From Sanskriti Khandelwal | Product & Market Analysis
Microsoft's Exit Offer: 60,000 Eligible, and What 3%, 5% or 10% Uptake Removes
On this page
Microsoft's second voluntary exit offer reaches about 60,000 employees, roughly 27% of its workforce, and the company expects fewer than 5% of them to take it. At that rate the Microsoft buyout removes about 3,000 people. That is close to the 2,600 who took April's far richer retirement deal, and the whole programme costs less than three days of Microsoft's current capital spending.
Key takeaways
- Microsoft's October exit offer is 7 times wider than April's retirement programme. About 60,000 staff at level 70 and below in four countries are eligible, against about 8,750 long-tenured US staff in April.
- Uptake of 4.3% would match April's headcount effect exactly. April drew about 2,600 acceptances at roughly 30%. The October offer needs only 4.3% of 60,000 to remove the same number of people.
- At 3%, 5% and 10% uptake the offer removes 1,800, 3,000 or 6,000 people. That is 0.8%, 1.3% or 2.7% of the 223,000 full-time staff Microsoft reported at 30 June 2026.
- The cost is a rounding error against capex. Even priced at April's richer per-head charge, the 10% case costs about $2.1 billion, under 5 days of the $39.4 billion Microsoft spent on property and equipment last quarter.
The short answer
Microsoft's voluntary exit offer, announced 26 October 2026, lets about 60,000 employees at level 70 and below in the US, Canada, the UK and India leave on 31 December with a severance-style package. Microsoft expects under 5% uptake, which means roughly 1,800 to 3,000 departures.
What the Microsoft voluntary exit offer actually covers
Chief people officer Amy Coleman announced the voluntary exit offer in a memo on 26 October 2026. She described it as a way of "creating greater flexibility and, where possible, more ways for employees to shape their next chapter," and tied it directly to the spring programme. "This builds on what we learned earlier this year through the Voluntary Retirement Program," she wrote, according to CNBC's report on the memo.
The offer is not company-wide. Each business decided whether to take part, and Coleman was direct about that. "Some organizations are providing the Voluntary Exit Offer, and some are not," she wrote, according to The Register.
Who is eligible
Eligibility runs to level 70 on Microsoft's internal scale, in participating teams in the US, Canada, the UK and India. Microsoft's levels start at 59 for entry-level staff, so the offer spans almost the entire individual contributor and management ladder. Level 70 reaches general managers and some employees with vice president titles, though not corporate vice presidents.
There is no age or tenure test this time. That single change is what turns an 8,750-person retirement scheme into a 60,000-person exit window. A source told Reuters that AI experiences, Copilot, finance, HR, Xbox and Windows are among the participating groups, as relayed by Fortune on 30 October.
Sales is the clearest exclusion. A Microsoft spokesperson said some data centre, sales and marketing employees are not eligible, and some engineering teams also opted out. Employees must decide in early December, and those who accept leave on 31 December 2026.
The terms are thinner than April's. Microsoft described the package to KUOW as comparable to its standard severance: base pay for a period tied to tenure, equity, healthcare and career services. Implicator reported a floor of at least 8 weeks' pay in the US, which Microsoft has not confirmed on the record.
The Microsoft buyout against April's retirement programme
The two programmes look similar from a distance and are built for opposite purposes. April targeted a small, senior, long-tenured group with generous terms. October targets a broad population with ordinary terms and lets the market decide who goes.
| Term | Voluntary Retirement Program (April) | Voluntary Exit Offer (October) |
|---|---|---|
| Announced | 23 April 2026 | 26 October 2026 |
| Countries | United States only | US, Canada, UK, India |
| Level ceiling | 67, senior director | 70, includes some vice president titles |
| Age or tenure test | Age plus years of service of 70 or more | None |
| Eligible | About 8,750, roughly 7% of US staff | About 60,000, roughly 27% of all staff |
| Cash severance | Up to 39 weeks of base pay, by level and tenure | Standard severance, reportedly at least 8 weeks in the US |
| Healthcare | Up to 5 years, first year fully paid | A fixed period, length not disclosed |
| Uptake | About 30%, roughly 2,600 people | Microsoft expects under 5% |
| Where October is the better deal | Closed to anyone under the age-plus-tenure line | Open to younger, shorter-tenure staff who got nothing in April |
April terms are from The Verge's reporting of the package, relayed by Techloy on 7 May 2026. Microsoft has not published either package in full. The last row matters: for most of the 60,000, October is the only offer they have ever been eligible for.
