Prem AI seeks $100M Series A for private enterprise AI

Why it matters

Prem AI is testing whether investors will fund private, verifiable AI infrastructure as a core enterprise category, not a security add-on.

Prem AI, the Swiss private AI infrastructure startup led by founder and CEO Simone Giacomelli, is raising a $100 million Series A round at a target valuation of at least $500 million, Bloomberg reported Thursday.

The key word is "target." Prem AI has not closed the financing. Bloomberg reported that Prem AI expects to close the round in the third quarter of 2026, citing a company statement and Giacomelli. No lead investor, participant list, committed capital amount, or final valuation was disclosed.

That makes this a live fundraising story, not a victory lap. Giacomelli is trying to price Prem AI as one of the more valuable early-stage companies in the private AI deployment market before the round is done, and he is doing it at a moment when enterprise buyers are still trying to reconcile two facts: employees want access to frontier AI systems, and regulated companies cannot simply pour client files, trading strategies, medical records, or legal work product into third-party systems they do not control.

Bloomberg described Prem AI as helping companies, including hedge funds and law firms, run AI models on their own private infrastructure. Prem AI does not name those customers in the Bloomberg report, so the claim should be read as category-level traction rather than a list of reference accounts.

Giacomelli's bet is not another chatbot

Giacomelli is identified by Bloomberg as Prem AI's founder and CEO. The financing pitch leans less on founder profile and more on a market wedge: enterprises want AI performance without surrendering control of their data, model behavior, or infrastructure boundary.

Prem AI's own product language leans into that wedge, emphasizing privacy, verifiability, and sovereignty for enterprise deployments. The company markets a private AI workspace and confidential compute offerings rather than another general-purpose assistant.

In a May 20, 2026 blog post announcing Prem Confidential APIs, Prem AI said its API runs models inside trusted execution environments and argued that standard AI APIs often process sensitive prompts, files, and conversations in plaintext on provider servers. The company says customers should be able to verify the environment cryptographically.

That is the actual product story behind the financing target. Prem AI is not merely selling enterprises a private wrapper around AI. It is trying to sell a control plane for AI workloads where the buyer can keep data inside a private perimeter, run models through protected infrastructure, and verify that the infrastructure is what it claims to be.

The number says investors are being asked to underwrite control

A $100 million Series A would be large by normal software standards, but AI infrastructure is no longer being financed like normal software. Compute, security engineering, enterprise sales, compliance work, and model operations all burn capital before revenue quality is obvious from the outside. That is especially true for a company selling into hedge funds, law firms, healthcare, finance, government, or other sensitive environments, where procurement cycles are longer and security reviews are part of the product.

There is prior financing context. In April 2024, FinSMEs reported that Prem Labs raised $14 million in seed funding.

Prem AI is selling against the hidden cost of hosted AI

Prem AI's pitch works because the biggest enterprise AI bottleneck is not always model quality. In many high-value workflows, the harder question is where data goes, who can inspect it, how long it is retained, and whether the customer can prove what happened after an AI request left the company boundary.

That is why hedge funds and law firms are credible buyer categories for Prem AI even without named logos. Both sectors are built on information asymmetry, confidentiality, and auditability. A hedge fund does not want proprietary research, portfolio context, or trading logic exposed outside its controls. A law firm does not want privileged documents or client communications processed in a way that later creates a confidentiality problem. If Prem AI can give those buyers useful AI while reducing the data-exposure argument, Giacomelli has a sharper sales motion than another generic enterprise assistant.

Prem AI's GitHub organization also shows the team publishing open-source code and tutorials related to its stack.

The Series A target is therefore a test of whether the market believes private AI infrastructure is an urgent category, not a compliance feature. Giacomelli is asking investors to fund Prem AI before the financing is complete, at a valuation that assumes enterprises will pay for verifiable control as AI moves from pilots into production. The next fact that matters is not the headline size. It is who signs the term sheet.