Every portco CEO with one of those sponsors got a meeting they didn't book.
If your firm is part of the Anthropic JV or the OpenAI DeployCo announcement, your portfolio companies are about to find out what "embedded engineers" actually means. Six people in a Slack channel nobody set up. A roadmap nobody asked for. KPIs the CFO did not agree to. Invoices routed through a vendor the sponsor owns a piece of.
I've sat with enough CEOs three weeks into a sponsor-driven transformation to know how this ends. They are not enthused.
Grant Thornton just measured the actual state of AI in PE. Five percent of portfolio companies are fully integrating it. Five. After three years of "AI native" annual letters and AI operating partner job titles.
The constraint was never the model. The constraint is operating capability. The same CFO who could not finish a Salesforce rollout is now expected to absorb six Anthropic engineers and a workflow redesign. The same OP team that struggled to land a pricing project is the lead vendor manager on a multi-portfolio AI programme. What about hiring the JV changes that.
The firms that wrote the cheque have not solved the operating problem. They have outsourced one variable inside it. The talent gap, the board cadence, the change discipline, the willingness to replace a weak leadership team in under nine months. None of that is in the JV deck.
When the sponsor co-owns the consultant, the value creation math also moves. The fee flows out of the portco. The equity owns a piece of the vendor. The carry is calculated on the portco. Three economic vectors, one set of partners, all three sides of the trade.
The question for the next LP advisory committee is not whether your sponsor is doing AI work. It is what actually changed in your sponsor's operating capability between the last quarterly call and this one.
There is a slide doing the rounds at PE firms right now. Anthropic in the value creation column. Zero entries under either conflict of interest or operating readiness.
That is the slide that will age the worst.
