The SaaSpocalypse that wasn’t – how Salesforce, Booking and IBM are thriving with AI

The article argues the “SaaSpocalypse” AI-obsolescence fears are overstated and highlights three long-term AI beneficiaries: Salesforce, Booking Holdings, and IBM. It cites Salesforce’s Agentforce ramp to $1.5B in annual recurring revenue in 18 months (from $100M), 216T customer records processed this year, and a 104T records ingested last quarter, while Booking notes AI-sourced traffic remains <1% of room nights and IBM says AI now drives half of new consulting signings and contributed to IBM’s AI business more than doubling over the past year. Overall, it claims markets have punished these stocks for the wrong reason, shifting positioning toward AI moat-driven software rather than AI-driven obsolescence.

Analysis

The market is still treating AI as a margin destroyer for application software, but the more durable setup is vendor concentration: the platforms that already sit on identity, workflow, billing, and audit trails should gain pricing power because AI increases the value of trusted data and execution rails. That makes CRM the cleanest second-order winner: if agents raise activity volume, the monetization layer shifts from seat-count to usage, data retention, and workflow orchestration. IBM is a different version of the same trade—less multiple expansion, more backlog durability and mix improvement if AI services remain outcomes-based rather than labor-based.

BKNG is a beneficiary, but the edge is narrower than the narrative implies. AI may lift top-of-funnel discovery, yet the real monetization still depends on payment, dispute resolution, and last-mile inventory control; that means the biggest risk is not disintermediation, but rising traffic acquisition costs and a slower-than-expected transfer of AI-sourced demand. In the next 1-3 months, the key catalyst is commentary on AI-sourced bookings, attach rates, and whether enterprise clients are expanding budgets or merely reshuffling spend.

Contrarian view: the consensus is underappreciating how much of AI value accrues to whoever controls the customer relationship layer, but it may be overpaying for the story in stocks that already re-rated. CRM has the most asymmetric setup if agent adoption shows up in net retention and free-cash-flow conversion; BKNG looks more fully valued to the thesis unless AI traffic becomes material; IBM is a slower-burn operating leverage story. Falsifiers are simple: weaker cRPO/NDR at CRM, AI-driven booking traffic staying subscale at BKNG, or IBM backlog conversion failing to improve despite AI mix gains.

Key Decisions for Investors

• Long CRM on a 3-6 month horizon; best risk/reward if the next earnings print confirms AI attach is improving retention and upsell rather than merely boosting usage. Falsify if NRR/cRPO stalls or if management frames AI as margin-dilutive.
• Buy CRM Jan-2027 call spreads on pullbacks rather than chasing spot; the asymmetry is in multi-quarter workflow monetization, not a one-day rerate.
• Hold IBM as a slower-moving quality/AI infrastructure compounder; use weakness to accumulate if backlog mix continues shifting toward higher-margin AI services. Exit if consulting margins fail to expand despite AI mix claims.
• Avoid adding BKNG after strength; the thesis is real but likely closer to fully reflected in price. If AI-sourced traffic stays below a low-single-digit mix over the next two quarters, treat BKNG as a hold, not a fresh long.
• Relative-value idea: long CRM vs short a broad software basket/proxy if available, because the market is still over-penalizing workflow/data incumbents while underestimating their pricing power from AI.