Software – particularly software as a service (SaaS) and vertical market software (VMS) – have been excellent investments. However, on February 3rd, the markets punished these software companies indiscriminately, which was characterized as a 5-sigma move—an event with an estimated probability of roughly 1 in 30 million. Investment success requires a variant perception - a view that differs from consensus and ultimately proves correct. This note is to express Heron Bay’s point of view as owners of many of the businesses that experienced price declines amidst concerns that Artificial Intelligence (AI) will erode their competitive advantages. We hope this creates some comfort in long-term ownership of these companies despite the narrative.
Imagine returning to the technology of the 1980’s when a carbon paper imprint of a credit card was physically delivered to a bank for manual processing. By contrast, our economy transacts at light speed. Software is the business conduit that facilitates much of the speedy data transfers needed in our economy. The companies behind the software tend to be deeply embedded in customer workflows, creating high switching costs: mission-critical functions with proprietary data and process would create considerable disruption to the business by changing vendors. Due to high switching costs, there has been low customer churn (lost clients), recurring revenue streams, and annuity-like economics. Capital intensity is typically modest for these companies, allowing for significant operating leverage as scale increases. Growth has often followed a “land and expand” model, where initial adoption leads to broader enterprise penetration over time. Entrenchment effects are frequently even stronger in VMS, where products are highly customized to specific industries and use cases. New features within each software platform tend to create increasing price dynamics and average sales per customer tends to increase over time. This dynamic makes for consistently high profitability and a resilient business model.
However, capitalism is characterized by creative destruction. Large profit pools inevitably attract competition, and AI - along with “vibe coding” and agentic software - promises to materially lower the cost of developing and deploying custom solutions. A reasonable concern is that as software development becomes cheaper and more accessible, enterprises may reduce their reliance on third-party SaaS providers. While the fear that AI will deteriorate the protective moats of these companies, we believe that the indiscriminate market activity is too broad based to capture the nuances of each business. We believe that these software companies have been unfairly punished.
First, not all software serves the same function. Disruption risk varies significantly depending on what “job” the software performs.
Second, some software products function as systems of record (for example, in clinical trials, ERP, accounting, payroll, or compliance). In these contexts, errors can be costly, legally risky, or operationally catastrophic. This creates a high barrier to displacement that AI alone is unlikely to overcome. Usage-based pricing can be employed to counter any erosion in seats sold.
Third, software companies are themselves major users of AI tools. Like many enterprises, they are likely to improve productivity, streamline operations, and reduce headcount over time. This should translate into higher revenue per employee and stronger margins.
Fourth, many end users have little appetite to build custom software in-house, particularly in non-core areas of their business. Even where technically feasible, doing so may not be economically rational or introduce unnecessary risk.
In 7 Powers: The Foundations of Business Strategy, Hamilton Helmer identifies “cornered resources” as one of the key attributes of enduring franchises. In software, this often takes the form of proprietary data, unique standards, or compliance-critical infrastructure.
Consider Adobe (ADBE), one of our portfolio holdings. Adobe effectively owns the PDF standard, which has become the default format for legal, regulatory, and enterprise documentation. This represents a durable cornered resource in Document Cloud. In contrast, portions of Creative Cloud may face erosion among casual and SMB users as generative tools proliferate. For professional and enterprise users, however, Adobe’s file standards, workflows, and skill capital remain meaningful barriers to displacement, albeit potentially somewhat diminished at the margin.
FactSet (FDS) may, in fact, benefit from the rise of AI. Its true asset is not merely its terminal interface but its proprietary, normalized financial data infrastructure. Decades of institutional expertise in cleaning, structuring, and maintaining financial datasets are difficult to replicate. Many other platforms rely on FactSet data as an input, reinforcing its role as a foundational provider rather than a replaceable application layer. A similar dynamic applies to Salesforce (CRM), where the core customer data model and governance layer are likely to become more valuable—not less—as AI-driven applications proliferate.
The key point is that large SaaS and VMS companies are not monolithic. Within the same firm, some business lines may face pressure while others become relatively advantaged. Our task as investors is to distinguish between those outcomes—identifying which franchises have been unfairly penalized and which face structurally declining economics.
As a general rule, the most resilient software businesses are likely to exhibit one or more of these characteristics:
- Ownership of proprietary data or standards
- Deep embedding in customer workflows
- Material audit, regulatory, or compliance requirements
- High switching costs tied to systems of record rather than user interfaces
Conversely, the most vulnerable businesses are likely to be single-feature tools that sit at the application layer, lack proprietary data, and can be easily replicated by AI agents interacting with underlying systems.
Jensen Huang, Nvidia CEO recently commented "There's this notion that the software industry is in decline and will be replaced by AI," Huang said late Tuesday at an event hosted by Cisco, hours after a selloff in U.S. software stocks.
He argued it makes sense for AI to use existing tools to accomplish tasks, rather than reinventing them, using the examples of ServiceNow, SAP, Cadence and Synopsys. "Would you use a hammer or invent a new hammer?" he asked.
"There's a whole bunch of software companies whose stock prices are under a lot of pressure because somehow AI is going to replace them," Huang said. "It is the most illogical thing in the world."
There are few universal truths in investing, but one remains enduring: the intrinsic value of any asset is the present value of its future cash flows. Once we identify businesses with durable advantages, we can stress-test valuations using conservative assumptions around growth, margins, and capital intensity. Those that have been unduly punished by recent market moves may represent attractive opportunities. This is what we at HBCM have been busy doing. While the AI led software disruption may be unique – market dynamics where the collective judgement of the market gets clouded happens with fair regularity. We have historically taken advantage of such opportunities and will do so again.
The team at Heron Bay Capital Management is committed to communicating our thoughts and research with clients. If you have further questions or concerns, please don’t hesitate to reach out.