Harvard's Judgment Professor: Numbers Don't Make Decisions; People Do

Harvard Business School lecturer and serial entrepreneur Reza Satchu argues that executive judgment, rather than data or raw intellect, will become the key source of investment and business outperformance as AI commoditizes information. Satchu cited his decision to reject a roughly $1.1 billion offer for Alignvest student housing, later selling the business for $1.7 billion, as an example of calibrated risk-taking. He advises investors to assess management judgment through leaders' decisions and organizational responses during crises rather than relying solely on financial statements or polished public narratives.

Analysis

This is not a fundamental catalyst for the named equities; its investable implication is a governance screen as AI compresses information and implementation advantages. The market is likely to reward management teams that convert AI into faster product cycles, sales capacity, and lower service costs rather than merely reporting copilots or pilots. NET is better positioned than most software peers because sustained execution against large-enterprise security/network consolidation can validate AI as a distribution and operating-leverage tool; CRM needs evidence that AI monetization lifts net retention and subscription attach rather than raises compute costs.

The second-order risk is that AI commoditizes features faster than it creates pricing power. That favors platforms with proprietary workflow data, distribution, and low customer churn, while putting multiple pressure on application vendors whose differentiation is primarily user interface or generic analytics. For NVDA, the relevant management-quality variable is customer capital-allocation discipline: hyperscaler demand remains supportive near term, but a shift from experimentation to ROI-based inference budgets could reduce the market's tolerance for inventory, lead-time, or gross-margin disappointments.

Over the next 1-3 months, treat executive behavior during guidance resets, major outages, product delays, or M&A integration as a higher-signal input than prepared AI commentary. Over 6-18 months, private-market sponsors BX and TPG may benefit if AI-driven operating plans expand EBITDA at portfolio companies, but their realization values remain more sensitive to rates and exit-market liquidity than to operational rhetoric. Contrarian view: the market may be underpricing execution dispersion within software, but broad AI-beta longs are no longer a clean expression of that insight.

For GS, the takeaway is mildly negative at the margin: direct founder-led processes can pressure advisory fee pools in sponsor-backed transactions, especially where repeat buyers and owners have strong sector knowledge. This is structural but slow-moving; complex cross-border, public-company, and financing-heavy deals remain resistant to disintermediation.

Key Decisions for Investors

• Maintain a 6-12 month quality-AI software bias via long NET versus short an equal-dollar basket of lower-retention, AI-feature-led SaaS peers; use NET's next earnings to require accelerating large-customer growth and stable/improving gross margin. Exit if net retention weakens materially or sales-and-marketing intensity rises without bookings conversion.
• Do not add directional NVDA exposure solely on AI narrative strength. Add only after a results-driven pullback if hyperscaler capex guidance remains intact and NVDA demonstrates that inference mix is not diluting gross margin; a 10-15% downside from entry versus renewed earnings revisions offers a more favorable setup than chasing momentum.
• For BX/TPG, prefer BX on a 6-18 month horizon if credit issuance and realization activity improve; BX's scale in private credit and perpetual capital provides better fee durability. Falsify on widening private-credit losses/defaults or a renewed decline in realizations/distributable earnings.
• Flag GS advisory revenue and backlog at the next two reporting periods rather than initiate a short. A sustained decline in sponsor M&A fees despite improving announced-deal volumes would support a structural-intermediation short thesis; absent that evidence, the signal is insufficient.