$101M Per Firm. 6% Impact. The AI Proof Gap in Private Equity.
Apr 23, 2026
|
Key Takeaways ▶ PE firms plan $101M average AI spend, but only 6% report high impact ▶ The gap between AI activity and AI proof is the defining problem in PE operations ▶ Mega-funds (Cerberus, Apollo, Platinum Equity, KKR) already built operational infrastructure ▶ 70% expect high impact in 3-5 years - the window to build capability is now |
|
$101M average planned AI investment per PE firm (KPMG) |
6% of GPs report high AI impact today (Bain) |
€45m at-risk revenue identified in 6 weeks |
PE firms are planning to spend an average of $101 million on AI over the next twelve months. That is not a typo. KPMG surveyed PE leaders in February 2026 and the number came back at nine figures per firm. Meanwhile, 95% of funds report that their AI initiatives are meeting or exceeding the business case, according to FTI Consulting.
Read those two data points together and you would conclude that AI in private equity is working. The money is flowing. The self-assessments are positive. The conference panels are optimistic.
Now read the third data point. Only 6% of GPs report high AI impact today. And 39% do not expect material AI financial impact on their portfolio companies this year. Both figures come from Bain’s 2026 GP Outlook, published in the same quarter as the optimistic numbers above.
There is a name for the distance between $101 million in planned investment and 6% reported impact. We call it the AI proof gap. And it is the defining problem in PE operations right now.
The paradox is structural, not accidental
This is not a story about bad technology. The models work. The infrastructure exists. The talent market, while tight, is deeper than it was two years ago.
The proof gap exists because of how AI gets deployed in PE portfolios, not whether it gets deployed. The pattern is consistent across every mid-market fund we speak to: the firm allocates budget, the portfolio companies hire consultants or buy tools, pilots proliferate, and twelve months later nobody can point to a number on a P&L that changed because of AI.
|
95% vs 6% 95% say their AI initiatives meet the business case. 6% report high impact. The gap between those two numbers is the gap between activity and proof. Sources: FTI Consulting (March 2026), Bain Global PE Report 2026 |
Activity is a team in a portfolio company running a chatbot pilot. Proof is €45 million in at-risk revenue surfaced in six weeks at a European industrial distributor, with the system still running daily and the sales team relying on it. Activity generates slide decks. Proof generates EBITDA.
The mega-funds are moving. The signal is loud.
In March 2026, The Information reported that Anthropic committed between $200 million and $1 billion to a joint venture with Blackstone, Hellman and Friedman, Permira, and General Atlantic. The purpose: deploying AI into PE portfolio companies at scale.
That is not a pilot. That is not an “exploration.” That is four of the largest firms in alternative assets partnering with a frontier AI lab to build production infrastructure for portfolio-level AI deployment.
The signal matters because of who is involved. Blackstone manages over $1 trillion in assets. General Atlantic has been building operating capability for decades. Permira has invested heavily in technology-enabled businesses. These are not firms experimenting. They are building the operational infrastructure for the next era of PE returns.
And they are not alone. The mega-funds have been building this capability for years, through different models but with identical intent.
|
Cerberus (COAC) - A proprietary affiliate with 110+ full-time operating executives. A separate legal entity, solely dedicated to operational value creation across the portfolio. |
|
Apollo (APPS) - A centralised AI and data platform that creates cross-portfolio intelligence. The “Quant PE” model: proprietary data lakes, systematic analytics, and AI agents operating at the fund level. |
|
Platinum Equity (M&A&O) - Operations embedded in the deal process from day one. The operating partner is not consulted after the acquisition. The operating partner is part of the acquisition thesis. |
|
KKR (Capstone) - Dedicated operating team deployed across the portfolio with repeatable playbooks and sector-specific expertise. |
These firms did not start with AI. They started with operational capability, built the infrastructure to support it, and are now layering AI on top. The sequence matters. Capability first. Technology second. The firms that reverse the order get pilots. The firms that follow it get proof.
70% expect high impact in 3-5 years. The question is what happens now.
Bain’s data tells a revealing story about timing. Only 6% report high impact today. But 70% expect high impact within three to five years. That is an extraordinary gap between current reality and near-term expectation.
It means the industry knows AI will matter. It is not a question of conviction. Every GP we speak to believes AI will be material to portfolio returns within this fund cycle. The question is execution.
And execution is where the proof gap lives.
