BIT Research | AI Meets SaaS: Who Will Be the Tears of the Era, and Who Will Turn the Tables Against the Wind?
- Core Thesis: In early 2026, Wall Street experienced an AI-driven valuation restructuring in the software industry. The traditional SaaS per-seat pricing model is facing fundamental challenges, but the industry is not ending. Instead, it is shifting towards a new paradigm centered on proprietary data, outcome-based pricing, and AI governance.
- Key Elements:
- Following Anthropic's release of products like Claude Cowork, the SaaS sector saw $285 billion in market value evaporate within 48 hours, with the sector's forward P/E ratio compressing from 84x to 22.7x, resulting in cumulative market value losses of $1-2 trillion.
- The traditional SaaS per-seat pricing model is under pressure, as the core logic is that AI agents can replace a significant amount of human labor, drastically reducing the number of software seats required by enterprises.
- Bullish arguments include: proprietary data moats (e.g., Salesforce's CRM data), high switching costs (replacement takes years and millions of dollars), and essential compliance and governance needs.
- The fundamentals of the leading companies being sold off remain resilient: ServiceNow revenue growth is 22%, Salesforce revenue is $41.5 billion, and HubSpot growth is 19%, indicating the industry is not collapsing.
- SaaS companies have three major counter-strategies: building proprietary AI agents, shifting pricing models from "per-head" to "per-outcome" (what Goldman Sachs calls Results-as-a-Service), and becoming the AI governance layer.
- Half of ServiceNow's net new business is booked through non-seat-based pricing, a key data point for industry transformation; its Now Assist ACV target has been raised from $1 billion to $1.5 billion.
- Gartner predicts that by 2030, 35% of SaaS point tools will be replaced by AI agents, but ERP, HR, and compliance infrastructure will be more defensible due to high switching costs.
Key Data: Global SaaS Market (2025) ~$408 billion | SaaS sector market cap erosion ~$2 trillion | IGV Software ETF down ~22% YTD | Salesforce FY2026 revenue $41.5 billion | ServiceNow Q1 2026 revenue $3.77 billion
1. SaaSpocalypse: Defining the Historic Event of 2026
In early 2026, Wall Street witnessed the largest AI-driven valuation restructuring in software history.
Event Timeline:
- January 12: Anthropic launched Claude Cowork, a desktop AI product capable of autonomously executing multi-step workflows across applications
- January 30: Anthropic open-sourced 11 business plugins, covering legal, finance, marketing, sales, and customer support
- February 3-5: Market crash. Within 48 hours, the SaaS sector lost $285 billion in market capitalization
Core Logic: Traditional SaaS charges per seat. If 10 AI agents can do the work of 100 employees, a company only needs 10 Salesforce seats, not 100. Jason Lemkin's words circulated widely on Wall Street: "If 10 AI agents can do the work of 100 reps, you need 10 Salesforce seats, not 100."
Scale of Losses: Cumulative market cap loss reached $1 to $2 trillion (from peak calculations). Thomson Reuters recorded its largest single-day drop, LegalZoom plummeted nearly 20%. The forward P/E ratio of the software sector compressed from a peak of ~84x to 22.7x.
2. Is This the End of SaaS, or the Beginning of a Transformation?
Three Bullish Arguments:
Proprietary Data Moat: General-purpose AI agents cannot replace specialized agents trained on five years of a company's proprietary CRM data. Salesforce's data is within Salesforce; ServiceNow's ticket history is within ServiceNow—these are assets general AI cannot access.
Switching Costs Underestimated by the Market: Replacing a deeply embedded enterprise software suite means years of effort, millions of dollars in costs, and retraining thousands of employees. SaaStr founder Jason Lemkin notes that building a functional application with AI coding tools accomplishes roughly only 2% of the work required to operate an enterprise software platform.
Essential Compliance and Governance Needs: In regulated industries like banking, healthcare, and government, enterprise software's value is not just automation—it's audit trails, compliance records, and access control. General-purpose AI agents currently cannot replace this functional layer.
Key Data Contradiction: At the peak of the sell-off, ServiceNow beat its guidance for the ninth consecutive time, with revenue growth accelerating to 22%. Salesforce recorded $41.5 billion in annual revenue. HubSpot maintained 19% growth. These are not the numbers of an industry in collapse.
3. How SaaS Companies Are Fighting Back: Three Strategic Pillars
Building Proprietary AI Agents: Training exclusive agents on their own platform data, rather than waiting for third-party agents to replicate their functions. Agentforce runs on Salesforce's CRM data; Now Assist runs on ServiceNow's ticket data—an advantage general AI cannot replicate.
Pricing Model Transformation: Shifting from "per head" to "per outcome" pricing. In Q1 2026, half of ServiceNow's net new business was secured through non-seat-based pricing models—the most important structural data point for the entire sector. Goldman Sachs has termed the new model "Results-as-a-Service."
Becoming the AI Governance Layer: Large enterprises need a trusted platform to centrally manage, audit, and secure the behavior of all AI agents. ServiceNow's "AI Control Tower" and Salesforce's "Agentforce Trust Layer" are competing for this critical infrastructure positioning.
