Public and private SaaS companies without a credible AI story are trading at roughly 3.4 to 3.6 times revenue in 2026, down from the 15-to-20x peak of 2021. Companies with a credible AI-native story are commanding 15 to 30 times revenue, and foundation model companies considerably more. That is not a market-wide correction. It is a fork, and it happened fast enough that a meaningful number of PE firms got caught on the wrong side of it without fully realizing it yet.
Sponsors are responding, but only on the deals still in front of them. Software underwriting has become materially more cautious as firms reassess AI’s impact on target companies, according to Ropes & Gray’s 2026 market recap, and some sponsors, Vista Equity Partners among them, are reportedly raising the bar for new software investments from the traditional Rule of 40 to a Rule of 50 or 60, baking AI-driven margin expansion directly into how a deal gets priced going in.
That is the easier problem. New deals can be underwritten with AI already priced into the model. The harder problem is sitting inside portfolios firms already own.
Call it the Second Underwriting: the unplanned, more expensive re-underwriting event facing assets a fund already closed, priced, and financed under assumptions that predate the current AI repricing. Unlike a new deal, there is no clean entry point to bake AI into the model. The company has to be pivoted into an AI narrative mid-hold, on a valuation clock the fund does not control, competing for talent against a capital market that has already decided where it wants to place its bets.
That capital market signal is not subtle. AI-native startups captured roughly 80 percent of global venture funding in the first quarter of 2026 and 86 percent of US venture dollars by mid-year, according to PitchBook. And 2026 has brought the clearest evidence yet of where that capital eventually goes: SpaceX completed its IPO in June at a valuation near $2.1 trillion, Anthropic filed confidentially at roughly $965 billion, and OpenAI filed at roughly $852 billion with a targeted listing later this year. A senior AI leader is no longer just weighing a startup’s promise. They are weighing a visible, near-term path to a mega-listing against a PE fund’s standard offer: a management equity pool averaging around 9 percent of fully diluted equity, vesting on a 3-to-4x return hurdle, over a hold period that increasingly runs 5 to 7 years. That gap, the Compensation Ceiling described in prior coverage, is precisely what makes the Second Underwriting so hard to close. The company needs the AI narrative to protect its exit multiple, and the people who could build that narrative have a better offer sitting in front of them.
The instinct is to assume this talent problem will solve itself. It will not, and the layoff headlines are actively misleading firms into thinking otherwise. Tech layoffs hit roughly 142,000 in 2026, with profitable companies including Meta, Amazon, and Oracle among those cutting staff, largely to help fund a combined $700 billion in AI infrastructure spending. But those cuts are concentrated in a different labor pool than the one PE firms need. Roles in machine learning infrastructure, model evaluation, AI safety, and applied research remain in acute shortage even at the same companies announcing layoffs elsewhere, while the roles being cut skew toward routine software engineering, recruiting, and back-office functions.
Globally, AI talent demand now exceeds supply by roughly 3.2 to 1, with an estimated 1.6 million open AI-related positions against about 518,000 qualified candidates, and AI-specific roles commanding salaries roughly 67 percent higher than comparable traditional software positions. A portfolio company staring at a stack of layoff headlines and assuming AI talent is suddenly abundant is reading the wrong number. The abundance is in the skills AI is displacing. The scarcity is in the skills needed to make the displacement pay off at exit.
The Second Underwriting problem does not show up on a standard operating dashboard. Distribution yields across private equity have fallen to roughly 11 percent, down from more than 25 percent a decade ago, as hold periods stretch well past what most funds originally modeled. A company that cannot close its AI leadership gap is not simply missing an upgrade. It is on track to be marked and eventually sold in the 3.4x-to-3.6x tier of the market instead of the 15x-plus tier, at the exact point in the fund’s life when it needs to return capital to LPs, not explain why it can’t.
We are seeing PE firms bring us searches for AI-specific architectural talent with a different tone than a standard technical search, closer to urgency than to planning. In a number of these conversations, the underlying trigger is not a five-year roadmap. It is a live re-underwriting exercise, on an asset the firm already owns, where leadership has recently concluded the current valuation trajectory will not support the exit the fund modeled at acquisition. We also track venture funding daily. Over a recent ten-day window we recorded roughly $4.67 billion committed across 56 disclosed rounds, almost entirely into AI-native companies, the same capital pool pulling candidates away from every one of these searches.
Firms that treat this as a portfolio-wide talent audit, not a one-off hire when a problem surfaces, are the ones closing the Second Underwriting before the market closes it for them.
Verticalmove is a strategic talent consulting partner that helps organizations solve business problems related to talent attraction, selection, and retention. We work with PE-backed, venture-backed, mid-market, and enterprise companies to design talent strategies, strengthen leadership teams, and build the workforce capabilities required to achieve critical business objectives. When growth stalls, transformation accelerates, or organizational priorities shift, talent is often the constraint. We help companies identify, attract, assess, and retain the people who create competitive advantage.