Article
The first thirty opportunities from across the national lab ecosystem
Written by the COMPASS Team on July 31, 2026 · Updated September 29, 2026
COMPASS finds venture-scale technology inside America's national laboratories and builds the path from lab bench to funded company. This July we held Waypoints, our first lab-investor summit, at SLAC, where investors spent a day with the batch and the scientists behind it. More than 40 venture firms reviewed the opportunities.
The thirty opportunities below are what they reviewed: our first batch. Each one addresses a specific, expensive customer problem through work developed at a national laboratory.
Opportunities 01–04
The limits on computing and autonomy are increasingly physical: the energy a chip burns, the heat a package sheds, and the signal a machine loses underground. These four opportunities work on those limits.
Simulates the interfaces inside 3D chip packages that today's tools can only guess at.
Advanced chip packages fail at interfaces that conventional simulators cannot model, so packaging teams find out by building prototypes, and every failed build costs months. This software predicts how heat and current move through a specific material stack before anything is fabricated, and runs inside the simulation tools engineers already use.
Frontier-model inference at a fraction of the power
Power is now the limit on how fast data centers can add inference capacity, and much of a chip’s energy goes to moving model data. This venture is building an inference chip that spends far less energy moving data and is designed for high-volume manufacturing. Its first chips already match the accuracy of conventional digital hardware, and the next generation is designed to run a full frontier model on a single module.
A computing architecture that runs Super Intelligence on the device itself
Super Intelligence is moving into robots, drones, wearables, and sensors, and today’s chips use too much energy to run capable models on them. This venture is building a new computing architecture designed to run machine-learning models at a fraction of the energy of conventional chips, for device makers to build into their own chips.
A radio that carries signal, timing, and position through barriers that stop GPS and WiFi.
Underground, underwater, and inside large structures, autonomous systems lose contact the moment they lose signal. This is a compact radio system that keeps connectivity, precise timing, and positioning running through the barriers that block other wireless technology.
Opportunities 05–15
Better measurement is often what stands between a market and higher yield, safer operations, or autonomy. These eleven opportunities serve semiconductors, space, defense, medicine, and industrial monitoring.
A camera that sees through the welding arc and catches defects as they form.
Robotic cells lay millions of welds a day, and defects still surface only at inspection because conventional cameras cannot see through the glare of the arc. This system watches the weld as it forms and lets the cell correct a defect in the moment rather than after the part is finished.
Catches the infrastructure failure that has never happened before, with no training data.
Machine-learning monitors need examples of past failures, and in critical infrastructure the failure that matters most is usually the one with no history. This method finds it without labeled data, and has been validated across large scientific facilities and industrial machining.
A container wall that records any break-in, permanently.
Commercial tamper seals are routinely defeated in testing. This technology builds the evidence into the container material itself: a breach anywhere leaves a permanent visible record, and each unit carries a signature that cannot be swapped for a copy.
Software that lets a quantum sensor keep working through jamming and radio clutter.
Quantum sensors are extremely sensitive to their targets, and to everything else around them. This software lets a sensor pick out its signal under interference much stronger than the signal itself, with no change to the hardware.
A quantum magnetometer rugged enough to leave the lab.
The most sensitive magnetometers only work in shielded, vibration-free rooms, while navigation without GPS and monitoring at nuclear sites need them in the field. This sensor cancels vibration, drift, and interference in its own design instead of relying on shielding.
A laser for brighter EUV sources and compact accelerators.
Fiber lasers are the most efficient way to turn electricity into light, but they have not been able to deliver the very short, high-power pulses that EUV sources and compact accelerators need. This approach makes those pulses from fiber, at kilowatt power levels.
Catches slow drift in fab equipment before wafers are scrapped.
Fabs watch thousands of sensor streams with models that miss slow drift and have to be retrained whenever conditions change. This software catches drift early and runs inside the monitoring platforms fabs already use.
A security ink that authenticates the medicine itself, not the barcode on the box.
Conventional security inks are among the most counterfeited features in circulation. This ink prints several independent authentication checks into a single pigment, so a forger cannot pass it by matching its color. A major security-ink supplier is evaluating it.
A flight-proven sensor that measures the space environment from on board the spacecraft.
When a solar storm swells the upper atmosphere, drag changes and orbit predictions fall apart, just as low orbit is getting crowded. This small instrument has already flown in orbit and measures conditions from on board each spacecraft. Every unit in service adds to a live picture of the space environment.
Fast enough that fabs can measure every wafer at line speed.
The detectors inside many vacuum instruments have been stuck at the same speed limit for decades, so fabs measure samples instead of every wafer. This detector runs far faster without losing timing precision, and its first instrument targets in-line measurement of transistor films.
Memory and switching that work at cryogenic temperatures, beside the qubits.
