AI Will Create 'Weird' Economic Benefits: Manidis

Will Manidis, co-author of “The Commanding Heights” and SVP of Veradigm AI, argues that as AI becomes cheaper and more widely available, companies that best deploy it in the real world may capture the largest economic gains. He says markets may be mispricing AI’s benefits and that technology creators will not necessarily be its biggest winners.

Analysis

The investable distinction is not “AI producers versus everyone else,” but who can retain the surplus after implementation costs and customer bargaining. AI adoption can lift output per worker without lifting incumbent margins: if tools are easy to buy and rivals can replicate them, productivity gains may be competed away through lower prices. The more durable beneficiaries are likely to be firms that control distribution, proprietary workflow data, and the ability to redesign operations—not simply those that announce pilots. Conversely, model and infrastructure providers could face commoditization or heavy investment requirements if capability improves faster than pricing power.

The essay offers a useful framework, not evidence of realized earnings impact. Near term, sentiment may favor AI infrastructure and software; over 1–3 months, the test is measurable adoption and guidance, while 6–18 months are needed to establish whether savings accrue to shareholders or customers. The contrarian risk is treating diffusion as automatically bullish for all incumbents: implementation expense, security constraints, and organizational friction can delay or dilute returns. There is no company-level data here to justify a directional position or valuation claim.

Key Decisions for Investors

• No trade on the interview alone. Build a watchlist of labor-intensive businesses with repeatable workflows and direct customer distribution; require evidence of realized productivity or margin improvement before adding exposure.
• For earnings diligence over the next 1–3 months, separate AI-related revenue from cost savings, and check implementation expense, workforce changes, customer retention, and whether price concessions are absorbing productivity gains.
• Treat AI infrastructure and application vendors as distinct exposures: monitor pricing, customer concentration, and investment intensity for infrastructure; monitor renewal rates and paid usage—not pilot announcements—for application providers.
• Falsify the diffusion-bull case if adoption remains pilot-heavy, implementation costs persist without measurable unit-cost improvement, or productivity gains show up mainly as lower customer prices rather than stronger margins.