SUSE Is Cutting About 190 Jobs. Who Will Check the AI's Work?

SUSE will cut about 10% of its workforce, roughly 190 people, according to a memo CEO Dirk-Peter van Leeuwen sent to staff on Monday, Nov 2. Liam Proven of The Register saw the memo and reported it first, and The Next Web followed. SUSE's owner, the Swedish private equity firm EQT, puts the company's headcount at more than 1,900, which is how The Next Web arrived at its estimate of roughly 190.

Van Leeuwen gave his reasons plainly. He wrote that "the requirement for some existing roles has greatly diminished," and that modern tooling is changing daily workflows in ways that allow "flatter reporting lines." He also told employees the cuts reflect "shifting organizational needs, not your talent or commitment." I take that line at face value. The people hearing this news did nothing to cause it, and their CEO said as much in writing.

The memo did not name the affected roles, or where they sit. SUSE says the decisions turned on whether a person's role can "evolve," and on what that person can stop doing and start doing. The schedule is short. Affected staff are to be told by Nov 4, employee representatives are nominated the next day, one-to-one meetings begin the following week, and SUSE shares its fiscal 2027 strategy on Nov 16.

The sentence every software company will recognize

One line in van Leeuwen's memo will travel well beyond Nuremberg:

"Automation and AI are fundamentally shifting how software companies operate."

He is right about that, since AI tools now draft code and answer support tickets in seconds. The memo's phrase about flatter reporting lines points the same way, and a flatter chart can make sense when software absorbs coordination work. Review work still has to land on someone's desk, though, and a flatter chart gives it fewer desks to land on.

What the memo leaves open is the question I would want answered before any list is final: when the AI does more of the work, who inside the company checks it?

Who checks the machine

For a company like SUSE, that question carries real weight. Much of enterprise Linux is open source, so what customers pay for is support, and the confidence that someone at the vendor will know what to do when a system misbehaves. A model can draft a fix that looks perfect on the surface, but recognizing the one that is subtly wrong usually takes a person who has spent years inside the system and remembers how it failed before.

Most software companies can now rent the same models, and once nobody in-house can judge the output, a company ends up renting two things: the model, and the opinion about whether the model is right. Rented judgment costs far more than any invoice shows.

The model is for sale to everyone. The person who knows when it is wrong works for you. In a market where everyone rents the same intelligence, that person's judgment is the hardest advantage to copy.

The list of names

I do not know who is on SUSE's list, and nobody outside the company does yet. "Evolve" is a fair test, and SUSE has said what it weighed.

The question I would add applies to any leader holding a list of names this season: who on this list catches the model when it is wrong? Those people rarely carry a title that says so, because their best work appears as problems that never reach a customer, and an organizational spreadsheet has no column for prevented failures. Two follow-ups make it concrete. Who reviewed the last 10 AI outputs that reached a customer? If that person left on Friday, who would catch the next bad one on Monday?

I spent 20 years at Cisco and shaped its innovation work, including co-innovation centers on five continents. The people I came to rely on most were the ones who could look at a promising demo and say whether it would survive a real customer's network. They were hard to find, and they were seldom the most senior person present in the meeting.

Lose those people and a company takes on what I call the learning debt in my forthcoming book, The Great Relearning: the retraining it defers today, which comes due later with interest. The next AI transition will arrive within a few years. Someone will have to judge that one too.

A busy stretch for the same question

SUSE's memo lands in a crowded few weeks. I wrote recently about Barclays and American Express announcing AI-related cuts in the same week, and about who leaves when Microsoft makes a voluntary buyout offer. The AI job loss statistics hub on this site tracks the wider pattern as it moves.

For roughly 190 households, this week is a kitchen-table conversation first and a headline second, and the expertise those people built at SUSE leaves with them to benefit whichever company hires them next.

If you are more likely to read a memo like this than write one, this question belongs to you too. Take it into your next team meeting: when our AI tools get something wrong, who here is expected to notice, and how do they keep that skill sharp? Asking it makes you one of the people who check.

I am taking the same question into my own meetings. Somewhere this month, a leader will look at a list and ask which roles AI has made smaller. Before that list is final, who on it knows when the machine is wrong? If you are working through that for your own team, I'm easy to find.

Here is what makes Alex a credible voice on this topic: Alex Goryachev is the WSJ-bestselling author of Fearless Innovation and shaped a $1.1 billion innovation portfolio at Cisco. He advises enterprise leaders on workforce decisions in the age of AI, including which people a company cannot afford to lose.

Talk with Alex about your workforce strategy →