In Zurich, the terms are arithmetic: one month of salary for every year an employee has worked at Google, capped at twelve.
That formula is the Swiss version of a voluntary exit program Google has been running since June, Reuters reported on November 4. What began as a pilot in the U.S. and Britain now extends to Canada, France, Germany, Japan and Switzerland. DeepMind, the company’s AI research unit, and Google Cloud are excluded.
Alphabet’s search advertising business generates most of the cash that pays for the company’s AI buildout, and the buyouts are one way management has kept headcount flat while spending rises. Google has not said how many employees accepted. People who did have been leaving since the summer, and in some departments the voluntary offers were followed by involuntary cuts.
The design of such programs is familiar inside large technology companies. Rather than announcing layoffs, the employer offers a payment to people who choose to go, which reduces the number of forced departures and the legal exposure that comes with them. The cost lands in the quarter the offers are accepted; the benefit shows up later, in a lower salary base. Finance departments prefer the approach when a company needs to shrink without booking a restructuring charge large enough to alarm shareholders.
Excluding DeepMind and Cloud says more about strategy than the buyout terms do. DeepMind is where the company’s frontier models are built, and Cloud is where AI capacity is sold to other businesses. Both are being funded rather than trimmed. The offers are aimed at the parts of the company that support a business model built on search advertising and the products attached to it.
Europe complicates the mechanics. In France and Germany, works councils have a formal role in any reduction that touches a large number of staff, and consultations can take months. Japanese employers face different expectations about lifetime employment, and a buyout framed as a favor to employees is easier to sell there than a compulsory cut. The Swiss formula, with its flat twelve-month cap, is the simplest of the arrangements and the easiest for employees to price.
For an individual, the offer is worth less than it appears. Someone who leaves gives up unvested equity, and at a company whose shares have done well, that is often the largest part of a compensation package. The calculation favors people close to retirement, people who have already decided to leave, and people whose units are being shrunk anyway. It works against engineers in the middle of a vesting schedule who suspect their work is still needed.
That is roughly how the program appears to have played out. The departures began in the summer and have continued, and in some organizations the voluntary phase was followed by cuts, which is the sequence most companies plan for: take the volunteers first, count who is left, then decide how deep the compulsory round has to go.
The pressure behind the exercise is not hard to identify. Alphabet has been spending at a pace set by the data center buildout, and every dollar committed to buildings, chips and power is a dollar not available for headcount. Buyouts are the personnel-side counterpart to capital spending: both reduce the room for everything else. The difference is that capital spending buys capacity that can generate revenue, while a buyout buys a smaller cost base and, usually, a shorter list of products.
There is a cost on that side too, and it is harder to measure. Companies that cut through attrition and buyouts tend to lose people who have options elsewhere, which is not always the same group as the people management would have chosen. Excluding DeepMind and Cloud protects the units where the company believes the next decade of revenue is being built, and it tells you where management thinks a departure would be expensive.
The comparison with the rest of the industry is unavoidable. Technology companies across the sector have been reshaping staff around AI priorities for two years, hiring in infrastructure, model training and data center operations while trimming support functions, marketing and some product teams. Google’s program is unusual in its geography rather than in its purpose: a U.S. company using country-by-country terms to do the same thing everywhere it employs large numbers of people.
What the company has not done is describe the program as a response to weak demand. Search advertising has held up, and Cloud has grown. The framing inside Alphabet is that resources are moving toward AI, not that the business is contracting, and the absence of a public number for acceptances keeps that framing intact.
The Zurich formula gives employees something concrete to weigh against that framing. Twelve months of salary is a generous offer by local standards, more than many Swiss employers provide in a restructuring, and the queue at the door over the coming months will be the clearest measure of how many people at Google had already decided, for reasons of their own, that they would rather leave with a check than stay through the next reorganization.
