Microsoft Has $81 Billion in Revenue and Is Offering 8,750 People Money to Leave
There’s a detail in Microsoft’s voluntary buyout announcement that’s easy to miss because it’s buried in the financial context.
Microsoft isn’t struggling. Its last quarter produced $81.3 billion in revenue. Net income was $38.5 billion — up 60% year on year. The company is spending $145 billion on capital expenditure this fiscal year. Its shares are up more than 10% in April alone heading into earnings week.
Microsoft is offering thousands of employees early retirement not because it cannot afford to keep them. It’s doing it because it has decided it doesn’t need them in the same way it did before. That’s a different kind of thing, and it deserves a different kind of analysis.
Direct Answer: What is Microsoft’s buyout programme and who is eligible?
On April 24, 2026, Microsoft announced the first voluntary retirement programme in its 51-year history. Approximately 7% of its US workforce — roughly 8,750 people out of 125,000 — are eligible. The formula is called the “Rule of 70”: an employee must be at the senior director level or below, and their age plus their years of service at Microsoft must add up to 70 or higher. Someone who is 52 with 18 years at the company qualifies. Someone who is 45 with 20 years does not. Eligible employees will receive their specific offer details on May 7 and have a 30-day window to decide. Employees on sales incentive plans cannot participate. Microsoft’s Chief People Officer Amy Coleman framed the programme as giving eligible employees “the choice to take that next step on their own terms, with generous company support.” The company has not disclosed the financial terms of the package.
The 2026 Big Tech AI Workforce Restructuring Index
What the major tech companies are doing with their workforces as AI spending accelerates — and what it signals about who they think they need.
| Company | Workforce Action (2026) | AI Capex Commitment | Revenue Context | Framing |
|---|---|---|---|---|
| Microsoft | 8,750 buyout-eligible (7% of US) | $145B capex FY2026 | $81.3B quarterly, up 17% | “Voluntary retirement” — company first |
| Meta | 8,000 cuts planned (May 2026) | $115–135B capex 2026 | Record profits | ”Efficiency” restructuring |
| Amazon | ~30,000 cuts H1 2026 | $200B capex 2026 | AWS growing 17% YoY | Division consolidations |
| Oracle | Up to 30,000 (March 2026) | AI data centre expansion | Mixed results | 6am emails, no warning |
| Atlassian | 1,600 cuts + CTO replaced by two AI leads | Undisclosed | Growing revenue | Restructuring around AI |
| Google/Alphabet | Selective cuts, AI teams growing | $75B capex 2026 | Search revenue holding | ”Investing for the future” |
Across these six companies: roughly 80,000+ workers are exiting in 2026 Q1–Q2. Total tech sector cuts in 2026 have already exceeded 95,000 as of late April, with an estimated 44% linked to AI automation.
The Rule of 70 Is a Design Choice, Not Just a Formula
The eligibility criteria for Microsoft’s buyout are worth examining because they reveal what kind of employee the company is trying to offer a gentle exit to.
Someone who is 55 years old with 15 years at Microsoft qualifies. They’re 70 combined. Someone who is 40 with 28 years at the company also qualifies — though that profile is rarer. The profile the formula targets most densely is the mid-to-late career employee: someone in their late 40s to late 50s, with 15 to 25 years of accumulated institutional knowledge, at a level below senior director.
These are, in many organisations, the people who know how things actually work. Not the people who manage strategy from above, and not the newer employees still absorbing the culture — the people in the middle who understand why decisions were made, who remember why a particular system is built the way it is, who have the kind of operational knowledge that never makes it into documentation.
Satya Nadella said in an earnings call earlier this year that approximately 30% of Microsoft’s code is now written by AI tools. That statement is relevant context for the buyout. If AI can write 30% of the code, the value of a human who has spent 20 years accumulating context about a specific codebase is lower than it was. Not zero — but lower. The Rule of 70 is, in part, a formula for identifying the employees whose value proposition has been most directly affected by that shift.
Microsoft isn’t saying this explicitly. “Simplifying operations” and “maintaining pace as the company expands its AI and cloud infrastructure” are the official framings. Both are true. Neither is the complete picture.
