

Batch:
September
23
How Deep Checks Works
Founders apply on our site. We vet their decks for (1) being deeptech and (2) appearing venture backable. We don't conduct interviews. We filter with a light touch, preferring to minimize our own bias. Every three weeks, we email the best decks to you. You log into the page you're viewing now, browse the startups below, and click "Click to copy email" to reach out to founders directly. In short, we are not Y Combinator; we're a light deck filtering program. We work really hard to be efficient.
Invite your diligence team
Invite your associates to review decks with you by introducing them to [email protected]
Confidentiality is critical
Do not share these decks, or you'll be permanently banned. Keeping founders' decks confidential and within your fund is critical to us maintaining founders' trust.
Showing:
Solidec
Chemicals production on-site leveraging modular electrochemical reactors
Solidec builds modular reactors that produce high-purity chemicals like hydrogen peroxide directly on-site using only air, water, and electricity. This decentralized approach eliminates expensive distribution steps, protects companies from fragile supply chains, and significantly reduces emissions.
- The company has secured over $750,000 in paid pilots with major industry players, including NASA and the largest rare-earth producer outside of China.
- They have built a $75 million pipeline of interested customers across the water, critical minerals, and aerospace industries before shipping their first commercial system.
- Their technology shrinks a traditional seven-step chemical manufacturing process down to just one step, cutting energy use in half and reducing emissions by 90 percent.
- The commercial-scale reactor is proven to work for multiple essential chemicals beyond hydrogen peroxide and is backed by more than a dozen peer-reviewed papers in top scientific journals.
- The team brings over 150 years of combined experience in electrochemistry and has successfully built and scaled technology at more than ten previous startups.
Founder’s note on traction
Three paid pilots totaling $750K+: a completed pilot with a cooling tower company holding 25% of global market share, a $270K NASA-funded on-site deployment, and a $500K pilot in contract with Lynas Rare Earths, the largest rare-earth producer outside China. $75M customer pipeline across signed LOIs and MOUs. Commercial-scale MVP demonstrated, with the platform proven across six molecules including hydrogen peroxide (flagship), formic acid, acetic acid, and syngas. Technology is grounded in 12+ peer-reviewed publications including Science, out of the Wang Lab at Rice. $2.5M raised to date led by New Climate Ventures, and the team has built and scaled before at Nel, Ohmium, Fortescue, Hyundai, Bloom, and SACHEM.
Open Membrane Database, Activate, Rice University, Yale University, University of Cambridge, University of Arkansas
Displace
Displace builds completely wireless, battery-powered TVs powered by a local AI operating system that turns any screen into a private smart assistant. This technology transforms passive displays into proactive tools that help people manage daily tasks while keeping their personal data completely secure.
- The founder has a strong track record, having previously built and sold two hardware companies for around $100 million, and has personally invested $6.5 million into this new company.
- Their main product is a 4K TV that runs on batteries and uses a special vacuum mount to stick to almost any wall without the need for drilling.
- The software runs artificial intelligence directly on the device instead of in the cloud, which keeps user data completely private and secure.
- The company has already secured over $750,000 in pre-orders, won multiple awards at CES, and plans to sell on major platforms like Walmart and Temu early next year.
- While they currently make their own hardware, their long-term strategy is to generate most of their revenue by licensing their AI software to other screen manufacturers.
Founder’s note on traction
1. Product launched and won best of CES 2. $750K in revenue 3. Built the Local AI Intelligence Layer and shipping in Jan 2027 4. Selling on Walmart Marketplace and Temu Marketplace in Q1 2027 5. Founder has invested $6.5M
Displace, Frazy, RedKangaroo
GreenLIB
Module processing mixed black mass into specification-grade feedstock
Founder’s note on traction
GreenLIB has built an $84M commercial pipeline, with 7,500 t/y across customer pilots, LOIs, and a 5-year commercial agreement. We have 6 active pilots and 6 LOIs, with customers progressing from testing into commercial agreements. We have secured a lead investor for our $5M seed round and are now bringing in the final $1–2M ahead of close.
Nuum Robotics
Nuum Robotics builds snake-like robots that navigate deep inside aircraft engines to perform inspections and repairs. This technology prevents the need for expensive engine disassembly and helps airlines avoid millions of dollars in grounding fees caused by traditional tools getting stuck.
- Despite being only two months old, the startup already has a signed agreement with Aerotec Engines, is negotiating partnerships with two other aviation companies, and is building the product with veteran Air Canada technicians.
- Traditional inspection tools frequently get stuck inside engines, grounding aircraft and costing over $1.5 million, but preventing just one stuck tool pays for ten years of this robotic service.
- The robots use artificial intelligence to follow exact paths every time, which guarantees consistent inspections and helps make up for a projected global shortage of 710,000 aviation mechanics by 2040.
