Marc Benioff Freezes Software Engineer Hiring as Salesforce Plans USD 300 Million Anthropic AI Spend
Salesforce expects to spend USD 300 million on Anthropic tokens this year, following CEO Marc Benioff’s decision to freeze software engineer hiring for 2025. Benioff stated that internal AI tools have boosted developer productivity by 30%. While headcount growth has stopped, engineers are transitioning into supervisory roles to manage autonomous coding agents.
Marc Benioff-run Salesforce is projected to spend approximately USD 300 million on Anthropic artificial intelligence tokens this year, shifting its financial focus from expanding its human engineering workforce to investing heavily in advanced software automation. The multi-million-pound token expenditure follows a strategic decision by Chief Executive Marc Benioff to freeze the recruitment of traditional software engineers for 2025, capitalizing on significant internal productivity gains driven by AI coding assistants.
Despite halting engineering expansion, the company maintains that the technology is designed to augment human workers rather than eliminate them entirely. The transition reflects a broader structural trend across Silicon Valley, where enterprise tech corporations are rapidly reallocating capital to balance sheet items like API tokens, advanced models, and server infrastructure over traditional headcount growth.
Salesforce AI Token Spending and Anthropic Investment
Speaking on the All-In podcast, Benioff detailed that the bulk of the projected USD 300 million token budget will be tied directly to programming and software development tasks. Salesforce has already invested more than USD 300 million in Anthropic and holds a minority stake in the startup, using its Claude models to power a variety of services, including new automation features within the corporate communication platform Slack.
To manage the high operational costs associated with large language models, Salesforce is reportedly developing internal routing systems. These systems are designed to automatically direct complex engineering requests to larger, more expensive frameworks like Anthropic’s models, whilst diverting simpler routine tasks to smaller, cost-effective algorithms.
Marc Benioff Software Engineering Recruitment Freeze
The financial pivot towards token consumption comes after Salesforce achieved a 30% surge in internal engineering velocity. Benioff first announced the 2025 hiring freeze last year, arguing that tools like Agentforce had made the company's existing pool of approximately 15,000 engineers vastly more efficient, negating the need for further workforce expansion.
While tech development roles face a standstill, Salesforce has redirected its recruitment efforts toward client-facing departments. The company previously confirmed plans to hire between 1,000 and 2,000 additional sales professionals specifically tasked with demonstrating the concrete business utility of its evolving AI catalogue to enterprise customers.
Agentforce Revenue Growth and Digital Labour Shift
Benioff clarified that artificial intelligence models are currently incapable of autonomous software creation and still require strict human oversight. Engineers at Salesforce are increasingly stepping into supervisory positions, managing workflows and reviewing code generated by autonomous agents, OpenAI Codex, and developer tools like Cursor.
The structural adjustment appears to be yielding financial results, with Salesforce's AI-centric business unit, Agentforce, recently reaching USD 800 million in annual recurring revenue. Describing the transition as a "digital labour revolution," Benioff estimated that automated systems and AI agents now handle between 30% and 50% of Salesforce’s total global workload.
(The above story first appeared on LatestLY on May 18, 2026 11:35 AM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).
