Skip to main contentSkip to navigation
Veldarium
Work with us
The institution

America should be able to build difficult ships reliably.
Veldarium is building toward the shipbuilder that does.

Why does this company need to exist?

An American industrial company. The intent: build an American shipbuilder able to design, supply, build, inspect, integrate, and deliver difficult vessels reliably. The reason is national: skilled work, ships built right, and a fleet the country can build for itself. Earned in order: software, then physical capability, then vessels.

chris@veldarium.com What is proven today

Why the company exists, and why software first

Why it exists

The work stops where four departments meet.

Engineering releases the drawing, material receives the plate, production schedules the crew, inspection holds the weld. Each is right about its own fact; nobody holds the answer between them. A crew waits on a revision nobody sent, or welds to one that has moved. Veldarium starts there, on the floor.

Software over yard data is already being bought. In December 2025 the Navy announced a 448 million dollar investment in ShipOS, a program that applies Palantir’s software to shipbuilding data, run by its Maritime Industrial Base program with Naval Sea Systems Command. U.S. Navy, December 2025. Veldarium intends to go further: into the yard itself, its production cells and the automation that executes released work.

Why software first

The unit of improvement is the build, not the screen.

Governed build state is the one layer that can be proven before there is a yard: one partner’s work area, read-only, no steel at risk. Everything after acts on it. VeldBuild keeps build state governed. Bounded reasoning may propose; named people authorize. See the operating boundary.

What a modern shipbuilder has to be The long-form argument

The company's position

A company that starts in software says it will build ships. What does it propose to own?

Build the governing layer. Connect what already holds the facts. Defer the physical capacity until it is earned.

Another authority over engineering, material or inspection data would deepen the problem. Veldarium builds the layer between those systems and earns the rest.

From software state control to shipbuilding capability

The rungs in their capital phases, each opened only by evidence from the one beneath.

from the roadmap

Today proof level 3 of 8: synthetic system demonstration.

  1. Software foundation Target
    1. Rung 0Governed build state Synthetic
    2. Evidence stops hereRung 1Operational reasoning Target
  2. First deployments Target
    1. Rung 2Program-wide governed state Target
  3. Physical R&D Long-term
    1. Rung 3Bounded physical cell Long-term
    2. Rung 4Integrated production unit Long-term
  4. Production capability Long-term
    1. Rung 5Integrated shipbuilding capability Long-term
    2. Rung 6Vessels Long-term

Each rung, its evidence gate and how it fails

AuthorityWhat stays where it is
PLM, ERP, MES, QMS, CAD, scheduling and supplier systems keep their own facts. Veldarium works out how those facts relate and becomes the record for none of them.
When it failsWhat would make this dishonest
Describing a deferred capability as if it were current.
Physical consequenceWhat moving the line requires
Published proof from the rung beneath it. Capital or intent alone does not move it.

Who is building it

One founder, and the institution a shipbuilder needs.

1 of 37 seats filled today The founder’s. The founder came to this from operations work. The plan does not rest on one person: each rung names the seats and partners it needs. No openings are posted; the founder reads every introduction.

Christopher Azar-Brandes

Founder, Veldarium, in Connecticut, United States.

Legal entity
Veldarium Technology Systems LLC

How decisions are made

The founder holds every decision today. Each needs written evidence, fails closed without it, and moves to a named seat once that seat is filled.

The 5 decisions, what each requires, and the seat it moves to
  1. Publishing a claim
    Requires
    An entry in the ledger with evidence a reader can inspect, and a release gate that passes.
    Without it
    No evidence, no claim. A status is derived, so it cannot be raised by hand.
    Held by
    Founder today; moves to the quality seat
  2. Opening the next rung
    Requires
    The rung’s gate evidence, at the proof level the gate demands.
    Without it
    Until then the rung keeps the status its weakest stage gives it, whatever has been spent on it.
    Held by
    Founder today; moves to the program leadership seat
  3. Committing capital to a phase
    Requires
    The gate beneath the rungs that phase funds.
    Without it
    A phase waits on its gate. No amount is published for any phase.
    Held by
    Founder today; moves to the operations seat
  4. Holding a function in-house
    Requires
    An attempt on the rung from which the company has to hold that function itself.
    Without it
    Until then the function is reached through a partner, and the seat count changes only when a real hire is made.
    Held by
    Founder today; moves to the operations seat
  5. Reading a partner’s records
    Requires
    Written terms first: the work area, data rights, the baseline, and the success and stop conditions.
    Without it
    No terms, no access. A reached stop condition ends the work, and stopping is a recorded result.
    Held by
    Founder today; moves to the data and integration seat

