Anthropic, OpenAI, and Google: AI is Squeezing Software Companies from Both Ends
An English long post points out that the software industry is facing "dual pressure from above and below": at the top, Anthropic, OpenAI, and Google are entering the SaaS field directly with their own products, while at the bottom, small teams of 1 to 3 people are using AI to create usable software at lower costs and faster speeds. The author of the post believes this will put traditional software companies in the middle layer into a structural dilemma.
English materials also reinforce this judgment. Industry research describes the software competition in the AI era as a "system of action" struggle, emphasizing workflows rather than just functions; other analyses point out that AI enables small teams to accomplish development, operations, and delivery that previously required large teams, compressing the traditional SaaS moat.
Source: Public information
ABAB AI Insight
This judgment's value lies in elevating the impact of AI on the software industry from "replacement tools" to "industrial structural reorganization." The top model companies are focusing on platformization and vertical integration, while the small teams are capturing long-tail demands. The middle-layer SaaS companies are losing both functional premiums and scale efficiencies, with those relying on single-point products and channel lock-ins being the most affected.
This is a typical case of bidirectional compression: top companies control models, distribution, and default entry points, while bottom teams possess speed, flexibility, and low-cost experimental capabilities. Historically, similar situations have occurred many times, such as platform companies swallowing distribution layers and open-source forces breaking through old software profit pools, but the difference with AI is that it simultaneously weakens both the "cost of making software" and the "barriers to selling software."
On a deeper level, this will shift the software industry from "selling tools" to "selling results." Those who can embed AI into workflows, delivery chains, and decision-making chains will be closer to profit centers; those who merely add a bit of AI functionality to old products will be more easily squeezed out from both ends. This is not a short-term valuation fluctuation, but a repricing of software business models.