39% of GPs do not expect material AI financial impact this year. Not because they think AI is irrelevant, but because they know their current approach is not producing results fast enough. The pilots are running. The tools have been purchased. The consultants have delivered their assessments. And the P&L has not moved.
The firms in the 6% are not using better technology. They are deploying differently. They start with an EBITDA problem, not a technology assessment. They deploy production systems in weeks, not pilots over quarters. They measure success in revenue protected and margin expanded, not in AI tools adopted.
The gap between 6% and 70% is a three-to-five-year window. The firms that close it early will compound the advantage. The firms that wait will find the window has closed and the leaders have pulled too far ahead to catch.
What proof actually looks like
Proof has three characteristics that distinguish it from activity.
|
A number Not “improved efficiency” or “enhanced decision-making.” A specific, auditable number tied to a P&L line. Revenue at risk identified. Margin leakage quantified. Working capital impact measured. |
|
A timeline Proof is fast. If it takes eighteen months to show results, it is an IT project, not an operational intervention. The hold period is 6.6 years and 52% of PE inventory has already been held for four years or more. There is no time for slow proof. |
|
A user Proof means someone in the business uses the system daily to make better decisions. Not a dashboard that gets reviewed in a quarterly board pack. A production tool that the commercial team relies on, the way a sales team relies on its CRM. |
|
Case in Point One European industrial distributor. Six weeks. €45m in at-risk revenue identified. The system runs daily. The sales team uses it to prioritise accounts, flag churn risk, and protect revenue that was previously invisible. That is proof. A production system with a number attached, not a pilot with a slide deck. |
The cost of the gap is compounding
The proof gap is not a neutral problem. It has a cost, and that cost compounds with every quarter it remains open.
For portfolio companies, the cost is unrealised EBITDA. Revenue leaking through customer churn nobody can see. Margins eroding through pricing inefficiencies nobody is measuring. Knowledge walking out the door as experienced employees retire. Every month without production AI against these problems is a month of value creation lost.
For GPs, the cost is fundraising pressure. DPI is now 2.5 times more likely to be ranked the “most critical” LP metric compared to three years ago, according to PEI. 47% of LPs are monitoring GP AI adoption, per Ontra. 53% rank value creation strategy as a top-five criterion for selecting a manager, according to McKinsey. LPs are not asking whether their GPs are spending on AI. They are asking what the AI has produced.
|
31,000+ Unsold portfolio companies worth $3.7 trillion sitting in the backlog. Average hold period: 6.6 years. 52% of PE inventory held for four years or more - the highest proportion on record. Source: Bain Global PE Report 2026 |
The firms that can demonstrate AI-driven operational alpha across their portfolio will differentiate in fundraising and in exits. The firms that can demonstrate AI spending but not AI results will face increasingly difficult questions from LPs who are paying closer attention than ever.
This series is about closing the gap
The previous series on this blog, “The Death of Financial Engineering,” established that financial engineering is exhausted and operational value creation is the only path to premium returns. That argument is no longer controversial. The data from McKinsey, Bain, BCG, KPMG, and every major industry report published in 2026 confirms it.
This series answers the question that follows: if AI is the answer to operational value creation at scale, why is it not working yet?
Over the next five articles, we will examine:
|
The sprawl problem 97% of enterprises are exploring agentic AI. Only 12% have centralised management. The root cause of the proof gap is not technology. It is governance, measurement, and deployment discipline. |
|
The governance cliff The EU AI Act begins enforcement in August 2026. Penalties reach €35 million or 7% of global turnover. 78% of PE firms cannot pass an AI governance audit within 90 days. |
|
The performance divide Companies with integrated AI see 4x revenue growth versus those still piloting. PE-backed companies with systematic AI have roughly 2x return on invested capital. |
|
What the 6% actually do The firms reporting high AI impact follow a specific deployment model. It is not complicated. It is disciplined. |
|
The portfolio-level playbook How to move from scattered pilots to production AI across the portfolio, with a timeline, metrics, and a realistic sequence. |
The proof gap is real. The cost of leaving it open is compounding. And the firms that close it first will define the next era of PE returns.
Does your fund know the difference between AI activity and AI proof? That is the question this series is built around. Follow along.
