Direct answer: Global private equity deal count fell 10% in H1 2026 versus the same period last year, and EY's Q2 2026 PE Pulse report says the decline sits almost entirely in software. GPs are responding by loosening exit price expectations rather than waiting for multiples to recover. If you're an owner-operator building toward a sale, this is not a signal to panic.
It's a signal to check whether your business still commands a premium when the easy money leaves the room. That's what the Sovereignty Stack is built for.
I ran a nuclear reactor plant on a submarine before I ran a business. On a boat, you don't get a memo when the market shifts. You get an alarm, and you have about four seconds to know which system failed and why.
EY's Q2 2026 report is that alarm for software operators. Alternatives Watch reported that EY surveyed 50 general partners across the Americas, EMEA, and other regions. The finding was blunt: deal count down 10%, concentration almost entirely in software, GPs signaling more flexible exit pricing to get liquidity moving again.
Why Software Specifically Got Hit
Here's the part most owners are missing. This isn't a generic slowdown. It's a repricing of what software is actually worth when AI can replicate a feature in weeks instead of years.
PwC's deals team called it directly. The North American tech software index fell about 30% from its September 2025 peak after agentic AI tools launched capable of automating legal, financial, and data-services work (PwC, February 2026). Apollo's co-head of private equity, David Sambur, told Bloomberg the same thing in plainer terms. AI is compressing the time horizon for competitive advantage, and the old model of underwriting steady subscription revenue is breaking down (reported via Startup Fortune, April 2026).
The math backs it up. Windsor Drake's Q2 2026 SaaS valuation benchmark shows seat-priced horizontal applications compressed to 1.8x to 4.0x EV/Revenue, down from levels north of 6x two years ago. AI-native platforms with proprietary data and workflow lock-in still trade at 8x to 25x. That's not a small gap.
That's two different asset classes wearing the same SaaS label. So when EY says deal count fell 10% concentrated in software, here's what's actually happening. Buyers are walking away from commodity tools and staying hungry for anything defensible.
GPs are bending on exit price because they would rather take a discount on a durable asset. They'd rather do that than sit on dry powder waiting for a multiple that isn't coming back for the undifferentiated stuff.
The Owner-Operator's Real Question
If you're building a company to sell, in three years or five or whenever your number hits, the question EY's data forces on you isn't "is the market bad." It's "which bucket am I in."
Ask yourself three things, the way I'd have run a casualty drill on the boat:
- Can a competitor replicate my core workflow with an AI agent in under six months?
- Does my revenue depend on seats, or does it depend on outcomes my customers can't get anywhere else?
- If my top three customers left tomorrow, does the business still run?
Most founders answer those questions wrong because they're grading themselves on last year's playbook. I've spent two decades helping clients raise capital through Angel Investors Network, over $1 billion placed across those deals. The pattern that separates a 9x exit from a 2x fire sale has never been growth rate alone. It's whether the business runs on systems that don't depend on any single person, partner, or feature staying static.
Growth rate used to be the only number that mattered to a buyer. Not anymore. A software company growing 40% a year on rented infrastructure, thin data, and seat-based pricing looks fast right up until diligence starts.
Buyers in this market run a different drill. They ask what happens to that growth rate the day a frontier model ships a feature that does the same job for free. If the answer is "we don't know," the offer drops, or it walks.
I've watched both happen this year to founders who assumed their top-line number would carry the deal the way it did in 2021.
What the Numbers Looked Like Before This
Context matters here, because a 10% drop in deal count sounds abstract until you see where multiples came from. Hardman & Co's mid-2026 analysis of the software market pegged the damage at roughly $285 billion wiped from SaaS market capitalization within 48 hours. That happened after a major AI agent demo showed the tools performing financial analysis and project management work end to end.
That is not a slow bleed. That is a market repricing risk in real time, the same way a reactor SCRAM repriced my afternoon the instant an alarm tripped.
The mechanism behind it is simple, and it is the same mechanism eating seat-based pricing everywhere. If an AI agent can do the work of three analysts, the customer does not need three licenses. Per-seat revenue, the foundation of SaaS 1.0 economics, starts eroding from the inside.
PwC's deals team flagged this directly: winning companies are moving toward outcome-based or value-based pricing tied to impact delivered rather than headcount touching the software. That is harder to forecast. It is also harder to replace.
The Sovereignty Stack
This is the framework I built for exactly this moment. The Sovereignty Stack has three layers, and you need all three to hold value when multiples compress.
Layer 1 -- Data Moat. Proprietary data your customers can't get by switching to a competitor or prompting an LLM. Not "we have a database." A moat means the data compounds and gets harder to replicate the longer you operate.