Quantum computers send every bit out to room-temperature electronics because conventional memory and logic do not work beside the qubits. This is a transistor designed to switch at those temperatures with very little energy, so memory and logic can sit inside the cold system.
Opportunities 16–23
Domestic manufacturers compete on cost, quality, and yield, and a single process improvement can move all three. These eight opportunities cover chips, materials, testing, and qualification.
A commercial home for the radiation test every space and defense chip must pass.
Every electronic component bound for orbit or a defense platform must pass heavy-ion radiation testing, and there is no dedicated commercial facility for it, so programs wait in line for time at research facilities. This is a purpose-built test facility designed for commercial scheduling around the clock.
Leading-edge chip patterning on far less expensive equipment
Lithography is the most expensive step in making a chip, and only a handful of companies can afford to manufacture at the frontier. This venture is building a patterning tool that reaches advanced feature sizes on far simpler equipment, and the approach has been demonstrated in hardware. It enters through data-center photonics and builds toward leading-edge logic and memory.
Cheaper, stronger carbon fiber on the same production equipment.
Every producer sells essentially the same carbon fiber, so structures built from it carry extra material to meet spec. This technology produces fiber that processes faster on existing lines and performs better in composites, at a lower modeled cost per kilogram.
End-to-end data links for quantum computers and other cryogenic systems
Quantum computers, cryogenic sensors, and detector arrays all depend on wiring between room temperature and the cold stage, and that wiring already limits how large they can grow. This venture is building data links that carry those signals with far less heat and space, and early devices are already more energy efficient than the best components available today. The product is a complete module any cryogenic system builder can integrate, with quantum computing as the first market.
Watches the brain at work in real time, without the cryogenic vault.
Magnetoencephalography gives the most precise non-invasive view of brain activity, and few hospitals can offer it because today's systems need cryogenic cooling and a dedicated shielded room. These sensors reach comparable sensitivity at room temperature and are built to work outside a shielded room.
Makes the most expensive chip tools faster by upgrading the photoresists fabs already use.
Today's photoresists force chipmakers to trade sensitivity against resolution and durability. This process upgrades a standard resist in minutes per wafer so it performs better under EUV light, while the rest of the fab's qualified workflow stays the same.
Rubies and sapphires grown with concentrated sunlight, each traceable to its source.
Luxury houses want gem materials they can trace, and natural colored stones often pass through supply chains with sanctions exposure. These stones are grown with concentrated sunlight at a national laboratory, each logged to its production run, with supply limited by the process itself.
Qualifies new alloys as fast as machine learning can design them.
Machine-learning models propose new materials faster than anyone can qualify them, so defense and aerospace programs take decades to adopt better alloys. This platform automates the loop from design through production and testing, and returns qualification-grade data for the digital-engineering workflows new programs now require.
Opportunities 24–30
Energy security depends on domestic materials and a grid that holds up under stress. These seven opportunities cover storage, solar, carbon removal, critical minerals, and the infrastructure behind them.
Grid batteries that cannot catch fire, built from domestic materials.
Lithium-ion storage brings fire risk, foreign supply chains, and limited lifetime into every new grid project. This battery chemistry uses water-based cells built from domestic materials and has shown a many-fold gain in cycle life in lab testing.
A domestic, high-volume supplier of space-grade solar
Thousands of satellites now launch every year, and the high-performance solar cells the most demanding missions need are in short supply. This venture is building a domestic manufacturer of space-grade solar cells around a much faster production process. It runs on production-scale equipment and has produced independently certified cells comparable to space-qualified product.
Solar panels that pass the light crops need and power the greenhouse with the rest.
Rooftop solar takes the light a greenhouse crop needs to grow. These panels are tuned to each crop, letting through the light plants use and turning the rest into electricity, and early demonstrations showed healthy growth under working panels.
Seals the small well leaks that cement cannot reach.
Most well-integrity failures start in gaps too narrow for conventional cement repair. This method places a sealant directly in the leak and locks it in place there, and the same step helps protect the well casing from corrosion.
Engineered wood without formaldehyde, at the cost of the resins it replaces.
Emission rules are tightening on the formaldehyde resins that bind particleboard and MDF, and the alternatives cost more or run slower. This binder is made from an abundant byproduct of paper production and runs on existing lines at modeled cost parity or better.
Low-cost permanent carbon removal, hosted on infrastructure that already runs.
Direct air capture is expensive largely because of the machines that move air. This process runs on equipment geothermal power plants already operate, so it needs little new hardware, and analysis puts its cost well below most direct air capture.
Recovers critical materials from waste streams where they already flow.
Industry pays to dispose of waste streams that carry critical materials the country imports. This modular system recovers them on site and turns a disposal cost into domestic supply, with no central refinery required.
Companies, investors, and founders can reach the team behind any opportunity in the batch, whether to license the technology, develop it further with the lab, or build a company around it. Choose one, tell us what you have in mind, and we will connect you with the right people.
