Koyeb
Koyeb is a French serverless cloud platform enabling developers to deploy and scale AI inference, agents, APIs, and databases across GPUs and CPUs in 50+ locations; acquired by Mistral AI in February 2026 to anchor Mistral Compute.
- Company typePrivate
- Founded2021
- HeadquartersParis, France
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Koyeb does
Koyeb is a French serverless cloud infrastructure company founded in 2020–2021 by three ex-Scaleway engineers (Yann Léger, Edouard Bonlieu, Bastien Chatelard) and headquartered in Boulogne-Billancourt, France. It operates a high-performance serverless platform built on Firecracker MicroVMs, Kubernetes/Nomad orchestration, and a Kuma/Envoy service mesh with Anycast BGP and mTLS networking, exposing compute across CPUs, NVIDIA/AMD/Intel GPUs, and accelerators across roughly 50 locations and 10 datacenters on three continents. Core products include Serverless GPUs (supporting H100, A100, H200, B200, L40S, L4, RTX-A6000, and RTX Pro 6000), Koyeb Sandboxes for isolated AI agent code execution, Serverless Postgres with pgvector, Koyeb Workers for background jobs, and infrastructure-as-code providers for Terraform and Pulumi.
Koyeb serves AI/ML teams, software developers, and startups through a self-serve, product-led growth motion with no traditional sales force. Revenue is generated via usage-based compute pricing (CPU from $0.0022/hr, GPUs from $0.50–$1.60/hr) layered on tiered subscription plans (Pro, Scale, Enterprise), a freemium startup credit program (up to $30k), and managed services for Serverless Postgres. The company has raised approximately $8.6M in total funding, including a $7M seed round in November 2023 led by Serena with participation from ISAI, Samsung Next, and MongoDB. In February 2026, Koyeb entered a definitive agreement to be acquired by Mistral AI (valued at $13.8B), with the platform set to become the core component of Mistral Compute and the entire 16-person team joining Mistral's engineering division.
Koyeb firmographics
Firmographics- Name
- Koyeb
- Legal name
- Koyeb SAS
- Website
- https://koyeb.com
- Company type
- Private
- Founded year
- 2021
- Operating status
- Acquired
- Headcount range
- 11–50 employees
- Short description
- Koyeb is a French serverless cloud platform enabling developers to deploy and scale AI inference, agents, APIs, and databases across GPUs and CPUs in 50+ locations; acquired by Mistral AI in February 2026 to anchor Mistral Compute.
- Ownership category
- akta.pro rank
Koyeb industry classification
Industry- Product category
- Cloud Infrastructure
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Computer Systems Design and Related Services (54151)
- SIC
- Services-Computer Integrated Systems Design (7373), Services-Computer Programming, Data Processing, Etc. (7370)
- akta.pro primary industry
- Hybrid Cloud Compute & Virtualization Stack (HCI/VMs/Kubernetes) (HDABAMAF)
- akta.pro secondary industries
- Container Platforms & Orchestration (Kubernetes) (HDABADAF), Cloud-Native Application Development (Containers/Kubernetes/Serverless) (BPAEAFAE), AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers) (HDAAAAAG)
Keywords
Where Koyeb is headquartered
LocationHeadquarters
- HQ city
- Paris
- HQ country
- France
- HQ region
- Europe
Offices2 records
Markets served
Koyeb business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Infrastructure, Technology or R&D, Personnel, Marketing or Sales, Operations
Revenue model
- Compute Infrastructure (GPU/CPU): Usage-based pricing for GPU and CPU compute resources. GPU instances range from $0.50/hr (RTX-4000) to $1.60/hr (A100). CPU resources start at $0.0022/hr. Customers pay for actual compute consumption with autoscaling.
- Platform Subscription Plans: Tiered subscription plans (Starter, Pro, Scale, Enterprise) with varying support levels, feature access, and team member limits. Annual and monthly billing options available.
- Managed Database Services: Serverless Postgres with pgvector for AI applications, charged based on usage and storage consumption.