- Instead of just looking around, future versions will send multiple robots in at the same time to clean, weld, and fix parts from the inside, saving up to $10 million in teardown costs.
- The business model is highly profitable because each robot costs $25,000 to build but leases for $150,000 a year, leading to 81% gross margins as the company grows.
Founder’s note on traction
We started Nuum Robotics two months ago. We have one signed LOI with Aerotec Engines, are negotiating two design partnerships with Execaire Aviation and Dominion Dynamics, and are developing the product with two Air Canada technicians with 35+ years of inspection experience.
Velocity, Formula Nano, Mosaic Manufacturing, University of Waterloo, International Baccalaureate
TETMET
Lattice manufacturing company improving high strength aerospace components
TETMET combines AI, robotics, and lasers to 3D print high-performance metal structures without the need for traditional manufacturing tools. This patented approach greatly reduces material waste, energy use, and production costs for industries like aerospace and automotive.
- The company invented a patented method that builds strong metal parts out of rod-based meshes rather than solid blocks.
- This process cuts the weight of parts by 60 to 75 percent, which saves materials and energy for things like airplane seats, car frames, and space vehicles.
- They have already completed 35 pilot projects across six different industries, bringing in over $1 million in early revenue from major customers.
- Their first factory requires only $4 million to build but is designed to handle $14 million in annual revenue by eliminating expensive manufacturing tools.
- A team of 18 experts in Paris and Amsterdam runs the company, led by a chief scientist and a head of research and development who both hold doctorates.
Founder’s note on traction
We've sold > $1M in Pilots from our lab, and now we are spinning out of the lab into a first factory with a huge soft order book from pilot customers that love us do this at industrial scale and economics.
Brabantse Ontwikkelings Maatschappij (BOM), HEC Paris, University of California, Berkeley, Babson College, Nyenrode Business University
Plasma Orbital
Surveillance from orbiting edge computing providing near-real-time alerts
Plasma Orbital builds AI-powered microsatellites that monitor national waters to detect moving threats like illegal fishing and drug smuggling. By processing data directly in orbit, the system sends alerts in seconds so maritime operators can intercept vessels that hide in the blind spots of traditional surveillance.
- The startup has already secured $10 million in letters of intent and is working with major defense agencies like the US Navy, US Coast Guard, UK Ministry of Defence, and Indonesia's Ministry of Maritime Affairs.
- Instead of taking hours to send images back to Earth like traditional satellites, these satellites use onboard AI to send alerts in seconds so operators can respond immediately.
- The team has successfully built a ground prototype that runs their custom ship detection software on the exact computing chips they plan to send into space.
- The business model is highly scalable, where selling just three fleets of 16 satellites to allied nations would generate over $100 million in revenue.
- The leadership team combines engineers who built autonomous systems for the European Space Agency with veterans who have decades of experience stopping illegal maritime activity.
Founder’s note on traction
$10M in customer LOIs, active engagement with leading maritime defense agencies inc. US Navy, US Coast Guard, UK Ministry of Defence and Indonesia’s Ministry of Maritime Affairs (biggest archipelagic nation in the world)
Kickstart Global, Hardware is not Hard, The Residency, University of Leeds, Accel IT Academy | Cochin, Cochin University of Science and Technology
Inlyte Energy
Iron-sodium battery systems enabling high-density energy storage
Inlyte Energy builds high-density, iron-sodium batteries that provide clean, mission-critical backup power for data centers and industrial sites. This technology replaces polluting diesel generators with a safe, emissions-free alternative that speeds up deployment and meets emerging clean energy mandates.
- The company has built a $2 billion sales pipeline in under a year, which includes a $400 million reservation with Energy Vault and over $15 million in secured projects.
- The batteries run on an iron-sodium chemistry that is completely safe, requires no maintenance, and lasts for more than 20 years.
- The system is highly space-efficient, packing enough energy to replace 29 diesel generators or 156 lithium-ion battery packs into a footprint that fits on a standard diesel pad.
- By relying on abundant materials like iron and salt, the company avoids foreign supply chain risks and keeps manufacturing costs extremely low.
- Data centers currently face long delays due to grid limits and pollution concerns, and this emissions-free system allows them to bypass these roadblocks to get power faster.
Founder’s note on traction
We signed a pilot with NTS Colocation AG for a deployment at a Tier 4 data center in Switzerland, our first commercial demonstration in the C&I/backup power application. We also signed a paid offtake reservation worth up to $400M with Energy Vault for data center deployments. Finally, we have won $15+ million in projects with more than $120 million in LOIs (17+).
Activate, Stanford University, Harvard University, Iolani School
A* Labs
Edge-perception retrofit improving unattended lab run completion
A* Labs builds AI agents and camera sensors that upgrade existing scientific lab equipment to make it fully autonomous. This technology allows labs to automatically detect and fix physical errors without human help, speeding up research without the need to buy expensive new machines.
