Research & Commercial Insight · The Universal Owner · July 15, 2026
For a decade the universal owner's central problem was a surplus problem: too much capital chasing too few durable returns. That framing is too literal to survive this week's evidence — capital is still scarce or expensive for plenty of borrowers. The sharper, testable version is this: for a specific set of strategic megaprojects, the scarce input has shifted from money to the capacity that converts money into output. Four developments illustrate it; none alone proves a regime change, but together they define a frontier worth underwriting.
The enterprise budget
IBM's preliminary Q2 is the cleanest case. Clients pulled late-June spend into supply-constrained servers, storage and memory, and numerous large deals missed their expected close. That is a mechanism, not a verdict: software still grew 5%, Red Hat 11%, and management admitted execution failures. The falsifiable hypothesis is that AI capex is, at the margin, reallocating a finite enterprise budget rather than expanding it. The BIS estimates BDCs hold about $115bn of software loans — roughly a fifth of BDC lending — with AI-revenue risk not yet priced (though much is secured, at low leverage). A universal owner can finance the bottleneck (chips, power, data centres) and lend to the software revenue a reallocation may displace — two claims on one customer wallet.
Passage
Hormuz demands discipline. The U.S. proposed a 20% cargo toll, withdrew it within a day, and pointed to unspecified Gulf trade and investment deals — no counterparties, amounts or confirmations disclosed. The blockade applies to Iran-linked vessels, not neutral passage. "Gulf capital is the price of passage" is a hypothesis attributed to the President, not a fact. The operative signal is insurance, and it too needs precision: additional war-risk cover moved from about 2% toward 3% of hull value; one underwriter said difficult cover would probably cost at least 5% — a stressed quote, not a settled norm, quoted per seven-day exposure. The exposure to monitor is cover, freight, finance and the willingness to sail — not only crude.
Power
Korea's July 14 strategy raised its growth forecast and pledged to fast-track three megaprojects announced June 29 — a ~₩800tn chip build-out and ~₩550tn / 8.4 GW of data centres. These are mainly corporate plans (Samsung, SK, Naver), staged and conditional — not appropriations. The durable, owner-relevant point is the gap between announced capex and the capacity to execute it: power, water, land, interconnection and long timelines. But the counter-cases are real: on-site generation, storage and relocation can cut grid exposure; oversupply is a genuine risk; and scarce grid assets do not automatically earn excess returns when regulation caps them.
The commercial takeaway
Underwrite the conversion layer. Price insurance and freight as a distinct exposure, not a footnote to the oil view. Treat grid, water, permits and equipment as the binding constraints on AI and semiconductor theses, and separate announced capex from executable capex. Map the shared enterprise budget beneath the AI stack. And judge every sovereign or strategic vehicle by the correlation its funding creates. Capital commitments are abundant. The capacity to convert them into output — insurable passage, deliverable power, credible governance, a customer's finite budget — is not.