Why April drew 30%
April's package was designed to be taken. Its eligible group was, by construction, close to retirement, and it paired up to 39 weeks of pay with five years of healthcare access and six to 12 months of extra stock vesting. For someone at age 58 with 15 years of service, that is a bridge to Medicare, not a severance cheque.
I would not read the 30% as evidence that Microsoft staff are eager to leave. It is evidence that a well-priced retirement bridge sells to people already planning to retire.
Why October points much lower
A spokesperson told Fortune that "similar voluntary programs have around a 3% to 5% uptake," and added, "we anticipate an under 5% uptake on the 60,000 employees that are eligible." CFO Amy Hood reportedly told staff she expects "very few people" to accept. Those are Microsoft's own expectations, and I think they are about right for a standard-severance offer with no age filter.
Microsoft voluntary exit uptake: the 3%, 5% and 10% scenario table
Here is what each uptake rate means in people, as a share of the workforce, and against the two reference points Microsoft has already given us this year. The base is the 223,000 full-time employees reported in Microsoft's fiscal 2026 annual report, of whom 121,000 work in the US.
| Scenario | Uptake of 60,000 | Departures | Share of 223,000 staff | Against April's 2,600 |
|---|---|---|---|---|
| Low, bottom of Microsoft's peer range | 3% | 1,800 | 0.8% | 0.7 times |
| Break-even with April | 4.3% | 2,600 | 1.2% | 1.0 times |
| Microsoft's stated ceiling | 5% | 3,000 | 1.3% | 1.2 times |
| Equal to July's layoffs | 8% | 4,800 | 2.2% | 1.8 times |
| High, an outside analyst's assumption | 10% | 6,000 | 2.7% | 2.3 times |
Departures and shares are Zan Digital calculations: uptake multiplied by 60,000, divided by 223,000. The 10% case follows an estimate by Gian Estrada of TIKR, cited by Implicator, whose assumptions are not published. The 4.3% and 8% rows are solved backwards from April's acceptances and July's 4,800 layoffs.
The 10% scenario is the only one that changes the shape of the company, and I think it is unlikely. It would need twice the top of Microsoft's peer range on an offer with ordinary terms. Treat it as a stress case, not a forecast.
What the table does show is that October and April are not as different in effect as their headlines. One offered rich terms to few. The other offers plain terms to many. Both land near 3,000 people.
What the Microsoft buyout costs, as a ceiling
Microsoft has not priced the October offer. It reported no severance or restructuring line in its 28 October earnings release, and Implicator noted the company "has not put a number on" the cost or the expected savings. So any cost figure here is a bound, not an estimate.
The April per-head figure
The only priced reference point is April. Hood told analysts the retirement programme would cost about $900 million, booked in the quarter to 30 June 2026. Divide that by about 2,600 acceptances and you get roughly $346,000 per person.
That number overstates April's own cost. The $900 million was set before anyone accepted, and Microsoft's July results said the programme cost less than expected, without a final figure. It still serves as a ceiling, because April's package was the richer of the two.
At $346,000 per departure, the 3% case costs about $0.62 billion, the 5% case about $1.04 billion and the 10% case about $2.08 billion. The real figures should be well below these. October drops the five-year healthcare bridge and pays standard severance, not up to 39 weeks.
On the savings side, the TIKR estimate puts annual savings at about $1.9 billion at 10% uptake, or roughly $317,000 per departure per year. If both figures held, each exit would pay back in about a year. I would not plan on that precision, but the direction is clear: this is a cheap programme for Microsoft.
The trade: Microsoft buyout cost against capital spending
The more useful comparison is not cost against savings. It is cost against what Microsoft is spending on equipment, because that is where the company's money is actually going.
