What are the most and least promising categories to invest in, within SMB software?

Investor sentiment towards software has soured dramatically, creating a challenging but bifurcated landscape for new investments. The sector has experienced a significant capital rotation away from traditional Software-as-a-Service (SaaS) and into AI infrastructure and native applications [9, 17, 20]. This shift is evidenced by reports from early 2024 showing a **10-to-1 ratio** of selling versus buying volume in software stocks, with some labeling the sector "uninvestable" [6, 13]. The leveraged finance market has similarly pivoted, with software moving from a leading category for lending to being almost unfinanceable . This widespread sell-off, which has impacted both high- and low-quality companies , began with the emergence of large language models from firms like OpenAI and Anthropic , causing investors to fear disruption and wait on the sidelines to observe retention rates over the next two quarters before potentially reinvesting .

The primary determinant of a software company's promise now hinges on its relationship with AI, which creates both existential threats and greenfield opportunities. Narrow, single-function "point solutions" are considered **highly vulnerable** to having their functionality absorbed by larger platforms or replicated by simple AI tools . In contrast, a contrarian view holds that the defensibility of incumbent software platforms is underestimated [9, 27]. While some investors assert that software moats have never been real, particularly in dynamic fields like cybersecurity , others argue that barriers like deep enterprise integrations, customer trust, data governance, and security are becoming more important than code itself [9, 26, 30]. This suggests value may accrue both to well-positioned incumbents that effectively integrate AI and to new entrants building AI-native solutions for previously unserved markets . Investors are therefore advised to prioritize companies with deep vertical expertise and proprietary data moats, which are most defensible against disruption .

The most promising investment categories are those directly leveraging AI to enter new markets or provide foundational infrastructure. Analysts predict that AI will create **entirely new software categories** in verticals like legal, healthcare, and education, where dominant software incumbents may not yet exist [3, 15, 16]. Capital is flowing towards AI infrastructure and hardware companies, which are expected to outperform SaaS for the near future [14, 20]. Other promising areas include dual-use technology companies building software for critical infrastructure sectors like logistics and aviation . In this environment, traditional evaluation metrics are also being reconsidered, with some influential investors now asserting that gross dollar retention is a more reliable indicator of long-term health than the more popular net retention metric [10, 22].

Conversely, several software categories appear less promising or face significant headwinds. Traditional infrastructure software companies that would have been strong investments just two years ago are now struggling to raise capital as focus shifts to AI . The data streaming market has also been cited as a historically difficult category for venture capital investment . While the broader SaaS sector is out of favor, the indiscriminate sell-off may present contrarian opportunities in select, established public companies trading at historically low free cash flow multiples [9, 19, 27]. However, a significant risk remains that investors will permanently re-rate the enterprise SaaS sector to **much lower valuation multiples**, similar to the historical de-rating of PC maker stocks .