2 Sources
[1]
Swiss startup Prem AI is raising $100M so hedge funds and law firms can own their AI instead of renting it
Prem AI is raising $100M at $500M+ valuation. It builds private AI infrastructure for hedge funds and law firms. "Enterprise intelligence must be owned, not rented." Prem AI, a Swiss startup that helps hedge funds and law firms run AI models on their own infrastructure, is raising $100 million in a Series A at a valuation of at least $500 million. The company expects to close the round in the third quarter. CEO and founder Simone Giacomelli said the pitch is simple: "Enterprise intelligence must be owned, not rented." The Lugano-based company previously raised a $14 million seed in 2024, followed by a $6.1 million bridge at a $200 million valuation. Backers include Jim Breyer (Breyer Capital), Index Ventures, Marvel Studios founder David Maisel, and Sequoia Capital China co-founder Fan Zhang. Prem builds AI products for organisations that cannot send sensitive data to OpenAI or Anthropic. The model is self-hosted: the AI runs inside the customer's own environment. Prem combines security with what Giacomelli calls a "ready-to-use application layer," giving customers "absolute control without sacrificing usability." Alongside the raise, Prem announced Fluso, an encrypted AI workspace that runs agents and automates tasks. The product is designed for firms where data leaving the building is not an option, whether for regulatory, competitive, or security reasons. The timing maps to a growing concern among enterprises about the "rented AI" model. Companies paying OpenAI or Anthropic for API access are sending their most sensitive data, including legal documents, trading strategies, and client information, to third-party servers governed by someone else's terms of service. The sovereignty question applies to companies, not just countries. Prem's team currently has 35 people and is expanding. The company sits in the same category as Databricks' on-premise AI offerings and self-hosted open-weight models from Meta and Mistral, but packages the full stack, from model deployment to agent automation, as a product rather than a toolkit. For hedge funds, the value proposition is competitive intelligence that stays inside the firm. For law firms, it is privilege-protected work product that never touches a third-party cloud. For both, the alternative, renting AI from an American company that could change its terms, raise prices, or be compelled to share data with a government, is a risk that grows with every API call. Europe's cloud dependency is already a political risk. Prem is betting it is also a commercial opportunity.
[2]
Prem seeks $100M Series A as export bans boost sovereign AI demand
Prem seeks $100M Series A as export bans boost sovereign AI demand Swiss artificial intelligence startup Prem SA is raising $100 million in Series A funding at a valuation of at least $500 million, Bloomberg reported today. The round is expected to close in the third quarter. Prem sells software for running AI models on a company's own infrastructure, not a cloud provider's. Customers fine-tune models, analyze documents and run inference in private environments and the data never goes back to Prem. That pitch is built for buyers who handle information they cannot hand to a third party. Its early targets are hedge funds and law firms. The company dates to 2023. Its founder, Simone Giacomelli, had earlier helped start the decentralized AI network SingularityNET. Prem is launching the fundraise alongside Fluso, an encrypted workspace for running AI agents and automating work. It runs where the customer's data already sits, whether that is a private cloud, a virtual private cloud or an air-gapped on-premise system. Fluso uses open-weight models, so every parameter can be audited. None of the data touches a training pipeline, the company says. Timing helps. On June 12, the U.S. ordered Anthropic PBC to block foreign access to its newest Fable 5 and Mythos 5 models and the company disabled them the next day, saying that was the only way to comply. The pattern showed up elsewhere too. JPMorgan Chase & Co. cut Claude access for staff in Hong Kong after deciding Anthropic's licensing terms exclude Greater China and Goldman Sachs Group Inc. had done much the same earlier in June. For a regulated firm, that is the moment a model stops being a tool and starts being a dependency someone else controls. Sovereign AI budgets were rising before any of that. Gartner Inc. has forecast roughly $80 billion in global sovereign cloud infrastructure-as-a-service spending this year. European spending alone is projected to more than triple, from $6.7 billion in 2025 to $23.1 billion in 2027. Prem also sells Switzerland. The country has an EU adequacy decision, so data can move to and from the bloc even though Switzerland is not a member and a rewritten Swiss privacy law has applied since September 2023. That does not put a Swiss company beyond the reach of foreign governments. But for a buyer trying to cut its reliance on U.S. cloud providers, Swiss jurisdiction is at least something an auditor can sign off on. Plenty of rivals want the same customers. Mistral AI SAS has leaned into European infrastructure, including $830 million in debt financing for a data center near Paris. Aleph Alpha GmbH built its name on sovereign deployments for governments. The hyperscalers are pitching their own locality-bound options, with Amazon Web Services Inc., Google LLC, Microsoft Corp. and IBM Corp. all in the mix. The new funding isn't Prem's first raise. The company closed a $14 million seed round in April 2024, followed by a $6.1 million bridge at a $200 million valuation. Backers have included Fan Zhang, co-founder of Sequoia Capital China and David Maisel, founding chairman of Marvel Studios. Clear $500 million and the valuation is up roughly 2.5 times in about two years, with total funding pushing past $120 million.
