Skip to content

Programs

Your constraint in. A working solution out, measured.

Programs run on the platform. Most begin as a software answer your team builds on the same week; some run through a partner bench to a proven sample, and some all the way to a manufactured part in your device.

Your constraintThe searchThe benchThe measurementThe part
One program, on a loop: the constraint goes in, the search fans and closes, the bench makes it, the measurement pulls the curve it is judged on, and the part comes out.

01 · Where devices fail

Your device fails somewhere. Tell us where.

A device stops somewhere: at the field, at an interface, in a process, in the cold, on cost, at length, or somewhere we have not seen yet. Below are the six we see most, and what a program does about each. If yours is not there, send it anyway; that is the one we want.

Where a program stops

What a program does about it·The field

What the search decides
The conductor and the pinning that hold Jc at your field and temperature, ranked against the margin you named.
What the bench proves
A V–I curve at field and temperature on a short sample, on a rig named in the brief before it runs.
What you get
A route with Jc at your operating point on its first row, measured, and the margin under it.

What they look like as briefs

The bond

A superconducting film that has to survive a semiconductor process without losing what makes it work.

The wire

Leads that carry signal into a cryostat without carrying heat in with them.

The stack

A multilayer that works everywhere except where two of its layers meet.

The window

The deposition and anneal range a fab can actually hold, found before the runs are spent.

The recipe

A known material that needs one property moved: warmer, stronger in field, or cheaper per metre.

The length

A conductor that meets the spec in a short sample and has to meet it in a kilometre.

Schematic. In each plate, the heavy line is the part the brief is about.

02 · How a program runs

Nothing is paid for until both sides have read the same brief.

Two unpaid steps, then a defined program. Each stage after that ends at a review gate agreed before the first one starts, so a program that ought to stop is able to.

01No fee

A conversation

You describe the step you are stuck on. We say whether the model has anything to offer it, and where we would expect it to stall.

02No fee

A written brief

The constraint in numbers: field, temperature, current, the process it survives, the length, the ceiling. Both sides sign it, so the program can be judged against it.

03

The search

Every route that clears the brief, ranked, with the numbers under each and the trade-offs on the rest. Ours end to end, in about a week.

04

The bench

The top route made and measured by a partner named in the brief, against the spec, with the result coming home to the model whether it cleared or not.

Why there is no single number

The search is ours and takes about a week. The bench half runs on real furnaces, so we book them before the brief is signed, and we are moving synthesis in-house at Berkeley to shorten it. We publish the stages, not a total, because a total would hide which half moved.

03 · The benches

Run it yourself, or run it with us.

Your lab: the record, your bench, and our people on the call. The result is one your team integrates into the system it already runs.

Our network: contracted capacity your program buys time on, groups that publish under their own names, and the vendors, instrumentation partners and sourcing around them, scheduled by us. A vendor we already buy from is a month you do not wait; a rig we already book is a quarter. A two-person team gets a fab's cadence; a large lab plugs us into the bench it already runs.

The network around the benches

Procurement and sourcing

Targets, powders, tape and substrates, from vendors we already buy from. Weeks to the first run, not a quarter of purchase orders.

Instrumentation partners

The rigs a run needs and nobody on the program owns, including second-hand instrumentation groups. A rig that exists is a build you do not fund.

Scheduling

Every bench in the brief on one calendar, run by us, so a delay is seen the day it happens and not at the review.

UC Berkeley

Synthesis moving in-house at Berkeley, with the labs around it a call away, so a synthesis run is a booking rather than a queue.

A program names its benches in the brief, before anything is made, so nobody discovers afterwards which instrument a number came off.

04 · What a program looks like

Ask a bounded question. Get an answer from the software. Build on it.

Most programs are the first column. Some go further, and the brief says which before anything is paid for.

The recordThe sampleThe part010203

01

A software answer

The record: every route that clears it, ranked, in about a week.

What you ask, and what happens then
You ask
A bounded question about the stack you run, in numbers.
Then
Your team builds on it, or the brief goes one column right.
  • Fixed fee, or milestones with a review gate on each
  • Customer-funded experimental work
  • Prepaid validation credits
  • IP terms agreed before the first run

02

A pilot

The record, plus one bench run: a measured sample against your spec.

What you ask, and what happens then
You ask
The same brief, and a bench named in it.
Then
You integrate a proven result, or we run again.
  • Experimental credits
  • Reserved capacity held against your schedule
  • A program-management fee, and a premium for accelerated turnaround

03

Co-development

The route, the sample and the part, with the people who build it.

What you ask, and what happens then
You ask
A part your device needs and nobody makes.
Then
The part ships into your device, on the schedule the brief sets.
  • Upfront engineering fees
  • Technical and commercial milestones
  • Field-of-use options
  • Manufacturing-readiness and yield tooling
None of this is a rate card. What is written here is the shape an agreement takes, and the numbers in one are set per program.

Internal today, enterprise seats next. Ask a bounded question and the answer comes off the platform we run ourselves; the seat your team runs it from is the next release. Until then we are solving problems and building relationships, one brief at a time.

Some programs cost nothing.

If you see the same opportunity we do, we run it at no cost. Ask.

Everything on this page runs on superconductors that already exist. We chose to start there, and the search runs alongside it.

05 · IP and data

Agreed before the first run, not after the result.

Yours

The designs, materials and confidential inputs you bring.

Ours

The platform, the models and the physics under them.

What the program creates

Defined before work begins: use rights, field exclusivity, joint ownership or a licence.

What the platform learns

De-identified learning improves the model where the agreement permits it.

One constraint is enough to start, and the first two steps cost nothing.

Tell us what your device has to do.

Superconductors are the future.