Summary#
At Sunrise, one round of layoffs follows the next. Early this year the telecom group cut 147 positions, and now up to 450 more jobs could go. Sunrise is reviewing “a comprehensive realignment of its operating model” and wants to become more efficient with AI. Employees who spoke to NZZ am Sonntag suspect the cuts are meant to secure the executive bonus. Sunrise denies this.
Vincent Kaufmann, director of the Ethos investment foundation, points to the pressure from high debt and management incentives. He expects other companies, banks in particular, to announce similar plans soon. Mark Diethelm, equity analyst at Vontobel, sees telecom companies at an advantage because their networks and administration are already largely digitised. In growing markets, he says, companies use AI to generate more revenue per employee with the same headcount.
ETH labour economist Michael Siegenthaler has observed a notable rise in unemployment among highly qualified people in Switzerland since 2023, mainly in occupations heavily exposed to generative AI. The unemployment rate reached 4.9 percent in the second quarter.
My Quotes#
I see Sunrise as a beginning (quotes translated from German):
“I fear that in the coming months many companies in labour-intensive industries such as financial services or telecom will make similar announcements, and that Sunrise was only the beginning.”
When companies phrase their goals vaguely, AI washing is the obvious suspicion:
“Especially when you are listed on the stock exchange, it simply sounds much better to say ‘AI will make us more efficient’ than to say ‘we have grown fat and need to slim down’.”
The real efficiency potential lies where companies completely rethink or abolish processes that grew historically. But jobs depend on exactly those processes:
“That is very painful for the people affected, whose jobs depend on these processes.”
This creates massive resistance. Frustrated management then often resorts to job cuts, because the remaining workforce is forced to become more efficient and finally starts using the AI tools that are already there.