Can Software Survive AI?

Episode Summary

Executive Summary: The discussion argues that AI is a real disruption risk for software, but not an existential one for all incumbents. Winners will be software companies with deep domain data, modernized tech stacks, strong product roadmaps, and the ability to monetize AI. The sector may stabilize as fundamentals improve, competition normalizes, and investors rotate toward value and GARP names.

Main Topics: Why AI disruption fears intensified (Priority: 5/5): Investors shifted from AI enthusiasm to software selloff as coding tools improved, enterprise AI adoption accelerated, and competition from frontier AI players increased. Defensible moats in software (Priority: 5/5): Incumbents can still defend themselves through proprietary data, domain expertise, context, and long-built workflows, especially in areas like cybersecurity and enterprise platforms. What makes a moat durable (Priority: 5/5): Durability depends on replatforming legacy tech, investing in organic innovation, making smart acquisitions, and proving AI monetization and usage can create pricing power. Sector stabilization and fundamentals (Priority: 4/5): Software stocks could recover as ARR growth, LTV/CAC, and other metrics stabilize after several years of deterioration, with earnings reactions confirming improving investor focus. Investor base is shifting (Priority: 4/5): The cohort owning software is broadening from growth-only buyers to value, gap-earnings, and GARP investors who are more sensitive to margins, dilution, and cash flow quality. AI adoption inside software firms (Priority: 4/5): Established software leaders are using AI to boost internal productivity and selectively acquiring external innovation rather than trying to invent every frontier capability themselves.

Key Arguments: AI raised disruption risk because enterprise-grade coding tools and wrappers made generative AI accessible beyond developers, accelerating adoption across workflows. Incumbent software firms still have meaningful moats when they own rich proprietary data, deep customer context, and long-term domain expertise. A moat is only credible if the incumbent can prove better customer outcomes, better monetization, and stronger AI-powered product performance than plug-in alternatives or startups. The strongest leaders will modernize their stacks, reduce tech debt, keep engineering talent productive, and use M&A to fill roadmap gaps. Software fundamentals have weakened for years, but stabilization in metrics and positive post-earnings stock reactions suggest a potential recovery phase. Investor scrutiny is moving toward gap earnings, stock-based compensation dilution, and real pricing power rather than just revenue growth or non-GAAP optics. The investor base is widening because software growth has become scarcer, making valuation discipline and quality-of-earnings more important. Large incumbents are not standing still; they can adopt a 'drinking their own champagne' approach to AI and then acquire the best private-market innovations. Cybersecurity is cited as a model: active adversaries force continuous innovation, and M&A has helped platforms strengthen and customer acquisition costs decline.

Data Points: Timeframe of recent concern: last three months - Period when AI disruption fears and software stock sell-offs intensified Claude Code accessibility: January - Anthropic wrapped Claude Code for everyday knowledge workers CrowdStrike data-building period: 10+ years - Illustrates long-term data/experience moat in cybersecurity Software metrics deterioration: 4 years - ARR growth, LTV/CAC, and related metrics weakened after the 2020-2021 surge Digital transformation acceleration: 2020 and 2021 - Period when software demand and metrics were especially strong Median software growth rate then: north of 20% - Median company growth in 2020-2021 Median software growth rate now: closer to 10% - Current growth environment for software companies Okta free cash flow margin change: 0% to closer to 20% - Example of margin expansion over 18 months Okta margin expansion period: 18 months - Illustrates how investors pushed software firms toward better profitability Salesforce acquisitions: north of 10 acquisitions - Cited as recent example of M&A-led roadmap building Recording date: March 10, 2026 - Episode recording date stated at the end

Pivotal Quotes: "yes, selectively, the moats will endure" — Gabriella Borges: Direct answer on whether incumbents can maintain defensible advantages "the numbers need to contradict the narrative" — Gabriella Borges: Explains why improving fundamentals must override AI disruption fears "we finally at the point where you can actually start to do some interesting analysis around gap earnings power" — Gabriella Borges: Why investors are refocusing on earnings quality and stock-based compensation

Implications: Software is not doomed, but only differentiated incumbents are likely to win. Investors should focus on data, product integration, monetization, margins, and AI execution, while expecting sharper separation between leaders and laggards.