Salesforce Is Now Spending on Tokens, Not Engineers. That Is the Whole Future in One Line.

Summary: Salesforce hired zero new software engineers in its last fiscal year while planning to spend around $300 million on Anthropic AI tokens in 2026, much of it on coding. Sales headcount grew 20 percent. This is the token economy made literal: computation is replacing labor hours as the basic unit of production, and the budget is moving from payroll to compute. Here is what it means for how you plan, hire, and measure.

There is a single sentence buried in Salesforce's 2026 numbers that explains more about the future of work than a hundred conference keynotes.

The company hired zero new software engineers in its last fiscal year, and it plans to spend roughly 300 million dollars on Anthropic AI tokens in 2026, with the bulk of that tied to coding work. Read those two facts together and you are looking at a budget line that used to be called engineering salaries quietly renaming itself compute.

This is not a thought experiment about the future. It is a line item on a real income statement at one of the largest software companies in the world.

From headcount per output to tokens per outcome

For a century, the basic equation of business was simple. You needed more output, so you hired more people. Headcount was the lever. Productivity meant getting more out of each person, but the unit of production was always a human hour.

Salesforce just broke that equation in public. CEO Marc Benioff confirmed the company hired no new engineers while growing the sales team 20 percent, citing AI productivity gains of more than 30 percent. Fortune reported that across the company, almost no one is being hired except in sales. AI is now handling a large slice of the work that engineers and support staff used to do, and the company cut customer support headcount nearly in half.

I have a name for what is happening here. I call it the token economy. Computation tokens are replacing labor hours as the fundamental unit of organizational production. When a CFO at Salesforce wants more engineering output, the lever is no longer a req for a new hire. It is a bigger token budget. Efficiency stops being measured in headcount per output and starts being measured in outcomes per token.

That is a profound change in how a company is valued, too. For decades, a software company's worth was loosely tied to how many brilliant engineers it could attract and retain. In the token economy, valuation starts to reflect compute leverage, how much outcome a company can squeeze out of each dollar of compute, rather than how many badges are in the building.

Why sales grew while engineering froze

The detail that makes this story sharp is the asymmetry. Engineering went to zero new hires. Sales went up 20 percent. Why?

Because the work AI is best at right now is the structured, high-volume, specifiable work, and a large share of coding fits that description. Writing code against a clear spec, migrating a codebase, fixing well-defined bugs: this is exactly the kind of task that an AI coding agent can grind through at machine speed. Selling, by contrast, still runs on trust, relationships, reading a room, and the kind of judgment AI cannot yet own. So Salesforce did the rational thing. It let machines absorb the work machines are good at and it doubled down on humans for the work that still needs humans.

This maps almost perfectly onto the Hive Structure I teach. The bees, the AI agents, took over the volume coding work. The beekeepers, the humans, moved toward the parts of the business where direction, relationship, and judgment matter most. Salesforce did not just cut engineers and call it a day. It reallocated its human capital toward the roles AI cannot fill while pointing compute at the roles it can. That is the move. Most companies are only doing the first half.

The Replacement Exercise, running at corporate scale

I often tell audiences to run what I call the Replacement Exercise on themselves. Constantly hand off the automatable parts of your job to AI, so that you become irreplaceable on the parts that remain. Move yourself from bee to beekeeper inside your own role. Salesforce is running that exercise at the level of an entire org chart. It is replacing the automatable layer of its engineering function with compute, and pushing its people toward the irreplaceable layer.

The uncomfortable part is what this does to the bottom rung. If you never hire the junior engineer because AI does the junior work, where does the next generation of senior engineers come from? This is the real risk hiding inside the Salesforce story, and it is not unique to them. The whole industry is automating away the entry-level tasks that used to be how people learned the craft. Leaders who only see the cost savings are not looking far enough down the road.

What to actually do with this

You do not run Salesforce, but the logic now applies to you, and pretending otherwise is how you get caught flat-footed. Three moves.

First, start tracking compute the way you track payroll. If AI is doing real work in your organization, it has a cost, and that cost belongs on a managed line, not buried in a SaaS bill nobody reads. Pull your last AI invoice and ask what unit of work it actually bought. If you cannot answer in terms of outcomes, you are flying blind.

Second, separate the bee work from the beekeeper work in every function, not just engineering. Which tasks are high-volume and specifiable enough to hand to AI? Which require human judgment, trust, or accountability? Build your hiring plan around protecting and growing the second category.

Third, protect your pipeline of future beekeepers. If you automate away every entry-level task, you save money this year and starve yourself of senior talent in five. Decide deliberately how your people will still learn the craft when the machines are doing the practice rounds.

Salesforce just showed the market what the token economy looks like when a real CFO signs the budget. The money moved from salaries to tokens, and the company kept growing. The question is no longer whether this happens to your industry. It is whether you will design the move on purpose or have it forced on you.

So here is what I would ask your finance team this quarter. If someone handed you a board deck today, could you tell them your cost per outcome in tokens, or do you still only know your cost per head?

Sharon Gai is an AI transformation strategist, keynote speaker, and author of How to Do More with Less Using AI. She advises Fortune 500 companies on AI adoption and organizational redesign.