- The founders have decades of experience building autonomous systems and computer vision for high-stakes environments, including work on surgical robots at Intuitive Surgical and AI at NVIDIA.
- They have already interviewed over 40 lab automation leaders, resulting in three signed pilot agreements and one active design partnership to test their system.
- Instead of making labs buy entirely new robotic systems, they use modular cameras and edge computers to upgrade the expensive machines that labs already own.
- The system runs entirely on local computers rather than relying on cloud connections, ensuring that sensitive research data stays safely inside the lab.
- It directly solves a major bottleneck where up to 60 percent of automated lab runs fail and require human help, as the AI can spot issues like a missing pipette tip and tell the machine to correct it.
Founder’s note on traction
We’ve spoken with 42 lab automation leaders, mostly directors and other buyers with purchasing power. From that, we have 3 signed pilot LOIs and one design partner we are actively building and testing with. Those conversations confirmed both the need and willingness to pay. We keep talking to buyers as we build so we don’t get ahead of the customer again.
ArtifeX AI, Horizon Surgical Systems, Noah Medical, University of California, Los Angeles, University of Hawaii at Manoa, Sharif University of Technology
MusiQ Bio
MusiQ Bio develops a non-invasive drug delivery platform that uses targeted microbubbles and ultrasound to safely cross the blood-brain barrier. This allows treatments for diseases like Huntington's and Alzheimer's to reach specific areas of the brain without the need for surgery.
- The platform works like a two-step lock and key where patients receive an IV injection of microbubbles and doctors apply ultrasound to a specific brain area to release the drug exactly where it is needed.
- Early tests in mice showed a 10-fold increase in gene silencing, proving the technology can successfully deliver therapies directly to inflamed brain regions while leaving healthy areas alone.
- Instead of being limited to one drug, this system can carry many different types of treatments, like gene therapies and small molecules, to address multiple brain diseases.
- The science is backed by peer-reviewed papers in top journals and protected by two patents exclusively licensed from UT Southwestern Medical Center.
- The company has already secured a 2.6 million dollar state grant and 1 million dollars in seed funding to help push the technology toward human clinical trials.
Founder’s note on traction
MusiQ Bio has secured a $2.6M CPRIT SEED grant award from the state of Texas & ~$1M committed toward our $2M seed round. Our platform is supported by peer-reviewed publications and preclinical validation across oncology and brain delivery. We are in discussions with pharma companies interested in applying our platform to their therapeutic payloads and are generating partnership-enabling datasets.
Bioinformatics Research Network, Enventure, Nucleate, Rice University, The University of Texas at Austin, Austin Community College
Voyra Therapeutics
AT2R peptidomimetics improving IPF outcomes
Voyra Therapeutics designs long-acting peptide drugs to heal damaged lung tissue, starting with a terminal disease called Idiopathic Pulmonary Fibrosis. This approach targets the root cause of lung scarring to offer a safer, more convenient treatment for patients who currently rely on strict daily pill regimens with harsh side effects.
- The founding team has deep experience in drug development, having previously led a pulmonary fibrosis program to clinical trials and secured a major licensing deal.
- Instead of starting from scratch, the company improves upon a clinically proven biological target, removing the severe drug interactions and fasting rules associated with current treatments.
- Their computational design engine created a drug that lasts much longer in the body, allowing patients to take it just once a week or once a month instead of twice a day.
- Early tests in human lung tissue and mouse models show the drug successfully reduces scarring and performs just as well as existing therapies but with a cleaner safety profile.
- The company has already filed its first patent to protect its drug designs, validated early interest with pharmaceutical companies, and has a clear plan to reach human clinical trials within three years.
Founder’s note on traction
We built a peptide-focused computational design pipeline that already delivers high-affinity peptides. We validated our peptide approach in human diseased tissue (IPF) and in a mouse model of lung fibrosis. We filed in early September 2026 our 1st provisional patent with Goodwin (exactly 1 year after our founding). We've had initial convos with pharma (CVC) to validate customer interest.
Xaira Therapeutics, Genentech, Harvard Business School Online, Northwestern University - Kellogg School of Management, Conservatoire National des Arts et Métiers, CNAM - Paris
Aelora
Aelora uses an AI platform to discover small molecule drugs that reprogram immune cells from an inflammatory state to a healing state. This approach creates lasting treatments for autoimmune diseases, specifically helping the large number of patients who do not respond to current therapies.
- Instead of just suppressing the immune system, the company targets specific cells to retrain the immune response directly at the site of the disease.
- Their AI engine maps different cell states and predicts which drugs can permanently lock these cells into a healthy mode.
