Swiss AI companies pursue an alternative to US and Chinese tech giants

Swiss AI firms are pushing into a market led by US and Chinese giants. On October 29, 2026, start-up Giotto announced that it was taking its sovereign AI operating system beyond enterprise customers and into the general public. The move follows deployments at Groupe Mutuel, RUAG and SAP. At the same time, the Swiss-developed language model Apertus has passed six million downloads in a year, overtaking France’s Mistral. These efforts give Switzerland a visible foothold in a field shaped by products such as ChatGPT and DeepSeek. Their appeal rests on more than a local label: Swiss developers are pitching control, transparency and efficient computing as ways to compete. For a country whose researchers and companies want a say in how AI develops, Giotto’s consumer launch and Apertus’s reach mark a bid to turn domestic expertise into tools used at home and abroad.

Giotto is carrying its enterprise software into the consumer market. The Swiss start-up has already deployed its AI operating system at Groupe Mutuel, RUAG and SAP, and now wants members of the general public to use it. That shift gives Giotto a route to test its sovereign AI proposition beyond large organisations. Co-founder Aldo Podestà argues that Switzerland needs to master AI as a fundamental technology, comparing control of it with control of electricity. He says Swiss university research teams offer expertise the country can build on. Giotto’s competitive plan also differs from a race for speed alone: Podestà says the company wants to develop high-performance models that use little energy. “We must focus on efficiency,” he told RTS. The strategy sets out a distinct pitch for a smaller Swiss player: combine local technical capability with lower energy consumption and a system customers can use under Swiss control.

Apertus has paired fast-growing reach with an open approach. The Swiss-developed language model passed six million downloads in one year, a total that took it past French rival Mistral. EPFL machine-learning professor Martin Jaggi says the team is working on two measures of success: performance and public trust. “We’re trying to build trust through complete transparency,” he told RTS. That emphasis matters in a climate where AI’s capabilities also provoke concern. A Politico survey found that two in three Americans believed AI could destroy humanity. Apertus’s open model has drawn attention for its potential in both beneficial and harmful uses, including cybersecurity. For Swiss researchers, openness can make the system easier to examine, while also raising questions about who might use its capabilities. The model’s public profile now gives Switzerland a prominent example of a locally developed alternative, as well as a live test of how transparency shapes confidence.

Swiss developers frame sovereignty as a matter of national capability. Apertus co-founder Antoine Bosselut says that without a model like Apertus, Switzerland would have “no control over the future of AI in this country”. He points to the dependence that can come with systems such as ChatGPT or Claude: their developers could cut off Swiss access, or decide that data sent through them goes to the US government. Bosselut says the same concern applies to Chinese models. Giotto co-founder Podestà likewise argues that Switzerland must master AI to sustain its development. Yet local ambitions face a financing hurdle. Charles-Henry Monchau, head of investments at SYZ Bank, described investor reluctance and said access to capital was a bottleneck. The Swiss effort therefore depends on both research expertise and the ability of companies to fund development. Giotto’s push for efficient models and Apertus’s open approach show how local players are trying to establish a position on those terms.