How Goldman Sachs’ AlphaAI Targets the K‑Shaped Split in US AI Stocks

AlphaAI’s purpose and shift in investment logic

Goldman Sachs Asset Management recently launched the AlphaAI platform to move from "theme investing"—holding AI‑focused ETFs and funds with about $40.5 billion of assets—to "stock‑selection investing" that aims to identify companies that can actually turn AI into profit. The platform is designed to bridge the gap between beta exposure and alpha capture.

Integrating public‑ and private‑market data

AlphaAI combines data from Goldman’s public‑market investment teams with insights from its private‑market teams and the operational experience of more than 100 scaled AI‑application cases. By breaking the traditional data silos between public‑ and private‑equity units, the platform uses private‑market "real‑world" experience to improve public‑market stock‑selection judgments.

Anticipating internal industry divergence

The platform argues that AI will create a larger performance split within each industry, a split that is not yet fully reflected in stock prices. Companies with strong AI‑application capabilities are expected to see valuation gaps widen, creating an active‑management window for excess returns.

The AI‑spiral closed‑loop

The article describes a circular flow of capital in the US AI market:

Nvidia : commits up to $100 billion to OpenAI (partly tied to deployment milestones); OpenAI purchases large volumes of Nvidia chips; Nvidia also invests in new‑cloud firms such as CoreWeave and xAI, which in turn rent GPU capacity to OpenAI. By 2026, Nvidia‑related transaction volume exceeds several trillion dollars.

Microsoft : has invested over $13 billion in OpenAI and holds roughly 27 % equity; OpenAI promises to buy about $250 billion of Azure cloud services.

Amazon / Oracle / AMD : Amazon expands its AWS contract with OpenAI; Oracle has a roughly $300 billion cloud‑infrastructure agreement with OpenAI; AMD supplies chips to OpenAI and may receive equity or warrants.

Other loops : Microsoft and Nvidia jointly invest in Anthropic, which then commits to using Azure and Nvidia compute; similar structures appear with CoreWeave, xAI, and others.

Bloomberg calls these "circular deals"—investments where Company A funds Company B, and Company B uses that money to purchase products or services from Company A. While not illegal round‑trip fraud, they amplify apparent demand and revenue.

AlphaAI’s timing and market signals

AlphaAI launched in July 2026, marking a transition from "infrastructure‑theme hype" to genuine winner‑selection within the AI sector. Goldman observed that AI‑stock correlation fell from about 80 % in early 2025 to roughly 20 % later that year, indicating the market is beginning to separate firms that truly convert AI into profit from those that merely enjoy thematic exposure.

By leveraging combined public‑private data and over 100 large‑scale use cases, AlphaAI seeks to capture this emerging split early, signaling a shift from capital‑driven expansion to productivity‑driven returns.

Outlook

The article concludes that as the AI‑spiral loop matures, the industry’s internal differentiation will intensify, creating clear winners and losers once the circular capital is re‑priced by the market.