- Early data in rheumatoid arthritis shows that the balance of healthy to inflammatory cells perfectly matches the progression and remission of the disease.
- The initial focus is on rheumatoid arthritis to help the 40 to 60 percent of patients who get no relief from today's best available drugs.
- The founder is an experienced platform builder with a strong background in cell reprogramming from top institutions like Cambridge.
Founder’s note on traction
We're at computational validation, and the signal is strong. Across a public dataset of 70,246 synovial macrophages, the resident:inflammatory ratio tracks disease exactly as our thesis predicts: high in healthy joints, collapsed in active RA, recovering in remission. The recovery trend is early (n=7) but directionally clean, and it points straight at the mechanism our platform is built to drug.
Cambridge Gravity, 5050, Shift Bioscience, University of Cambridge, Kumamoto University, Newcastle University, Université de Monastir
ImmuNovus
ImmuNovus builds an AI-powered digital twin of the human immune system to predict how patients will respond to different treatments. This allows pharmaceutical companies to design better drugs, find new uses for existing medications, and avoid costly clinical trial failures.
- The technology is backed by over $18 million in funding from major agencies like the NIH and DoD, proving its scientific credibility.
- Instead of using black-box AI, the platform relies on models that trace every prediction back to real biology, meeting strict new FDA guidelines for explainable software.
- The company is already proving demand through active contracts and pilot projects with top-tier pharmaceutical companies.
- The software has successfully identified new treatment combinations for respiratory infections and found new uses for existing drugs in autoimmune diseases.
- The leadership team combines deep scientific expertise with proven business success, including a CTO who previously built and sold a simulation software company to Siemens.
Founder’s note on traction
We are in contracting with one pharma company. More than $18M of non-dilutive R&D sits behind the work. We sell a mechanistic immune model to pharma R&D teams around specific drugs and indications. About $1M is already in on an open SAFE. Demand is shown by that contracting work, not by a consumer product.
ImmuNovus, Inc, UNMC College of Public Health, University of Nebraska–Lincoln, University of Nebraska Medical Center, University of Nebraska at Omaha, Brno University of Technology
Inapill
Inapill is developing a new class of oral medicines that stop immune cells from causing inflammation before it starts. This provides a safer, easy-to-take pill alternative to expensive injectable drugs that often fail or cause resistance in patients with autoimmune diseases.
- Current autoimmune treatments require expensive injections and often stop working, but Inapill is creating a simple pill that avoids drug resistance.
- The drug targets a specific pathway to block inflammation early, avoiding the toxic side effects associated with older medications.
- In animal testing, their lead compound successfully saved mice from a deadly autoimmune disease where standard treatments failed.
- The company has a fast and clear path to human trials, with a study design that can show clear results in just a few weeks.
- The startup has already secured a term sheet for a 3 million dollar funding round and is backed by top scientists from Harvard, UC Berkeley, and UC San Diego.
Founder’s note on traction
UC Berkeley postdoc ideas attracted General Inception support (Deep Tech venture studio) / lead small molecule w/in vivo efficacy in rare autoimmune entry clinical indication / signed termsheet on $3M round w/Freeflow (Deep Tech VC) & LongGame (Longevity VC) to reach 1st DC + build out multi-asset platform / >2/3rds of round soft-circled / + feedback from both pharma strategics + patient advocates
University of California, San Francisco, UC Berkeley, University of California, Berkeley, Nebraska Wesleyan University
Nilsson
X-Silicon designs specialized computer chips and software that allow drones and robots to process artificial intelligence and understand their physical surroundings in real time. This technology helps autonomous machines operate longer and make faster decisions on their own without draining their batteries or relying on constant cloud connections.
- Current robots rely on separate chips for different tasks, which wastes power and slows them down, but this new design combines these functions into a single system to save energy and speed up reaction times.
- The platform allows machines to keep working safely and navigate their environments even if they lose GPS or internet connections.
- Early tests show their prototype uses over three times less energy and processes data much faster than standard graphics chips.
- The company is raising a $3.8 million seed round to validate their hardware, with $1.5 million already committed by the founders themselves and early licensing interest from global companies.
- The leadership team includes former executives from major tech companies like Qualcomm and AMD who have previously designed chips used in billions of devices.
Founder’s note on traction
X-Silicon is creating a real-time Physical AI & Spatial Intelligence sfw and chip platforms for robotics, drones, mobile & edge servers to operate & learn autonomously. Pre-revenue. Raising $3.8M seed w/$1.5M committed by founders to validate the patented architecture in FPGA hdw. Several global companies waiting. 2 LOIs in the EU & India for drone & robotics licensing. Chip production '28.
