The Great SaaS Valuation Reset: Why SaaS valuations have compressed, what recent declines reveal, and what management teams can do next.

By Brad Kayton

The last six months have been brutal for many public software companies in terms of market cap and private companies in terms of access to capital and valuation. An illustrative basket of five widely followed public company software stocks was down between roughly 20% and 45% over the last six months (for this analysis, from September 12, 2025 to March 12, 2026). The repricing is primarily based on investor concern that AI is compressing the value of generic application software but also reflects a mix of higher capital discipline, a renewed preference for companies that own proprietary data, mission-critical workflows, and measurable economic outcomes.

The repricing is visible both in index-level multiples and in individual names. Using recent market prices, Zoom is down about 87% from its prior peak, Twilio about 72%, and DocuSign about 85%. Those are not marginal corrections; they reflect a full re-underwriting of future cash flows and competitive positioning.

At the same time, the market has not written off SaaS altogether. Shopify is a useful counterexample: after a painful drawdown, it rebuilt credibility through operating discipline, payments and commerce depth, and a clearer AI-and-productivity narrative. Its stock remains below its late-2025 high, but far less so than many pandemic-era SaaS darlings.  The lesson is important: the sector is no longer being valued as one basket. Strong operators with real workflow ownership and improving profitability can still command premium outcomes.

This is why the phrase “SaaS collapse” and “SaaS apocalypse” is directionally true but incomplete. What has really collapsed is the old assumption that subscription revenue alone deserves a premium. The new market asks harder questions: Is the product embedded in a mission-critical workflow? Does it own truth, or just display data from somewhere else? Can AI improve outcomes without destroying gross margin? And can the company compound intelligence across the enterprise, rather than bolt point features onto a fragmented stack?

What happened?

For years, SaaS businesses were valued as if three assumptions would hold indefinitely: revenue would compound at premium rates, retention would stay unusually durable, and recurring subscription models would deserve a structural valuation premium. That framework has now been challenged. Investors have become less willing to pay high multiples for software businesses that look like interchangeable application layers, particularly when AI tools can automate portions of the work those applications once mediated.

The recent selloff has been severe enough that Reuters described it as a roughly $1 trillion rout in software stocks after new AI automation tools intensified fears that parts of the sector could be disrupted. The pressure did not hit only speculative names. Larger, more established software companies were repriced as well, suggesting that the market is no longer asking whether a company is “SaaS,” but whether it owns something hard to replace.

A six-month lens captures the speed of the market’s change in conviction. In the basket above, HubSpot was down about 44.9%, Constellation Software about 40.4%, Workday about 40.0%, ServiceNow about 39.2%, and Salesforce about 19.5% over this very short period. That range is wide, but the message is consistent: software valuations have become far less forgiving.

Why the rerating became so severe

Three forces have converged. First, AI has changed the debate from “how much can software expand?” to “how much of the software stack is actually defensible?” Generic workflows, standardized data fields, thin user-interface layers, and lightly differentiated analytics are all being scrutinized more aggressively. As Reuters reported in March, investors are increasingly focusing on whether software companies own years of exclusive data and deeply embedded operating processes that AI cannot easily replicate.

Second, the cost of capital remains far less forgiving than it was in the zero-rate era. Even if rates eventually move lower, boards and investors are valuing near-term cash conversion more heavily than distant, hypothetical upside. That hurts businesses whose valuation depended on long-duration growth narratives.

Third, the sector is being forced back toward fundamentals. Growth at any cost has faded. Investors want better sales efficiency, credible free-cash-flow conversion, shorter payback periods, and evidence that AI is improving the product faster than it is commoditizing the category. The old shortcut—recurring revenue equals premium multiple—has weakened materially.

Two case studies in valuation compression

Case study 1: Constellation Software (TSX: CSU)

Constellation Software is not a pure-play single-product SaaS company. It is a Canadian-listed operator of vertical-market software businesses, which is precisely why its decline is instructive. If even a company with deep acquisition discipline, a long operating record, and a strong reputation can sell off sharply, then the market’s reset is broader than a simple rejection of lower-quality names.

