Hard Tech Startups funded by Y Combinator (YC) 2026

Neuromorphic is building an embodiment-agnostic “brain” that makes robots fast to deploy and easy to manage through natural interfaces like Slack, email, and phone. Our goal is to make deploying robots in wet labs and other physical industries as easy as onboarding a new employee.

The technologies that define this century, including data centers, rockets, satellites, and robots, are built from complex metal parts. As these systems advance, the parts get harder to make. Past a certain complexity, casting and machining stop working, and the advanced methods that remain are slow, expensive, and hazardous. Ultrasonium has invented a new manufacturing method to make these critical parts 15x faster and 75% cheaper than the methods they replace, in the U.S. and at scale. Solid metal feedstock goes in and a finished part safely comes out, controlled by a physical AI layer and novel control processes. The systems that matter most, in aerospace, energy, and compute, are demanding metal geometries that can't be machined or cast and are too expensive and slow to make with incumbent methods, at a moment when domestic manufacturing capacity is a national priority. Our founding team has built and innovated some of the most complex engineering systems known to man: quantum computers, superconductors, nuclear reactors, and novel aerospace alloy processing methods.

Applied Electrodynamics builds imaging systems that see through solid matter. Our team previously co-invented the iPhone touchscreen and Nest learning thermostat. We’re starting in construction, helping teams map structural elements and utilities inside walls and ceilings. The same platform extends to security screening and cargo inspection. Our mission is to make the hidden physical world machine-readable across the entire electromagnetic spectrum.

We want to push machines to where humans cant go. To do that machines need to sense and understand their environment. Synapse Semiconductor is the first company to integrate all components of the computer vision stack into one substrate. Modern computer vision requires the complex orchestration of storage, compute, processing, and sensors, inherently limiting the flexibility of its applications. These must be consolidated.

We are building nuclear power delivery ships to power our floating compute barges at sea. Our systems are fully standalone, have better cooling, and avoid all the resource competition that leads to local community opposition. Our centralized manufacturing allows us to deploy anywhere on Earth, four times faster than on land.

Daqstra is the AI-native orchestration layer for physical R&D. It saves engineers in Aerospace, Energy and Robotics hundreds of man hours managing test infrastructure by connecting fragmented test hardware, bringing live and historical test data into one place, and by giving engineers an API to orchestrate their entire test setup. We enable teams to spend less time rebuilding test infrastructure and more time improving the hardware they test.

Geopolitical conflicts and natural disasters frequently expose the vulnerability of the American chemical industry. The reason? Production is centralized and fossil fuel-dependent. The solution? Localized production using an undervalued resource: biogas. The US produces enough biogas to make over $20 bn/y worth of chemicals, but 60% is wasted in low-margin applications or flared altogether. Rise Reforming is commercializing a modular process to transform this underutilized biogas on-site into cheap, supply-secure chemicals. Our beachhead market is dimethyl ether (DME), used as an aerosol propellant and a low-carbon fuel additive in propane. Our target market is methanol: an established billion $ chemical and the future of low-carbon shipping. We're doing our pilot project at a Chicagoland wastewater plant. This will be the first wastewater biogas-to-methanol demo in the Americas. We're hiring for a senior chemical process engineer. Apply here: https://www.ycombinator.com/companies/rise-reforming/jobs/wJ9Q9nv-senior-chemical-process-engineer

Robocurve builds open-source tools and independent benchmarks to measure how well robots can do real-world jobs. Instead of relying on unverified demo videos from frontier labs, we score their models on reproducible benchmarks that anyone can trust. Today there are no well-run, standardized robotics benchmarks. Labs evaluate in-house, and no independent group has stepped in to run continuous benchmarking as a service. The result is that no one actually knows how good anyone else is, or where the real frontier sits. Good benchmarks require operating and maintaining physical hardware and real-world setups. Simulation only goes so far, since models that look strong in sim can show large performance gaps once deployed in the real world. Building real-world benchmarks means coordinating job-domain experts, evals engineering, and hands-on robotics all at once. Frontier labs are targeting general-purpose robotics by 2028, yet the field of robotics evals barely exists. Whoever builds the trusted measure of robot capability becomes the reference everyone relies on. We combine backgrounds in AI evals and robotics to build and grow this field as fast as possible. We already shipped v1 of our open-source framework, Inspect Robots, and ran our first pilot scoring a frontier model on a real robot.

83 Sciences uses AI to mine discarded experimental data and drive scientific breakthroughs. Working alongside labs, we discover the materials that will power the new Industrial Revolution.

