The sell-off in software stocks that has erased roughly $2 trillion in market value over the past year may be based on the wrong question, according to a new op-ed from Yale researchers. The so-called “SaaSpocalypse” narrative — that AI agents will render traditional software obsolete — misses the real driver of value in the AI economy: data, not models.
Writing in a CNBC op-ed, Jeffrey Sonnenfeld, Steven Tian, and Stephen Henriques argue that the companies poised to win the AI race are not those building frontier large language models, but those that own the proprietary data AI agents cannot operate without. As they put it, “Lower cost of intelligence increases value of incumbent data,” echoing analysts at Wells Fargo.
Salesforce’s blowout quarter
Salesforce’s recent earnings report illustrates the shift. The company’s stock soared more than 20% after reporting adjusted earnings of $5.90 per share, nearly double consensus estimates of roughly $3.27. Revenue grew 11% to $11.35 billion, and current remaining performance obligation rose 14%. Management raised full-year guidance to as much as $46.4 billion.
Non-GAAP operating margins hit 34.1%, defying expectations that pricing power would evaporate. CEO Marc Benioff dismissed the SaaSpocalypse narrative on CNBC, noting that net new average order value growth was the strongest in four years, seats grew across Agentforce sales and service and Slack, attrition was near its lowest level ever, and bookings more than doubled quarter over quarter.
The data flywheel
The core of the argument is that AI agents need trusted data to function. Salesforce’s Data 360 ingested 104 trillion customer records in the quarter, up 355% year over year, while Agentforce delivered 3.2 billion units of agentic work, nearly double the prior quarter. That creates a flywheel: agents generate work, work generates data, and data deepens the moat.
Salesforce’s combined AI and data annual recurring revenue has reached $3.9 billion, more than tripling in a single year. Agentforce alone grew from $100 million to over $1.5 billion in ARR within 18 months of launch, up 240% year over year.
Frontier AI labs are customers, not competitors
The most telling evidence, according to the authors, is that the companies building the most advanced AI models are themselves paying for Salesforce. Nine of the top 10 AI companies now run on Salesforce and Slack, with their combined spending up 435% year over year. Anthropic even partnered with Salesforce on “Claudeforce,” a combination Benioff described as “The No. 1 AI in the world, Anthropic, and the No. 1 CRM, Salesforce, coming together.”
The authors argue that as frontier models become increasingly commoditized — leapfrogging each other every few months and facing competition from cheap Chinese open-source models — the value migrates to scarce complements. For AI agents, that complement is trusted proprietary data.
Benioff is betting big on this thesis, announcing a $25 billion buyback, the largest in company history, and declaring that “Salesforce is first and foremost in the data business.”
Winners and losers beyond Salesforce
The test generalizes across the software sector, the authors write. Companies with proprietary data valuable to AI agents should prosper, while those offering little beyond basic functionality could struggle. Balance sheet strength also matters: firms with strong free cash flow and low leverage have more room to reinvest, while heavily indebted companies may be forced to divert cash to debt service.
The direction of power in the AI economy is becoming clearer, they conclude: away from companies whose advantage is manufacturing intelligence, which grows more abundant by the month, and toward those that own the scarce assets intelligence cannot function without. Increasingly, that means data.
Source: www.cnbc.com — https://www.cnbc.com/2026/08/29/op-ed-salesforce-revealed-next-battleground-in-ai-its-not-models.html
This article is for informational purposes only and does not constitute financial, investment, tax, or legal advice. Do your own research and consult a licensed professional before making financial decisions.
