Berget AI

What Berget AI does

Berget AI is a Stockholm-based (founded 2024) sovereign AI infrastructure provider that operates LLM inference services on Swedish-owned, Swedish-hosted data centers. Its platform serves open-source language models — including Llama, Mistral, Gemma, DeepSeek, GLM, Qwen, Magistral, and specialized code, embedding, reranking, and Swedish-optimized speech-to-text models — through an OpenAI-compatible REST API (https://api.berget.ai/v1) designed to enable near-instant migration from OpenAI or Anthropic with a one-line code change. The underlying stack runs on AMD MI300x GPU accelerators (192 GB HBM3 per GPU, with 8 GPUs in production totaling 1.5 TB GPU memory) using the ROCm open-source platform, vLLM inference engine, and GitOps-managed Kubernetes infrastructure; no customer data is retained and no data leaves the EU, making the platform explicitly outside the reach of the US Cloud Act and FISA.

The company's product surface comprises Serverless Inference (pay-per-token API across tiered plans from €0/Free with €5 credits through €500/month Enterprise), Dedicated Inference (reserved GPU capacity for custom and fine-tuned models, listed as Coming Soon), the Berget AI Platform dashboard, and Berget Code — an agentic coding assistant priced at €150 per seat per month with a 256,000-token context window running on open models (Kimi K2.6, Gemma 4, Mistral Medium 3.5) and integrating with OpenCode and Pi agent harnesses. Go-to-market combines a self-serve, product-led motion (console.berget.ai, free tier, OpenAI-compatible endpoints) with a direct enterprise sales motion targeting regulated industries and public-sector procurement, including Peppol/invoice billing and custom contract terms. Compliance posture includes GDPR, NIS2 alignment, and an in-progress ISO 27001 certification.

Berget AI targets European public-sector bodies, regulated industries (financial institutions, healthcare), and developer/tech teams that require EU data residency and cannot route sensitive workloads through US hyperscalers. Reference customers include Swedish municipal operator Telge Återvinning and digital-transformation consultancy Digitalist; the company is also referenced by unnamed Swedish public-sector bodies and unnamed financial institutions. The company was co-founded by Christian Landgren (CPTO) and Andreas Lundmark (former BCG Gamma Nordic lead and global AI-at-Scale lead, now CEO), raised a €2.1 million seed round in February 2026 led by Luminar Ventures with Wellstreet and Norrsken Evolve, and operates with 1–10 employees.

Name

Berget AI

Legal name

Berget AI AB

Website

https://berget.ai

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

1–10 employees

Short description

Berget AI is a Stockholm-based sovereign AI infrastructure provider delivering OpenAI-compatible LLM inference on Swedish-hosted AMD MI300x GPUs, serving European public-sector bodies, regulated industries, and developers requiring full EU data residency.

Ownership category

akta.pro rank

Product category

Cloud AI Infrastructure / Sovereign LLM Inference Services

NAICS

Software Publishers (513210), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182)

SIC

Services-Prepackaged Software (7372), Services-Computer Programming, Data Processing, Etc. (7370)

akta.pro primary industry

Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) (HDAEANAC)

akta.pro secondary industries

Model Hosting, Serving & Inference Platforms (HDAAACAB), Model Deployment, Serving & Inference Platforms (HDAAABAF)

Keywords

• Sovereign AI infrastructure
• LLM inference services
• Open-source model hosting
• Agentic coding platform
• EU AI compliance

Headquarters

HQ city

Stockholm

HQ country

Sweden

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Operations, Marketing or Sales

Revenue model

1. Serverless Inference API — Pay-per-Token: Serverless inference on open language models billed per token consumed. Customers prepay credits (€5 free on Free tier, €25–€55 monthly top-up on paid plans) and consume tokens against their balance. Usage beyond included allocations billed at pay-as-you-go rates. Resource-based services charged hourly for reserved capacity.
2. Berget Code — Seat-Based Subscription: Agentic coding platform (Berget Code) priced at €150 per seat per month. Billing is flat per seat, non-usage-based, providing predictable costs for enterprise budget planning. Designed for compatibility with fixed public procurement processes.
3. Enterprise Plans — Dedicated Capacity: Enterprise tier at €500/month providing maximum rate limits, dedicated support, and invoice/Peppol billing. Includes €550 monthly credit top-up. Supports custom contract terms, Kubernetes clusters, and dedicated GPU capacity for high-demand workloads.
4. Credits and Prepayment System: Credits-based prepayment system where customers load prepaid credits into their account wallet. Credits are non-refundable and non-transferable, valid for 12 months. Auto-top-up available to prevent service interruption. Unused credits roll over monthly.
5. Kubernetes / Dedicated Inference — Resource-Based: Dedicated GPU clusters and Kubernetes services charged based on reserved resources (hourly billing for resource-based services). Custom model hosting, fine-tuned model deployment, and dedicated capacity for enterprise workloads.

Pricing tiers

Go-to-market motion2 records

Distribution channels5 records

Marketing channels8 records

Core offering

Berget AI operates a sovereign AI infrastructure platform hosted entirely in Swedish data centers, providing serverless LLM inference on open-source models (Llama, Mistral, Gemma, DeepSeek, etc.) via an OpenAI-compatible REST API, dedicated GPU capacity for custom and fine-tuned models, the Berget AI Platform for development and deployment management, and Berget Code — an agentic coding assistant running on Swedish infrastructure with a 256,000-token context window.

