AI disrupts SaaS? Citigroup: The software industry will enter a "winner-takes-all" era of major differentiation

Artificial intelligence is launching a disruptive impact on the software industry, and the traditional Software as a Service (SaaS) business model is facing severe challenges.

According to news from the Chasing Wind Trading Desk, Citigroup analysts Tyler Radke and Fatima Boolani stated in a report released on September 2 that AI will not simply end the "Software as a Service" (SaaS) model, but will usher in an era of "winner takes all" differentiation, with the market clearly divided into a few high-growth winners and many stagnating laggards.

Citigroup believes that although the market is generally concerned about the disruptive impact of AI, leading to pressure on the software sector, this sell-off may be "generally excessive." However, the risks are real, especially for application software companies that rely on a seat-based business model.

AI will become a watershed, creating a significant valuation gap between different software companies. Data shows that since 2022, software companies with growth rates exceeding 20% have seen their median enterprise value/revenue (EV/Revenue) valuation nearly double to 11.7 times; while companies with growth rates below 10% have seen their valuations stagnate around 4.9 times, close to the lows during the 2016 "SaaS massacre."

Citigroup also provided an "all-weather AI portfolio," which includes Microsoft, MongoDB, Snowflake, and others.

Three Possible Impacts of AI: From Disruption to Empowerment

Citigroup's report outlines three scenarios to project the potential impact of AI on existing software vendors. These three paths depict a broad spectrum from the worst disruption to the best empowerment.

• Bear Case: In this scenario, the development of AI and large language models (LLMs) will make it exceptionally easy to create customized applications, directly impacting or even replacing existing software vendors. At the same time, AI will accelerate the decline in the number of knowledge workers, posing a heavy blow to the seat-based SaaS model. In this case, application software companies like ADBE, CRM, NICE, and backend office software companies like ASAN, WDAY will face the greatest risks. Meanwhile, "hyperscalers" providing database, computing, and other services, as well as data infrastructure companies (such as MSFT, MDB, SNOW, ORCL), will become the main beneficiaries.
• Base Case: This is the scenario that Citigroup believes is most likely to occur. Existing giants with innovative capabilities can successfully develop AI products and commercialize them, with new revenues sufficient to offset the pressures faced by the seat-based model, thereby maintaining overall revenue growth. This is neutral to slightly positive for most SaaS models, while hyperscalers and data infrastructure providers remain winners
• Bull Case: In this most ideal scenario, existing software giants will lead AI innovation, achieving a renewed acceleration in revenue growth through a new product cycle. Even if "seats" come under some pressure, they can successfully shift to a results-based or usage-based charging model. This will be a significant boon for many existing software companies (such as MSFT, CRM, NOW) and may spur more merger and acquisition activities targeting successful AI startups.

"Winner Takes All": Valuation Divergence Has Become a Reality

The market has sensed the winds of change and is pricing future winners and losers with "real money." Data shows that there has been a sharp divergence in software company valuations.

According to Factset, as of August 26, 2025, software companies with a growth rate exceeding 20% have a median expected price-to-sales ratio (EV/Revenue NTM) of 11.7 times, nearly doubling from the low point in 2022. In contrast, companies with a growth rate below 10% have a median valuation of only 3.5 times, which is similar to the low point during the 2016 "SaaS-acre" (SaaS massacre) period.

Free cash flow (FCF) valuations also show the same divergence trend: the expected EV/FCF median for high-growth companies is as high as 56.2 times, while low-growth companies only stand at 16.0 times. This clearly indicates that capital is concentrating on a few high-growth companies deemed capable of riding the AI wave, while other companies face valuation pressure.

The report compares the current sentiment to the 2016 "SaaS-acre," when market concerns over slowing growth and business models led to a valuation crash. Citigroup believes that while overall valuations have not yet reached the absolute lows of 2016, the pressure is mainly concentrated on the low-growth group, while high-quality high-growth companies maintain or even expand their valuation premiums.

Seeking "All-Weather" Winners: Citigroup's AI Investment Strategy

In the face of a complex outlook, Citigroup has proposed an "All-Weather AI Basket" consisting of 12 companies aimed at identifying those that possess strong competitiveness regardless of the scenario.

• Applications and Data Software: Microsoft is seen as a core winner, with its presence in AI infrastructure (Azure), AI applications (Copilot), and AI search. MongoDB (MDB) and Snowflake (SNOW), as leading data management platforms, will benefit from AI-driven data volumes and product cycles. AI infrastructure provider CRWV is also favored due to its business model being decoupled from upper-layer applications.
• Infrastructure and Cybersecurity: Datadog and Dynatrace, with their mature "pay-as-you-go" model, can effectively avoid the risks of "seat-based charging." CrowdStrike, Palo Alto Networks, and Rubrik benefit from the critical nature of cybersecurity, regardless of how AI developsCybersecurity budgets will remain a high priority.
• Back-office software: Intuit has built a strong moat with its vast data footprint and distribution network. Pegasystems, on the other hand, occupies a unique position in the market by providing the underlying architecture for AI workflows through its unique AI Blueprint approach