Shares of several leading software companies were hammered this past week amid increased concern that AI will erode their businesses.
We asked Seeking Alpha analysts Louis Gerard and Brett Ashcroft Green for their thoughts on how much of a threat AI poses to the software sector.
Louis Gerard: I believe that there is a structural threat, mainly being the lower barrier to entry as AI acts as a force multiplier, creating a threat for established firms. Small teams can now build products that used to require dozens of engineers, meaning that the moat for many SaaS companies has been lowered. Plus, if software becomes significantly cheaper and faster to produce, the premium pricing models of the past may very well become relics.
Looking at specific companies, I would say that Microsoft (MSFT) will be a winner as it owns the infrastructure that integrates GitHub Copilot and Azure AI. The company actively turned the threat into a subscription revenue machine. ServiceNow (NOW) is likewise using AI to automate relatively boring enterprise workflows, meaning that the threat is turned more towards manual consultancy firms and not the software itself.
Companies at high risk are Adobe (ADBE) and Figma (FIG). Adobe is facing competition that is more lightweight and doesn’t have the bloat of 30 years of legacy code. For Figma, the threat would be the erosion of seat-based pricing as demand for licenses for 50-person design teams shrinks.
Brett Ashcroft Green: This is a topic I’ve contemplated pretty thoroughly as of late. Systems like Replit, Alphabet’s (GOOG) (GOOGL) Gemini/Google AI Studio, and Anthropic’s (ANTHRO) Claude can now all make apps from vibe-coding prompts. I’ve made some screeners in the past and some volatility trading algo apps that I’m sure would have cost me thousands if outsourced to a developer. This also comes at little cost to the user and is even more of a drop in the bucket on the corporate side.
While companies can now make their own custom enterprise software, the backend making sure consumer data is safe is still not a walk in the park and is costly. SaaS corporations have already put tons of capital into that backend, unseen part of the business. I also doubt that many companies outside the tech space would seek to replace their enterprise software with custom software. Even if it could theoretically be done by an amateur, many companies would not want to take the reputational risk of their vibe-coded software having flaws or backdoors they were unaware of.
Furthermore, IP is real. People are used to Adobe (ADBE), DocuSign (DOCU), and other apps they have implemented. Flipping the switch and going from brand name to generic (which is basically what you do when using your own app vs. a known app) could also turn off clients, as the software platforms being used often times build as much confidence in the consumer as the business itself. Short-term SaaS sell-off, but long-term, I think AI just helps these established brand names build and improve trusted software better and faster.