Sources & References
| KPMG: PE AI Investment Survey, February 2026 - PE leaders planning avg $101M AI investment next 12 months |
| FTI Consulting: AI in Private Equity Survey, March 2026 - 95% of funds report AI initiatives meeting/exceeding business case |
| Bain & Company: Global PE Report 2026 - Only 6% of GPs report high AI impact today, 70% expect high in 3-5 years, 39% don’t expect material AI financial impact in 2026, 31,000+ unsold portfolio companies, $3.7T backlog, 6.6yr hold, 52% held 4+yr |
| The Information: March 2026 - Anthropic $200M-$1B JV with Blackstone, Hellman and Friedman, Permira, General Atlantic for PE portfolio AI deployment |
| Cerberus Capital Management: Industry research - COAC: 110+ full-time operating executives as separate legal entity |
| Ontra: 7 PE Trends 2026 - 47% of LPs monitoring GP AI adoption |
| McKinsey & Company: Global Private Markets Review 2026 - 53% of LPs rank value creation top-5 manager selection criterion, 6.6yr average hold, 52% held 4+yr |
| PEI: LP Metric Rankings - DPI 2.5x more likely to be ranked “most critical” LP metric vs three years ago |
| G3NR8: Deployment data - €45m at-risk revenue identified in 6 weeks, European industrial distributor, production system in daily use |
Frequently Asked Questions
|
What is the AI proof gap in private equity? The AI proof gap is the distance between AI investment and activity in PE and the measurable impact on portfolio company P&L. PE firms plan to spend an average of $101M on AI (KPMG, February 2026), and 95% report their initiatives are meeting or exceeding business case (FTI Consulting). Yet only 6% of GPs report high AI impact today, and 39% do not expect material financial impact this year (Bain, 2026 GP Outlook). The gap between those numbers is the gap between activity and proof. |
|
How much are PE firms spending on AI? KPMG surveyed PE leaders in February 2026 and found firms are planning to spend an average of $101 million on AI over the next twelve months. Meanwhile, Anthropic committed between $200M and $1B to a joint venture with Blackstone, Hellman & Friedman, Permira, and General Atlantic for deploying AI into PE portfolio companies at scale. The investment is significant - the question is whether it is producing results. |
|
Why do 95% report success but only 6% report impact? The paradox is structural. 95% of funds report AI initiatives meeting or exceeding business case (FTI Consulting), but only 6% of GPs report high AI impact (Bain). The gap exists because of how AI gets deployed, not whether it gets deployed. Firms allocate budget, portfolio companies hire consultants or buy tools, pilots proliferate, and twelve months later nobody can point to a number on a P&L that changed. Activity generates slide decks. Proof generates EBITDA. |
|
What did the Anthropic and Blackstone joint venture signal? In March 2026, Anthropic committed between $200M and $1B to a joint venture with Blackstone, Hellman and Friedman, Permira, and General Atlantic. This is not a pilot - it is four of the largest firms in alternative assets partnering with a frontier AI lab to build production infrastructure for portfolio-level AI deployment. Blackstone manages over $1 trillion in assets. The signal is that mega-funds are building operational infrastructure for the next era of PE returns. |
|
What does AI proof look like versus AI activity? Proof has three characteristics. A number: a specific, auditable figure tied to a P&L line, not “improved efficiency.” A timeline: results in weeks, not eighteen months. A user: someone in the business using the system daily, not a dashboard reviewed quarterly. One European industrial distributor deployed in six weeks and identified €45m in at-risk revenue - a production system used daily by the sales team. That is proof. A pilot with a slide deck is activity. |
|
How is LP scrutiny changing around AI? LP scrutiny is intensifying. DPI is 2.5x more likely to be ranked the “most critical” LP metric compared to three years ago (PEI). 47% of LPs are monitoring GP AI adoption (Ontra). 53% rank value creation strategy as a top-five criterion for selecting a manager (McKinsey). LPs are no longer asking whether GPs are spending on AI. They are asking what the AI has produced. |
|
What is the cost of leaving the AI proof gap open? The cost compounds at three levels. For portfolio companies: unrealised EBITDA from customer churn, pricing inefficiencies, and knowledge loss. For GPs: fundraising pressure as LPs demand proof of AI impact, not just spending. For the industry: credibility, with 31,000+ unsold companies, 6.6-year average holds, and 52% of inventory held 4+ years. 70% of GPs expect high AI impact in 3-5 years - the firms that close the gap early will compound the advantage. |
This is Part 1 of 6 in the AI Proof Gap series - examining why PE firms are spending record amounts on AI but only 6% report high impact, and what the firms closing the gap are doing differently.
Keep informed with the newsletter for PE operating partners and the portfolio companies they back.
Get operational insights and trends, AI frameworks, resources and real deployment stories.
