Microsoft posted a tenth record quarter days after offering exits to 60,000 staff
No leading AI company is announcing that agents are replacing people. They don’t need to. The numbers say it for them.
Ignore the why and ask yourself how.
Every round of job cuts comes with a reason: removing layers, efficiency, restructuring. The reasons are not the interesting part. The interesting part is how a company grows revenue by 18% in a year while its headcount falls. Someone, or something, is doing that extra work.
This is not a small, subtle change. In a single year the biggest technology companies have stopped resourcing work with people in the way they always have, and recruitment has fallen to a fraction of what it was.
The numbers: growing the business, shrinking the team
The pattern is the same at every leading AI company: the business grows while the workforce shrinks or stands still.
Microsoft
Headcount fell from about 228,000 to 223,000 in fiscal 2026 (July 2025 to June 2026), its first annual fall since 2016. In October 2026 it offered voluntary exits to about 60,000 staff, roughly 27% of the workforce. Over the same period revenue reached $331.8bn, up 18%, its largest one-year rise, and its latest quarter (July to September 2026) was a tenth record quarter in a row at $92.3bn, up 19%. (10-K report, AOL)
Oracle
Up to 30,000 jobs are planned to go, with 21,000 already cut in fiscal 2026 (to 31 May 2026), taking headcount from about 162,000 to 141,000, a 13% fall. It plans around $70bn of capital spending this year, mostly on cloud and AI, and its own filing says AI “has resulted, and may continue to result, in reductions” to its workforce. (TechRepublic)
Amazon
About 14,000 corporate roles went in October 2025, then 16,000 more in January 2026. Internal plans reported by The New York Times aim to avoid hiring more than 600,000 US workers by 2033 through automation, and CEO Andy Jassy expects the corporate workforce to shrink “as we get efficiency gains from using AI”. (Yahoo Finance, Computerworld)
Salesforce
Customer support went from about 9,000 people to about 5,000, and AI agents now handle around half of all customer conversations. (Storyboard18)
Graduate hiring
New graduates are now just 7% of hires at the 15 largest tech companies, and graduate hiring is down by more than half since 2019. The jobs that used to train the next generation are the first to go. (SignalFire)
The forecasts point the same way. Goldman Sachs now estimates AI could displace up to 9% of US workers, about 15 million people, over a decade (OECD.AI). Anthropic’s CEO, Dario Amodei, has warned AI could remove half of all entry-level white-collar jobs and push unemployment to 10 to 20% within one to five years (Axios).
So who, or what, is doing the work?
Take Microsoft. With 2% fewer people it earned 18% more. That is roughly 20% more revenue per employee in a single year, about $1.49m per person against $1.24m the year before (our calculation from its published figures). People did not suddenly work 20% harder. The extra capacity came from somewhere else.
It came from AI: agents that write code, answer customers, reconcile accounts, draft content and run checks, around the clock, at a fixed cost. Salesforce has said it out loud for customer support. Oracle has put it in a filing. Most companies simply say “efficiency” or “fewer layers”.
You will not see a press release headed “We are replacing a large share of our staff with agentic teams”. It is not a PR event anyone wants to hold. But the leading AI companies are no longer resourcing work with people in the way they did a year ago, and the gap between their growth and their headcount is the evidence.
Two hard choices, one for businesses and one for staff
What starts at the leading AI companies spreads to everyone else. That paints a bleak picture for how people are resourced as AI adoption widens, and it leaves two uncomfortable choices.
- For businesses: invest in agentic AI, or keep scaling by headcount alone and stop being competitive against the companies that do. A competitor that answers customers in seconds and reports in real time, at a fixed cost, is hard to beat with more hiring.
- For staff: step up, or risk being replaced. The roles that are disappearing first are the routine, repeatable ones, and the entry-level jobs that used to teach them.
Not every forecast is this stark, and many economists expect a slower shift than AI leaders predict. But the direction is not in doubt, and waiting to see is itself a choice.
The better option: use AI now, so you are not replaced by it next year
There is a third way, and it is the one we believe in. The people the leading companies keep are the ones who direct the agents. So the best protection for your team is to make every one of them good at using AI now, while there is still time to learn.
What that looks like in practice for a UK business:
- Give staff safe, approved AI. If people are pasting company data into free chatbots, you have a data risk and no learning. A private AI that runs on your own hardware or private cloud keeps your data in the business.
- Train on real work. Start with the tasks your team does every day: emails, reports, customer replies, data entry. The skill to build is directing AI, checking its work and owning the result.
- Give everyone an AI assistant that knows your business. An assistant with access to company knowledge makes each person faster, rather than making them optional.
- Hand the routine work to agents, and move people up. Let agents take the repeatable tasks, and move people to judgement, relationships and the hard cases.
- Set the rules. A clear AI use policy, and a human sign-off on anything sensitive, from refunds to legal matters.
At Total Group we are doing exactly this ourselves. From Monday our own team gets Wingman, an AI assistant that learns how we work and knows our company. We help other businesses do the same with private AI that they own, so their people grow with AI instead of being replaced by it.
Don’t rent intelligence. Own it.
Where does your team stand? Find out with a free AI Readiness Review.
Frequently asked questions
Are big tech companies replacing staff with AI?
Few say so directly, but the numbers show it. Microsoft grew revenue 18% in fiscal 2026 while headcount fell 2%, Oracle has cut 21,000 jobs and says AI has reduced its workforce, and Salesforce cut customer support from about 9,000 to 5,000 people as AI agents took on half of conversations.
Will AI replace jobs in UK small businesses?
Routine, repeatable work is the most exposed, and the shift that starts at the largest AI companies spreads to everyone else. Businesses that give their staff safe AI and train them to direct it protect both their competitiveness and their people.
How can my team prepare for AI?
Give staff approved, private AI, train them on their real daily tasks, give each person an AI assistant that knows the business, hand routine work to agents and set clear AI use rules. A free AI Readiness Review from Total Group shows where your team stands today.
