Buyers still pay for growth. What’s changed is that growth on its own no longer determines what a software company is worth. Today’s buyers want evidence that the revenue behind it lasts and turns into cash sooner rather than later.
For much of the past decade, investors were willing to look several years into the future and give software companies credit for margins they had not yet produced. Recurring revenue, high gross margins, and confidence in long-term demand made distant cash flows feel unusually predictable.
That confidence has since weakened. While rising interest rates have changed the price of long-duration growth, AI has introduced a more fundamental question: How durable are the competitive advantages that supported premium software valuations in the first place?
In part 1 of this series on software valuation, we argued that revenue multiples were always a shortcut for future cash flow. In part 2, we argued that retention decides whether the shortcut is safe to take. This final piece is about why the shift to profitability occurred in 2026.
Uncertainty has a direct cost
A company’s value depends on two broad factors: The cash flows investors expect it to generate and the discount applied to those cash flows based on their timing and risk.
AI can affect both sides of that equation. It may reduce expected future cash flow if investors assume slower seat growth, weaker pricing, higher churn, greater product-development requirements, or additional infrastructure expense. It may also increase the return investors demand because the range of possible outcomes has widened.
These questions don’t need to be answered negatively to reduce valuation. Uncertainty itself is enough. A dollar of EBITDA generated today is observable. A dollar of EBITDA expected five years from now depends on a chain of assumptions about retention, pricing, competition, product relevance, and execution.
The more fragile that chain becomes, the less investors will pay today for the optimistic outcome.
Why AI is different from a normal product cycle
The concern isn’t just that every software product will be replaced by a chatbot. The deeper issue is whether AI changes the duration and strength of the economic advantages that once justified premium revenue multiples.
Investors and acquirers are now asking:
- Can AI reproduce important features more quickly and cheaply?
- Can a larger platform bundle a point solution’s functionality?
- Will customers require as many paid seats if AI allows fewer employees to complete the same work?
- Will value shift from application vendors to model providers or cloud platforms?
- Will agents make data migration and workflow reconstruction easier?
- Will inference and model costs reduce gross margins?
- Will usage-based pricing replace predictable per-seat subscriptions?
- Does the company control proprietary data or merely provide an interface?
- Is the product a system of record or a replaceable productivity layer?
For the first time in years, the central software underwriting question isn’t just how quickly the company can grow. It’s how long the company’s market position will remain valuable.
The market is paying less for uncertain revenue
PitchBook’s public enterprise SaaS comp sheet puts the median EV to trailing revenue multiple at 6.2x at the end of 2024, 4.9x at the end of 2025, and 3.3x on March 31, 2026. More than half the decline landed in one quarter, which PitchBook attributes to fears that agentic AI can run SaaS workflows directly, compounded by soft Q4 2025 earnings. Margins went the other way over the same period. PitchBook’s Q2 2026 update puts the median 2026 EBITDA margin estimate at 23.3%, up from 20% in 2025.
That combination is telling. Investors are paying less for each dollar of uncertain future revenue while demanding more current earnings from the companies they own.
The shift doesn’t mean every software business will suddenly be valued only on an EBITDA multiple. High-growth companies with clear AI advantages, durable platforms, or exceptional economics can still command premium revenue valuations. But traditional SaaS companies face a higher burden of proof, and mature software assets are increasingly underwritten through the earnings and cash flow they produce today.
Why EBITDA matters more when the future
is less certain
Current EBITDA provides information that a revenue multiple can’t. It shows that the company can convert customer demand into operating profit. It supports leverage. It gives the buyer the ability to collect cash, fund acquisitions, or reinvest selectively. Most importantly, it reduces the number of optimistic assumptions required to earn an acceptable return.
A high-growth, cash-burning company often requires a narrow set of outcomes. Growth must continue. Margins must improve. Financing must remain available. Competitive pressure must remain manageable. The terminal multiple must remain supportive.
A moderately growing, profitable company has more strategic flexibility. If growth slows, the business still generates cash. If the market remains uncertain, the buyer can underwrite a return from earnings rather than relying on multiple expansion. If an attractive opportunity emerges, cash flow can fund renewed investment.
Cash generation does not eliminate AI risk. It reduces the proportion of value that depends on accurately forecasting an uncertain future.
AI is creating a bifurcated software market
The market is splitting in two. Businesses viewed as systems of record, infrastructure providers, and owners of proprietary data have generally been more defensible than thin application-layer products with limited switching costs.
The distinction is increasingly based on what the company controls. A durable software business may control:
- The authoritative data set
- The transaction
- A regulated or mission-critical workflow
- A proprietary network
- Customer distribution
- Embedded integrations
- Industry-specific logic
- Security and compliance infrastructure
- A workflow that would be operationally dangerous to replace
These control points are more durable than feature breadth alone. A product can have excellent features and still be vulnerable if those features can be replicated, bundled, or delivered by a broader platform.
Customers don’t just buy code. They buy reliability, support, integrations, auditability, security, implementation knowledge and confidence that the system will continue to operate. Those factors create friction that AI may change, but does not automatically eliminate.
What buyers are underwriting now
The best software businesses still combine growth and profitability. Buyers are simply examining the quality of both with greater precision.
Growth quality
- Gross and net revenue retention
- Customer concentration and contract length
- Pricing power and expansion within existing accounts
- New-logo acquisition efficiency
- Exposure to AI-driven seat compression or commoditization
Profitability quality
- Free-cash-flow conversion
- Capitalized development costs and stock-based compensation
- Implementation, support and hosting intensity
- AI inference expense
- The R&D and sales investment required to maintain retention and growth
Strategic durability
- What the company controls
- Why the product is difficult to replace
- Whether the company owns proprietary data
- Whether AI strengthens or weakens its market position
- Whether an adjacent platform can reproduce or bundle the product
The new standard is defensible, durable growth
It would be wrong to conclude that investors now want software companies to stop growing and maximize short-term profit. A company with no growth may generate cash today while losing relevance, market share or pricing power.
The emerging standard is not growth at all costs, and it is not profit at all costs. It is efficient, defensible and durable growth.
Investors want companies that can grow without relying indefinitely on outside capital. They want retention that proves the product remains essential. They want margins that demonstrate operating leverage. They want cash flow that creates downside protection. And they want competitive advantages that can survive technological change.
Growth still matters because it expands the future cash-flow opportunity. Profitability matters because it proves that the opportunity is real. Durability matters because it determines how long those economics can persist.
Growth at all costs is done because the market is no longer willing to assume that every dollar of software revenue will eventually become valuable cash flow. That conversion now has to be demonstrated through retention, earnings and a defensible position in the AI era.