Our mission is to build a city on Mars. Cosmic-1, the first equipment in our fleet, is an autonomous robot capable of heavy-lift manipulation. Its been deployed on some of the largest solar farms in the US, more than doubling labor productivity across tens of thousands of panel installs. We're scaling in solar, expanding in data centers, and building lunar construction robotics under contract with NASA.

Nori builds home robots that cost less than $2,000 and that anyone can teach. Out of the box, the NORI L3 does household chores: loading the dishwasher, folding laundry, tidying up. But because anyone can train it, it also learns your specific tasks, like washing paintbrushes if you're an artist or flipping sheet music while you play guitar. It's a general-purpose robot, not an appliance. We launched pre-orders 6 weeks ago and have already booked $100K in revenue. Humanoid robots exist, but they cost $20,000 or more and are demoed, not shipped. We got to $1,688 without sacrificing capability. Instead of expensive actuators, we use low-cost servos and get more out of them through smarter control and design. A wheeled base with a lift gives full floor-to-counter reach without the cost and fragility of legs, and we streamlined manufacturing from day one, assembling in the US. The teachability is the moat. Every skill a user trains can be shared, so each customer makes every robot more capable. Hardware is the entry point; the skill ecosystem is the recurring revenue. We're robotics researchers from Columbia labs like RoboPIL (robot learning) and CGUI. Working with $100K research robots that struggled with everyday tasks convinced us the problem isn't capability, it's cost. Good robots don't need to be expensive. Households already spend $2,000+ on single-task machines: a dishwasher, a washing machine, a vacuum. We sell one robot at $1,688 that does the work of all of them and learns more. At 5% of US households and $25 a month for skills, that's $2B a year in recurring revenue, on top of hardware.

Our software turns any robot we deploy into an everyday independent worker at your facility. They are able to walk sites, handle instruments, overcome obstacles, collect measurements, and generate auditable compliance records, replacing hours of mandatory manual surveys and inspections. Built by Los Alamos robotics veterans and deterministic-manipulation researchers, we're starting in nuclear and expanding to other hazardous industries.

Praxis captures the data physical AI is bottlenecked on: egocentric video, 3D scans, and multimodal capture from inside real industrial and residential environments. Embedded within publicly listed and unicorn-scale conglomerates, we reach 60k workers across 5 continents and 150+ environment types, supplying the real-world training data frontier labs and humanoid companies can't source anywhere else.

Arlo Industries is building a network of sensors that precisely tracks stealthy drones and aerial threats better than radars. This capability does not exist yet because traditional radars are built on past war doctrines designed to protect single, centralised assets. Modern warfare, however, demands wide-area, persistent coverage that is both passive and economically scalable. By utilising a mesh architecture, Arlo Industries unlocks unprecedented asymmetric economics. The network is fundamentally built to scale: the cost of adding sensors grows linearly, but the tracking accuracy increases exponentially, delivering highly resilient, persistent coverage for a fraction of the cost of legacy systems. Deo founded Arlo Industries after spending over 6 years in the Israeli defence ecosystem, where he lived through multiple conflicts and experienced the Iron Dome and other systems in action against ICBMs and the infamous Shahed drones. Learning exactly how these defence systems could be improved after a Shahed exploded close to his apartment, he built Arlo Industries around the core philosophy that conflict should be concise & precise. The people behind Arlo1 includes a PhD in drones and autonomy and military experts. Arlo Industries has actively showcased this technology directly to Ukrainian militaries on the frontlines, as well as to operators across Europe and the USA.

Ornadyne builds flapping-wing drones for reconnaissance that look, fly, and sound like birds. Today’s drones are easy to detect acoustically, visually, and electronically. As counter-UAS systems improve, traditional quadcopters and fixed-wing UAS systems are becoming increasingly ineffective for close-range surveillance. Flapping-wing aircraft have existed for years, but have lacked the performance, endurance, and control required for real-world deployment. We’re changing that by engineering a high-performance, stealth-optimized system built from first principles for efficiency, autonomy, and reliable field use. The final frontier of surveillance won’t look like machines; it will look like wildlife already in the sky. Maybe birds aren't real?

We turn natural Language into Multi-Drone Mission execution. Degla is built around a simple idea: once you’ve expressed your intent, the system owns the problem.You don’t babysit drones. You don’t micromanage failures. You don’t rewrite plans. You state the goal, Degla figures out how.

