If your board asks why you plan a layoff when Microsoft offered 60,000 people a buyout, start with scope: Microsoft limited it to selected teams at level 70 and below, and let employees choose. That design is copyable. Its risk is that the people who say yes are the ones with the best outside options, and Microsoft has not said whether it can refuse them.
On October 26, Microsoft's chief people officer Amy Coleman announced a Voluntary Exit Offer for "eligible employees in participating teams in Canada, India, the United Kingdom and the United States at level 70 and below." Microsoft confirmed to TechCrunch that around 60,000 employees are eligible. On November 18, Satya Nadella explained the offer publicly at the New York Times DealBook Summit, and the reason he gave was an AI-era change in how teams are built.
What Microsoft Actually Offered
Microsoft offered a time-limited, opt-in exit to more than a quarter of its workforce, with eligibility set team by team. Microsoft had about 223,000 full-time employees at June 30, 2026, so 60,000 eligible is more than a quarter of the company.
The memo itself is short on terms. It says the offer is "time-limited," that it includes "financial support and career planning resources," and that eligible people would get an email with their details within hours. It names no teams. What it says about scope is this: "Some organizations are providing the Voluntary Exit Offer, and some are not. Participation varies based on each business's priorities, customer commitments and workforce needs."
The rest comes from reporting, and you should weigh it that way. Bloomberg reported, via TechCrunch, that Xbox, Windows, Copilot and Finance are participating, while some employees, including those in data centers and sales and marketing, are not. TechCrunch describes the eligible band as vice president level and below on select teams, and reported that severance terms would be similar to Microsoft's standard package.
Business Today went further: at least 60 days of base pay, rising to 39 weeks depending on seniority and tenure, an acceptance deadline in the first week of December, and exits by the end of 2026. Microsoft has not published those terms.
This was Microsoft's second buyout this year. In April, Microsoft offered buyouts to about 7% of its US workforce, roughly 8,750 people by Bloomberg's arithmetic, through a voluntary retirement program. Bloomberg later reported that more than 30% of those eligible accepted, according to TechCrunch. Coleman's October memo says the new offer "builds on what we learned" from that program. In between, Microsoft cut about 4,800 jobs in July, tied to an Xbox overhaul. So the buyout sits inside a year that also had a conventional layoff.
Why Nadella Says Teams Have to Change
Nadella's argument is that AI has changed the shape of a software team, and some employees will not want the new shape. At DealBook, interviewed by Andrew Ross Sorkin, he said the buyout was meant "to make sure that at least everyone felt like they had agency in making choices," according to TechCrunch's account.
He described software engineering as having gone from doubt that AI could write code, to a moment of "why do we need software engineers?", to "rediscovering that software engineering is very much alive." The change, in his words, is that "the way the team is organized and the work that a software engineer does has now got a new operating point."
His example was Cowork, the agent in Microsoft 365 apps, which TechCrunch reports is built by just 11 developers. "They build a software factory that builds the product. They don't build the product," Nadella said, adding that "that level of direction is what, by the way, is going to happen in all jobs." He also said the shift is "going to take a bit more time than anyone expects."
On the Pragmatic Engineer podcast the week before, Nadella was blunter about teams that lag. "I don't want to be stuck in a team that's not transforming, even inside of Microsoft," he said, as quoted by Business Today, and "if you're in a team where you feel stuck professionally, you can vote with your feet." TNW reports that in the same interview he did not say whether the offer was a prelude to layoffs, and that he said teams must first reorganize and build their "software factories" before hiring early-career engineers.
Read those together and the sequence is clear enough to plan against. First you redraw team boundaries around a smaller group that directs agents. Then you give the people outside the new shape a paid way out. Junior hiring comes after.
Why a Buyout Beats a Layoff, and Where It Costs More
A voluntary program costs the company more per exit but costs the employee less, and the best public evidence says it wins on total cost. RAND's study of the US Defense Department's civilian buyout program found that involuntary separation saves more budget over five years, but once you count the loss to the people separated, "VSIP generates more net savings and is more cost-effective at the margin than involuntary separation." RAND's model also estimated that raising the payment cap from $25,000 to $41,000 would increase voluntary separations by about 45%, which tells you the take-up rate is a price you set.
There are legal reasons too. In the US, voluntary separations generally do not count as employment losses under the federal WARN Act, according to Morgan Lewis, so a buyout can avoid notice obligations that a reduction in force of the same size would trigger.
