The AI Agent Shake-Up: Which Software Stocks Survive?

Software stocks have taken a hit this year.

The S&P Software & Services ETF is down about 7% year-to-date, considerably underperforming the S&P 500, which is up roughly 9% over the same stretch. The selloff is being driven by the rise of agentic AI.

The concern is that AI agents could fundamentally change how enterprise software is used and monetized. Instead of employees logging into applications and navigating dashboards, AI agents can increasingly complete tasks directly through application programming interfaces. That makes the user interface less central to where value is created.

More importantly, it challenges the economics that have underpinned enterprise software for decades. Most SaaS companies charge customers based on the number of employees using their products. But if a single AI agent can automate work previously done by several people, customers need fewer human licenses. That weakens the traditional per-seat pricing model and pushes vendors toward consumption- or outcome-based pricing, such as charging per resolved support ticket or completed workflow. Besides this, software that cannot integrate seamlessly with other applications also risks being bypassed altogether.

Customers spent $1.25 trillion on software last year, per Gartner. We don’t think that money is going anywhere. Some of it could get captured by the frontier model players and companies building agents themselves.

A real slice of it could also land with software companies that adapt fast enough to become the layer those agents run on.

Which companies look like winners here, and at what price?

The Data Owners

Software companies that own critical enterprise data stand to benefit most from agentic AI, as effective agents rely on rich, real-time access to information stored in these platforms to make smart decisions and take actions.

Salesforce (CRM) (-34% YTD) is a leading CRM provider that helps businesses manage customer relationships, sales, service, and marketing through its cloud platform. Because customer data resides in Salesforce, AI agents must access its platform to understand customer context and take action, making it more valuable as automation increases. With Agentforce, Salesforce is shifting toward a “consumption + subscription” model, charging based on agent activity and completed work rather than just human logins.

ServiceNow (NOW) (-27% YTD) specializes in digital workflow software for IT, HR, customer service, and enterprise operations. Its platform orchestrates enterprise workflows, making it the execution layer AI agents depend on rather than replace. Now Assist enables agents to autonomously resolve issues and coordinate work across departments, while premium AI capabilities create new monetization opportunities beyond traditional seat licenses.

Palantir (PLTR) (-21% YTD) provides software that integrates and analyzes proprietary enterprise data for complex government and commercial decisions. Its value lies in organizing enterprise data that AI agents need to reason and act, making the platform more important as agent adoption grows. Its Artificial Intelligence Platform (AIP) enables high-stakes use cases, from mortgage approvals to supply chain optimization, and is sold through long-term enterprise contracts tied to business value rather than per-user licenses.

SAP (SAP) (-33% YTD) is the dominant ERP provider, running finance, HR, procurement, and supply chain systems for many of the world’s largest companies. Because it owns the data behind these core business processes, AI agents must connect to SAP to access information and execute workflows. At Sapphire 2026, it introduced Joule Studio and the Autonomous Suite, with more than 50 domain-specific AI agents built on its Business Data Cloud and Knowledge Graph. The strategy is compelling, but investors remain cautious as AI infrastructure costs are weighing on margins before AI revenue has meaningfully ramped.

See how growth and margins of major SaaS players stack up 

The Infrastructure & Platform Play

These companies provide the tools for other businesses to build their own agents.

Microsoft (MSFT) (-19% YTD) is a leading provider of cloud computing (Azure), productivity tools (Microsoft 365/Office), and enterprise software. With Copilot Studio, it monetizes AI agents through Azure consumption charges and premium licenses for building and deploying autonomous agents across the Microsoft 365 ecosystem. Our concern is that Microsoft largely controls distribution rather than the underlying frontier models, which could limit its ability to capture value if model leadership becomes the primary competitive advantage.

Snowflake (SNOW) (+21% YTD) is a cloud data platform specializing in data warehousing, analytics, and sharing, allowing organizations to manage databases without traditional infrastructure hassles. Through its Cortex Code agent and the broader AI Data Cloud, it monetizes the agent execution layer by powering data-grounded agents directly on its infrastructure, driving revenue from increased compute usage for planning, analysis, and execution workflows.

While these companies appear better positioned than most software vendors to navigate the shift toward agentic AI, they still face execution and valuation risks. For investors seeking a more balanced path to outperformance, Trefis High Quality (HQ) Portfolio invests in a diverse group of 30 stocks that have collectively delivered stronger upside with reduced volatility compared to the broader indices. Discover the methodology behind these smoother, higher returns by checking the HQ Portfolio performance data.