4. Key Public Companies to Watch
1. Salesforce (CRM) – "Betting on Agentforce"
- FY2026 Revenue: $41.5 billion, up 10% YoY
- Agentforce Standalone ARR: $800 million, up 169% YoY; over 29,000 cumulative deals signed
- RPO (Remaining Performance Obligations): $72.4 billion, up 14% YoY, proving customers are not churning
- Authorized $50 billion stock buyback
- Key Question: Can Agentforce independently drive organic acceleration in FY2027—after stripping out the $1.1 billion contribution from the Informatica acquisition?
2. ServiceNow (NOW) – "The AI Control Tower"
- Q1 2026 Revenue: $3.77 billion, up 22% YoY (beat guidance for the ninth consecutive time)
- Now Assist ACV target raised from $1 billion to $1.5 billion, a single-quarter increase of 50%
- Renewal Rate: 97%, stable for six consecutive quarters
- Half of net new business secured through non-seat pricing
- CEO McDermott Quote: "It will be over $1.5 billion, and we need to raise it by another $500 million. It's just incredible."
- Key Question: A template-level validation of pricing model transformation; the most critical benchmark data source for the entire sector.
3. HubSpot (HUBS) – "Holding the Mid-Market Fort"
- FY2025 Revenue: $3.13 billion, up 19% YoY; stock down 70% to 80% from highs
- Bull Case: Mid-market customers are less likely to build their own AI; HubSpot's integrated ease of use remains a differentiating advantage
- Bear Case: Klarna has publicly announced replacing Salesforce contracts with AI. If the trend spreads to mid-market enterprises, structural pressure will be hard to avoid.
4. Workday (WDAY) – "The HR Data Moat"
- Employee data, payroll, talent profiles—any AI planning for human resources needs Workday's data
- Key Question: Compliance and regulatory requirements make HR software one of the most difficult SaaS categories to replace.
5. The 2026 Pricing Revolution: The End of the Seat Era
Three models are currently competing across the industry:
- Consumption-Based Pricing: Charging per query/task/token; revenue is more flexible but also more volatile
- Outcome-Based Pricing (Results-as-a-Service): Charging per completed ticket, reviewed contract, or generated lead—Goldman Sachs considers this the end-state model
- Hybrid Pricing: Seat license retains platform access, overlaid with incremental charges for AI work units—currently the most adopted approach
The Most Important Leading Indicator: Who will be the first to report a quarter where AI outcome revenue truly surpasses the seat revenue it replaces? This will be the historic data point defining the entire sector's future valuation logic.
6. Investment Risk Warnings
Not All SaaS Will Survive: Project management, document tools, simple marketing automation—these are the repetitive, rule-based tasks AI agents will conquer first. ERP, HR, compliance infrastructure—these are significantly more defensible due to switching costs and regulatory requirements. Gartner predicts that by 2030, 35% of SaaS point solutions will be replaced by AI agents, while 65% will survive, but in an evolved form.
Valuation Compression May Not Be Over: The forward P/E ratio of the software sector has compressed from 84x to 22.7x, but if disruption outpaces adaptation, there is still further downside. Distinguishing between "the sector is cheap" and "the sector *should be* cheap" is the most critical judgment to make now.
Build vs. Buy Threat: AI coding tools are significantly increasing the feasibility for large enterprises to build custom software in-house. Klarna's case is not an isolated incident; it is a trend signal worth continuous tracking.
Summary
Market participants with a preference for a more stable approach can observe changes in the fundamental resilience of the sector through ServiceNow or the IGV ETF. Those with a growth-oriented perspective can focus on tracking whether Salesforce's Agentforce business can achieve sustainable organic growth momentum. In 2026, SaaS is no longer about selling seats, but about whose platform makes a company truly indispensable—whether the workers are humans or agents.
BIT's US stock business is directly connected to licensed brokers, covering all core individual stocks and ETFs in the US market. It supports deposits and withdrawals via stablecoins, helping crypto users capture the 2026 AI stock dividend with one click. Services may vary due to jurisdictional restrictions and are not available in certain regions (including, but not limited to, Hong Kong).
Data as of April 2026. Sources include: Salesforce Inc. (SEC Form 8-K, February 25, 2026), ServiceNow Inc. (SEC Form 8-K, April 22, 2026), HubSpot Inc. (SEC Form 8-K, February 11, 2026), FinancialContent, Taskade, NxCode, Humai Blog, Goldman Sachs "Results-as-a-Service" research report, JPMorgan Software Sector Analysis, Gartner IT Spending Forecast, Precedence Research, Cirra AI, Fortune, 24/7 Wall St., Redevolution, TechStartups.
Disclaimer: This report is written by Jun, a special analyst for BIT's US stock business, for informational purposes only. The stocks and ETFs mentioned are used solely as industry case studies and for the analysis of public financial data. They do not constitute any investment advice, stock recommendations, or trading inducement. Historical performance and institutional forecasts are for reference only and do not represent future market performance or return expectations. Past performance is not indicative of future returns. Investing involves risks, including the potential loss of principal. Clients should consult a qualified financial advisor before making any investment decisions.
