Context and Report Objective

This report is intended to assess how artificial intelligence (“AI”) is reshaping the software investment landscape and its implications on private equity and venture capital portfolios with general exposure to software and technology. This commentary reflects Canterbury’s synthesis of PitchBook research and perspectives shared by select private equity and venture managers investing in the space.

Framing the Software Reset

The technology landscape is undergoing a meaningful reset, reflected in a widening divergence between private and public market activity. PitchBook data shows that private equity exposure to software reached a record 18% of U.S. deal value in 2025, even as public software valuations fell to multiyear lows. Public software multiples declined to more than one standard deviation below their eight-year average as investors repriced the risk of AI-driven disruption. Whereas prior cycles often treated software as a durable growth sector with inflation-resistant characteristics, the current environment has shifted the debate toward the long-term resilience of the traditional SaaS model. Even so, enterprise security requirements, switching costs, and the operational burden associated with retraining users may limit the pace of broad-based displacement across incumbent platforms.

Software and AI Across Private Equity

According to the Pitchbook, the current period can be viewed as the early stage of a long-duration technology transition with the potential to reshape enterprise workflows. In 2025, AI venture capital deal value reached a record $243.9 billion, although funding remained concentrated in a relatively small number of large rounds involving foundation model providers. More recently, capital appears to be broadening beyond horizontal platforms toward vertical applications, which led AI segments in both deal value and transaction count in late 2025. Within private equity, software portfolios are beginning to differentiate between businesses with stronger defensibility, including systems of record and companies with proprietary data, and application software companies with greater exposure to pricing, product, or competitive pressure. Historically, periods of relative dislocation in technology-focused private equity have created attractive deployment opportunities, particularly where valuation compression has outpaced any corresponding deterioration in underlying fundamentals. PitchBook’s broader conclusion is that lower development costs and faster application-layer
Innovation is increasing competitive pressure across portions of the legacy software landscape.

AI and Software Value Capture

AI is changing the cost, speed, and scope of software development, which is beginning to alter where value accrues across the software stack. Four perspectives appear most relevant.

1. Expansion of Software’s Addressable Market

Managers broadly argue that, while legacy software models are under pressure, AI expands software’s opportunity set by enabling automation of increasingly complex outcomes and allowing software to capture spend that historically sat outside traditional IT budgets.

• AI is not an extinction event for software, but it is likely to reallocate economics toward orchestration tools that control how work is performed rather than simply acting as systems of record.
• AI should enhance product capability, expand addressable markets, and position next generation software companies to capture new AI related budgets tied to mission critical workflows that were previously manual.
• AI driven automation is expected to generate substantial productivity gains across the global economy, potentially resulting in more than $10 trillion of economic value creation over the next decade

2. Defensibility Through Vertical Depth and Data Moats

Competitive advantage is increasingly shifting toward proprietary datasets and ownership of the business logic layer embedded within specialized
workflows.

• Defensibility should be assessed against the cost of workflow failure, particularly in deterministic domains such as payroll, tax, and compliance where enterprises have limited tolerance for error. Durable moats are increasingly built on proprietary data gravity and deep integrations that general purpose AI cannot easily replicate, as illustrated by assets such as Orca AI’s visual maritime sensor data.
• The most defensible layers of the software stack are data architecture and business logic, while workflow and user interface layers are more exposed to AI native competition.
• Vertical software serving SMB customers may remain relatively insulated because these customers typically lack the internal technical capabilities to develop their own AI tools and continue to rely on specialized software for domain specific tasks.

3. AI Native Performance Benchmarks and Growth Efficiency

AI driven productivity gains are resetting traditional performance benchmarks, with greater emphasis now placed on growth velocity and operating leverage rather than headcount supported expansion.

• Traditional KPIs are giving way to metrics such as autonomous resolution rate, which more directly measures the extent to which a platform can complete tasks without human intervention. Investors should also monitor leading indicators of competitive vulnerability, including employee attrition to AI native competitors and customer pushback on contract duration or price.
• Companies with modern codebases are better positioned to adopt agentic solutions quickly, which may allow them to scale more efficiently and take share from slower moving legacy incumbents.

4. Technological Risk and Obsolescence

The pace of AI development requires a more disciplined approach to underwriting technological durability and managing obsolescence risk.

• The value of frontier technology is tied not only to innovation, but also to a company’s ability to architect, transfer, and manage risk effectively. Without that capability, scalability may be limited.
• Companies created through consolidations and roll ups may be especially vulnerable because leverage and multiple legacy codebases can distract from the AI first transformation now required. In contrast, modular, API first architectures are better positioned than monolithic systems to support safe, permissioned AI driven actions.
• The primary risk in the sector is technological obsolescence, where a company’s core machine learning capability can be displaced quickly by a superior model or algorithm developed elsewhere.
• Niche software businesses with limited differentiation face significant pressure, making founder quality increasingly important in a market where horizontal AI capabilities remain directly in the path of foundation model expansion.

Conclusion

Based on these insights, the current market environment is best understood as a repricing and redistribution of software value rather than a broad collapse in software relevance. Public markets have moved quickly, and perhaps too aggressively, to reprice AI related disruption, prompting dramatic rhetoric such as “SaaSopalypse” and “SaaS megaddon,” while private markets have remained active and are beginning to separate durable businesses from more vulnerable assets. Across private equity and venture capital, the emerging picture is not one of broad software impairment, but of value being redistributed toward companies with stronger data assets, deeper workflow ownership, and greater ability to adapt.

Key Takeaways

Against that backdrop, several key implications for portfolio construction and underwriting are emerging:

• Portfolios are beginning to bifurcate between durable winners and vulnerable assets.
• Systems of record, orchestration layers, vertical software, and proprietary data-rich businesses appear best positioned.
• AI is expanding addressable markets by shifting value from labor to software.
• Defensibility is moving toward data, business logic, integrations, and mission-critical workflows.
• Performance expectations now favor velocity, leverage, autonomy, and architectural readiness.
• Obsolescence risk is becoming more central to underwriting.
• The reset may create an attractive deployment window for the right assets.
• Success will depend on separating AI beneficiaries from AI exposed businesses.

This document is confidential and prepared specifically for clients of Canterbury Consulting. No part of this document may be published, reproduced, or distributed without the prior written consent of Canterbury Consulting. The opinions provided reflect the views of the managers and do not necessarily represent the views of Canterbury. This presentation is not intended to be a solicitation to engage Canterbury, nor is it a recommendation to invest with any of the managers listed in this presentation.