NSDN / LJ1D LLC Consulting, Dallah Trans Arabia, Zones, LLC
ItsWare
ItsWare provides a real-time operating system for factory floors that uses sensors to automatically track parts, tools, and assemblies. This eliminates manual data entry and creates a live digital twin of operations, helping manufacturers reduce waste and meet strict defense compliance standards.
- The company has strong early traction with seven live locations and two multibillion-dollar aerospace manufacturers as paying customers.
- They have built a near-term pipeline of 8 to 12 million dollars in deals that are in the procurement stage or later, including a potential eight-figure military contract.
- The software proved its value in a real-world case study with an aerospace supplier by reducing missing tools by 90 percent and saving thousands of dollars a month.
- The platform runs entirely on local, disconnected networks, making it a perfect fit for defense contractors who need to securely track parts and meet strict government rules.
- They grow revenue through a land-and-expand model, starting with a single pilot program and multiplying that revenue by 13 times when rolling out across an entire facility.
Founder’s note on traction
7 locations live, two multibillion-dollar aerospace manufacturers as paying customers, including one of the world’s largest drone companies, with a $120K PO expected and a potential $1.5M expansion. We are in procurement with 2 other multibillion-dollar drone companies, a major airline, and are in the final stage to be awarded an 8 figure military depot contract. Our near-term pipeline is $8-12M.
Hindsight Optical, UCI Neuromorphic Machine Intelligence Lab, Osaka University, University of California, Irvine, Santa Monica College
Pioniq Technologies
Pioniq builds solid-state batteries using quantum electrolytes to replace traditional lithium-ion technology. This creates a safer and highly scalable energy storage solution that completely eliminates the need for critical materials like lithium and cobalt.
- The technology relies on a newly discovered state of matter called 1D-water, which allows protons to move freely and power the battery without liquid components.
- These batteries are completely free from critical materials like lithium and cobalt, avoiding the fragile supply chains that currently dominate the industry.
- Unlike traditional batteries that cause tens of thousands of fires each year, this solid-state design offers absolute safety and removes the risk of catching fire.
- The company is finalizing a paid contract with a major European defense company and negotiating with a leading smart-eyewear brand for alternative battery supplies.
- The team is led by veteran scientists from top French research institutes and has already secured 2 million euros in seed funding from a major deep tech investor.
Founder’s note on traction
We are pre-revenue. We are finalizing a paid SOW with a major European defense and electronics company and are in advanced discussions with a leading connected-eyewear company looking for an alternative battery supplier as it scales. We are also working with several other industrial customers on testing and qualification.
Indépendant, Happy Electron x Kinetik, Université de Paris, Université Paris Cité, Sorbonne Université, Université Paris-Saclay, Polytech Paris-Saclay
Momentum Transfer
Momentum Transfer uses powerful synchrotron light sources to analyze mineral samples with extreme speed and precision. This allows mining operators to accurately predict ore recovery rates, optimize their extraction processes, and significantly reduce waste.
- Mining operations currently lose 5 to 25 percent of their value because they cannot accurately predict how much mineral they will actually recover from their ore.
- By measuring samples at synchrotron facilities, their method is a million times brighter than standard lab X-rays, delivering 200 times better data quality and a thousand times the capacity.
- Customers simply ship standard samples to the lab, where the team handles all preparation, measurement, and data interpretation to return results in a matter of days.
- The company secured 12 mining customers in just five months, including a major project with BHP that quickly expanded into a 1.6 million euro proposal.
- While their current focus is the critical minerals market, their platform also serves the broader materials industry with over 120 paying partners including Merck, BASF, and Bayer.
Founder’s note on traction
12 mining customers in five months, including BHP: a €79k first project that turned into a €1.6m follow-on proposal for 10,000 samples. Bookings grew from €187k to €275k half over half at roughly 65% contribution margin. We still serve the materials side we started from — BASF, Evonik, Merck, Bayer and more than 120 paying partners — but we now focus on mining. We measure thousands of samples a mo
Consultant, AugurSignals.com, MESI, University of Georgia, Department of Physics and Astronomy, Manager Tools Academy
Thunderstone
Thunderstone develops an in-situ mining technology that uses electrical voltage to extract copper directly from deep underground without digging traditional mines. This approach unlocks massive untapped metal reserves to help meet the growing global copper deficit while eliminating the billions of dollars and decades of time typically required to build a conventional mine.
- Traditional mining requires moving hundreds of tonnes of rock to get a single tonne of metal, but this method leaves the rock in the ground and only extracts the valuable copper.
- Building a new copper mine usually costs billions of dollars, whereas this well-based approach costs less than a million dollars for an initial setup and can be ready in days.
- The process completely prevents mine waste and tailings, avoiding massive environmental damage and the billions of dollars in liabilities that come with traditional mining.
- Copper ore is normally too tight for fluids to pass through, which has stalled other projects, so the company applies high voltage to crack the rock and low voltage to guide the dissolved metals out.
