Goldman Sachs chief strategist Ben Snider warned investors on Monday that uncertainty over artificial intelligence-driven disruption will continue to weigh on growth stock valuations for several quarters, possibly years, and that broad sector exposure is no longer a viable strategy.
AI 'Seat Compression' Crushes SaaS Business Models
According to a report by Yahoo Finance's Brian Sozzi, ServiceNow has fallen 48% year-to-date, Salesforce 36%, and DocuSign 42%. The declines are driven by 'seat compression'—a scenario where a single AI agent replaces multiple human users, eroding the per-user licensing revenue that underpins SaaS companies.The software sector has lost approximately $2 trillion in market capitalization this year, Snider noted. Until clear evidence emerges that AI is not displacing existing business models—such as strong earnings surprises and improved unit economics—stock prices in exposed sectors are unlikely to bottom.
Goldman's Deep Dive: 'Will AI Eat Software?'
The warning builds on Goldman's March 2026 report titled 'Will AI Eat Software?,' which concluded that while AI will not fully replace software, it will force major architectural shifts around large language models and autonomous agents. Incumbents hold advantages in proprietary data and workflows, but adaptation windows are finite. Citigroup analyst Tyler Radke echoed the concern, noting that fears over 'software architecture, business model sustainability, and terminal value' may deepen in coming months.Meta, Amazon, and Alphabet were singled out as partial exceptions, expected to recover first thanks to strong earnings forecasts for 2026 and 2027. In contrast, the broader 'Magnificent Seven' is struggling—JPMorgan strategist Mislav Matejka said the group no longer acts as a safe haven relative to the S&P 500; Tesla is down roughly 23% year-to-date.
Public Panic: 80% of Americans Fear AI
Wall Street's caution mirrors public sentiment. A Quinnipiac University poll of 1,397 U.S. adults found 80% are worried about AI, and 70% believe it will reduce job opportunities—a sharp rise from 56% in April 2025. Trust in AI-generated information remains low: 76% said they 'almost never' or 'sometimes' trust AI output. Separately, 75% of Americans oppose building AI data centers in their communities, driven by fears of higher electricity costs (72%) and water usage (64%). While 51% have used AI for research (up from 37% in 2025), adoption outpaces trust. Goldman's forecast suggests the AI cycle is entering a phase where skepticism, not enthusiasm, drives the narrative—keeping valuations under pressure for years.