In the quarter to 30 September 2026, Microsoft's additions to property and equipment were $39.4 billion, up from $19.4 billion a year earlier. Over 92 days, that is about $428 million a day. Including finance leases, Hood put quarterly capex at $45.2 billion and said she expected more than $50 billion in the December quarter.
Headcount down, capex doubling
Over fiscal 2026, Microsoft's additions to property and equipment rose from $64.6 billion to $115.9 billion, a rise of 80%. Over the same year, full-time headcount fell from about 228,000 to 223,000. Fortune put it plainly: a future with fewer workers and more equipment.
I disagree with the reading that the buyout is funding the build. At 5% uptake the ceiling saving is about $1 billion a year against more than $150 billion of annual equipment spend at the current pace. The buyout does not pay for anything. It signals which way the company's marginal dollar goes, and the pattern holds across peers in the four-company comparison of capex per employee.
The cash cost of that build is visible elsewhere. Operating cash flow of $58.0 billion less $39.4 billion of equipment leaves about $18.6 billion of free cash flow for the quarter, down from about $25.7 billion a year earlier. That pressure is tracked across the group in the analysis of hyperscaler capex against operating cash flow.
Microsoft layoffs and buyouts: the 2026 sequence
Read the two programmes as one sequence with the July layoffs in between. Coleman said April's turnout reduced the size of the July cuts, according to Implicator. That is the model HR teams call buyouts first, layoffs second, and Microsoft has now run it once.
What Microsoft says the offer is not
Microsoft told The Register it is not classifying the offer as a buyout, because no one is being incentivised to leave. A spokesperson told KUOW that AI is not driving it. And on the numbers, the company said, "We don't expect 60,000 people to walk out the door, and we would never want that."
I take the first claim less seriously than the other two. A package of pay, equity and healthcare offered on a deadline is an incentive to leave, whatever the label. The better reading of the label is legal and reputational, not economic. How companies word these announcements is covered in the review of how firms frame AI-era layoff statements.
The AI framing is harder to settle from the outside. Copilot and AI experiences teams are reportedly inside the offer, which cuts against a simple story of AI replacing staff in other teams. The wider evidence on whether 2026 cuts are about AI or about correcting earlier hiring is weighed in the analysis of AI layoffs as an overhiring correction.
Where this analysis is weakest
Three parts of this post rest on thinner ground than the rest, and you should discount them accordingly.
The uptake range is borrowed, and the counts are secondhand
The 3% to 5% range is Microsoft's description of other companies' programmes, not a forecast built from its own staff. The Register links it to Google's 2025 buyouts. Microsoft's own April result was 30%, so the company's single data point sits far outside the peer range. The 2,600 acceptance count comes from Implicator, not from a Microsoft filing, and the "more than 30%" figure comes from a July post I have only seen quoted.
The cost ceiling is crude
Dividing an expected charge by a reported acceptance count mixes a forecast with an outcome. Microsoft said April came in lower than expected, so the true per-head cost was below $346,000. October's package is also poorer. The dollar column could be overstated by half or more, and I have no source that lets me narrow it.
The eligible pool is not uniform
Level 59 engineers in India and level 70 general managers in Seattle face completely different choices. A blended uptake rate hides that. If senior US staff in AI-adjacent groups leave at 10% while Indian teams leave at 1%, the headcount effect is small and the talent effect is large. Nobody outside Microsoft will see that split.
What the Microsoft buyout means if you run a team
Two practical points follow for anyone managing headcount against an AI budget, at any size.
First, a voluntary window is a pricing decision. Microsoft priced April to sell and got 30%. It priced October at standard severance and expects under 5%. If you run one, decide the number of exits you need first, then set the terms that will produce it. Do not set terms and hope.
Second, the people most likely to accept a broad offer are the people with the best outside options. That is the main cost of the voluntary route, and it does not appear in any per-head calculation. If you sell seats to large Microsoft-style buyers, expect fewer users on renewal and price for it, a pattern covered in the analysis of seat compression in SaaS pricing.
Frequently asked questions
How many Microsoft employees are eligible for the voluntary exit offer?
