Investor relations

AI creates intelligence.Who owns it?

Every conversation, code review, and decision your team makes with AI generates context - and today it evaporates when the session ends. Fathym builds cognitive development tools that compound that intelligence and keep it yours.

The thesis

Cloud. Data. Now intelligence.

Companies fought to own their cloud, then their data. AI starts the next fight: the intelligence your people create working with it.

Today that intelligence accumulates inside tools you don't own, and it leaks when a session ends, a model changes, or a teammate leaves.

The companies that control their AI context will compound intelligence. The companies that don't will leak it.

That is Intelligence Sovereignty - and it is the category Fathym is building.

The product family

One progression, increasing stickiness.

Personal

fAI

CLI workbench; vaults compound your AI context.Wedge. Free and open source. Developer adoption.
Visibility

Mosaic

The ambient cognitive interface; see what your AI knows.Retention. Team-level. Q3 preview.
Team

OpenX

The workshop where workbenches compose; connect and own your data, crystallize shared sources of truth, build apps and MCP.Scale. Per-seat. Team entry.
Vertical

OpenIndustrial

Industrial IoT + AI - telemetry, SCADA, workspace management.Verticalization. High ACV.

GTM tiers, not scale rungs - the architecture scale journey is Workbench -> Workshop -> Studio -> Platform; Mosaic and the Cognitive Loop run across all of it.

Positioning

Not another foundation-model provider.

Fathym doesn't compete with OpenAI, Anthropic, or Google on models. We compete on what happens after the model responds - context ownership, intelligence that compounds, architecture that composes.

Foundation-model providers
Sell model access
We build on top of any model
AI wrappers
Prompt templates
We compound intelligence, not prompts
Agent frameworks
Chain prompts
We compose workbenches with real architecture
Enterprise AI
Vertical AI features
Horizontal tools plus vertical packs

Why it matters

Your context is your IP.

Enterprises are realizing their AI context is proprietary intelligence. If that context lives in a vendor's cloud, they don't own it.

Fathym's vault architecture keeps context where the customer chooses - local in ~/.fai/ or in their own infrastructure. It's a buying criterion that didn't exist 18 months ago.

  • Context stays in the customer's infrastructure, not a vendor cloud
  • Vaults are owned files - portable across tools and models
  • MCP hosting means any AI client connects; never locked to one provider

The moat

Intelligence that compounds.

Evolve-forever isn't one mechanism - it's three compounding loops that feed each other.

Context

The knowledge vault gets smarter every session through synthesis.

Capability

The factory produces better capabilities from real usage and feedback.

Scale

Patterns discovered at team level flow back down as better individual defaults.

The market

Yes-and on Context Engineering.

Context Engineering is the emerging category around getting the right context to AI. We don't fight it - we occupy it, then go further: owning and compounding that context.

Gartner sizes Context Engineering at $6.27B in 2025, growing to $28.45B by 2030.

Gartner: Context Engineering $6.27B (2025) -> $28.45B (2030).

Traction

Shipping, not slideware.

  • Two products live: fAI (the CLI workbench) and OpenX (the team workshop).
  • 1,000+ tests passing across the stack.
  • Works with Claude, Copilot, Cursor, Codex, Windsurf, and Ollama - multi-model by design, not single-model routing.
  • EU AI Act enforcement lands August 2026 - context ownership shifts from preference to compliance need.

Team

Who's building it.

CEO

Matt Smith

Founder; leads vision, product, and the Intelligence Sovereignty thesis.
CTO

Michael Gearhardt

Leads architecture and engineering across the Fathym stack.

See the numbers.

The thesis is public. The model, projections, pricing, and named pipeline live in the data room.