- The team has proven the core steps in the lab using real ore and is raising 2 million dollars to reach a commercial pilot in 14 months, with major mining companies already asking to test the technology.
Founder’s note on traction
Pre-revenue, no LOIs yet; a signed pilot partner is our month-14 milestone. Demand is proven by in-situ copper stalling on the exact problem we solve: Gunnison is permitted and built but publicly stalled on low flow rates. Miners and mining CVCs incl. Rio Tinto, Vale and Mitsubishi asked to see it work in their rock, so we offer 6-8 week core testing. Copper faces a ~10 Mt deficit by 2040.
Deep Science Ventures, Bean B2C Recycling, ETH Zürich, Imperial College London, North Carolina State University
US Carbon
US Carbon transforms abundant American coal into critical carbon materials like graphite in hours instead of weeks. This creates a secure, low-cost domestic supply of essential materials needed for electric vehicle batteries and defense systems, breaking reliance on foreign supply chains.
- The company uses a patented process to turn cheap domestic coal into high-grade graphite in hours rather than the weeks it takes traditional methods.
- It solves a major national security issue by creating a local supply chain for graphite, a material crucial for batteries and defense that is currently dominated by China.
- By starting with inexpensive raw materials and recovering over 95 percent of the processing chemicals, they aim to cut production costs by 75 percent compared to standard synthetic graphite.
- The technology is already proven at a small scale to produce highly pure graphite, and they are currently testing samples with defense and battery customers.
- Their Wyoming facility is already permitted and partially equipped, allowing them to skip years of red tape and quickly scale up production to meet demand.
Founder’s note on traction
Pre-revenue, validated at kg scale: >95% graphitization, <0.1% ash, cycling confirmed against commercial anodes. Demand proof: a defense/battery customer sampling toward a nine-month qualification, an anode developer under CDA, and rich pipeline. Seed funds a 100 MT/yr pilot up to 2,000MT with $2M of launch-priced revenue at 100MT scale.
Debut, University of North Carolina at Chapel Hill, PPG Industries, University of Wisconsin-Madison
Metari Technologies
E-waste processing turning discarded boards into high-purity metals
Metari uses artificial intelligence to optimize the recycling of electronic waste, turning old circuit boards and chips into pure, reusable metals like copper and gold. This helps close the growing gap in global metal supply while tackling the massive amount of electronic waste generated by the booming AI hardware industry.
- Electronic waste contains over $91 billion worth of metals each year, but less than a quarter of it is currently recycled.
- The company uses machine learning to monitor and adjust the recycling process in real time, which extracts more metal and reduces costs compared to older methods.
- The rapid growth of AI computing is creating millions of tons of new electronic waste, making this a fast-growing and highly valuable source of materials.
- Instead of building massive central plants, they plan to create modular recycling units that can be easily franchised and set up anywhere in the world near the waste sources.
- The founding team includes former engineers from Fuse, and alongside £300,000 in committed funding, they have already signed letters of intent for supply and sales while negotiating a site for their first pilot plant.
Founder’s note on traction
We are in the process of building our first prototype, we signed for our first workshop and have identified our first pilot. We have £300k of funding committed and are looking to find a lead to close our pre-seed round. Please email me for the deck, it didn't upload
Fuse Energy, Supercritical, The University of Manchester, The University of Salford
AIRA
AIRA builds and deploys mobile AI robots to perform manual, repetitive tasks in manufacturing environments through a flexible subscription model. This approach directly addresses severe labor shortages in the industrial sector by providing on-demand robotic workers that reduce labor expenses by over 30%.
- They already have a $400,000 purchase order with a $150,000 down payment and are starting a pilot with Lear Corporation that could bring in 1.2 million pounds in annual revenue.
- The company has a clear path to grow with a 20 million pound sales pipeline and letters of intent from customers who want to expand these robots across dozens of factories.
- Instead of selling expensive hardware, they offer a subscription model where customers pay for the labor the robots provide, making it much easier for factories to adopt the technology.
- The founding team has deep experience in the field, having previously built and deployed robots that successfully completed over 5,500 construction projects at a past startup.
- They have been accepted into several accelerator programs, including one in Oregon that grants them direct access to the United States aerospace and defense sectors.
Founder’s note on traction
We're in revenue with a $400,000 PO and $150,000 down payment from our first customer. We're developing further pilots with adjacent customers in Automotive sector. We have a clear scale target from the first customer and LoIs committing to scale. We've been accepted to several accelerators including UAS in Oregon granting us access to US aerospace and defence sectors.
q-Bot, Ventive, Green Structures, University of Cambridge
Cohesive Robotics
Cohesive Robotics builds smart robotic systems powered by artificial intelligence to automate complex manufacturing tasks like welding and polishing. This technology helps factories overcome severe labor shortages and clear production backlogs by making it easy to automate highly variable, made-to-order parts.
