The Investment Thesis · In 10 Seconds
NOVOThe Airbnb of AI Inference
The Market
AI companies need ever more compute — yet capacity is fragmented across data centers, clouds and specialist providers.
The NOVO Model
NOVO becomes the Airbnb of AI Inference: the platform connects demand with global compute capacity and workloads are routed to the best execution option by cost, availability, latency, region and compliance.
NOVO does not build its own data centers. Saudi Arabia can serve as an anchor-supply market for structurally attractive energy economics; global partners provide latency, resilience and data residency.
Target: $0.49 per 1M processed tokens in the defined reference product and >50% long-term gross margin — after contractual validation of COGS.
The Problem · A Growing Market Without a Simple Access Layer
AI Needs More and More Compute — Procurement Remains Complicated
Agentic AI Increases Consumption
A conventional AI request may trigger one model call. An agent plans, uses tools, checks results and calls models repeatedly — turning one task into many inference steps.
Compute Capacity Is Highly Fragmented
Capacity sits across different clouds, data centers and inference providers — with different prices, regions, hardware classes and contracts.
The Enterprise Challenge
Enterprises do not want to manage this complexity themselves. They need reliable access, predictable capacity, data protection, billing and clear SLAs.
The Solution · One Access Layer Instead of Ten Infrastructure Contracts
NOVO Connects Available Compute Capacity with Growing AI Demand
Capacity originates
Data centers and compute partners provide available or reservable inference capacity.
NOVO aggregates
NOVO sources from multiple partners and unifies pricing, regions, availability and commercial terms in one platform.
Enterprise Access
AI companies get reliable access through NOVO instead of managing many separate infrastructure contracts.
$0.49
target price per 1M tokens
1 API
for multi-provider capacity, routing and enterprise policies
Asset-light
secure supply contractually instead of building data centers
Why Now · Three Developments Are Converging
The Right Time for a New Inference Layer
AI Usage Is Growing Rapidly
After the training boom, everyday AI usage is becoming a permanent infrastructure market. Every app, copilot and agent creates ongoing compute demand.
AI Agents Multiply Usage
AI agents do not execute just one request. They plan, check, use tools and call models repeatedly — sharply increasing inference volume.
New Capacity Is Emerging Worldwide
The GCC, Europe and specialist AI data centers are adding new capacity. Saudi Arabia offers attractive conditions for building and sourcing compute capacity.
The Economics · Simply Explained
Buy Efficiently.
Sell Competitively. Scale.
Procurement
Procurement target: no more than $0.20–0.25 total cost per 1M processed tokens for a defined reference case. The model uses $0.22. Important: this is not yet a confirmed supplier price and must be substantiated by concrete quotes and contracts.
Sales
Target selling price: $0.49 per 1M processed tokens in the reference product. Large enterprise customers with committed volumes could receive prices of roughly $0.39–0.44 depending on volume.
Margin
Target economics at scale: $0.49 revenue minus $0.22 total cost = $0.27 gross profit, or 55.1% gross margin. During the build-out phase, actual margin may be materially lower because of lower utilization and capacity reserved in advance.
The Moat · A Compound Moat, Not a Feature List
With Every Stage of Scale NOVO Becomes Harder to Replace
Supply + Energy
Anchor supply in energy-efficient regions plus multiple global providers. The advantage only counts once NOVO can contractually prove delivered power cost, $/GPU-hour and availability.
Routing Data + Enterprise Lock-in
Every production workload creates proprietary data across model × hardware × geography × price × latency × utilization. Integrations, policies, commitments and SLA history increase switching costs.
Financing + Commitments
Customer minimums, supplier financing and Ijara/infrastructure financing are designed to scale capacity without funding every incremental compute dollar with equity.
Compliance + Residency
Zero-persistent-retention, data residency and enterprise security are built as auditable architecture. Claims become proven only after technical validation and external review.
Market & Competition · Preview of the Competitive Landscape
Where NOVO Intends to Compete
NOVO target rate: $0.49 per 1M tokens. The detailed analysis benchmarks it against public reference pricing from Groq, Together AI and other providers.
Multiple compute sources are intended to form one available capacity layer, with routing across providers and regions.
One access point for capacity, routing, residency policies and SLA abstraction. Asset-light, without building owned data centers.
Inside the detailed analysis: hyperscalers, Together AI, Fireworks, Groq, DeepInfra and OpenRouter · price references · feature matrix · positioning logic
Open Detailed Competitive Analysis ↗Go-to-Market · From a Few Large Customers to Predictable Revenue
Not Millions of Small Customers — but Large AI Workloads
LAND
Target customers: AI-native companies, agentic-AI platforms and large enterprise AI workloads with high, recurring inference demand.
EXPAND
Annual minimum commitments make demand more predictable. As volume grows, additional models, regions and capacity tiers are added.
Start with a clearly defined, high-volume inference workload at AI-native and enterprise customers.
Multi-year minimum commitments and reserved capacity make demand and procurement more predictable.
Add models, regions and workloads to grow volume per customer — without acquiring millions of small users.
Go-to-market logic: NOVO optimizes for a small number of strategic enterprise customers with recurring compute demand, not small transactions. Contract value, expansion and long-term utilization of aggregated capacity drive scale.
Scale · The Path to a Platform Company
From the Inference Wedge to a Global Compute Marketplace
Production-ready core product, first supplier and customer contracts; management target: ~$3–5M annualized commitments.
Multiple infrastructure partners, enterprise security and repeatable sales. Management case: ~$15–25M revenue.
Management case: ~$50M+ revenue and >50% target gross margin as utilization grows.
Management case: ~$250M revenue, derived from growing customer count × average contract value × expansion into additional workloads/regions. To put potential enterprise value in context: at roughly $250M of annual revenue, an 8–15× revenue multiple would imply approximately $2.0–3.75B of enterprise value. This is a scenario, not a forecast.
The valuation is not a forecast. The multiple is a sensitivity, not a claim: it depends on growth, realized gross margin, retention, customer concentration, capital intensity and platform differentiation. The unicorn case exists only if these metrics are proven operationally.
Strategic Financing · Up to $35M · Milestone-Based
One Strategic Capital Partner. Up to $35M Released Against Clear Proof Points.
Capital Framework at Full Deployment
The NOVO Principle
The up-to-$35M amount is a total financing framework, not a day-one payment. Capital is intended to be released in tranches only after pre-agreed, objectively measurable milestones are achieved. The allocation shown represents the target framework at full deployment; as demand grows, larger compute blocks should increasingly be matched with customer commitments, supplier financing and specialized infrastructure financing. The financing is designed as a long-term partnership with one strategic capital partner. Following successful expansion, that partner can provide additional growth capital to support NOVO through global scale and minimize the need for additional equity investors.
2+ independent capacity sources, a defensible COGS bridge and reproducible production benchmarks.
Pilots, LOIs/MSAs and first annualized customer commitments establish the basis for the next release.
Further capital funds capacity and go-to-market only after demand, economics and operational delivery are validated.
NOVO Is Not Building the Next Data Center.
NOVO Is Building the Airbnb of AI Inference.
Own the Customer Relationship. Aggregate the Infrastructure.