Nine Fives produces modern radio frequency test equipment and solutions to disrupt the $130B test equipment industry. By combining modern electronics, AI ready test automation, and strategic supply chains, we’re equipping our customers to cut their time to market in half.

Dispatch is building refurbishable reentry vehicles that host and return payloads for companies making ultra-high-value materials (semiconductors, biotech, and pharma) that can only be manufactured in space. Right now, these companies have no practical way to manufacture and get their products back to Earth. We solve this bottleneck. We just returned from the Mojave Desert testing our first full-scale reentry heat shield designed to protect the satellite as it returns from space at Mach 20+, built in an apartment for 100x cheaper than competitors. To test, we held it behind a rocket engine to simulate the forces of atmospheric reentry. IT WORKED! Watch here - https://youtu.be/dhlQmnIoptg To scale, we anchor permanent infrastructure in orbit - building high power Autonomous Space Stations designed from day one for "lights-out" manufacturing serviced by the reentry vehicles.

Ndea is building frontier AI systems that blend intuitive pattern recognition and formal reasoning into a unified architecture. # AI for Scientific Advancement Unlike all life before us, humanity's ascent is a story of ingenuity, not just biological evolution. Our progress has been driven by the curiosity to acquire knowledge, the ability to pass it on, and an intrinsic drive to innovate. We build technology which gives us leverage beyond our biology. We stand at the top of a knowledge and technology colossus that we collectively created over the past ten thousand generations. Scientific progress has helped us overcome burdens that long defined human life — famines, plagues, and widespread illiteracy. Science will continue to redefine the boundaries of the human condition. Today, the acceleration of scientific progress hinges on one factor: AI capable of independent invention and discovery. This capacity is the gateway to advancements beyond our wildest imagination. # A New Research Lab We're starting Ndea — an AI research and science lab. The name - like 'idea' with an 'n' - is inspired by the Greek concepts ennoia (intuitive understanding) and dianoia (logical reasoning), capturing our first goal to merge deep learning with program synthesis. Ndea is entirely focused on developing and operationalizing AGI to realize unprecedented scientific progress in our lifetime for the benefit of all current and future generations. Building AGI alone is a monumental undertaking, but our mission is even bigger. We're creating a factory for rapid scientific advancement — a factory capable of inventing and commercializing N ideas. From our vantage point today, we see many 'known' frontiers like self-driving vehicles, drug discovery, sustainable energy, robotics, and space exploration. While AGI will benefit all of these, the most exciting adventure lies in the 'unknown'. AGI promises discoveries and progress we cannot imagine today.

Up to 80% of the most advance AI compute chips sits idle because of slow networking. Piris Labs builds photonic based networking solutions to enable fast, scalable, and low cost AI workload execution infrastructure.

Visibl makes custom chip development dramatically faster and lower cost for hardware companies. Today, most hardware companies that need a custom chip have bad options: sign a massive contract with an incumbent like Broadcom, pay a traditional design house millions of dollars and wait 18+ months, or keep using off-the-shelf chips and FPGAs that are not optimized for their product's cost, power, performance, size, or manufacturability. We handle the full path from architecture and design through verification, foundry management, and production. That makes custom ASICs accessible to robotics, drones, industrial, IoT, and edge hardware companies that were previously priced out of the market. As more hardware products need differentiated silicon, we think custom chip development should be available to far more companies than it is today.

Zephyr Fusion is building the first in-orbit fusion power source: a megawatt-class reactor enabling large-scale industry in space, at a fraction of the cost of equivalent solar. Founded by former physicists from Lawrence Livermore and Oak Ridge National Laboratories, Zephyr leverages decades of fusion research to deliver a compact magnetic confinement system designed to take advantage of the space environment.

Thesis is an AI research lab automating AI R&D itself. We are building Darwin, a hill-climbing machine capable of searching vast combinatorial spaces of experiments. By using the outcome of every experiment to optimize the next, Darwin turns machine learning research into a compounding system rather than a sequence of manual attempts. We are directing this machine toward the grand challenges of biology and materials science. Our mission is to build the machine that discovers the next Transformer, the next AlphaFold, and the breakthroughs that push science and engineering to the absolute frontier.

We build brains for robots. We work with manufacturers like Unitree to make robots listen, speak, and act. You can program your robot out of the box by literally talking to it, no coding involved. Robots running Lightberry are emotionally intelligent, always on, and fully autonomous. Just like in Star Wars!