What Makes This Different From a Layoff
Microsoft eliminated over 15,000 positions in 2025 — 9,000 in a single July round and 6,000 in May. Those were layoffs: involuntary, immediate, with severance packages that employees received without a choice window.
The voluntary buyout is structured differently, and the difference matters both practically and symbolically. Eligible employees get information on May 7 and 30 days to decide. They can consider their specific package terms, talk to a financial advisor, evaluate their options in the job market. If they decide the offer isn’t right for them, they stay. If they decide it is, they leave on their own terms — which is a meaningfully different experience from being handed a termination notice.
The practical effect from Microsoft’s perspective is similar in terms of headcount reduction. The experience from the employee’s perspective is substantially different. Employment lawyers are already noting that the voluntary structure reduces Microsoft’s legal exposure considerably compared to involuntary layoffs: under the Older Workers Benefit Protection Act, employers must provide specific disclosures and waiting periods when offering severance to workers over 40 in exchange for waiving age discrimination claims. A voluntary programme that targets the Rule of 70 population — heavily skewing toward workers over 45 — requires careful compliance with OWBPA, and the May 7 notification and 30-day window is likely designed with those requirements in mind.
The “voluntary” framing also serves Microsoft’s internal culture. The employees who don’t take the offer and stay need to believe they’re staying because Microsoft wants them, not because Microsoft couldn’t afford to let them go. A voluntary programme preserves that narrative in a way that mass layoffs don’t.
The Actual Numbers Behind the Decision
Microsoft’s financials make the “we have to cut to invest” argument harder to sustain than it might be at a company under real pressure.
Q2 FY2026 results: $81.3 billion in revenue (up 17%), $38.3 billion in operating income (up 21%), $38.5 billion in net income (up 60%). The company is printing money. Its Azure business, which includes AI services, grew 33% year on year. Microsoft Copilot had 15 million paid seats as of January, with 450 million Microsoft 365 subscribers in the install base. The AI investment is working.
The capex context: Microsoft is spending $145 billion in capital expenditure this fiscal year. That money goes to data centres, servers, networking, and the GPUs and custom chips inside them. It is a massive number — but Microsoft’s quarterly operating income alone is over $38 billion. The company can afford both the investment and the workforce without choosing between them.
The choice to offer buyouts is not financial necessity. It is strategic allocation. Microsoft has decided that the marginal dollar is worth more invested in compute infrastructure than in retaining experienced employees whose roles have been partially or wholly automated. That calculation may be correct from a shareholder return perspective. It represents a specific view about what kind of company Microsoft wants to be in five years — one that runs leaner on headcount and heavier on AI-powered productivity for those who remain.
There is nothing dishonest about that decision. But calling it a response to “inevitable” AI-driven change, rather than a choice made by profitable company leadership, understates the agency involved.
The Pattern Across the Industry
Microsoft is not doing this alone, and the collective pattern is worth seeing clearly.
Oracle eliminated up to 30,000 employees in March, delivering the news via 6am emails with no advance warning, to fund AI data centre expansion. Meta plans to cut 8,000 workers in May as part of AI restructuring while simultaneously doubling AI infrastructure spending to between $115 billion and $135 billion. Amazon is signalling roughly 30,000 cuts across Alexa, AWS, and Prime Video in H1 2026. Atlassian cut 1,600 jobs and replaced its CTO with two AI-focused executives.
By late April, more than 95,000 tech workers had lost their positions in 2026 across 249 companies — an estimated 44% of those reductions linked directly or indirectly to AI automation.
The nature of the link varies. Some roles are eliminated because AI can do the specific task cheaper. Others are eliminated because the company is redirecting budget toward AI infrastructure, and headcount is the most controllable variable. Still others are eliminated because the company has used AI to flatten its management structure, and mid-level managers who previously coordinated between individual contributors now have less work to coordinate. All three are happening simultaneously across the industry, which is why the numbers are large and the explanations are varied.
What’s consistent across all of them is that the companies are profitable and growing. This is not a recession-era workforce reduction. It is an AI-era reallocation — a deliberate reshaping of who these companies employ, toward what, and away from what. The people affected are learning this distinction the hard way.