Share
Copy Link
Swiss startup Prem AI is raising $100 million in Series A funding at a valuation exceeding $500 million to build private AI infrastructure for hedge funds and law firms. The company's pitch centers on owning enterprise intelligence rather than renting it from third-party providers. Alongside the funding round, Prem launched Fluso, an encrypted AI workspace designed for organizations where data sovereignty and regulatory compliance are non-negotiable.
Swiss startup Prem AI is raising $100 million in a Series A funding round at a valuation of at least $500 million, with the round expected to close in the third quarter
1
2
. The Lugano-based company previously raised a $14 million seed in April 2024, followed by a $6.1 million bridge at a $200 million valuation, bringing total funding past $120 million2
. Backers include Jim Breyer from Breyer Capital, Index Ventures, Marvel Studios founder David Maisel, and Sequoia Capital China co-founder Fan Zhang1
. The valuation represents roughly 2.5 times growth in about two years, signaling strong investor confidence in the sovereign AI market.Prem AI builds private AI infrastructure for enterprises including hedge funds and law firms that cannot send sensitive data to OpenAI or Anthropic
1
. CEO and founder Simone Giacomelli, who previously helped start the decentralized AI network SingularityNET, frames the company's mission simply: "Enterprise intelligence must be owned, not rented"1
2
. The company's model is self-hosted, meaning customers can run AI models on private infrastructure within their own environment, whether that's a private cloud, virtual private cloud, or air-gapped on-premise system1
2
. This approach gives organizations absolute control over their data without sacrificing usability, combining security with what Giacomelli calls a "ready-to-use application layer"1
.
Source: SiliconANGLE
Alongside the Series A funding round, Prem announced Fluso, an encrypted AI workspace that runs agents and automates tasks for firms where data leaving the building is not an option
1
2
. Fluso uses open-weight models so every parameter can be audited, and none of the data touches a training pipeline2
. For hedge funds, the value proposition centers on competitive intelligence that stays inside the firm. For law firms, it means privilege-protected work product that never touches a third-party cloud1
. The product addresses organizations facing regulatory compliance requirements or those handling information they cannot hand to a third party2
.The timing of Prem's raise maps to growing concerns about data sovereignty and geopolitical dependencies in AI. On June 12, the U.S. ordered Anthropic to block foreign access to its newest Fable 5 and Mythos 5 models, and the company disabled them the next day
2
. JPMorgan Chase cut Claude access for staff in Hong Kong after deciding Anthropic's licensing terms exclude Greater China, and Goldman Sachs had done the same earlier in June2
. For regulated firms, these incidents transform AI models from tools into dependencies controlled by others. Companies paying OpenAI or Anthropic for API access send their most sensitive data—including legal documents, trading strategies, and client information—to third-party servers governed by someone else's terms of service1
.Related Stories
Gartner forecasts roughly $80 billion in global sovereign cloud infrastructure-as-a-service spending this year
2
. European spending alone is projected to more than triple, from $6.7 billion in 2025 to $23.1 billion in 20272
. Prem's Swiss jurisdiction provides an EU adequacy decision, allowing data to move to and from the bloc even though Switzerland is not a member, with a rewritten Swiss privacy law applied since September 20232
. For buyers trying to cut reliance on U.S. cloud providers, Swiss jurisdiction offers regulatory clarity that auditors can approve. The company's 35-person team is expanding to meet this demand1
.Prem faces competition from several players targeting the self-hosted secure AI systems market. Mistral AI has leaned into European AI infrastructure, including $830 million in debt financing for a data center near Paris
2
. Aleph Alpha built its reputation on sovereign deployments for governments2
. The hyperscalers are pitching their own locality-bound options, with Amazon Web Services, Google, Microsoft, and IBM all competing for the same customers2
. Prem sits in the same category as Databricks' on-premise AI offerings and self-hosted open-weight models from Meta and Mistral, but packages the full stack from model deployment to agent automation as a product rather than a toolkit1
. The alternative for enterprises—renting AI from an American company that could change its terms, raise prices, or be compelled to share data with a government—represents a risk that grows with every API call1
.Summarized by
Navi
[1]
16 Mar 2026•Business and Economy

21 Aug 2025•Business and Economy

12 Jun 2026•Startups

1
Policy and Regulation

2
Policy and Regulation

3
Business and Economy
