Total cost of ownership

Where AI cost comes from.

The model is one part. The machine, the people who run it, the data and the record make up the rest. With Core, one partnership carries most of it, for a set-up fee and a monthly fee.

An illustration of how each kind of AI cost grows as a firm adds people and asks more: per-seat tools step up with every hire, per-token APIs rise with every question, and Core is a set-up fee then a flat monthly fee, with questions on the firm’s own machine not metered.

The five parts of the cost.

  • The machine

    What it takes
    A GPU machine or server, kept running and patched
    With Core
    A machine you own, or a DGX Spark we supply. We keep Core running on it
  • The model

    What it takes
    Chosen, served and replaced as better ones come out
    With Core
    We serve it, and move your Core to better models with you
  • The people

    What it takes
    Engineers for models, infrastructure, data and security
    With Core
    Our engineers, as part of your partnership
  • The data

    What it takes
    Filings, prices, notes and email, connected and kept flowing
    With Core
    Connected at set-up, and kept flowing
  • The record

    What it takes
    An answer log and an inspection pack a regulator can read
    With Core
    Built in, on your machine

Three ways to bring AI to your desk.

  • Runs on a machine you own

    Vibe-coded
    Partly
    Off-the-shelf
    No
    Core
    Yes
  • Every figure checked against its source

    Vibe-coded
    No
    Off-the-shelf
    Partly
    Core
    Yes
  • An answer log and inspection pack for your regulator

    Vibe-coded
    No
    Off-the-shelf
    Partly
    Core
    Yes
  • Mapped to the SFC, MAS and FSC rules on AI

    Vibe-coded
    No
    Off-the-shelf
    Partly
    Core
    Yes
  • Set up for your kind of firm

    Vibe-coded
    Partly
    Off-the-shelf
    Partly
    Core
    Yes
  • Your choice of model, open or under your agreement

    Vibe-coded
    Partly
    Off-the-shelf
    No
    Core
    Yes
  • Not priced per seat

    Vibe-coded
    Yes
    Off-the-shelf
    No
    Core
    Yes
  • No team to hire

    Vibe-coded
    Yes
    Off-the-shelf
    Yes
    Core
    Yes
  • Engineers who keep it running and move it to better models

    Vibe-coded
    No
    Off-the-shelf
    Partly
    Core
    Yes
  • Keeps working when the person who built it leaves

    Vibe-coded
    No
    Off-the-shelf
    Yes
    Core
    Yes

What can still change the bill.

Cloud models
If you use Claude, GPT or Gemini, the provider bills its tokens under your own agreement
Hardware
A machine you already own costs nothing more. A DGX Spark we supply is quoted in your proposal
Add-ons
API and MCP access, and AlphaLenz market data, if you choose them
Another Core
A second office or a new kind of work on its own machine is scoped as its own Core

Questions.

Why don’t you publish prices?

A desk of five and an enterprise with many offices need different things from Core. Each Core is scoped to the firm, and you get a set-up fee and a monthly fee in writing after a demo.

Does usage change what we pay?

Not on your own machine: Core is not priced per seat, and questions are not metered. With a cloud model, the provider bills its tokens under your own agreement.

What does the hardware cost?

Nothing more if Core runs on a machine you already own. A DGX Spark we supply is quoted in your proposal.

Who owns the hardware?

You do, whether it is a machine you already have or a DGX Spark we supply. In a private cloud, we run a server for your firm alone.

Explore more of Core.

Get a proposal.

Book a 30-minute demo. We scope your Core with you and send the proposal in writing.