Actionable Steps: If You Work in Tech
If you are a Microsoft employee who received the May 7 notification: Read the full offer terms carefully before the 30-day window opens. The financial package — the specific months of severance, the equity treatment, the benefits continuation period — determines whether the offer makes sense for your situation. The voluntary structure means you have time; use it. Consult a financial advisor and, given the OWBPA requirements that apply to workers over 40, potentially an employment lawyer before signing anything.
If you work in tech and haven’t received a buyout offer: The 95,000 cuts in 2026 span 249 companies. The restructuring pattern — experienced workers, mid-career, roles that touch systems AI now assists with — is consistent enough that it’s worth auditing your own situation. What percentage of your current work could be done by an AI tool in 18 months? What skills do you have that AI cannot replicate? These aren’t hypotheticals anymore.
For hiring managers and HR teams: Voluntary programmes like this one create an opportunity to hire experienced, senior people who have taken packages from major tech companies. The Microsoft buyout eligibles include people with 15–25 years of institutional knowledge and domain expertise — exactly the profile that’s often the hardest to find at startups. Watch for the post-May announcements.
For sovereign and privacy-first workplaces: The AI reallocation happening at major tech firms creates an unusual moment in the job market for privacy, security, and sovereignty-focused roles. As big tech consolidates toward AI-native structures, specialists in compliance, data governance, privacy engineering, and security — whose value isn’t easily automated — may find the talent market less crowded than it was.
FAQ: Microsoft’s Buyout and What It Signals
Q: What is the “Rule of 70” and who does it target? The Rule of 70 is Microsoft’s formula for buyout eligibility: an employee’s age plus their years of service at Microsoft must equal 70 or higher. An employee who is 55 with 15 years at Microsoft qualifies (55 + 15 = 70). An employee who is 48 with 22 years also qualifies. The formula targets mid-to-late career employees — typically those in their late 40s through late 50s with significant tenure — rather than newer or very senior employees.
Q: Why is Microsoft doing this while reporting record profits? Microsoft’s explanation is that it is “simplifying operations” to direct resources toward AI and cloud infrastructure expansion. The more direct explanation is that the company has decided its optimal workforce structure for an AI-intensive future is smaller and differently skilled than its current structure. It can afford to offer buyouts because it is profitable, and it is willing to do so because it has concluded the resulting headcount reduction serves its strategic direction.
Q: How does this compare to the 2025 layoffs? Microsoft laid off roughly 15,000 workers in 2025 — 6,000 in May and 9,000 in July — in involuntary rounds with immediate effect. The 2026 buyout is the first voluntary programme and is explicitly framed as an option, not a requirement. The mechanisms are different; the strategic direction they both serve is consistent.
Q: Will all 8,750 eligible employees take the offer? Almost certainly not. Voluntary buyout take rates in the tech industry typically range from 30% to 60% of eligible employees, depending on the package terms and the alternatives available. Until eligible employees see the specific offer on May 7, it’s impossible to estimate the actual headcount reduction from this programme.
Q: What happens to employees who don’t take the offer? According to Microsoft’s framing, they continue in their current roles. Nothing in the announcement indicates that declining the buyout has adverse consequences. Whether that remains true in subsequent restructuring rounds is a separate question — Microsoft has run involuntary layoffs in the past and may do so again.
Q: Is this a trend or a one-off? It is clearly a trend. Meta, Oracle, Amazon, and Atlassian have all reduced experienced headcount in 2026 while increasing AI capex. The specifics vary — Microsoft’s buyout is more humane in design than Oracle’s 6am emails — but the direction is the same across the industry. Companies that are profitable and growing are choosing to deploy capital toward AI infrastructure rather than maintain workforce sizes built for a pre-AI operational model.
Related Articles
- Meta Is Keystroke-Logging 70,000 Employees to Train the AI That Will Replace Them
- Anthropic Just Took $100 Billion From Its Three Biggest Competitors
- Best Open-Source AI Models in April 2026 — Ranked by Sovereignty
- Google Just Solved the Biggest Problem in AI Training — Without Building New Data Centers
- SpaceX Just Bet $60 Billion on Cursor — and It Changes AI Coding Forever
