Lenders price AI risk as sponsors keep paying for AI upside.

Ridgeview Partners, a San Francisco technology private equity firm, agreed a recommended take-private of Pinewood Technologies, which trades as Pinewood.AI and supplies cloud software to car dealers and manufacturers. Shareholders will receive £4.48 in cash per share, valuing the company at approximately £545 million on a fully diluted basis, with a rollover alternative into unlisted Ridgeview vehicles ([1]). Lithia & Driveway, Pinewood.AI's largest shareholder and customer, supports the deal ([2]).

The stated thesis matters more than the price. Ridgeview said the business's next phase needs a step-change in technology investment, particularly in data and AI-driven innovation, and that private ownership can enhance that investment ([3]). In effect, the buyer is saying the AI build is easier to fund in private hands than on the public market. The value creation plan will now have to show the payback on that spend, in a UK mid-market deal where the exit buyer will ask for evidence.

Not every software asset earns that treatment. Fortune's Term Sheet reported that private equity investor Robert F. Smith had told a Fortune Brainstorm Tech audience that a small but real portion of his software companies no longer have a right to exist amid AI-driven change ([4]). One venture investor estimated that 10-20% of his portfolio feels very vulnerable on top of normal attrition ([5]). The two categories investors named as having some insulation were regulated licence businesses and businesses with proprietary data that is difficult to access ([6]).

For deal teams, that gives a usable screen. Buyers are still paying for software with embedded data and distribution, and lenders are charging more for the rest. Pinewood.AI's platform runs sales, aftersales, accounting and CRM for dealers in over 36 countries, with partnerships with over 50 manufacturer brands ([7]): the kind of workflow and data position the screen rewards. Diligence should test how much of a target's revenue rests on assets like these, and how much on features a foundation model now does for free.

Goldman Sachs research published on 19 August found that industries more exposed to AI automation have generally seen slower job openings growth since the second half of 2022 ([1]). Employment in call centres, software publishing, management consulting and advertising has fallen sharply below its historical trend across developed markets ([2]). Call centre employment is 39% below trend in the US, 33% in Canada and 27% in Germany ([3]).

The effect falls hardest on new entrants. Across more than 800 occupations, a 10% occupational exposure to AI was associated with only a 0.1 percentage point drag on annual headcount growth overall, but more than 0.2 percentage points for entry-level workers in the US and more than 0.6 points in Australia ([4]). Goldman's combined survey data puts AI adoption at roughly 15% to 20% in major developed markets, with France, the US, the Netherlands and the UK leading ([5]).

For sponsors holding contact centres, consultancies, agencies and IT services, this is cross-country evidence that the disruption has moved from forecasts into the labour data. The junior layer is where pyramid-shaped services businesses earn their margin, billing graduates at a multiple of cost. If clients expect AI to do that work, per-hour and per-head pricing comes under pressure first. Value-creation leads should track the share of revenue on outcome or transaction pricing, and the ratio of junior to senior staff, as leading indicators.

Sponsors are also buying into the shift. A private equity firm signed a letter of intent for US insurance-defence law firm Wood Smith Henning & Berman at an enterprise value of roughly $700 million, about 18 times 2025 adjusted EBITDA of $38.2 million, through a management services organisation that houses the non-legal operations ([6]). Billing rates are typically constrained by insurers and margins can be modest, so technology, offshoring and centralised operations are the main profit levers ([7]). An 18x entry for a business with capped prices only works if those levers are sized, phased and evidenced in the plan from the first hundred days.

A Pearl Meyer survey of 116 board members, CEOs, C-suite executives and senior managers found only 34% of C-suite executives say it is consistently clear which executive or team makes calls about AI, against 53% of board members and 57% of managers below the C-suite ([1]). Asked whether strategic goals will require significant change to how the organisation operates within three years, 88% of CEOs agreed but only 42% of directors did ([2]). Confidence that AI will deliver significant gains within 18 months held at about 50% at every stage of maturity, including among companies that had not started ([3]).

EY is acting on the same problem inside its own firm. It is creating an AI Value Realisation Office to govern AI spending, monitor usage, decide which initiatives to scale and oversee how AI reshapes jobs ([4]). EY-Parthenon research found that 75% of the potential enterprise value from AI comes from value streams that cut across functions, against 25% from projects inside a single function ([5]). An AI router that sends staff to the best model for each task has cut EY's token consumption by 60% since April, and 98% of 534 senior decision-makers in an EY survey said token spending had made them reconsider their approach ([6]).

For a value-creation team the lesson is structural. Departmental budgets produce pilots inside functions, while the value EY describes sits in flows that cross them, such as quote to cash. Each portfolio company needs one named owner for the AI line in its value creation plan, a budget that follows cross-functional use cases rather than org charts, and a cost discipline on model spend.

Boards are the weak point. If directors believe the structure is settled while management expects major operating change, the AI line will be underfunded until results disappoint. Pearl Meyer's Brad Jayne said he worries about "finger pointing" when spending has to be justified ([7]). Sponsors control portfolio boards directly, so they can close that gap faster than any listed company: agree the owner, the measures and the review cadence now, and the evidence will be in place by exit.

PitchBook research on private equity's zombie problem found that at the end of 2025 roughly 40% of US PE-backed companies, representing more than $860 billion of net asset value, had been held for more than seven years, the highest level since 2016 ([1]). Funds that began investing in 2021 have returned just 0.14x investors' initial commitments ([2]). Of the 6,437 US PE-backed companies held for more than five years, more than half have completed no deal since the end of 2021: no add-ons, recapitalisations or refinancings ([3]).

"I think the three- to five-year hold as a hard rule is gone now," Elm Capital's Hamza Khaldi told PitchBook ([4]). That changes the planning horizon for AI. A programme that needs 18-24 months to reach EBITDA looks too slow late in a five-year hold. In a hold stretching to seven years or more, it is often the main operating lever left.

Most of these companies were bought between 2018 and 2022, before generative AI tools were widely available, so their original value creation plans priced AI as neither an upside nor a threat. They will be sold to buyers who now test both, and to lenders who, as this week's LCD data shows, already charge for AI exposure in software. An aged asset that has banked no AI savings, or cannot show how exposed its revenue is, gives the next buyer an easy reason to cut the price.

Sponsors with long-held assets should run an AI exposure and opportunity review across them this quarter. The output is a short list: assets where a funded AI programme can lift EBITDA before sale, assets where the exposure is better sold than fixed, and assets where continuation vehicle buyers will want the AI case written down before they price it.