Product overview

Berget AI offers a sovereign AI infrastructure platform for European organizations, consisting of Serverless Inference API (OpenAI-compatible pay-per-token service), Dedicated Inference (reserved GPU capacity for custom models), and Berget Code (agentic coding assistant). The platform is built entirely on open-source models and Swedish infrastructure, providing GDPR-compliant AI services without data leaving the EU. Pricing tiers range from free (€5 credits) to Enterprise (€500/month), with Berget Code priced at €150/seat/month.

Differentiator

Problem solved

Functional benefit

Brands

• Berget Code: Agentic coding service hosted in Sweden built on open-source software and open models for developers needing data sovereignty.

Products and services

• Serverless Inference OpenAI-compatible serverless LLM inference API on Swedish infrastructure, billed pay-per-token with €5 free starting credits, scale-as-you-go model, supporting chat completions, embeddings, reranking, and streaming responses across open-source models. Targeted at developers and engineering teams.
• Dedicated Inference Dedicated GPU instances running on AMD MI300x accelerators for any model including custom fine-tuned models, with customizable instances, high-demand workload support, dedicated resources for predictable performance, and hourly resource-based billing. Targeted at enterprises and organizations with reserved capacity needs.
• Berget AI Platform Unified platform providing intuitive dashboard, resource management, cost transparency, and team collaboration tools for developing and deploying AI applications on Swedish sovereign infrastructure. Targeted at developer teams and enterprises.
• Berget Code Agentic coding assistant hosted entirely in Sweden using open-source software and open models, providing AI coding agents for fullstack, frontend, backend, DevOps, app development, and security tasks. Supports OpenCode and Pi agent harnesses, 256,000-token context window, and flat €150/seat/month pricing. Targeted at European developer teams and organizations requiring data sovereignty for proprietary code.
• Berget AI API OpenAI-compatible REST API providing chat completions, embeddings, reranking, and streaming responses against open-source models hosted on Swedish infrastructure. Enables developers to migrate from OpenAI or Anthropic with minimal code changes. Targeted at developers and engineering teams.

Quantifiable outcome

• CO₂e tracking per API response — transparent environmental impact for every inference call
• +3 more outcomes

Named customers4 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration3 records

AI capability10 records

Feature8 records

Partnerships

Two partnerships are on record, tiered minor.

• DigitalistminorStrategic or Co-development PartnerStrategic partnership announced for delivering sovereign AI solutions to enterprise clients undergoing digital transformation. Berget AI provides the infrastructure layer while Digitalist contributes digital transformation consulting expertise to joint enterprise clients.
• Opper AIminorStrategic or Co-development PartnerAnnounced partnership between Berget AI and Opper AI. Nature of partnership appears to be co-marketing and potentially integration/GTM collaboration between the two AI-native companies serving the European market.

Scale indicators5 records

Recent moves6 records

Expansion highlights7 records

Direct peers

• Replicate: Open-model inference API running a wide catalog of open-source models with usage-based pricing and a developer-first console. Closely comparable on product surface, although lacking dedicated capacity and European sovereignty positioning.
• Fireworks AI: Inference platform for open-source and proprietary models with OpenAI-compatible endpoints, serverless and dedicated deployment options, and per-token pricing. Direct product-and-model parity with Berget's serverless/dedicated offering.
• Together AI: Open-model inference platform with OpenAI-compatible API, pay-per-token pricing, and dedicated GPU deployments. Direct functional competitor on product, business model, and developer GTM — but US-based and without sovereign positioning.

Broad incumbents

• Scaleway: French sovereign cloud provider expanding into GPU instances and AI inference for European customers. Comparable in serving EU data residency and regulated workloads, but at a broader cloud IaaS layer rather than a focused AI inference platform.
• OVHcloud: European cloud provider with AI inference offerings and explicit sovereign cloud positioning. Targets the same EU public-sector and regulated-industry customers, but as a broad hyperscaler alternative rather than a specialized AI inference provider.
• CoreWeave: Large-scale GPU cloud provider offering both raw GPU instances and managed inference for AI workloads. Comparable on the dedicated-capacity side of Berget's business, but at massive scale and without sovereign positioning.
• Mistral AI: Paris-based foundation model and inference platform serving European enterprise and government customers. Strong sovereign AI positioning and OpenAI-compatible endpoints, with much larger scale and a proprietary model stack that Berget serves as open-model alternative.
• Aleph Alpha: European sovereign AI company based in Heidelberg, offering Luminous and Pharia models with EU data residency for regulated industries and public sector. Most direct European sovereign AI peer; broader portfolio and more capital, but similar sovereignty thesis and target customers.

Emerging players

• OpenPipe: YC-backed startup focused on fine-tuned open-model inference and agent infrastructure with usage-based pricing. Comparable in targeting developers and enterprises seeking open-model alternatives to OpenAI, though primarily US-based.

Regional players

• Silo AI: Helsinki-based European AI lab (acquired by AMD) offering open-source LLMs (Poro, Viking) and AI infrastructure for Nordic and European customers. Regional peer with strong Nordic overlap and explicit alignment with EU AI sovereignty and sustainability themes.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• LinkedInlinkedin.com/company/bergetai
• Xx.com/BergetAi87840

Compliance3 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles7 records

Funding overview

Funding rounds1 record

Investors3 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Berget AI