.

Bull case (less likely, but plausible):

  • AI demand and orders surprise to the upside; supply chains (power, racks, advanced packaging) ease faster than expected; AI software monetization inflects beyond tech into vertical industries, improving breadth.

Bear case (tail risks):

  • Capex moderation arrives sooner than forecast; major deployment bottlenecks (power constraints, permitting); policy/regulatory shocks; or macro slowdown compressing multiples.

Bottom line (6 months): Leaders > market > early adopters, but with higher volatility around earnings/guide updates. The market’s narrow leadership introduces path dependency.

5) Two–three year view (to 2027–2028): diffusion vs. bottlenecks

Thesis: Over multi-year horizons, performance should converge toward companies that (a) sell the inputs to AI at scale, (b) monetize AI natively in software/platforms, or (c) prove measurable productivity uplift in non-tech sectors.

Supportive dynamics

  • Research houses continue to estimate multi-trillion-dollar productivity potential as AI diffuses across workflows; if even a fraction is realized, earnings power for AI leaders should compound faster than the market. 
  • Capex durability (even with growth slowing from 2025 peaks) still implies a large, recurring spend base as models scale and refresh cycles occur. 

Risks to watch

  • Hardware saturation or faster obsolescence compressing returns for some suppliers;
  • Power/energy constraints limiting datacenter expansion;
  • Regulatory constraints on data/model use;
  • Competitive dynamics (open-source vs. closed, model commoditization) shifting value from chips to software—or vice-versa.

Bottom line (2–3 years):

  • AI leaders: Higher probability of above-index EPS growth if they continue to control choke points (compute, platforms, distribution).
  • S&P 500 overall: Outcomes hinge on diffusion—the more non-tech sectors turn AI into measurable productivity, the more index earnings broaden beyond today’s narrow leadership.
6) What investors and operators can do now
  1. Segment exposure explicitly: “AI inputs” (semis/compute, power, networking), “AI platforms” (cloud/model APIs), and “AI adopters” (end-markets).
  2. Underwrite with evidence: Favor names showing line-of-sight monetization (units shipped, contracts, AI ARPU uplift, attach rates) vs. slide-ware.
  3. Mind concentration risk: The S&P 500’s dependence on a handful of firms cuts both ways. Position sizing and hedging matter. 
  4. Track capex and earnings cadence: Hyperscaler capex updates and quarterly guides have been the highest-signal data in 2024–2025—and likely remain so through 2026.