AI value in private equity comes down to proof, not promises.
Retail investors chase AI exposure.
A fresh sign of the AI gold rush emerged as tech analyst Dan Ives launched a first-of-its-kind fund to give the public exposure to private AI companies. The $200 million closed-end vehicle, named the Ives Ultra AI Opportunities fund, will trade on the NYSE ([1]). It aims to buy stakes in late-stage, unlisted AI leaders, a space traditionally accessible only to venture and private equity funds.
This move highlights both the excitement and the challenges around capturing AI’s value. Sponsors have poured capital into AI-driven start-ups and portfolio initiatives for years, anticipating transformative growth. With this new fund, Ives is effectively betting that some of those bets will pay off so spectacularly that even public market investors will want in. Yet the backdrop is sobering: while AI has dominated headlines, many portfolio company projects have yet to deliver meaningful profits. Sponsors are now facing a dual imperative, continue investing in AI for future upside, while also demonstrating concrete returns in the present.
ROI reality check in the portfolio.
A string of new data underlines a widening gap between AI’s potential and the profits on the books. FTI Consulting’s latest Private Equity AI Radar finds that an overwhelming 95% of PE firms are meeting or beating the business cases for their portfolio AI projects ([1]). However, only 17% say these initiatives are significantly exceeding expectations ([2]). In other words, incremental improvements are common, but game-changing results remain rare.
This “AI value paradox” is forcing a rethink of value-creation playbooks. Many firms jumped into generative AI pilots hoping for quick wins in cost or revenue. Now operating partners and deal teams are scrutinising why those pilots haven’t moved EBITDA or cash flow as much as hoped. The consensus: real ROI demands more than surface-level tweaks. The technology can’t just be layered on top of old processes; it has to be integrated deeply into how a business operates.
The upshot is a maturing approach to AI in value-creation plans. Gone is the blank-cheque optimism. In its place, sponsors are instituting more rigorous milestones and ROI tracking for AI projects. They are focusing on fewer, higher-impact use cases, for example, automating customer service workflows or augmenting sales teams with AI-driven insights, and insisting on evidence of value early. As one operating executive put it recently, “We have gone from vanity pilots to true EBITDA discipline”, reflecting a shift towards AI initiatives that directly contribute to margin or growth instead of just showcasing cool tech.
AI haves and have-nots at exit.
If there’s a silver lining for private equity, it’s that the few portfolio companies which do crack the code on AI stand to reap outsized rewards at exit. New research from McKinsey shows that PE-backed businesses which broadly embrace AI, not just in back-office automation, but in their products and strategies, achieve valuation multiples more than double those of peers sticking to narrow, tactical uses of AI ([1]). In a global dataset of 471 companies, those embedding AI deeply into offerings commanded median revenue multiples around 20×, versus ~14× for those limiting AI to internal operations ([2]). Investors are effectively paying a premium for what they see as future-proof, AI-enabled growth.
However, buyers have become far more discerning about how that growth is achieved. As M&A activity picks up in AI-related sectors, technical due diligence is weeding out the pretenders. One tech investment bank warned that simply slapping “AI-powered” on a product without real substance is “essentially worthless” once experts look under the hood ([3]). Impressive demos mean little if there are no live deployments or paying customers. Instead, trade buyers and PE acquirers are rewarding targets with defensible data advantages, high customer retention and scalable AI solutions. Proprietary data “moats” and proven AI use cases have become key discussion points in deal rooms, especially as firms brace for a wave of AI-driven disruption in software and services.
Regulators and LPs demand accountability.
It’s not just commercial forces driving a more hard-nosed approach, regulators and investors are raising their expectations too. In Europe, the AI Act’s initial requirements quietly took effect in August, and enforcement is now ramping up. As of this week, any company deploying generative AI or chatbots in the EU must implement new transparency measures (like content disclosures and user notifications) by year-end ([1]). The European Commission’s new AI Office can investigate and fine providers of general-purpose AI models up to 3% of global revenues for violations ([2]). This regulatory push means PE firms with portfolio exposure to AI, whether as developers or heavy users, need robust governance around data and model risks. Compliance and risk teams at funds are working with portfolio companies to audit AI systems, update policies and ensure they can meet impending requirements.
Meanwhile, limited partners are increasingly scrutinising GPs’ approach to AI. Recent industry surveys show many LPs remain unconvinced that all this AI spending will translate into realised returns. Some large investors are now pressing PE managers for concrete examples of value created by AI, and for assurances that risks, from biased algorithms to cybersecurity, are being managed. In response, top-tier firms have begun formalising their AI strategies: hiring dedicated data science advisors, setting up AI centres of excellence and integrating AI checkpoints into due diligence and 100-day plans. The goal is to demonstrate to both regulators and shareholders that AI is being used responsibly and profitably. As one mid-market dealmaker summed up: *“We won’t pay extra for an ‘AI story’ at exit unless we see the impact in the numbers.*”