The case against is adverse selection: your strongest people have the best outside offers, so they are the likeliest to take the money. The evidence here is mixed, and it is worth knowing. RAND's study of the Army's 1990s drawdown found that lower-quality personnel were more likely to accept the separation offer than high-quality personnel: the offer more than doubled the separation probability for the lower group, from about 13% to more than 29%, and raised them by 10 percentage points for the high group. So the program did not drain the top first. It still raised exits among the people the Army most wanted to keep.
A software engineer at Microsoft has a far more liquid job market than an enlisted soldier in 1992, so expect a stronger pull on your best people than the Army saw. The Morgan Lewis guidance covers the defence: reserve the right to reject applications, and use objective, non-discriminatory criteria such as tenure rather than first-come limits. Microsoft's public memo does not say whether managers can decline an acceptance. If you copy the program, decide that before you announce it.
What Microsoft's Design Gets Right
The part worth copying is the team-level scoping. A company-wide buyout invites exits from the teams you are trying to grow. Microsoft let each business decide whether to participate based on "priorities, customer commitments and workforce needs," which keeps the offer away from areas where it needs people. The reported carve-outs for data center and sales staff fit that pattern.
The level cap does similar work. Capping at level 70 keeps the most senior executives, who would run the redesigned organization, out of the pool while still reaching the management layers that agent-directed teams tend to thin. That is the same delayering that Canva, American Express and BMW pursued this autumn through involuntary cuts.
Sequencing matters as well. Microsoft ran a smaller retirement buyout in spring, measured the take-up, and then went bigger. Coleman's memo says as much: "The Voluntary Exit Offer is new for us. We'll learn from it," as quoted by TechCrunch. A pilot on one population before a company-wide offer is the cheapest way to learn your own take-up rate.
Scoping does not fix the review load. An 11-person team directing agents needs fewer people, but it also needs review capacity, and review is where AI coding output piles up. The Harvard and Jellyfish study found AI coding agents raised pull requests by 35.7% and moved the bottleneck to code review. Cut headcount before you fix that and you get the reversal Meta went through when its 50-to-1 manager ratio failed to hold.
What to Do Before Your 2027 Org Plan Locks
The action is to decide your scope, your refusal rule and your legal timeline before anyone mentions the word buyout in a planning meeting.
This Week:
- Ask each business leader to answer Microsoft's own question in writing: does this team participate, and why, given its priorities, customer commitments and workforce needs. Any team that says no should name the roles it is hiring into.
- List the 20 people per participating team you would least want to lose. If more than a handful would plausibly accept, that team is not ready for a voluntary program.
This Month:
- Get employment counsel to confirm your US timeline. For employees 40 and over, Morgan Lewis advises allowing at least 45 days to consider the offer and a seven-day revocation period after signing, plus written disclosures of the job titles and ages of everyone in the decisional unit. Build your announcement date backward from that.
- Write the rejection rule now, using objective criteria such as tenure or critical-skill designation, and get legal sign-off. A rule you invent after the best engineer applies looks targeted.
- Price the take-up. Use RAND's estimate that a higher payment would raise exits by about 45% as a reminder that the package size decides how many people leave. Model two or three package levels against the headcount you need to remove.
Before Q1 Planning Closes:
- Redraw one team around a small group that directs agents, measure its review queue and defect rate for a quarter, and only then extend the shape. Tools such as GitHub Copilot raise output faster than they raise review capacity.
- Decide in advance what happens if take-up is too low. If the honest answer is a layoff, say so internally, because Microsoft's own year (a buyout in spring, roughly 4,800 cuts in July, a larger buyout in October) shows employees will read the sequence anyway.
The Bottom Line
Voluntary programs are an old tool. The US military leaned on them in its 1990s drawdown, and RAND's numbers show they can be the cheaper path once you count the cost to the people leaving. What Microsoft added is a stated reason tied to how AI changes a team's shape, and a scope drawn team by team. That matters because most companies are still adding people alongside agents: Genpact's survey found 92% kept or grew headcount. For most of you, the 2027 question is which teams get redrawn around agents and who in them gets a paid exit.
If you borrow the design, borrow the scoping and the pilot first, and write down who you will refuse before the offer goes out.
Continue Reading
- Canva Reorg Brings Product, Growth and Platform Into One Org
- Meta Tried 50 Employees per Manager. Now It Wants Managers Back.
- AI Coding Agents Lift Pull Requests 35.7%, Harvard Study Finds
- ServiceNow Says AI Cut Sales Ops as Quota Sellers Grew Double Digits
- Amex Cuts About 5% of Staff and Books a $336M Charge
- They Fired Workers for AI. 55% Now Admit It Was a Mistake.
