A correlation collapse in the public AI stock universe is the market condition Goldman Sachs Asset Management is betting AlphaAI can exploit. In the first half of 2025, price correlation among large public AI hyperscalers ran at roughly 80 percent. By late 2025 it had dropped to approximately 20 percent, a shift Goldman's own research tracked. The market had stopped treating AI as a single trade and started distinguishing between companies that could demonstrate revenue benefits and those that could not.
Goldman Sachs Asset Management, the investment management division of Goldman Sachs that oversees more than $4 trillion in assets under supervision as of Q2 2026, has launched AlphaAI, a platform that applies active research backed by private market data to identify specific sector-level beneficiaries of AI adoption rather than holding broad AI exposure. Lou D'Ambrosio, a Goldman partner who founded the firm's Value Accelerator in 2018, will chair the platform with the title Chairman of Artificial Intelligence for Asset Management, according to an internal memo seen by Reuters.
The $40.5 billion category of US-domiciled AI-themed ETFs represents the investment structure AlphaAI is positioning against. That category holds primarily the infrastructure and enablement layer of AI: chipmakers, cloud providers and software platforms. When correlation among AI stocks runs at 80 percent, that construction captures the trade. When correlation falls to 20 percent, whether a company in healthcare or financial services has actually deployed AI and begun realizing productivity gains becomes the variable that drives returns, and that variable does not appear in any public market data available at the time investors need to act on it. A peer-reviewed analysis of AI-themed ETF performance published in 2025 found that most funds exhibit return characteristics that closely track broad US technology indices, with limited differentiation.
D'Ambrosio told Reuters he expects greater dispersion within sectors that is not yet reflected in market prices. Goldman's 2026 technology investment outlook named selectivity as the essential response to continued divergence across the AI ecosystem. AlphaAI is the organizational structure built to act on that view.
The platform's stated data advantage rests on two inputs. The first is a combined view of public and private market investments. Private market data on AI-driven cost reductions, revenue changes and margin trajectories reaches Goldman before any of it surfaces in public disclosures. The second is operational intelligence from portfolio companies that Goldman said already have more than 100 scaled AI use cases. The Value Accelerator that D'Ambrosio built embedded a network of more than 100 senior operating executives as partners to portfolio company management, giving the platform ground-level visibility into how AI is actually changing a company's cost structure and sales motion before it reaches public filings.
Goldman shares rose approximately 3.2 percent in Thursday's regular-session trading on the day of the announcement. Darius Adamczyk, who joined Goldman in April 2025 as D'Ambrosio's co-leader of the Global Portfolio Operations Group, will take global leadership of the Value Accelerator as D'Ambrosio moves to the AlphaAI platform.
No assets committed to the platform, financial targets or performance benchmarks have been disclosed. Goldman is also co-leading the planned Anthropic IPO alongside Morgan Stanley and JPMorgan, targeting a listing as early as October 2026.
