SUMO-TX

CPU-Distributed AI Execution Platform — Governed, Multi-Node AI Workloads Under Your Control

Raising £3M Seed Round

Confidential — For discussion purposes only  ·  Q2 2026

What is SUMO-TX?

SUMO-TX is not a GPU training platform or hosted AI wrapper. It is a CPU-distributed, multi-node AI execution platform built natively for Microsoft Azure — enabling organisations to run governed, auditable, and cost-controlled AI workloads on their own infrastructure, with full sovereignty over data, execution, and outcomes.

🔀

Multi-Model Orchestration

One subscription unlocks 50+ Azure AI Foundry models. Run GPT-4o, Llama, Mistral, Phi-4, Sora, and more — in parallel, side-by-side.

📊

Benchmarking-as-a-Service

Independent, reproducible model evaluation across accuracy, latency, cost, hallucination, and safety — without relying on vendor self-reporting.

🧠

SUMO-TX Execution Platform

CPU-distributed, multi-node execution layer with deterministic orchestration, audit traceability, and sovereign deployment — without vendor lock-in.

Platform in Action

Beta screenshots — Q2 2026

SUMOTX One-Click Azure Deployment Dashboard
Distributed Execution Dashboard — orchestrate AI workloads across multiple CPU nodes with explicit failure isolation and backpressure control on Microsoft Azure
SUMOTX Orchestration & Audit Traceability
Orchestration & Audit Traceability — deterministic task graph execution with full traceability, compliance replay, and controlled workflow progression

Six-Layer Platform Architecture

6
Governance & Control Cost · SLA · Policy · Compliance  |  4–8 nodes
5
State, Memory & Audit Event Logs · Snapshots · Traceability  |  Sharded across nodes
4
Execution Worker Pools CPU-first · Task execution · Model calls  |  Up to 32 nodes per pool
3
Distributed Message Bus Event streams · Queues · Task routing  |  32+ nodes (horizontal scale)
2
API-First Control Layer Versioned OpenAPI (Swagger) interface — automation, integration, and custom enterprise tooling
1
Microsoft Azure Foundation Azure Batch · Azure Storage · Azure Identity · CPU Compute · Networking

Subscription Tiers

Tier Monthly Price Key Capability
Starter £5 Any model from Azure AI Foundry catalogue; 100K tokens/month
Pro £19 All 50+ models; run up to 3 simultaneously; vision, audio, agents
Unlimited £39 High-volume access; up to 10 parallel models; video (Sora); SUMOTX reports
Enterprise Custom B2B Benchmarking API; private deployment; SSO; EU AI Act audit trail

The Benchmark Flywheel — Our Defensible Moat

Investors fund flywheels. This is ours.

1
Users query multiple models

Every subscriber request runs through the orchestration engine across foundation models

2
SUMOTX provisions distributed execution infrastructure

CPU node pools, storage, networking, and identity are configured within the customer's Azure subscription

3
Best answers become structured dataset

High-quality (task, winning-response, reasoning trace) tuples are curated

4
Execution jobs run across distributed CPU nodes

Repeatable, deterministic workload execution with full visibility into outputs and audit traces

5
Better routing attracts more users

Trained models are registered and promoted across environments → enterprise trust → more subscribers

↺ Repeats — compounding advantage
No competitor can replicate this loop without the same user base. Every deployment deepens SUMOTX's sovereign AI capabilities and strengthens customer lock-in on their own infrastructure.

Market Opportunity

Segment Size (2025) CAGR
Global AI/ML Platform Market ~£40 B 38%
LLM API & Inference Market ~£9.5 B 62%
AI Benchmarking & Evaluation Tools ~£2 B 55%
Enterprise AI Safety & Compliance ~£3 B 48%
SUMOTX SAM (sovereign AI infrastructure + training + operations) ~£8 B 50%
SUMOTX SOM (Year 3 target) ~£400 M

UK government and EU AI Act compliance requirements are generating urgent, funded demand for sovereign AI infrastructure — precisely SUMOTX's deployment and training layer. SUMOTX is the platform positioned to fill the UK's sovereign AI infrastructure gap.

Five-Stream Revenue Architecture

💳

B2C SaaS Subscriptions

Recurring monthly revenue from developers, researchers, and teams across Starter, Pro, and Unlimited tiers.

~68% gross margin
🏢

Enterprise Licensing

Multi-seat agreements with custom SLAs, private deployment, and priority engineering support.

~75% gross margin

Metered Overage Usage

Pay-as-you-go overage charges capturing high-usage power users beyond their monthly tier.

~65% gross margin
📦

Benchmark Dataset Licensing

Evaluation datasets and comparative benchmark studies sold to enterprises, labs, and government AI teams.

~85% gross margin
🛒

Model Marketplace Commissions

Revenue share from third-party fine-tuned models listed on the SUMOTX platform.

~90% gross margin
📈

Unit Economics

  • Blended B2C ARPU: £22/mo
  • Enterprise ACV: £15K/yr
  • Blended Gross Margin: ~72%
  • CAC (PLG): ~£40
  • LTV (18-mo avg.): ~£396
  • LTV : CAC: 9.9×

Financial Projections

Year B2C Subscribers B2B Contracts ARR Gross Margin EBITDA
2026 2K 5 £520K 65% –£1.4M
2027 15K 25 £4.1M 69% –£0.6M
2028 80K 80 £21M 71% £3M
2029 280K 200 £74M 73% £19M
2030 850K 450 £210M 75% £62M

Milestone Roadmap

Q1 2026 ✅
Platform architecture live on Azure AI Foundry; multi-model routing implemented
Q2 2026
Public beta; parallel model response UI launched; benchmark v0.1 published
Q3 2026
Paid tiers active (£5–£39); multimodal (vision, audio) available on Pro+
Q4 2026
Video model access (Sora) live on Unlimited; B2B benchmarking pilots begin
Q1 2027
10K paying subscribers; first enterprise benchmarking contracts signed
Q3 2027
SUMOTX v1.0 — first sovereign model training pipeline complete
Q4 2027
50K subscribers; B2B benchmarking ARR > £1M; Series A close
2028
SUMOTX v2.0; multi-cloud; £20M ARR target

£3M Seed Round — Use of Funds

An infrastructure and engineering raise — not a SaaS marketing budget. Capital funds the distributed CPU compute infrastructure and engineering talent to build the SUMOTX execution platform and sovereign AI governance layer.

35%
Azure Compute Infrastructure — SUMOTX distributed execution platform (CPU node pools, storage, networking) £1.05M
30%
Engineering Hiring — ML engineers, backend, ML research (×4) £0.90M
20%
Go-to-Market — B2B sales and developer growth £0.60M
10%
Operations & Compliance — Legal, SOC 2, EU AI Act readiness £0.30M
5%
Reserve £0.15M
Runway: 18–24 months to Series A milestone: 10K paying subscribers, SUMOTX v1.0 complete, 10+ B2B benchmarking contracts signed.

Ready to Discuss the Investment Opportunity?

Contact OrkinosAI to learn more about SUMOTX and the seed round.