What are investors describing as the criteria for 'category-creating' software?

Investors are defining category-creating software by its ability to perform entire jobs, rather than merely enhancing human productivity [5, 8]. This represents a fundamental shift from software as a tool to software as an agent capable of executing complex tasks [3, 16]. The primary criterion for this new category is the targeting of a "non-software TAM," where the competition is not another software product but human professional services in verticals like legal, healthcare, and education [7, 10, 13, 27]. By automating work previously done by people, this new class of software aims to capture budget from massive services markets, creating opportunities for companies to reach valuations in the tens or even hundreds of billions of dollars [6, 29]. This expansion into services is seen as the key driver for creating entirely new market categories where no classic software incumbent exists [10, 13].

This functional shift necessitates a complete overhaul of traditional software business models and metrics. The per-seat licensing model of SaaS is considered ill-suited for software that performs a job; instead, investors are looking for companies built on outcome-based pricing, where value is tied to a result like a sale made or a problem solved [2, 9, 16]. This transition presents a significant change-management problem for incumbents, creating an opening for new entrants [2, 20]. Consequently, financial evaluation is also evolving. The classic "triple, triple, double, double" growth trajectory is being replaced by expectations of explosive, non-linear adoption curves for successful AI products . Some investors are now prioritizing **gross dollar retention** as the single most critical metric, viewing it as the truest indicator of product-market fit over net retention, which can mask customer churn [11, 18, 19]. Others point to the change in net new bookings, or the "second derivative of growth," as a more powerful predictor of future success than overall revenue growth [14, 22].

There is a notable tension among investors regarding the impact of this shift on incumbent software companies. One camp argues that AI is a disruptive platform shift, similar to the internet or cloud, that will cause a sell-off in traditional SaaS as capital rotates to foundational models and AI-native applications [1, 4, 25]. This view holds that incumbents are constrained by their existing business models and will struggle to adapt, creating a market that favors "best of breed" solutions from new challengers [2, 20, 26]. In direct contrast, another investment thesis posits that AI will be a **massive TAM accelerator** for established leaders [11, 12, 15]. This perspective suggests incumbents can leverage AI to make their complex, deeply integrated products more accessible to a wider audience, thereby expanding their market and solidifying their position [21, 23]. A third, more nuanced view suggests both can be true: incumbents will get bigger, but the market will also see the emergence of many new $10 billion to $100 billion companies in completely new categories .

Ultimately, the emerging investment thesis for category-creating software centers on the potential for massive scale and defensibility in an agent-centric world. Venture capitalists are applying a **"$100B public company test,"** favoring businesses with vast addressable markets and clear multi-product potential . While AI-native companies are expected to have structurally lower gross margins due to high inference costs, investors anticipate higher terminal operating margins driven by superior operational efficiency in sales and engineering . Defensibility is no longer just about features, but about owning a direct user relationship or a critical system of record that is difficult to displace . The core evaluation criterion is a company's ability to build for a future where software agents are the primary users and operators, a paradigm that redefines the nature of software itself [3, 28].