Salesforce CEO Marc Benioff’s “AI models are commodities; data is the new gold” argument came during the company’s fiscal 2025 first-quarter earnings call on May 29, 2024, covering the quarter ended April 30. It was a strategic thesis and a pitch for Salesforce’s data products—not an independent finding that all AI models are interchangeable. The company reported $9.13 billion in quarterly revenue, up 11% year over year, while management also described customers scrutinizing budgets and delaying or shrinking some deals.
What Benioff meant by calling AI models commodities
On Salesforce’s fiscal 2025 Q1 earnings call, Benioff argued that models and user interfaces can be replaced or changed, while a company’s customer relationships and the data describing them can persist. “The models and the UI are not the critical success factors,” he said. He also said models “are just commodities now,” arguing that they do not know a company’s customer relationships.
His point was about where Salesforce believes enterprise AI value can come from: not only the model generating an answer, but also the proprietary business context available to it. That is Benioff’s view and Salesforce’s commercial positioning. The call does not establish an industry-wide consensus that models are commodities or prove that data matters more than models in every AI application.
Why Salesforce called customer data “the new gold”
Benioff said customer data and metadata—the information that describes or gives context to data—can help AI produce more relevant business outputs. He said Salesforce managed more than 250 petabytes of customer data. That was a company-reported scale figure on the call, not an independently audited statement of the storage scope.
#1 Best Overall
Salesforce’s Data Cloud narrative is to bring structured and unstructured data together, harmonize it into unified customer profiles, and connect sources such as Snowflake, Redshift, BigQuery, and Databricks. The company says this can ground AI responses in business and customer context. These are Salesforce’s descriptions of its product and benefits, not an independent assessment of their performance.
Data, integration, and trust are separate parts of the thesis
Having data alone does not guarantee useful AI. Benioff’s argument depends on access to relevant customer context; in practice, data quality and integration also affect whether that context is usable. Privacy and trust controls matter because enterprise data can be sensitive. Salesforce describes its Einstein Trust Layer as including data masking, a zero-retention architecture, and an LLM audit trail. Those are vendor-stated controls; the available call and company material do not independently verify their implementation or effectiveness.
The fiscal 2025 Q1 numbers behind the AI argument
The figures below are Salesforce management’s disclosures for the historical quarter, not current results. Data Cloud activity measures reflect the company’s reported product activity; they should not be read as revenue or as proof of business impact.
| Measure | Salesforce’s reported figure |
|---|---|
| Revenue | $9.13 billion, up 11% year over year |
| Subscription and support revenue | Up 12% year over year, or 13% in constant currency |
| Customer data managed | More than 250 petabytes |
| Data Cloud records ingested | Eight trillion, up 42% year over year |
| Data Cloud records processed | Two quadrillion, up 217% year over year |
| Data Cloud activations | More than one trillion, up 33% year over year |
Salesforce also said Data Cloud was included in 25% of deals worth more than $1 million and that it added more than 1,000 Data Cloud customers for a second consecutive quarter. Those are period-specific company disclosures, not measures of present-day product performance.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #3
Customer spending caution tempered the AI story
Executives described elongated sales cycles, deal compression, budget scrutiny, smaller closes, and decisions pushed out of the quarter. Those comments reflect management’s account of Salesforce’s selling environment, not a universal measure of enterprise spending. The remarks also included differences by geography and sector, so the caution should not be generalized to every customer or business.
The contrast matters when reading the call: management presented AI and Data Cloud as areas of opportunity while also acknowledging that some customers were taking longer to make purchases or committing less. The earnings call alone does not establish that AI products had overcome those spending pressures.
What the “new gold” metaphor does—and does not—show
The metaphor captures a strategic distinction: a model may be available to many companies, while a company’s own customer history, relationships, and operational context can be harder to replicate. But this does not make data automatically valuable. Its usefulness depends on whether it is relevant, well-integrated, appropriately governed, and usable for the task at hand. Nor does the metaphor settle how much value belongs to a model, a data platform, or the systems connecting them.
Salesforce’s earnings-call remarks are best read as a historical explanation of how the company wanted investors and customers to think about its AI strategy in May 2024. They are not a current earnings update, independent validation of Salesforce’s product claims, or proof that the same balance between data and models applies across the AI industry.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
