Microsoft and Meta Are Firing Thousands

In Brief

Sarah Franklin runs Lattice, an HR platform that helps companies measure and develop their people. In April 2026 she sat down on Bloomberg Technology with host Ed Ludlow and said what a lot of people are thinking but few executives say out loud: Microsoft and Meta are cutting thousands of jobs to free up capital for AI, but the cuts barely move the financial needle, and the real price is higher than the spreadsheet shows.

Ludlow asked the questions that needed asking. Does cutting at this scale actually free up enough capital to matter? And what is the real difference between Meta's forced layoffs and Microsoft's voluntary buyouts?

Franklin's answer: almost none. Both companies are investing in severance, not in the skills of their people. That's a short-term gain on the spreadsheet and a long-term loss of knowledge in the workforce.

Tokenmaxxing is not a performance metric

One of the new measurements appearing in tech companies right now is tokenmaxxing: counting how many tokens (the basic units an AI reads and writes) an employee uses in their work, and treating that number as a proxy for productivity.

Franklin thinks the premise is wrong. Her question:

"Would you look at the person at your company that sends the most email as the most productive? It's a new thing that we can use to measure. It's not necessarily the correlation to performance."

Sending a lot of messages to an AI is not the same as doing good work. It measures activity, not value created. Revenue per employee is a more established metric that actually says something real, but tokenmaxxing doesn't tell you who is contributing and who is just prompting more.

The cuts don't free up what you think

Microsoft has announced roughly 10,000 layoffs. Meta around 8,000. Both companies are trimming headcount ahead of quarterly earnings, and the narrative is that it frees up capital for AI investment.

Franklin is skeptical of that math:

"It doesn't really relieve the pressure, and it also deeply impacts culture. Whether it's one person or a thousand people or 10,000, it is a human. And people care about what happens to other humans."

There is a real cost on the spreadsheet, and there is a real cultural cost. Experienced people leave. Some go on to found companies of their own. The colleagues who stay can see what happened, and they feel it.

Layoffs and buyouts: the same thing underneath

Ludlow draws a distinction: Meta is doing forced cuts, while Microsoft is offering voluntary buyouts. Is there a meaningful difference?

Franklin's answer is direct: not really.

"It's still an investment in severance and not in the skills and your people, and that's the big difference."

Both approaches spend money to move people out of the company. Neither spends money helping people develop the skills they need for the AI era.

The real shift

Franklin uses an analogy that names what employees are actually experiencing. You cannot tell someone who has spent decades building a career that they need to change overnight. It is like telling someone to be in Europe in an hour.

"Unless you know how to teleport, it's not possible. But you can get there in a day."

Transformation is possible, but it takes time. It requires leaders with the courage to invest in their people, not just in hardware. It requires individuals to take ownership of their own development. And it requires companies to think long-term: losing expertise is expensive, even when it looks cheap on next quarter's results.

The real question Franklin is asking is not whether AI is taking jobs. It is whether companies are choosing to bring their people along into the new era, or leaving them behind on the way out.

Glossary

Sources and resources