From the End of SaaS to the AI Profit Margin Super Cycle: The Fundamental Restructuring of Software Investment Logic

The Paradox of the Software Industry: Superficial Downturn, Hidden Turbulence

In the first half of 2026, the advanced software sector experienced a dramatic adjustment. Publicly traded software stocks underperformed the broader market, leveraged loans hit new lows, SaaS doomsday panic spread, and private equity funds were forced to set redemption thresholds. However, in its latest "2026 Advanced Software Launch Report," PitchBook presents a counterintuitive judgment: this downturn masks the launch of a historic 'AI Profit Margin Super Cycle.'

The core contradiction lies in: the market is pricing a completely new business model using the valuation framework of the old model. The traditional SaaS profit model based on seat subscriptions is collapsing, while the AI agent-based digital labor model is reshaping the economics of software. This is not the pain before a recession, but the pain of a structural turning point.

Why Is It Happening? The End of SaaS and the Beginning of AI

The advantage of traditional SaaS lies in predictable recurring revenue and high gross margins. But over time, issues such as product homogenization, rising customer acquisition costs, and pressure on renewal rates have made the 'growth first, profitability later' model unsustainable. The market selloff in 2025-2026 is essentially a question about the sustainability of this model.

• At the same time, the maturity of agentic AI has fundamentally changed the value creation logic of software. When AI can autonomously perform complex tasks and replace human labor hours, the pricing basis of software shifts from 'per user per month' to 'per task/result.' This brings about two changes:
• Profit Margin Explosion: Digital labor has extremely low marginal costs. Once deployed, revenue growth no longer requires linear increases in labor or server costs, and profit margins can break through the ceiling of traditional SaaS.
• Moat Reconstruction: By combining AI reasoning capabilities with proprietary enterprise data, the resulting 'enterprise IP' becomes a competitive barrier that is difficult to replicate. PitchBook calls this a moat of 'scaling human judgment.'

Who Will Benefit? Active Management Investors and Fast Adopters

The winners of this super cycle are not those managers who passively wait for the value of existing SaaS assets to return, but the activists who actively promote AI transformation. Two cases in the report are highly representative:

Case 1: The Warning of Medallia's Equity Wiped Out On June 17, 2026, customer experience management company Medallia was acquired by a loan consortium, resulting in the complete wipeout of its original shareholder Thoma Bravo's $5.1 billion equity. The trigger was Medallia's slowness on the AI roadmap. The new investors immediately committed $150 million to accelerate the adoption of agentic AI. The clear signal from this event is: in the agentic era, passive management is a huge liability. Capital is willing to use extreme measures to cleanse old management and shareholders, just to switch faster to AI-driven growth models.Case 2: Blackstone and Anthropic's Cross-Stack Alliance In May 2026, Blackstone, Anthropic, and OpenAI Deployment Company respectively established joint ventures. This "capital + AI model + deployment" consortium marks the upgrade of survival standards from single product innovation to cross-stack intelligent ecosystem collaboration. By integrating capital, top-tier models, and industry deployment capabilities, they are building a new generation of software delivery platforms.

• Beneficiaries include:
• Active management faction in private equity: PE institutions that can identify and promote their portfolio companies to adopt AgentOps infrastructure and restructure business models.
• Enterprises with unique data assets: Their data, combined with agentic AI, can be transformed into high-barrier industry solutions.
• AI infrastructure service providers: Including model layer, agent orchestration, monitoring and security tools, etc.

• Those under pressure are:
• Investors still relying on traditional SaaS valuation logic: If they continue to view software companies using low-growth discounted cash flow models, they will miss opportunities for asset value reassessment.
• Enterprises with aging management and hesitation in AI transformation: They may become the next Medallia.

Implications for Investors: From Holding to Participation, From Valuation to Ecosystem

This report is essentially an investment action guide. The core conclusion is: The rules of the game for software investment have changed.

First, valuation methodology needs updating. The traditional EV/Revenue multiples for SaaS are no longer applicable. Investors need to assess a company's "agentic readiness": AI revenue proportion, penetration of outcome-based pricing contracts, and the depth of the AI moat (proprietary data + model fine-tuning).

Second, capital allocation must be proactive. PitchBook emphasizes that "extremely active management" is necessary. This means investors cannot just buy and hold; they need to deeply participate in their portfolio companies' AI strategy formulation, introduction of ecosystem partners, and business model migration. The Blackstone-Anthropic joint venture is the ultimate manifestation of active management.

Third, ecosystem collaboration becomes a competitive barrier. It is difficult for a single company to simultaneously possess top-tier models, industry data, and large-scale deployment capabilities. Cross-stack alliances will dominate the future. The role of private equity funds is shifting from financial investors to platform builders.

Outlook for the Next 3-5 Years1. Software as a Service (SaaS) will evolve into Software as Labor. Most enterprise software will no longer be charged per user, but based on actual output (e.g., number of tickets processed, lines of code generated, customer service sessions completed). This will significantly boost revenue for top-tier companies while squeezing the survival space for small and medium-sized enterprises. 2. AgentOps becomes a new IT infrastructure. Similar to the role of DevOps in the cloud computing era, agent operations, monitoring, and security will give rise to a new batch of unicorns. 3. North American tech capital continues to concentrate on AI infrastructure. Private equity and venture capital will allocate a large proportion to agentic AI-related areas, especially platforms that can deliver "result guarantees." Policy dividends from the CHIPS Act and IRA will further accelerate this trend. 4. Corporate M&A will become more aggressive. Mid-sized software companies with data assets or industry know-how will become targets for both PE firms and tech giants, with valuation logic shifting from revenue-based multiples to AI synergy-based ones.

Conclusion: The Super Cycle Window Is Brief, Act Fast

PitchBook's report title states bluntly, "The window of opportunity is short." The current valuation trough is a rare window to buy into AI transformation opportunities, but only if investors possess the ability to identify and drive change. Medallia-style purges will continue to unfold, and only those players who embrace agentic AI, build ecosystem alliances, and reshape their business models can seize the initiative in the new cycle.