- Institutional Conviction: Goldman Sachs maintains that the 2026 AI rally is fundamentally different from the 1990s dot-com bubble, underpinned by unprecedented earnings per share (EPS) growth rather than mere speculation.
- The Infrastructure Pivot: Investment logic has shifted from primary chipmakers to the “Third Wave” of AI: utilities, liquid cooling providers, and specialized power grid hardware.
- Agentic Economy ROI: Markets are now prioritizing “Empowered Users”—companies successfully deploying autonomous agents to slash operational costs and expand margins.
The skepticism that once characterized the high-valuation era of artificial intelligence has largely been silenced by the balance sheets of 2026. While the “bubble” narrative resurfaces with every market fluctuation, institutional heavyweights—led by Goldman Sachs—argue that we are witnessing a structural realignment of the global economy, not a transient fad. This rally is no longer fueled by the promise of what AI might do, but by the measurable ROI of what it is currently doing to corporate margins across the S&P 500.
The ROI Mandate: Why 2026 Valuations Hold Firm
Unlike the speculative fervor of the early 2020s, the current market climate demands proof of performance. Goldman Sachs’ analysts point to a critical divergence: while P/E ratios remain elevated, they are closely tracked by aggressive earnings growth. The “AI ROI Gap” that concerned investors in 2024 has been bridged by widespread enterprise adoption of agentic workflows.
For instance, the rise of machine-to-machine economies has allowed startups like Natural, which recently raised $30M for AI agent payments, to challenge legacy financial rails. This shift toward autonomous transactional systems is a core reason why Goldman believes the software and “Empowered User” sector has significant room for expansion. The focus has moved from training models to executing tasks, turning AI from a capital expenditure into a primary revenue driver.
The Three Pillars of the 2026 Rally
- Earnings Quality: Top-tier AI beneficiaries boast higher net margins and stronger balance sheets than the median S&P 500 company.
- Capital Expenditure Resilience: Hyperscalers (Microsoft, Alphabet, Amazon) continue to increase Capex, signaling long-term conviction in the underlying infrastructure.
- Scarcity Value: As energy and cooling become bottlenecks, companies with secured supply chains command a massive premium.
The Shift to Energy and Infrastructure Bottlenecks
In the 2026 landscape, the conversation has evolved beyond just GPUs. Goldman Sachs has expanded its “AI Basket” to include the silent enablers of the revolution: the power grid and thermal management. As data centers consume an ever-increasing share of the global energy supply, utilities like NextEra Energy and hardware specialists like Vertiv have become as vital to the AI trade as the semiconductor giants themselves.
This physical infrastructure requirement creates a “tech moat” similar to what we see in high-end entertainment and manufacturing. Just as Imax utilizes proprietary laser projection to dominate the premium cinema market, AI infrastructure providers are leveraging specialized liquid cooling and modular power units to lock in long-term contracts with hyperscalers.
Goldman’s Updated “AI Winners” List
| Category | Key Players (2026) | Strategic Focus |
|---|---|---|
| Semiconductors | Nvidia, Marvell, ARM | Custom silicon & chiplet architecture. |
| Cloud & Power | Microsoft, NextEra, Vertiv | Energy-efficient compute & grid stability. |
| Empowered Users | Salesforce, Adobe, Meta | Agentic AI integration & ad-tech optimization. |
Sovereign AI and the Geopolitical Premium
A burgeoning factor in Goldman’s bullish outlook is “Sovereign AI.” Governments worldwide are no longer content to outsource their intelligence needs to a handful of Silicon Valley firms. Massive state-sponsored investments in domestic data centers and local-language models are creating a secondary layer of demand for hardware and consultancy.
This localized growth is mirroring trends in other high-priority supply chains. For example, as logistics giants race for cold storage to support the GLP-1 boom, AI infrastructure is seeing a parallel surge in specialized, geographically diverse facilities. Goldman argues that this diversification reduces the risk of a single-point failure in the AI trade, providing a “floor” for valuations even during broader market volatility.
Conclusion: A New Phase of Outperformance
The core of the Goldman Sachs thesis is that we are in the “Phase 3” of the AI cycle. The first phase was the discovery (Nvidia’s initial surge); the second was the infrastructure build-out (Cloud providers); and the current phase is the integration of productivity. According to official Goldman Sachs Global Investment Research, the potential for AI to boost labor productivity could eventually add nearly 1.5 percentage points to annual US GDP growth over a 10-year period.
For investors, the message is clear: while the easy money in “pure-play” AI may have been made, the broader structural rally is just beginning. By focusing on companies that bridge the gap between digital intelligence and physical infrastructure, Goldman Sachs believes the path for continued outperformance is not only justified but mathematically probable.
