Voluntary Buyouts Let Your Best People Leave

A voluntary buyout is the one kind of layoff where your best people get to decide whether they leave.

Microsoft just had a monster year. Per The Register, Azure is up more than 40 percent, Copilot has 30 million paid seats, and Redmond returned $10.2 billion to shareholders. Its response to all that is to invite 60,000 employees to “consider continuing their careers outside the company.” Rupert Goodwins opens his column with “Capitalism can seem so perverse”: a company dripping with cash showing the door to the people who built the pile.

That’s the strongest version of the outrage, and it deserves a straight answer. Here’s mine. The fact that Microsoft is cutting isn’t the problem. The way it’s cutting is.

Start with the steelman. Voluntary redundancy normally makes sense when a company is running out of money. Microsoft isn’t. So why shed people at the top of the cycle? It looks like greed dressed up as strategy: margins pushed up for shareholders, paid for by employees.

It falls apart on one point. Cash on hand is not a reason to keep a payroll, and Microsoft’s cash is already spoken for. Goodwins reports that capex came in light, and not for lack of money. What Microsoft wanted to buy was “Out of stock. More expected soon.” Analysts say Microsoft is hinting it will go into “negative free cash flow” territory in 2028 and take out massive loans. Goodwins ties the buyouts straight to that: “The cost of borrowing depends on the financial health of the borrower.” So this is a company cutting payroll to fund an AI build-out it will have to borrow for. Every big platform transition reprices labor against capital. Mainframes to PCs did it. On-prem to cloud did it. AI is doing it now, and doing it faster. Trimming payroll before you go to the debt market is what you do when you can see the shift coming and want to finance it on good terms. Cutting in a downturn is what you do when you didn’t see it. Of the two, cutting early is the smarter move.

So the cut itself isn’t perverse. The tool is badly chosen.

Goodwins makes the sharpest point in his own column, almost in passing: “The biggest knock-on of such schemes is cultural, with only the very confident moving on out, a disproportionate loss of talent and role models.” He’s right, and it’s worse than it sounds. Ask who actually takes a voluntary package. Not the median employee with a mortgage and an average résumé. The person who takes the check is the one with a better offer waiting. That’s the staff engineer who already has a recruiter from Anthropic in her inbox. Or the product lead who’s been quietly raising a seed round. Or the infra specialist every hyperscaler is fighting over. The people with the best outside options are the most likely to say yes, and the people with none are the least likely. You end up paying a premium to lose your top decile and keep the people nobody else is trying to hire. Economists call this adverse selection. Operators call it watching your best team leave on the company’s dime.

Why do it anyway? Because power structures matter more than stated policy. A forced layoff is a decision management has to own: names, criteria, headlines, lawsuits. A voluntary program turns that decision into a choice employees made themselves. It’s cheaper for the reputation and easier on the executives’ consciences. Management gets less blame, and in return it gives up control over who goes.

Here’s where I’d take Goodwins’s point further. The AI angle makes this worse, not better. The whole pitch for Copilot is leverage: one engineer with good tools does the work of several. Leverage doesn’t make people equal. It multiplies whatever is already there. Give a great engineer and a median engineer the same assistant and the gap between them grows. So in an AI-first company, every top performer is worth more than they were three years ago, and losing one costs more than it used to. A company betting its future on AI leverage is running a program that, by design, encourages its highest-leverage people to leave first.

There’s a signal here too, and this is my read, not Microsoft’s statement. A company doesn’t invite 60,000 people to think about leaving unless it believes it can do the same work with fewer of them. In effect, Microsoft is telling you what it thinks Copilot is worth inside its own walls. That’s more credible than any keynote slide, because payroll is where companies put their real money.

The stakes go beyond Redmond. Every large company with an AI story and a build-out to finance will face the same choice in the next eighteen months. Most will reach for the voluntary program because it’s the easiest one to defend. Plenty will find out a year later that their AI roadmap left with the people who could have shipped it.

Cut if the math says cut. Just don’t let the people with the best options pick who goes.