ChatGPT Alternative,
EU Sovereign AI from France

European controller, European staff

VOLTAGE EI is a French company headquartered in Solaize, near Lyon. SIREN 943 808 824. Founder and primary engineer based in France.

Model in an Intel TDX enclave

Every prompt is processed by a model inside an Intel TDX trust domain, whose memory the CPU encrypts with a key per trust domain. The inference provider is named in the DPA and its host sits outside the trust domain. Our relay forwards prompts in clear.

Native GDPR Article 28 DPA

Data Processing Agreement is the default contract on every plan, free or paid. No negotiation required.

Memory encrypted by the CPU

The model runs with memory encrypted by the CPU, with keys held in the processor, so its host cannot read it. Our relay, outside the enclave, sees prompts in clear while forwarding them.

No training on your data

Requests go through our API to the model, which runs inside an Intel TDX enclave at our inference provider; nothing is used for training. GET /v1/attestation proves that the instances serving a model are TDX guests, not that a given reply came from them.

No US parent

No corporate parent in the US. Our web host and our inference provider are US companies, documented in the DPA and the sub-processor register.

FREE

50 messages per month. Model in Intel TDX. No credit card.

PLUS

2,000 messages on Qwen3.5-397B-TEE (256K context). Native GDPR Art. 28 DPA. Model in Intel TDX.

PRO

5,000 requests a month on Qwen3.5-397B-TEE plus DeepSeek-V3.2-TEE reasoning, 10 seats.

Is VoltageGPU actually a European company?

European company. VOLTAGE EI is a French sole proprietorship registered with SIREN 943 808 824, based in Solaize, France. Founder Julien Aubry. There is no US parent and no US holding. Our web host and our inference provider are US companies documented in the DPA; the model runs in an Intel TDX enclave whose memory its host cannot read, and our relay handles prompts in clear while forwarding them.

Why is the EU AI sovereignty conversation important?

Forrester estimates European organizations will spend €1.5 trillion on cloud and AI infrastructure by 2030, with most of that flowing to three US hyperscalers and OpenAI. That creates strategic dependence comparable to energy imports. EU alternatives, including VoltageGPU, Mistral, OVHcloud, Scaleway, exist to give European companies real choice.

How is VoltageGPU different from Mistral Le Chat?

Both are EU-based and refuse training on user prompts. The difference is where the model runs: here, inside an Intel TDX enclave at our inference provider, with memory encrypted by keys the CPU holds, so that host cannot read it. Our relay still handles prompts in clear while forwarding them.

Will my existing ChatGPT prompts work?

Yes. The chat experience is intentionally familiar. Most ChatGPT prompts work without modification on Qwen3.5-397B-TEE. We also provide an OpenAI-compatible API endpoint, so developers migrating from gpt-4o or gpt-4o-mini only need to change the base URL and model name. Requests pass through our relay, which sees them in clear (TLS ends at our host). To keep plaintext off our servers, run your own inference server in a Confidential VM and terminate TLS inside its trust domain.

Does VoltageGPU comply with the EU AI Act?

Yes. We publish transparency notices, model cards for every TEE model, and provide human oversight features. Because we are an EU-located controller and our core business is regulated AI infrastructure, we are positioned for the AI Act's obligations on general-purpose AI providers and high-risk deployers.