About 60,000 employees are eligible, roughly 27% of the 223,000 full-time staff Microsoft reported at 30 June 2026. The offer covers participating teams in the US, Canada, the UK and India at level 70 and below. Some sales, marketing, data centre and engineering teams chose not to participate, so being at the right level does not guarantee eligibility.
What is the deadline for the Microsoft buyout?
Eligible employees must decide in early December 2026, and those who accept leave Microsoft on 31 December 2026, according to CNBC's reporting of the 26 October announcement. Microsoft has not published the exact decision date publicly. Employees in participating teams received the terms through internal channels.
How many Microsoft employees are expected to take the voluntary exit offer?
Microsoft expects fewer than 5% of the 60,000 eligible employees to accept, which means up to about 3,000 people. A spokesperson said similar programmes elsewhere see about 3% to 5% uptake, which would mean 1,800 to 3,000 departures. That compares with roughly 30% uptake, about 2,600 people, for April's richer retirement programme.
What is the difference between Microsoft's voluntary retirement program and the voluntary exit offer?
The April retirement programme was US-only, capped at level 67 and limited to staff whose age plus years of service reached 70. It offered up to 39 weeks of pay and up to five years of healthcare access. The October exit offer has no age or tenure test, reaches level 70 in four countries and pays terms comparable to standard severance.
Is Microsoft's buyout caused by AI?
Microsoft says no. A spokesperson told KUOW that AI is not driving the offer. The timing sits alongside a doubling of Microsoft's quarterly equipment spending to $39.4 billion, and Copilot and AI teams are reportedly among the participating groups. The evidence supports a shift in where the money goes, but it does not prove AI is replacing the people who leave.
How much will the Microsoft voluntary exit offer cost?
Microsoft has not said. The only priced reference is April's retirement programme, which Microsoft expected to cost about $900 million, or about $346,000 per acceptance. Using that as a ceiling, October costs at most about $1.04 billion at 5% uptake. The real figure should be lower, because the October package is less generous.
Where to start
If you are one of the 60,000, ask for two numbers in writing before early December: total weeks of base pay and the healthcare end date. Compare them with what a layoff package at your level paid in July, because that is the real alternative being priced.
If you sell software into Microsoft or a buyer like it, pull your seat count by team now and mark which teams sit inside the participating groups. That is your renewal exposure for 2027, and you will want the list before the customer has it.
References
- CNBC, Microsoft offers voluntary buyouts to 60,000 employees as AI leads to efficiency push, Jordan Novet, 26 October 2026. Used for eligibility, deadlines, the Coleman memo and the July uptake statement.
- CNBC, Microsoft plans first-ever voluntary employee buyout for up to 7% of U.S. workforce, Jordan Novet, 23 April 2026. Used for April eligibility rules.
- Fortune, Microsoft's 60,000-person buyout offer hints at a future with fewer workers and more equipment, Steven John, 30 October 2026. Used for spokesperson uptake quotes, participating teams and Hood's comment.
- The Register, Microsoft offers parachutes to 60k employees if they jump by the end of the year, Brandon Vigliarolo, 26 October 2026. Used for the buyout classification and the Google peer comparison.
- Implicator, Microsoft buyouts reach 60,000 workers, Marcus Schuler, 26 October 2026. Used for the 2,600 of 8,750 acceptance count, the 8-week floor, the effect on July layoffs and the TIKR estimate. Secondary source.
- KUOW, Microsoft offers voluntary buyouts to a quarter of its workforce, Monica Nickelsburg, 26 October 2026. Used for the standard severance description and the AI statement.
- Techloy, Microsoft's first-ever voluntary retirement package includes healthcare, cash and stock vesting, 7 May 2026, relaying The Verge. Used for April severance, healthcare and vesting terms. Upgrade to The Verge original.
- Microsoft, FY27 Q1 earnings release, 28 October 2026, and 2026 Annual Report. Used for additions to property and equipment, operating cash flow and revenue, the 223,000 headcount, the US split and fiscal 2026 property and equipment additions.
Weakest point in the source base: acceptance counts and package terms for both programmes come from press reporting of internal memos, not from Microsoft filings. Last verified 12 November 2026. Re-check once Microsoft reports its December quarter, which should carry the first charge for the exit offer.
Related reading