Apollo: David Sambur said AI has complicated the valuation of software companies

Artificial intelligence is changing not only the products themselves, but also the logic of deals in the software market. David Sambur, co-head of the private equity direction at Apollo, says that because of AI it has become more difficult to evaluate software companies, although M&A activity has not disappeared even against the background of new geopolitical turbulence.

Why it's harder to count

Sambur's thesis sounds simple: the old benchmarks for evaluation no longer provide the same confidence. In the classical model, an investor could rely on revenue growth, customer retention, margins, and the predictability of SaaS subscriptions. But AI is rapidly changing the structure of expenses, the speed of feature rollout, and the competitive landscape itself. A company that looks expensive today could in a year become either a new leader or a product without notable protection. That's why the same numbers are now read differently than they were two years ago.

The problem is not only the buzz around generative models. AI simultaneously creates new revenue sources and puts pressure on old advantages: basic functions become cheaper, they are copied faster, and customers begin to expect more for the same money. Therefore, comparing software businesses by old multiples becomes riskier: past results explain future value less well. Especially when the market has not yet decided who will capture the main value — platforms, applications, or infrastructure.

Deals haven't frozen

At the same time, the deal market itself, according to Apollo's top manager, has not stopped. In an interview with Bloomberg in The Close program, Sambur noted that even the uncertainty related to the conflict around Iran did not kill the desire to make deals. Yes, markets are more comfortable in a calm and clear environment, but private equity cannot wait indefinitely for a perfect window. There is money on the market, but tolerance for uncertainty is distributed unevenly, and this itself affects the pace of negotiations.

"Fortunes are made on volatility,"

Sambur said, explaining why turbulence doesn't always slow down buyers. For funds, this means stricter selection and a longer asset verification process. When there's a lot of noise around — from geopolitics to AI expectations — the cost of error grows. But at the same time, the probability of buying a strong asset when the market cannot quickly agree on a fair valuation also grows. It is in such zones, judging by Sambur's words, that there remains space for large deals. This is where funds try to see the price before consensus appears in the entire market.

What they're looking at now

If we translate Sambur's words into practical terms, software investors are now looking not just at familiar SaaS metrics. It's much more important to understand whether AI gives the company a sustainable advantage or just adds a fashionable layer to an old product. There are more questions, and almost all of them come down to the quality of the business model over several years. The buyer needs to understand what exactly will remain in the business when the initial interest in the topic wanes.

• Can AI really improve the product and customer retention, not just presentation to the market
• What business functions risk becoming commoditized quickly due to cheap models and new competitors
• How will margin and unit economics change if the cost of inference and product development continues to decline
• Will the company be able to protect its price when competitors release similar AI features faster
• What in revenue growth is related to real demand and what is temporary hype

For buyers, this is no longer just a matter of technology. It's about how quickly the team is restructuring the product, sales, and pricing for the new reality. If AI reduces barriers to entry, then the premium goes not to every company with a loud AI label, but to the one that turns technology into a repeatable financial result. This requires not fashionable positioning, but demonstrable execution discipline and clear economics of implementation in each customer segment.

What this means

The software deal market is entering a phase where standard valuation formulas are not enough. Winners will be investors who can separate real AI effects from overpriced expectations, and companies that can prove it with numbers. For founders, this is a signal: one word "AI" in a presentation is no longer enough — you need to show how technology changes revenue, margin, and market position. This is likely where the new boundary will pass between an expensive asset and an overvalued story.