What Software Investors Are Underwriting Now

Software valuations have moved around a great deal this year, and much of that movement comes down to one question. For any company preparing a raise of $20 million to $100 million, investors want to know whether AI makes the business more valuable or starts to erode what customers are paying for.

There are two sides to that question. AI can improve a product, reduce the cost of delivering it, or make proprietary data and established workflows considerably more valuable. It can also put pressure on pricing, reduce the number of seats a customer needs and make parts of a product easier to replicate.

A market with two tiers

That uncertainty is visible in public valuations. Across larger software companies, median enterprise value to revenue was about six times at the end of 2024, fell to roughly three times by the spring of this year, and recovered to about four and a half times over the summer.

The median only tells part of the story, because the gap between the strongest and weakest businesses has become very wide. Some of the best public software companies still trade at 15 to 25 times revenue, and more ordinary companies sit at one to three times.

The same selectivity shows up in private markets. A typical lower middle market software business might be valued at around four to five times ARR today, and higher-quality assets can reach seven to nine times or more. Investors are drawing a much sharper distinction between average businesses and high-quality ones, which raises the practical question of what drives that difference and how a company demonstrates it.

Quality and durability of revenue

Most of the answer lies in the quality and durability of revenue. Growth still matters, but investors are spending far more time on how efficiently it is generated and how predictable the underlying revenue base is. That is why the Rule of 40 has become more important again. Only a relatively small share of public software companies clear it, and those that do generally trade at a meaningful premium.

Retention is the other major part of the picture. Net revenue retention of around 120% tells an investor a great deal, because it shows that customers are staying and spending more over time. A business with that profile does not have to replace its revenue base each year in order to grow.

Investors will also look closely at customer concentration and contract durability, since both bear directly on how much of today's revenue can be relied on in three to five years.

The AI question

Once investors are comfortable with the quality of revenue, they return to AI and ask whether that revenue stays durable as the technology develops. Any company preparing for a raise needs a clear answer as to why AI does not erode its economics.

That answer will differ from one business to the next. It may rest on proprietary data, a mission-critical workflow that is hard to replace, regulatory complexity or high switching costs. In some cases AI simply makes the product considerably better. What matters is that there is an answer, and that it can be supported with evidence.

Seat-based pricing is a useful example. If a company charges per user, and AI means its customers eventually need fewer people doing the same work, investors will naturally ask what happens to revenue per customer. A seat-based model is not a bad business, but management does need to explain how the company keeps capturing value if a customer can achieve the same output with fewer seats.

Capital remains available

There is still plenty of money for the right assets. Private equity is sitting on more than $2.5 trillion of dry powder and sponsors have remained very active in software, particularly where they believe public markets have mispriced a good business. We have seen a number of large take-privates this year.

The same selectivity applies here too. Sponsors are being more careful where a business carries significant AI exposure, and more careful still where the transaction requires a large amount of financing. Businesses with defensible revenue continue to attract strong attention, and vertical software is one of the clearest examples. A product that owns a workflow inside a particular industry, that customers rely on every day, with strong retention and meaningful switching costs, remains a very attractive profile.

Raise size shapes the process as well. At $20 million to $35 million, the round usually depends on a named institutional lead, with a syndicate forming behind it. From $35 million to $60 million, growth and crossover investors price on unit economics and a defined path to liquidity. From $60 million to $100 million, capital is highly concentrated and the underwriting threshold is higher still.

Before going to market

A company raising capital needs to understand how investors will categorize it before it starts presenting. That means knowing the right public and private comparables, where the business is stronger or weaker on growth, margins and Rule of 40, what retention, expansion and concentration look like, and how investors are likely to view its exposure to AI, including both the benefits and the risks.

Those questions ultimately determine whether an investor sees a high-quality, durable asset that deserves a premium or a standard software company that should trade close to the broader market. That distinction can currently be worth several turns of revenue.

Transactions are getting done and investors will still pay very good prices for the right business, but the underwriting has changed, and it is far better to understand how before the first conversation.