Anthropic, OpenAI Can Both Win AI Race: T. Rowe Price

T. Rowe Price partner Emma Norchet says Anthropic is moving into “act two” of its growth story, expanding from coding toward an AI layer for businesses to execute entire workflows. Norchet says T. Rowe Price has multi-billion-dollar positions in both Anthropic and OpenAI; the article excerpt provides no further details on investment size or market impact.

Analysis

The investable implication is not simply “more AI demand”: if enterprises move from coding assistants to workflow execution, value could shift from model access toward orchestration, permissions, data integration, and reliable completion of business tasks. That could expand monetization per customer, but also raise deployment, inference, and liability costs; adoption is only economically attractive if measured labor savings exceed those costs. Incumbent software vendors may be distribution winners if they embed models into existing systems, while standalone tools risk being bundled or displaced. Cloud providers could benefit from greater usage, although model-provider bargaining power and compute efficiency determine how much economics they retain.

The near-term signal is weak: the article provides an investor’s thesis, not evidence of customer conversion, recurring revenue, unit economics, or realized productivity. The cited private positions do not establish exposure on T. Rowe Price’s consolidated balance sheet; vehicle, ownership, and valuation details need verification. Over 1–3 months, watch enterprise deployment disclosures and evidence of paid workflow usage. Over 6–18 months, the key test is whether deployments scale beyond coding without deteriorating inference economics. A credible workflow shift would support AI software and infrastructure exposure; weak renewal, security incidents, or falling cost savings would reverse it. No direct public-equity trade is justified from this excerpt alone.

Key Decisions for Investors

• Do not treat the interview as a standalone buy signal for AI equities or as evidence of T. Rowe Price balance-sheet exposure. Verify which investment vehicles hold the positions and how they are valued before drawing implications for T. Rowe Price.
• Put enterprise software and cloud names on a watchlist, but require evidence of paid workflow deployments, renewal/expansion, and customer productivity before adding exposure. Favor distributors able to integrate models into existing workflows over undifferentiated model-access claims, subject to company-specific evidence.
• Monitor quarterly disclosures and customer case studies over the next 1–3 months for workflow usage, revenue contribution, and inference-cost trends. Falsify the thesis if deployments remain pilots, customers do not expand usage, or security/reliability issues delay production adoption.
• Keep any AI exposure sized for private-market valuation and commercialization risk: the excerpt offers no entry price, valuation, or independently verifiable financial metrics, so a direct Anthropic/OpenAI valuation conclusion is not supportable.