- The company has three deployed systems that have generated over $400,000 in revenue over the past 18 months by processing hundreds of parts per week on factory floors.
- They have secured seven letters of intent and paid pilots worth $89 million at full expansion, which includes a deal with a major global energy technology manufacturer.
- Early customers are seeing real results, with one manufacturer doubling their welding capacity and buying a second system within eight months.
- They use a clear business model that pairs a $195,000 upfront hardware sale with a $35,500 recurring software plan for ongoing support and AI updates.
- The founding team has deep industry experience, featuring a CEO with 15 years in robotics and a CTO who has led robotics projects at major companies like Tesla and Blue Origin.
Founder’s note on traction
We have 3 deployed revenue-generating systems, manufacturing hundreds of parts per week, growing utilization 48+% per month. This represents over $400K revenue in the past 18 months. Additionally we have 7 signed LOIs and paid pilots worth deployment revenues of $89M at full account expansion. This includes a leading global gas turbine manufacturer ($130B market cap).
D.J.A. Imports, Ltd, Oceaneering, University of Maryland College Park, University of Pennsylvania, Trinity College-Hartford
Freestyle Robotics
Pharma-lab automation enabling integrated robotics and AI
Freestyle Robotics builds AI-powered robotic hands and vision systems to automate repetitive physical tasks and strict auditing in pharmaceutical labs. This frees up highly trained scientists to focus on actual research while eliminating costly errors and speeding up drug development.
- The company is finalizing an agreement with a major global pharmaceutical partner and expects to generate its first revenue within the next two months.
- Instead of rigid legacy systems that take months to set up, their AI platform adapts to new workflows in days and manages delicate tasks like handling vials or pills.
- Their custom robotic hand combines human-like flexibility with industrial strength, capable of lifting 60 pounds while being gentle enough to handle fragile glass samples.
- They sell the robots as a service with no upfront costs, saving customers an estimated $131,000 per year per robot with a six-month payback period.
- The founding team includes robotics and AI experts with advanced degrees from Carnegie Mellon and MIT, plus experience at NASA, Amazon, and Meta.
Founder’s note on traction
We're finalizing an LOI with a Tier 1 pharma partner, expecting first revenue in 1-2 months. Discovery with 12+ Tier 1/generics labs revealed a costly problem: highly trained scientists do repetitive, high-mix/low-volume tasks where errors cost millions. We solve this by automating both the physical execution and the strict, manual auditability unique to pharma.
Bloomfield, Bloomfield Robotics, Ceres Robotics, Carnegie Mellon University, University of Illinois at Chicago, City Colleges of Chicago-Wilbur Wright College, Notre Dame High School
Sensate Robotics
Automation enabling multimodal grading of produce
Sensate Robotics builds robotic arms equipped with vision, touch, and electrical sensors to automate the final grading and packing of fresh produce. This technology replaces expensive manual labor and detects hidden defects like bruising, saving packing facilities money and recovering revenue lost to damaged goods.
- The robots combine cameras, touch sensors, and electrical imaging to feel how firm a fruit is and spot internal damage that normal cameras cannot see.
- The company makes money two ways on every installation, charging a monthly lease for the hardware and a subscription fee for the data analytics software.
- They have already secured research funding from NASA and the National Science Foundation, filed two patents, and signed their first pilot agreement with an orchard.
- By running 24 hours a day, these robots can cut packing labor costs in half while improving grading accuracy to over 95 percent.
- The founding team includes experienced robotics experts who have previously built and sold commercial robotic grippers.
Founder’s note on traction
Sensate Robotics is bringing a human-like sense of touch to robots. Our tactile-sensing gripper, paired with a robot arm and an integrated vision system, automates handling tasks that could not be automated before. The debut application sorts, grades and packs delicate fruit in a single step, in a self-contained cell that replaces manual handling and fits small and large packhouses alike. We have completed NASA and NSF SBIR Phase I awards, our NASA Phase II is awarded and kicking off, our NSF Phase II is pending, we graduated Techstars in June, and our first R&D pilot is underway. Customers come through founder-led direct sales to packhouses, sourced through grower associations and advisor referrals, with the system offered as robotics-as-a-service or outright purchase.
Voyager Technologies, OnRobot, Jet Propulsion Laboratory, U.S. Naval War College, University of Southern California, Old Dominion University, Tulane University
Kas Robotics
Laboratory automation using autonomous dual-arm robots
Founder’s note on traction
Our team built and operated 20 task-specific robotic systems in an accredited food-testing lab. That lab is our first customer. A proposed $708K annual agreement is expected to execute at financing close. We’re developing the Bench Agent to unify these capabilities. We’re also in early talks with testing labs and a large life-sciences company.
