Skip to content
Clexerl

MissionClexerl Technologies

We start companies where the map ends.

A parent company
Four divisions · One bench
Columbia, Maryland

Old maps ran out of facts at the edges and drew sea monsters instead. That margin is where we work: too early for a category, too big for one product, too strange for a quarterly plan.

Earth from orbit
Fig. 1The edge of the known map, from four hundred kilometres up

Why we exist

01 StartTake early ambition seriously,02 Scalecarry it all the way to industry scale,03 Meanswith whatever technology the problem actually needs.

We judge ourselves by what is running, not by what is announced: a fleet on public roads, an agent that clears a week of admin, a model that still remembers last quarter.

Twenty-seven orders of magnitude.

The smallest and largest problems, with the same rigour. Scroll, or grab the ruler.

  1. 10−6m

    A single transistor

    A rounding error, a race condition, one bad lidar return. The smallest problems are where the large ones start.

    A circuit board up close

  2. 100m

    One person’s day

    The hour lost to scheduling. The task nobody owns. Agents do the coordination so people keep the judgement.

    A man sitting at a table talking on a phone

  3. 104m

    A city that moves

    Every street a machine has to cross safely, in the rain, at night, every time. Perception and control from the sensor up.

    Glass towers seen from below

  4. 107m

    A thinking planet

    Models that read text, images, audio and video as one picture, remember for months and act through tools.

    City lights across the Earth at night

  5. 1021m

    Whatever comes next

    If you can imagine it, we are probably already building it. If you can’t, we definitely are.

    A bubble nebula in deep space

What we believe

  1. 01

    Start from what is true

    Physics and economics first, category conventions never. Most of our best ideas sounded wrong for about a week.

  2. 02

    Same rigour at every scale

    A spreadsheet and a car crossing a city get the same care. We write the safety case and read the post-mortem, whatever the size.

  3. 03

    Nothing starts from zero

    Every venture inherits our engineers, capital and distribution on day one, and keeps the freedom to reshape itself as it learns.

  4. 04

    Measure, then move

    Speed is the point, so we instrument everything. If we cannot measure it, we have not finished building it.

One parent, four orbits

CLEXERL

Four bets, one parent

Each division owns a frontier and a colour. They share one bench, so what one learns the others use the same quarter.

Hover a planet to stop the system and read it.

How a venture is born

Step 1 of 4 · drag the handle

Prototype

A small team on the shared bench builds the first working version, usually in weeks, never in decks.

What we won’t do

Speed is the point, so the lines we will not cross for it are written down where anyone can check.

  1. 01Ship what we cannot measure.
  2. 02Race past a safety case to make a date.
  3. 03Sell the data people trust us with.
  4. 04Call it finished because the demo works.
  5. 05Hype what is not real yet.

Horizons

  1. Now

    Ship the first wave

    Clexerl Go out of private beta, the autonomy fleet on public roads in Maryland, memory research in production.

  2. Next

    Make it compound

    Every venture on one bench, so a breakthrough in one division shows up in the others within a quarter.

  3. Later

    Industry scale

    Ventures that change how a whole sector works, built by teams small enough to still argue in writing.

Bring your own
sea monsters.

We are hiring across every division, and we start new ventures with outside founders too.

See open roles Start something with us