- Startup Program Credits: Early-stage startups can apply for up to $30k in credits to accelerate go-to-market with high-performance cloud infrastructure.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Subscription | Monthly | Starter Plan (phasing out) |
| Subscription | Monthly | Pro, Scale, Enterprise Plans |
| Usage-based | Pay-as-you-go | GPU Compute - Usage-based |
| Usage-based | Pay-as-you-go | CPU Compute - Usage-based |
| Freemium | Pay-as-you-go | Startup Credits Program |
Go-to-market motion3 records
Distribution channels4 records
Marketing channels7 records
Koyeb product offering
Product offeringCore offering
Koyeb operates a high-performance serverless cloud platform that enables developers and AI teams to deploy, run, and scale compute-intensive applications across CPUs, GPUs, and accelerators in 50+ global locations. The platform supports AI inference, model fine-tuning, AI agents, APIs, web applications, databases, and background workers through a unified runtime with autoscaling, scale-to-zero, and sub-200ms cold starts.
Product overview
Koyeb is a global, high-performance serverless cloud platform for deploying AI-intensive applications. The core offering is the Koyeb Serverless Platform, a unified serverless runtime that provides infrastructure for AI inference, APIs, inference endpoints, agents, and databases. The platform is built around GPU compute (via the Serverless GPUs sub-product), which supports NVIDIA GPUs ranging from RTX 4000 to H200 and B200. Surrounding the core platform are specialized sub-products: Koyeb Sandboxes for isolated AI agent code execution, Serverless Postgres with pgvector for AI-ready database workloads, and Koyeb Workers for background job processing. Infrastructure automation is handled through the Koyeb Terraform Provider and Koyeb Pulumi Provider. The Koyeb MCP Server enables coding agents to manage Koyeb resources programmatically via natural language. The Koyeb Startup Program provides up to $30k in credits as a go-to-market add-on for early-stage companies. The platform operates across 10 global regions on 3 continents and was acquired by Mistral AI in February 2026 to form the core of Mistral Compute.
Differentiator
Problem solved
Functional benefit
Products and services
- Koyeb Serverless Platform A unified serverless runtime that lets developers deploy APIs, inference endpoints, AI agents, workers, and databases globally via Git push, CLI, or Docker, with autoscaling and scale-to-zero across 50+ locations.
- Serverless GPUs On-demand GPU compute supporting NVIDIA H100, H200, B200, A100, L40S, L4, RTX-A6000, RTX Pro 6000, and AMD/Intel accelerators, billed per second with autoscaling and scale-to-zero for AI inference, fine-tuning, and training workloads.
- Koyeb Sandboxes Fast, scalable, fully isolated execution environments for AI agents that provide secure, sandboxed code execution with JavaScript and Python SDKs for agent fleet management.
- Serverless Postgres A fully managed Postgres database service with pgvector for embedding storage and vector search, designed for large-scale and AI workloads including RAG pipelines.
- Koyeb Workers Background job processing capability that enables deployment and scaling of high-performance worker processes for asynchronous task execution, job queues, and distributed system workloads.
Quantifiable outcome
- 80% cost savings compared to hyperscalers for GPU workloads
- +3 more outcomes
Companies that use Koyeb
Customer profileNamed customers9 records
Segments3 records
Ideal customer profiles3 records
Koyeb technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration11 records
AI capability6 records
Feature7 records
Koyeb partnerships and signals
Strategic signalPartnerships
Five partnerships are on record, tiered flagship and core.
- Mistral AIflagshipKoyeb entered definitive agreement to be acquired by Mistral AI. Koyeb's serverless platform, technology, and team will accelerate Mistral Compute offering. 16-person team including 3 co-founders joining Mistral's engineering division. Platform to become core component of Mistral Compute.
- Pruna AIcorePartnership where Pruna AI's optimization engine integrates with Koyeb's serverless GPU platform to achieve 5x faster inference speeds for AI workloads.
- VultrcoreStrategic partnership enabling teams to deploy serverless AI workloads across Vultr's extensive network of 32 global regions. Combines Koyeb's serverless platform with Vultr's infrastructure.
- NVIDIAcoreNVIDIA GPUs (H100, A100, H200, B200, L40S, L4, RTX-A6000, RTX Pro 6000) available on Koyeb platform. Featured at NVIDIA GTC events.
- TenstorrentcoreWorld-premiere availability of Tenstorrent n300s and Wormhole instances via Koyeb Serverless Platform for AI accelerator workloads.
Scale indicators8 records
Recent moves6 records
Expansion highlights5 records
Koyeb competitors and assessment
Company assessmentDirect peers
- Replicate: Replicate runs machine learning models in the cloud via a simple API, focusing on serverless inference for open-source and custom models. It targets the same AI developer audience as Koyeb's Serverless GPUs.
