Oryon
Portfolio indexActive · Seed · 2023
ActiveAI × crypto convergence

Prodia

CategoryAI · Infra
HeadquartersLondon
FoundersSneh Roy
— Overview

Prodia operates a GPU marketplace for AI inference with distributed supply. The platform lets developers access image generation, LLM inference, and multimodal model APIs at consistently lower cost than centralized cloud providers by aggregating idle GPU capacity from a global operator network.

The platform serves 12M+ image generations per day across an ecosystem of 40+ apps including gaming studios, creator tools, and consumer AI products. The enterprise tier launched in Q1 2026 targets model providers who need cost-efficient inference at scale.

Prodia's economic bet is that model providers will progressively disaggregate from compute providers, and that the marketplace layer between them becomes a durable business — similar to how AWS separated compute from application logic in the previous cloud cycle.

— By the numbers
12M
Images/day
3.2K
GPU nodes
80+
Models supported
2022
Founded
— Investment
Round
Seed 2023
Our role
Lead
Check size
$4M
Co-investors
Index Ventures · Framework Ventures
— Our thesis

GPU marketplace for AI inference with decentralized supply. We wrote the seed check on the thesis that model providers will disaggregate from compute providers.

— Team
SR

Sneh Roy

CEO & Founder
KP

Kavi Patel

CTO
AS

Anika Sharma

Head of Product
MR

Marco Ricci

Head of GPU Operations
— Timeline
  1. 2022Founded by Sneh Roy in London
  2. Q2 2023Oryon leads seed
  3. Q4 202410M daily image generations
  4. Q1 2026Enterprise tier launch
— Latest update

Serving 12M images/day across an ecosystem of 40+ apps. Enterprise tier launched Q1.

— In the press

Prodia is the AWS Lambda for AI inference — cheap, fast, and disaggregated from any one cloud.

The Information, 2025