Software sector rebounds 21% as AI agents reshape SaaS pricing

Three months after traders coined "SaaSpocalypse" for the $285 billion wiped from software stocks in 48 hours, the iShares Expanded Tech-Software ETF (IGV) surged 21% in May — its best month since October 2001 — as investors re-rated the sector company by company on a single question: does this business sit on the path autonomous agents must travel, or is it priced for what agents replace?

"The February selloff was a product announcement shock, not a fundamental deterioration," said Tal Liani, analyst at Bank of America, who on May 18 reinstated coverage of ServiceNow at Buy with a $130 price target and Salesforce at Underperform with a $160 target on the same day. "Agentic AI deployments elevate the need for orchestration, permissions, approvals, policy enforcement and auditability — directly aligning with ServiceNow's core capabilities."

Snowflake supplied the fuel. On May 27, the data platform reported first-quarter fiscal-2027 revenue of $1.39 billion, up 34% year over year, and raised full-year product-revenue guidance to $5.84 billion from $5.66 billion. The company also announced a five-year, $6 billion commitment to AWS for Graviton compute and AI infrastructure and acquired Natoma, a startup building enterprise governance for the Model Context Protocol, for about $110 million. Snowflake surged roughly 37% after the report. Constellation Research analysts framed the move as "selling shovels" — the argument that data infrastructure companies benefit when agents consume data rather than replace human seats.

The market is now sorting winners from losers based on pricing model exposure. Companies whose revenue scales with agent activity — data platforms like Snowflake, observability players like Datadog, and workflow and security platforms like ServiceNow and CrowdStrike — have rebounded sharply. Seat-denominated systems where AI can reduce human-user counts face more skepticism. Gartner projects 35% of point-product SaaS tools will be replaced by AI agents by 2030. Seat-based pricing adoption among vendors has already dropped from 21% to 15% in roughly 12 months, according to Bain & Company and Deloitte reports.

The Salesforce Question Remains Unanswered

Salesforce's first-quarter fiscal-2027 results, reported the same evening as Snowflake's, complicated the bear case without resolving it. Revenue of $11.1 billion beat the $11.05 billion consensus, and adjusted earnings per share of $3.88 crushed the $3.12 estimate by nearly 24%. Agentforce annual recurring revenue reached $1.2 billion, up 205% year over year, with 29,000 deals closed since launch. Yet the company is running three pricing models in parallel — per-conversation, flex-credit, and per-user — a sign it is still testing rather than executing. Full-year revenue guidance of $45.9 billion to $46.2 billion, up 11%, slightly undershot expectations. Salesforce closed May down roughly 30% for 2026 despite the earnings beat.

Liani's three pressure points on Salesforce — muted net-new customer additions, limited upsell potential, and an underwhelming monetization path for Agentforce — remain unresolved. The structural question is whether consumption revenue from Agentforce will scale fast enough to offset compression in the legacy seat base. One quarter of acceleration does not answer it.

The Infrastructure Thesis Has Limits

The rebound is partly mechanical mean reversion, partly a genuine re-rating. Snowflake is up roughly 50% on the week of earnings; HubSpot is up 10% in a single session but still down nearly 46% on the year. The median software company is still climbing out of a deep hole. Natoma's MCP gateway had been deployed at only a handful of enterprises as of the acquisition announcement, and the bulk of agentic workloads inside Fortune 500 IT departments remain pilots rather than production systems.

For investors, the cost of being wrong on a single ticker is far higher than the index suggests. Snowflake trades at elevated multiples on the infrastructure thesis; ServiceNow recovered roughly 33% from its 2026 lows on the orchestration argument. But if agent adoption proves slower than expected, the infrastructure names carry the most downside risk from current levels. The February funeral was premature. The May triumphalism may be too.

This article is for informational purposes only and does not constitute investment advice.