- Modal Labs: Modal provides serverless cloud infrastructure for AI and data workloads, offering on-demand GPU compute, autoscaling, and a Python-native developer experience. It is the most direct competitor to Koyeb's serverless GPU and inference positioning.
- RunPod: RunPod offers on-demand GPU cloud compute for AI training and inference with serverless endpoints, container-based deployments, and global regions. It competes head-to-head with Koyeb for AI developers seeking flexible GPU access.
- Fly.io: Fly.io is a developer platform for running full-stack apps and APIs close to users with global deployment and autoscaling. It shares Koyeb's PLG, multi-region, container-based deployment approach though without GPU focus.
- Together AI: Together AI provides a cloud platform for generative AI workloads with serverless inference endpoints and dedicated GPU clusters. It overlaps with Koyeb on AI inference infrastructure serving AI/ML teams.
- Railway: Railway is an application platform that lets developers deploy code and databases with minimal configuration, autoscaling, and a developer-friendly experience. It competes with Koyeb for the same PLG developer audience building APIs and apps.
Emerging players
- Vast.ai: Vast.ai operates a marketplace for GPU compute across a distributed network of providers, offering lower-cost GPU instances for AI workloads. It overlaps with Koyeb's cost-competitive GPU positioning though via a marketplace model.
Broad incumbents
- AWS Lambda: AWS Lambda is the dominant hyperscaler serverless compute service, now extending into container and GPU-adjacent workloads. It is a broad incumbent that competes with Koyeb for serverless application deployments and increasingly for AI inference.
- CoreWeave: CoreWeave is a large-scale GPU cloud provider offering dedicated clusters and Kubernetes-based infrastructure for AI workloads. It competes with Koyeb in GPU compute but focuses on larger dedicated deployments rather than serverless.
- Lambda Labs: Lambda provides GPU cloud instances, clusters, and on-demand servers for AI training and inference. It is a broader GPU cloud incumbent with overlapping but more traditional infrastructure offerings versus Koyeb's serverless model.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights7 records
Customer concentration
Koyeb social profiles
Digital presenceKoyeb financial estimates
Financial estimateRevenue estimate
Valuation estimate
Koyeb leadership team
Management profileNumber of profiles
Profiles3 records
Koyeb funding detail
Funding detailFunding overview
Funding rounds3 records
Investors9 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Koyeb M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Koyeb
What does Koyeb do?
Koyeb operates a high-performance serverless cloud platform that enables developers and AI teams to deploy, run, and scale compute-intensive applications across CPUs, GPUs, and accelerators in 50+ global locations. The platform supports AI inference, model fine-tuning, AI agents, APIs, web applications, databases, and background workers through a unified runtime with autoscaling, scale-to-zero, and sub-200ms cold starts.
Is Koyeb a public or private company?
Koyeb is a private company. It is classified as corporate owned and is currently acquired.
When was Koyeb founded?
Koyeb was founded in 2021. It employs 11 to 50 people.
Where is Koyeb based?
Koyeb is headquartered in Paris, France, in the Europe region.
How does Koyeb make money?
Four revenue lines are on record. Compute Infrastructure (GPU/CPU) is the primary driver. The others are platform Subscription Plans, managed Database Services and startup Program Credits.
Who are Koyeb's main competitors?
Direct peers on record are Replicate, Modal Labs, RunPod, Fly.io, Together AI and Railway. Vast.ai is listed as an emerging player. Broad incumbents are AWS Lambda, CoreWeave and Lambda Labs.
Does Koyeb have an API?
Yes. Koyeb provides a public API enabling developers to automate resource management, deploy applications, and interact with Koyeb programmatically. The API is documented at /docs/api and supports all platform features including deployments, instances, apps, services, regions, scaling, volumes, secrets, and more. The Koyeb MCP Server is also available, allowing coding agents and AI assistants to interact with Koyeb resources in natural language via the Model Context Protocol. Developer documentation is at www.koyeb.com/docs/api.
What industry is Koyeb in?
Koyeb's product category is Cloud Infrastructure. Its primary akta.pro industry code is HDABAMAF, Hybrid Cloud Compute & Virtualization Stack (HCI/VMs/Kubernetes), with a secondary code of HDABADAF, Container Platforms & Orchestration (Kubernetes). Its NAICS code is 5182 and its SIC code is 7373.