Using publicly surfaced historical market data, Constellation fell from about C$4,392.79 on September 12, 2025 to C$2,619.09 on March 12, 2026, a drop of about 40.4% in six months. Reuters had already captured the logic of this kind of move in 2025 when Lorne Steinberg said that these stocks had been “priced for perfection.” That phrase remains useful because it describes what happens when premium software franchises stop receiving the benefit of the doubt.

The deeper lesson is that operational quality and valuation immunity are no longer the same thing. Constellation may still be admired for discipline and execution, but if investors become nervous about multiple compression across software, even admired compounders can reprice quickly. That matters for private SaaS boards because public comps often set the tone for strategic and sponsor conversations in the private market.

Case study 2: Workday (NASDAQ: WDAY)

Workday offers a more direct SaaS case study because it sits at the center of HR, payroll, and finance workflows. Yet Reuters reported in February 2026 that the company’s weak sales outlook and AI-related concerns pushed the stock to a more-than-five-year low, with the shares down about 40% year-to-date at that point. The selloff was not just about one quarter; it was about what that quarter suggested regarding deal velocity, customer caution, and category defensibility.

There was also a sharp message in the analyst reaction. Reuters cited Piper Sandler analysts saying that, amid AI debates, investors were scrutinizing “every metric” for application-layer software names. That is a revealing phrase because it captures the shift from narrative investing to evidence investing. Management can no longer rely on broad confidence in cloud software as a category. It must prove why its workflow, data position, and economics deserve premium treatment.

Workday’s own defense was equally revealing. Reuters quoted co-founder and CEO Aneel Bhusri arguing that “no amount of vibe coding” will reproduce an HR or ERP platform. The market has not fully rejected that argument, but it clearly wants more proof. In the current environment, embeddedness alone is not enough; companies must show that AI strengthens their moat rather than narrowing it.

What the market seems to be rewarding now

Recent commentary has not been uniformly bearish. Reuters reported on March 10, 2026 that Deutsche Bank upgraded software to overweight, arguing that the months-long rout had likely gone too far. That matters because it suggests the market is not rejecting software outright. Instead, it is differentiating more aggressively between vulnerable software and defensible software.

The strongest defenses appear to cluster around a few ideas: proprietary data, deep workflow integration, clear switching costs, AI deployed inside production workflows, and financial discipline. Rebecca Wettemann of Valoir told Reuters that Oracle’s flexibility and data position give customers meaningful choice as the AI ecosystem evolves. James St. Aubin of Ocean Park put the point even more directly: “Proprietary data is the deepest moat by far.” For SaaS leaders, that is the strategic brief in one sentence.

Solution path one: the 5 steps

If generic software commands a lower multiple, the practical response is to become less generic. One workable framework is a five-step progression from application vendor to AI embedded decision platform.  For SaaS companies to make the leap to an enterprise AI-first platform, there is a five-step methodology that can be adopted to achieve this transition.

Step 1: Own the system of record. A thin application layer is easier to displace than a platform that holds the live operational record—transactions, workflow state, historical context, permissions, and cross-functional relationships.

Step 2: Build the business-logic layer. The moat is not raw data alone. It is the rules engine behind the data: calculations, exceptions, workflow sequencing, compliance logic, domain constraints, and the meaning that turns raw information into action.

Step 3: Unify data and interoperability. SaaS companies with the best survival odds are increasingly those that orchestrate across fragmented systems. The winning product becomes the convergence layer rather than another isolated module.

Step 4: Embed AI into workflow execution. The durable winners are not merely adding a chatbot. They are using AI to complete work, accelerate decisions, reduce clicks, recommend actions, and improve throughput inside the operating process itself.

Step 5: Monetize outcomes and prove ROI. The final step is commercial. Companies that can demonstrate lower labor cost, faster close cycles, higher conversion, reduced churn, better compliance, or measurable productivity gains are more likely to regain premium positioning.

Two additional solutions SaaS companies should pursue now

1) Rebuild pricing and packaging around measurable value. Too many software companies still price as if seats and feature bundles are the natural commercial unit. In an AI-sensitive market, that is increasingly weak. Pricing tied to throughput, automation delivered, dollars saved, revenue influenced, risk reduced, or cycle time compressed is more defensible because it maps directly to customer economics and outcomes rather than software form factors.

