Giotto launches Solo and Alto: Swiss sovereign AI that runs on one GPU — or eight B300s

Switzerland has a new answer to the sovereignty question in AI — and it comes in two sizes. On October 27, 2026, Lausanne-based lab Giotto announced Giotto Solo and Giotto Alto, a family of Swiss-developed reasoning models built around a single idea: advanced AI should be a public good that people and organisations can use on their own terms, on infrastructure they control.

The pitch is efficiency as sovereignty. More efficient models need less compute, energy, and infrastructure, Giotto argues, which makes it practical to run advanced AI in very different places — from a workstation to a multi-GPU server rack — instead of only inside a handful of hyperscale clouds. "Europe must offer more than a different flag on the same dependency," said Aldo Podestà, Giotto's CEO and co-founder. "Sovereignty has to belong to the people and organisations using AI, and not just to the companies building it."

Two models, two footprints#

Giotto Solo is designed to run on a single GPU. Giotto describes it as the smartest model to run on one GPU, aimed at users who want capable reasoning on a workstation rather than a cluster. Giotto Alto, the family's flagship for demanding reasoning and enterprise workloads, runs on eight B300 GPUs. Alto is made available for free to everyone through Giotto's shared cloud.

Giotto is no newcomer to reasoning benchmarks. Founded in Lausanne in 2017 by EPFL mathematics alumni, the lab has worked on machine-learning interpretability, language models, and reasoning systems, and its results placed 11th in the 2024 ARC-AGI competition and second in 2025, among roughly 1,500 teams.

Three ways to run it#

Across the family, users choose from three deployment options. Shared cloud is hosted and operated in Europe. Dedicated cloud runs in isolated environments at partner data centres in Switzerland or elsewhere in Europe. On-premises installation runs on compatible hardware through partner integrators, from a one-GPU workstation for Solo to multi-GPU server racks for Alto.

Giotto's cloud services are GDPR-aligned and, the company stresses, do not depend on US cloud infrastructure — the "no US dependency" line is the launch's sharpest political edge, aimed squarely at European governments and enterprises rethinking where their AI runs.

Free access, flat-fee API#

Shared cloud access is free. API access on the shared cloud uses a flat fee for all-you-can-use requests, limited to one concurrent request at a time, with token packages available for multi-concurrency — a deliberate break from uncapped per-token pricing, giving users more cost visibility. Dedicated cloud instances, starting at one GPU, are rented monthly and include the Giotto software licence. On-premises deployments are licensed annually or over multiple years, with pricing based on the number and type of GPUs.

Alongside the models, Giotto is launching its app for iOS and Android, plus a free anonymous chat that requires no account — the consumer-facing door into the same sovereign stack.

Why it matters#

Sovereign AI has mostly been a story of big national programmes and billion-euro compute builds. Giotto's bet is the opposite direction: make the models efficient enough that sovereignty becomes a deployment choice, not a construction project. If a capable reasoning model genuinely runs well on one GPU, a hospital, a ministry, or a mid-size company can keep its AI inside its own walls without building a data centre.

The open question is the one Giotto's own framing invites: whether efficiency at this level can hold up against frontier models running on vastly more hardware. Giotto's ARC-AGI results suggest the reasoning chops are real, but the market test — enterprises actually moving workloads onto Solo and Alto — starts now. "Our ambition is not to make the world dependent on Europe," Podestà said. "It is to show that more intelligence does not have to mean less independence."