Amazon Web Services Chief Executive Matt Garman said investor anxiety over artificial intelligence undermining traditional software companies has likely gone too far, even as the sector posts one of its steepest pullbacks in years.
“Look, my own opinion is that much of the fear is overblown,” Garman told CNBC’s Jon Fortt on Thursday, addressing concerns that generative AI platforms could erode the dominance of large software-as-a-service providers.
Technology shares, particularly enterprise software names, have fallen sharply in 2026 following the rapid commercialization of AI tools built on models from companies such as OpenAI and Anthropic. The selloff reflects mounting concern that AI-native applications could commoditize existing SaaS offerings or reduce enterprise spending on legacy systems.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
The iShares Expanded Tech-Software Sector ETF is down 24% this year, putting it on track for its worst performance since 2022, when inflation and rising interest rates forced companies to trim technology budgets after a surge in pandemic-driven digital investment.
Market analysts have described the correction as a “SaaS apocalypse,” citing slowing growth rates, compressed valuation multiples, and uncertainty about how AI will reshape software consumption patterns.
Cloud Strength and AI Infrastructure Demand
The turbulence in software contrasts with AWS’s own results. Parent company Amazon reported that fourth-quarter revenue at its cloud infrastructure division rose approximately 24% year over year to $35.6 billion, exceeding analyst expectations. AWS posted a 35% operating margin, slightly higher than the previous quarter, underscoring sustained profitability in its core business.
The divergence highlights a structural distinction in the AI value chain. While application-layer software companies face questions about displacement, hyperscale cloud providers supply the compute, storage, and networking infrastructure required to build and deploy AI systems at scale.
“There’s a huge disruption,” Garman said. “AI is absolutely a disruptive force that’s going to change how software is consumed and how it’s built. And I would argue that the systems of record, as you call them, the SaaS providers and the large players of today have an inside track to winning that business.”
Systems of record — enterprise platforms managing financial data, human resources, compliance, and customer relationships — are deeply embedded within corporate workflows. Replacing them entails operational risk, data migration complexity, and integration challenges, which can slow wholesale displacement.
AWS generates revenue from established vendors, including Adobe, Intuit, and Zillow, while also benefiting from AI model developers expanding compute usage. In November, AWS disclosed a $38 billion spending commitment from OpenAI, reflecting the scale of infrastructure required to train and run large language models.
“Our perspective is that our customers are going to consume more compute technology and more infrastructure than they ever have,” Garman said, arguing that whether companies build AI internally or buy AI-enabled SaaS, overall infrastructure demand should rise.
Slowing SaaS Growth and Broader Spillover
Even as infrastructure spending climbs, growth among major SaaS firms has moderated. ServiceNow, an AWS customer, recently reported fourth-quarter revenue growth of 20.7% year over year, down from nearly 26% two years earlier. While still strong relative to many sectors, the deceleration has weighed on valuations that were priced for sustained hypergrowth.
Investor concern extends beyond enterprise software. Florida-based Algorhythm Holdings said Thursday that an AI-powered product enabled logistics clients to quadruple freight volumes without increasing headcount. Shares of C.H. Robinson Worldwide fell about 23% in midday trading, reflecting fears that AI-driven efficiency gains could pressure revenue models tied to transaction volume or labor-intensive processes.
The underlying debate centers on whether generative AI will cannibalize traditional software categories or expand total addressable markets by unlocking new use cases. Historically, major computing transitions — from on-premises infrastructure to cloud, and from desktop software to web applications — have produced both displacement and expansion. Companies that adapted architecture and pricing models often retained leadership; those that failed to evolve lost relevance.
The current correction represents a repricing of growth expectations and risk premiums for investors. For AWS, the calculus differs. As long as AI development requires large-scale compute, hyperscale cloud providers remain positioned to capture incremental spending, regardless of which application-layer companies ultimately prevail.
Garman’s remarks suggest confidence that AI will alter the shape of enterprise software without necessarily shrinking its economic footprint. The market’s volatility indicates that investors are still determining where value will accrue in an ecosystem being rapidly rewritten by artificial intelligence.