2) Restore credibility through capital discipline and strategic optionality. The market is rewarding discipline again. That means tighter expense control, better sales efficiency, more visible free-cash-flow conversion, and less dependence on distant narrative upside. It also means boards should evaluate optionality actively: partner where distribution matters, divest non-core modules where focus matters, and assess whether private ownership, strategic sale, or merger may unlock more value than insisting on independence.

A practical operating framework for management teams

Boards and executive teams should translate the valuation debate into a practical operating agenda. First, identify where the company is vulnerable to AI substitution. Second, identify which data assets are proprietary and which are standardized. Third, measure how much customer value is convenience versus true mission-critical dependency. Fourth, redesign the roadmap so AI performs work, not just assists with it. Fifth, restate the go-to-market narrative around hard ROI and cash generation.

The important point is that a valuation rescue rarely begins in investor relations. It begins in product architecture, pricing strategy, capital allocation, and proof of economic value. Public investors eventually notice those changes, but management teams need to build them before they can market them.

Where this leaves private SaaS companies

Private SaaS companies should not assume that the public-market drawdown is someone else’s problem. Lower public multiples affect sponsor appetite, debt capacity, M&A framing, and what strategic buyers are willing to pay. They also increase the burden of proof in fundraising. A company that previously could raise capital on category excitement now needs sharper evidence on retention quality, margin durability, customer concentration, AI readiness, and path to cash flow.

That does not mean the opportunity is gone. It means the winners will be more specific. The next premium SaaS companies are likely to be those that control important operating data, encode hard-to-replicate domain logic, insert AI into the daily flow of work, and commercialize against measurable outcomes. In other words, the valuation premium is not disappearing; it is being reassigned.

Conclusion

The collapse in SaaS valuations is best understood not as the death of software, but as a reset in what the market believes software is worth. Over the last six months, the repricing has been sharp, visible, and psychologically important. Investors have become less interested in software as a category and more interested in whether a specific company owns durable economic leverage.

For management teams, the answer is not nostalgia for the old multiple environment. It is reinvention. Follow the 5 steps to transition to an AI-first company. Rebuild pricing around measurable value. Restore capital discipline and preserve strategic optionality. The SaaS companies that do those things best are the ones most likely to emerge from this reset with stronger customer relevance, stronger economics, and eventually stronger valuations.

Selected source notes

• Reuters, February 5, 2026: U.S. software stocks slammed on mounting fears over AI disruption; reported the sector had shed about $1 trillion in market value since January 28.
• Reuters, February 25, 2026: Workday hits over five-year low as sluggish sales forecast sparks AI disruption fears; included the Piper Sandler commentary and Aneel Bhusri remarks.
• Reuters, March 10, 2026: Deutsche Bank turned software ‘overweight’ after the selloff.
• Reuters, March 12, 2026: Software companies fight back against fears that AI will kill them; included comments from Mike Sicilia, Rebecca Wettemann, and James St. Aubin.
• Reuters, April 10, 2025: TSX ends lower as tech shares slide; included Lorne Steinberg’s ‘priced for perfection’ comment on highflying software names.
• The Five Steps to an AI-first transition came from the company Apex, a company that built a “SaaS bridge” for AI transition.
• Historical stock-price points for Salesforce, ServiceNow, Workday, HubSpot, and Constellation Software were compiled from publicly surfaced historical price pages returned by web search on March 12, 2026.

About the Author

Brad Kayton is the U.S. Regional Director for the CFO Centre, the world’s largest and oldest fractional CFO practice.  Brad has been a fractional CFO for the last 12 years (4+ years in his position at the CFO Centre), a C-level executive for 20+ years, a CEO three times for VC-backed firms, a CMO/VP of marketing, COO twice, and a general partner for an angel capital fund. As an entrepreneur for 20+ years, he has consistently been a key member of successful technology companies, big and small. Beginning as a computer programmer out of college and working his way up, Mr. Kayton has held positions and titles in companies in most functional areas, from the technical side, to marketing and sales leadership roles, to financial and administrative areas, as well as board member including board governance and board secretary roles. He has co-founded four SaaS / high-tech firms over the years, two with large exits. As a financial advisor and professional he has extensive experience with SaaS companies and is considered an expert in growth SaaS ventures. Mr. Kayton holds an MBA and MA, a Series 65 license as a Registered Investment Advisor, and a Certified Exit Planning Advisor (CEPA) designation.