Aliensense
Hardware ecosystem enabling deterministic sensor-to-compute synchronization
Aliensense builds pre-integrated motherboards and smart sensors that allow engineering teams to assemble a working robot in a single day. This eliminates the months of wiring, calibration, and hundreds of thousands of dollars typically spent on custom integration for new robot programs.
- The company has already manufactured its first batch of 1,000 units in Taiwan and started shipping them, with over 30 robotics teams evaluating the hardware and six running formal six-month trials.
- They sell the hardware at an affordable price upfront and make recurring revenue through an app store where customers buy software for robot perception and autonomy.
- Recent US government bans on Chinese robotics have created an immediate need for their technology because their entire hardware stack is built without any Chinese parts.
- The founders funded the initial development and manufacturing by securing 1.2 million dollars in custom engineering contracts, allowing them to build the platform without giving up equity.
- The leadership team has deep experience building mission-critical hardware, including cameras for the International Space Station, and recently secured their first approved US patent.
Founder’s note on traction
The first production run of 1,000 NXS smart sensors was built in Taiwan and began shipping on 20 August 2026, with 50 units reserved out of that batch. Silicon and supply partners on the current designs: Sony, Intel, SiMa.ai, Lattice, Microchip and onsemi. 6 teams are running six-month evaluations under signed agreements
Microavia, Emerging markets IT consultancy, Startup Leadership Program, IESE Business School, International University of Nature, Society and Man "Dubna"
Altaris Space
Altaris builds swarms of artificial intelligence-powered robots to map the lunar surface and construct essential infrastructure like landing pads. This technology provides space agencies and commercial companies with the accurate ground data and safe landing zones needed to establish a permanent human presence on the moon.
- NASA needs accurate data about the moon's south pole and has previously spent $10 million for this exact type of information.
- Aerospace company Firefly is already reviewing the startup as a potential subcontractor and has advanced them to a formal technical review.
- The founding team brings engineering and technical leadership experience from top aerospace organizations like NASA, Blue Origin, and Lockheed Martin.
- Their first product is a robotic scout that uses artificial intelligence and space-rated computers to explore the moon, map the terrain, and guide spacecraft to safe landing spots.
- Beyond just mapping, these robot swarms will eventually work together to build physical structures on the moon, allowing the company to make money by selling data, charging for construction, and collecting landing fees.
Founder’s note on traction
No closed revenue or a signed LOI yet, but demand is showing up on both sides of the market: NASA leadership says they have no ground truth on the lunar south pole, and they already paid $10M for exactly this kind of data. Firefly has advanced us to a formal payload questionnaire and technical review under NDA showing interest in us as a subcontractor.
Blue Origin, BLUE ORIGIN, Maxar Technologies, University of Colorado Boulder, Boston University, University of Vermont, Colby College
ExLabs
Mission-as-a-service platform enabling autonomous, modular spacecraft
ExLabs builds commercial spacecraft and mission infrastructure to make deep space exploration accessible beyond Earth's orbit. By offering a repeatable platform and a mission-as-a-service model, they allow organizations to participate in deep space missions without the massive cost of building their own programs from scratch.
- The company offers a single, modular spacecraft design that can carry up to 5,000 kilograms of cargo across multiple orbits, including deep space.
- Customers can choose to either buy the spacecraft outright or simply pay for space on an upcoming mission, making space access much more flexible.
- By partnering with NASA's Jet Propulsion Laboratory, they combine top-tier space expertise with commercial efficiency to get missions ready in just 12 months.
- They already have proven demand with $22 million in signed contracts and another $500 million in agreements currently closing.
- The leadership team and advisors include a former NASA administrator and experienced engineers from major aerospace companies like SpaceX, Boeing, and Planet.
Founder’s note on traction
ExLabs is building the GEO+ spacecraft and mission infrastructure layer beyond Earth orbit, enabling government agencies, universities, research institutions, and commercial customers a practical way to participate without owning an entire mission program. We combine our SERV spacecraft platform with recurring mission services, turning historically bespoke, government-led missions into repeatable commercial access. We are acquiring customers through direct government, institutional, and international partnerships, with more than $500M in hardware and mission-services MOUs, LOIs, and active contracts.
Inqubator LLC, Rocket Motorsports, Yonder, New York Film Academy
The Good Drone Company
Drones enabling long-range middle-mile cargo delivery
Founder’s note on traction
14 signed LOIs across 4 countries, $4.45M in Kenya alone. Key customers include Freight in Time (UPS contractor, 8 East African countries), ELEVATE Kenya for healthcare and tea transport (sponsored by the Office of the President), and Xenopia for radiopharmaceuticals. We hold a Kenyan ROC, one of 3 granted by KCAA in 2025, and dedicated BVLOS airspace for certification flights with Konza.