McKinsey partners: you need to protect successful new ventures from the core business

AI is reducing venture experimentation costs and accelerating scaling: successful ventures now reach $10 million in revenue in 31 months on average versus 38 months previously, while breaking even with 40% less capital. Companies launching three or more ventures simultaneously can generate up to 30% higher revenue growth, and CEO-prioritized venture building can contribute nearly 20% of enterprise revenue within five years. Examples include Honeywell Connected Enterprise, which generated about $1.5 billion in 2023 sales and grew roughly 3x faster than Honeywell overall, and BCP's Yape, which has scaled beyond 18 million users.

Analysis

The investable implication is not generic AI upside but a widening valuation gap between incumbents that can monetize proprietary operational data through recurring software and those that merely absorb AI costs. HON has the installed base, domain workflow ownership, and distribution to turn AI-enabled offerings into higher-value service attach rates; the key upside would be software mix driving incremental gross-margin expansion and a higher-quality revenue multiple, rather than near-term AI revenue alone. The risk is that industrial customers treat AI features as bundled functionality, limiting price realization and making the spend defensive rather than accretive.

For HON, the next 1-3 months hinge on evidence of bookings, renewal rates, and margin contribution from its software/digital portfolio in earnings commentary—not venture-growth claims. Over 6-18 months, the strategic risk is disintermediation: hyperscalers and industrial-software rivals such as AVEVA/SCHNEIDER and Siemens can use broadly available models to commoditize the application layer unless Honeywell retains control of critical workflow data. A useful falsifier is digital revenue growth decelerating toward the legacy-company rate while segment margin fails to expand despite AI investment.

BCP’s optionality is more asymmetric but less directly attributable to public equity earnings: a scaled payments ecosystem can lower customer-acquisition costs, deepen deposit relationships, and create fee pools in merchant services, credit, and advertising. The counterpoint is that super-app scale is not automatically monetization; regulatory constraints, fraud losses, and competition from fintechs can convert user growth into elevated compliance and incentive expense. This is a watch item rather than a standalone AI trade absent disclosure on wallet take rate, transaction growth, credit losses, and funding intensity.

Consensus may over-credit companies for launching internal AI ventures while underweighting governance execution. Faster experimentation should raise the failure rate of weak projects as well as the number of winners; disciplined capital reallocation is therefore a positive signal, while rising corporate venture spend without disclosed commercial KPIs should be treated as margin and multiple risk.

Key Decisions for Investors

• Maintain or initiate a 6-12 month long HON position only on evidence that digital/software bookings and recurring revenue are accelerating faster than the core while segment margins expand; target a 10-15% upside from mix-driven multiple support, with a stop/review if digital growth converges with core growth for two reporting periods.
• Use HON versus a broad industrial proxy (long HON / short XLI) rather than an outright beta position if industrial cyclicality is the concern; the thesis requires software/service mix to outperform capital-equipment exposure over the next 2-4 quarters.
• Do not add BCP solely on reported wallet scale. Set an earnings-data alert for payment volume, merchant monetization, deposit retention, credit-loss trends, and operating-cost intensity; a sustained improvement in take rate without rising loss provisions would justify a 6-18 month long.
• Screen industrial peers including Siemens, Schneider Electric, Rockwell Automation, and Emerson for disclosed AI-linked recurring revenue and pricing. Favor firms reporting paid workflow adoption and renewal metrics; avoid those framing AI primarily as productivity investment without revenue or margin attribution.