Salesforce is getting ready to spend round $300 million on Anthropic tokens in 2026 because the enterprise software program firm deepens its funding in synthetic intelligence instruments whereas sustaining a freeze on software program engineer hiring.
Talking on the All-In podcast, Salesforce Chief Government Officer Marc Benioff mentioned a big portion of the projected spending would help coding-related work powered by AI programs and software program brokers.
The spending plans underline a broader shift underway throughout the expertise business, the place firms are redirecting investments from workforce growth in the direction of AI infrastructure, coding assistants and automation instruments.
Benioff described AI coding brokers as “superior” and mentioned they have been serving to decrease software program growth prices whereas bettering engineering velocity and productiveness.
Hiring freeze tied to AI productiveness good points
Benioff first introduced in 2024 that Salesforce would cease hiring software program engineers in 2025 after inner AI instruments improved engineering productiveness by greater than 30 per cent.
“We’re not including any extra software program engineers subsequent 12 months as a result of we now have elevated the productiveness this 12 months with Agentforce and with different AI expertise that we’re utilizing for engineering groups by greater than 30%,” Benioff had mentioned on the time.
In response to the CEO, AI integration considerably elevated growth velocity throughout Salesforce engineering groups.
Nonetheless, Salesforce did proceed recruitment in different divisions, significantly gross sales.
Benioff beforehand acknowledged that the corporate deliberate to rent between 1,000 and a pair of,000 gross sales workers to assist clarify AI merchandise and their enterprise functions to prospects.
Engineers more and more work alongside AI brokers
Whereas Salesforce has paused extra software program engineering hiring, Benioff clarified that the corporate doesn’t imagine AI can but totally exchange human engineers.
As a substitute, he mentioned engineers are more and more working in partnership with AI coding instruments and autonomous software program brokers.
Salesforce’s engineering groups at the moment work with:
Benioff mentioned Salesforce’s roughly 15,000 engineers have been evolving into supervisory roles the place they more and more oversee AI-generated coding workflows.
“Once they begin to use these fashions, they’re now working not solely with the AI, however brokers to assist them code,” he mentioned.
He added that engineers stay needed as a result of AI programs can’t but operate totally autonomously.
“We’re not at that degree but of AI,” Benioff mentioned.
Anthropic partnership expands throughout Salesforce merchandise
The projected token spending displays Salesforce’s rising integration of Anthropic’s AI fashions throughout its enterprise software program ecosystem.
Salesforce has already invested greater than $300 million in Anthropic and reportedly holds a stake within the synthetic intelligence firm.
Anthropic’s Claude fashions are more and more being built-in into Salesforce platforms, together with:
- Slack
- Agentforce
- Enterprise AI workflows
- Coding and developer instruments
Benioff additionally indicated that Salesforce was constructing programs able to routing AI duties between bigger and smaller language fashions relying on complexity and value effectivity.
The strategy is designed to scale back operational expenses whereas bettering AI efficiency throughout customer-facing and inner instruments.
AI enterprise turns into central progress engine
Salesforce’s AI-focused enterprise unit, Agentforce, has emerged as a significant income contributor for the corporate.
In response to Benioff, Agentforce has already reached roughly $800 million in annual recurring income.
The corporate has positioned AI as a central pillar of its future enterprise mannequin, with Benioff describing the broader transformation as a “digital labour revolution”.
He mentioned synthetic intelligence now accounts for between 30 per cent and 50 per cent of Salesforce’s general workload.
The corporate’s evolving technique displays a wider development throughout Silicon Valley, the place AI instruments are more and more reshaping engineering operations, software program growth economics and workforce planning.