Salesforce CEO Marc Benioff has said the software giant expects to spend nearly $300 million this year on tokens from Anthropic, with most of the expenditure aimed at coding-related work.
Benioff made the
remarks during the “All-In” podcast on Friday, where he reportedly described AI coding agents and Anthropic as “awesome” while outlining Salesforce’s growing reliance on artificial intelligence tools for software development.
In the AI ecosystem, tokens are the units used by Large Language Models (LLMs) to process and generate text. AI companies typically charge customers based on token usage, making large-scale deployment expensive for enterprises heavily dependent on AI-generated outputs.
Salesforce Expands AI-Led Coding Strategy
Benioff said the spending would ultimately reduce software development costs across Salesforce. The announcement also reinforces the company’s broader shift toward AI-assisted engineering and automation.
The Salesforce chief had earlier announced that the company would not hire additional software engineers in 2025 because AI tools were significantly improving productivity. At the time, he cited internal efficiency gains driven by Salesforce’s autonomous AI platform, Agentforce, along with other AI systems.
The latest comments suggest the company is increasingly relying on AI coding agents rather than expanding traditional engineering teams.
Hiring Freeze, Layoffs And Fresh Graduate Push
Despite slowing hiring in some functions, Benioff has maintained that AI is transforming jobs rather than eliminating them entirely. Last month, he said Salesforce planned to hire 1,000 fresh graduates to help the company adapt to the AI-driven transition.
Over the past two years, Salesforce has combined AI-related restructuring with selective recruitment efforts.
In February 2026, the company reportedly cut nearly 1,000 positions across marketing and product management as automation expanded. Earlier, in September 2025, Salesforce reduced around 4,000 customer support roles, again citing AI-driven efficiencies.
At the same time, Salesforce increased hiring for sales roles. Benioff had previously said the company urgently needed 1,000 to 2,000 additional sales representatives to explain increasingly complex enterprise AI products to clients.