Layer 2 -- Workflow Lock-In. Your product sits inside a customer's operating process, not next to it. If ripping you out means retraining twelve people and re-mapping three integrations, you're locked in. If it means canceling a subscription, you're a feature.
Layer 3 -- Owner Independence. The business generates cash and closes deals without the founder in every room. Buyers discount founder-dependent businesses hard, and 2026's tighter diligence process makes that discount worse, not better.
I learned layer three the hard way at Hartford, before Munich Re acquired the division I ran. The unit that survived integration cleanly wasn't the one with the best quarter. It was the one where the acquiring team could read the org chart and see the business would run the same on day 91 as it did on day one.
That's the frame buyers use now. EY's GPs bending on price isn't charity. It's triage. They're finding the businesses worth paying for and discounting everything else to move it off the shelf.
What Flexible Exit Pricing Actually Means for You
Don't read "GPs are bending on exit pricing" as good news across the board. Read it as GPs separating two piles.
Pile one holds durable software with a real moat, still commanding premium multiples, buyers competing for it. Pile two holds everything else, where sellers are accepting earnouts, seller notes, and multi-year structures just to get a deal done.
McKinsey's review of 2025 deal flow found PE-led transactions now make up 52% of global buyout-backed software activity (Economic Times, 2026). That means financial buyers, not strategics, are setting the going rate. Financial buyers price risk more conservatively than strategics chasing a growth story. That's the environment you're selling into.
If you're eighteen months from a target exit, this quarter's report is your 90-Day Bottleneck Audit trigger. Don't wait for your banker to tell you which pile you're in. Find out now, while you still have runway to move from pile two to pile one.
I've run that audit with founders who came to me convinced their business was recession-proof because revenue was still climbing. Revenue climbing and enterprise value holding are not the same fact. A 90-Day Bottleneck Audit forces you to isolate the three or four points in your business where a buyer's diligence team will push hardest: customer concentration, founder dependency, data defensibility, and pricing model exposure to AI substitution. Ninety days is enough time to fix at least one of those before you're sitting across from a GP who has already read this quarter's PE Pulse report and knows exactly which questions to ask.
The Capital Is Still There
One thing EY's report does not say matters as much as the headline number. Private equity is not retreating from software. It is getting selective.
Dry powder sitting on the sidelines globally is measured in trillions, not billions, and that capital has a clock on it. LPs do not fund GPs to sit still. Every quarter that dry powder stays uncommitted is a quarter a fund manager has to explain to their own investors.
That is good news for owner-operators who do the work now. A market with capital looking for a home and fewer qualified targets is a market where the qualified targets get bid up, not down.
The EY data confirms the targets are getting scarcer in the commodity bucket. It does not mean the whole market went cold. It means the bar to clear went up, and the reward for clearing it went up with it.
Doctrine Connection
Systems beat slogans. A growth story gets you a pitch meeting. A system that runs without you, holds data no one else has, and locks into a customer's workflow gets you a check that clears, even in a quarter when deal count is down 10% and the buyer across the table is watching every dollar.
FAQ
Q: Does the EY report mean all software companies are worth less now? A: No. It means the spread between defensible software and commodity software just got wider. Windsor Drake's Q2 2026 data shows AI-native platforms still trading at 8x to 25x revenue while seat-priced tools sit at 1.8x to 4.0x. Know which side of that line you're on.
Q: What does "more flexible exit pricing" mean in practice? A: GPs are more willing to accept earnouts, seller financing, and staged payouts rather than walking from deals over headline multiple disagreements. It gets capital moving again, but it shifts risk onto the seller if the business underperforms post-close.
Q: How do I know if my business is in the durable bucket or the commodity bucket? A: Run the three-question test in this article. Can AI replicate your workflow in six months? Does revenue depend on seats or outcomes? Does the business run without you in the room? Two or more weak answers means you have work to do before you list.
Q: Is this a bad time to sell a software company? A: It's a bad time to sell an undifferentiated one. It's a fine time to sell one with a real data moat and workflow lock-in. GPs are actively hunting for exactly that asset because their own LPs are asking hard questions about aging 2021-2022 vintage portfolios.
Q: How long does it take to move from the commodity bucket to the durable bucket? A: In my experience helping owners prep for exit, twelve to eighteen months is realistic if you start now, building the data asset, deepening workflow integration, and reducing founder dependency. Waiting until the banker's engagement letter is signed is too late.
*Jeff Barnes is the founder of demg.ai and CEO of Angel Investors Network. The views expressed are his own and do not constitute professional advice. demg.ai provides marketing education and systems for owner-operators. Past results do not guarantee future outcomes.*