The seats to fill, rung by rung

Today1 filled

  • Founder

  1. Rung 04 of 37

    Governed build state

    • Industrial softwareSoftware and reasoning
    • Data and integrationSoftware and reasoning
    • Configuration managementPrograms and engineering
    • Production control and planningProduction

  2. Rung 15 of 37

    Operational reasoning

    • AI and reasoningSoftware and reasoning

  3. Rung 27 of 37

    Program-wide governed state

    • Supply chainSupply
    • QualityQuality

  4. Rung 316 of 37

    Bounded physical cell

    • Manufacturing engineeringProduction
    • Welding engineeringPrograms and engineering
    • ControlsMachines and facilities
    • RoboticsMachines and facilities
    • SafetyLeadership
    • InspectionQuality
    • MetrologyQuality
    • Machine operatorsSkilled trades
    • MaintenanceMachines and facilities

  5. Rung 426 of 37

    Integrated production unit

    • Production engineeringProduction
    • Industrial engineeringProduction
    • Production supervisionProduction
    • Nondestructive testingQuality
    • WeldersSkilled trades
    • FittersSkilled trades
    • PipefittersSkilled trades
    • Robotics techniciansSkilled trades
    • Material and logisticsSupply
    • OperationsLeadership

  6. Rung 535 of 37

    Integrated shipbuilding capability

    • Naval architecturePrograms and engineering
    • Systems engineeringPrograms and engineering
    • Structural engineeringPrograms and engineering
    • Electrical engineeringPrograms and engineering
    • Supplier developmentSupply
    • ElectriciansSkilled trades
    • MachinistsSkilled trades
    • FacilitiesMachines and facilities
    • Program leadershipLeadership

  7. Rung 637 of 37

    Vessels

    • Vessel programsPrograms and engineering
    • CommissioningProduction
Each seat stands at the rung from which the company has to hold it itself. Until then a partner who already holds the function supplies it. Highlighted: the 4 the first gate needs.
Introduce yourself to the founder

Operating principles

The rules the software follows are the rules the company follows.

  1. 01Evidence before claim.

    In this companyEvery status on this site is computed from the capability ledger. The site’s release gate fails if any rung reads stronger than its weakest stage or the rung beneath it.

    Enforced by one guard in code.

  2. 02A named person decides.

    In this companyEvery public claim, control and proof has one named owner: the founder, not departments that do not exist yet. The software holds the same line: it offers approval only when every consulted domain is clear.

    Enforced by two guards in code.

  3. 03Fail closed.

    In this companyReadiness has a third answer, cannot establish, and it never reads as ready. The company holds itself to the same rule: no cycle-time figure until a physical unit is measured.

    Enforced by two guards in code.

  4. 04Say what would prove it wrong.

    In this companyEvery stage on Progress names the finding that would falsify it, and every reasoning recommendation names the gate that would overturn it.

    Enforced by two guards in code.

Why this shape can get there

A team built in the order the evidence asks for it.

  1. The first gate tests the whole thesis, cheaply.

    Rung 0 needs 4 of the 37 seats in-house and one partner’s work area, not a yard. If governed state cannot change a real release decision, the company learns it before buying any steel.

  2. No gate asks for less proof than the one beneath it.

    The gates climb from proof level 4 (real records, live decision) to level 8 (vessel delivery), each above the level 3 reached today. Capital phases fund the rungs in the same order: money follows evidence, not the calendar.

  3. The team arrives with the rung that needs it.

    7 seats are held before the first physical rung and 30 after it. Until a rung needs a function in-house, a partner who already holds it supplies it.

  4. Nothing is thrown away between rungs.

    9 of the 20 things the rungs stand on come from more than one rung down. A decision record with named owners, left by rung 0, still carries weight at rung 6.

  5. What this does not show.

    None of it is evidence that the plan will work. It is the order in which a wrong assumption is cheapest to find, with every gate published before it is attempted.

Each rung, what it stands on and how its gate is judged

The current state

What is proven, and what is not yet.

Today that means software: governance engines tested in Veldarium’s code, and public demonstrations on synthetic records. Nothing has run on a yard’s records; everything shown runs on one synthetic vessel.

Demonstrated, on synthetic records
  • Tested software for readiness, change impact, evidence and authority
  • A synthetic vessel and build hierarchy to run it on
  • Three answers only: ready, not ready, or cannot establish
Not yet
  • A run on a customer’s real records
  • A connector to any yard system
  • The standalone application
  • A team beyond the founder, a funding round, a customer or a facility

Check the workProgress Roadmap Trust Systems map