Wardstone builds a low-cost, modular interceptor platform for air and missile defense. Our platform enables the warfighter to right sizes their interceptors against the threat. This allows us to defeat air-breathing and ballistic missile threats at a fraction of the cost of exquisite interceptors, adding affordable magazine depth to existing air defense systems.

We are scaling foundation models that are curious about the world. We’ve seen general intelligence appear exactly once, and its hallmark has been curiosity: a dissatisfaction with what we already know about the world, an ability to learn from world experience, and a disposition to share knowledge with other individuals. Across generations, these human traits have built modern science, technology, and culture. We think superintelligence will come from models that work the same way: learning from their own observation, updating as they go, growing what they know. These models will start from what humanity knows and discover past us. We believe that scaling curious intelligence is a defining problem of our time. If solved, we will live to see a new era of technological advances, scientific understanding, and human flourishing.

SigmanticAI is building an AI-native hardware development assistant that automates the entire RTL design flow, from natural language to synthesizable HDL and testbenches, inside a seamless VSCode fork. Powered by fine-tuned Verilog LLMs, reinforcement learning, and real compiler feedback, SigmanticAI iteratively refines code until it compiles and passes synthesis, no matter how complex the design. It’s like the Cursor for HDL design. In addition to code generation, SigmanticAI can generate token-level annotations and onboarding documents to accelerate ramp-up for new engineers. With support for both cloud and on-prem deployment, it integrates with existing EDA tools to dramatically reduce debug time, improve collaboration, and accelerate hardware design across teams

Check out the project: https://github.com/cactus-compute/cactus

Boost Robotics is building autonomous mobile manipulation robots to remotely perform manual tasks in data centers. The surge in AI applications is pushing data centers to scale faster than they can manage, leading to suboptimal operations and staffing shortages amid rising power and compute demands. We save data centers millions in insurance premiums and SLA violations by increasing the frequency of inspection and maintenance while reducing the time it takes to triage and address failures.

PowerMatrix provides the MOST efficient and compact next-generation high-performance power supplies for emerging technologies, such as GPUs, xPUs, AI server, datacenters, drones, EV, aerospace, etc. Based on patented circuit topology, embedded packaging, and novel magnetics, our power supply systems can reduce energy loss by up to 50%. We have developed several prototypes and are collaborating with world-leading customers.

Photonium is building software to automate optical system design. We supercharge optics experts with intelligent tooling to reduce costs and deliver faster. We handle the full design stack — from optimization, verification, sourcing, to prototyping — for AR/VR, quantum, biotech, metrology/chip fab, LiDAR, and more.

What if every experiment you wanted to run could be modeled using AI? Godela is an AI-powered physics engine that gives engineers faster, cheaper replacement to simulations and physical prototypes. No set up, no wait. Just ask—and get simulation quality result.

Cascade Space is building the commercial alternative to the Deep Space Network. Our first ground station, Earth Station Zero, will be online and downlinking data from the Moon in 2026, with plans to scale to a full global network by 2028.

Instinct is making the Moon the new LEO. Our small lunar lander is designed to be launched as a regular LEO rideshare, enabling higher mission cadence at a lower cost.

Maritime Fusion is building fusion reactors for ships and off grid energy markets. En route on this endeavor we’re commercializing High Temperature Superconducting (HTS) cable technology for power distribution while advancing the physics basis of our HTS tokamak to decarbonize global shipping. Breakeven fusion is coming soon, but the first-of-a-kind (FOAK) reactors will be costly, high maintenance, and have low capacity factors, leading to 5–10x higher electricity cost on the grid. Maritime Fusion is pursuing the most widely accepted approach to fusion, the tokamak, but specifically designed for the marine environment where the challenges that arise between breakeven and a commercially viable energy source are alleviated. The market we're targeting requires 15x less power, lower up-time, and costs the same as alternative fuels but without any emissions. Since fusion does not use highly radioactive fuels or materials – unlike fission – we sidestep the vast majority of regulatory challenges and safety risks associated with nuclear energy. Our team brings over a decade of industry and research experience in plasma physics, nuclear engineering, and electrical engineering with backgrounds at SpaceX, Tesla, Columbia University, and University of Pennsylvania.

Optifye.ai does AI performance monitoring for factory workers We place cameras in factories and use vision AI to tell supervisors who's working and who's not in real time. The shop floor has historically been a black box. With Optifye, manufacturing companies can now accurately measure worker output and boost efficiency!

We are making reusable satellites for zero-g manufacturing & Hypersonic Technology Development. Our first orbital recovery mission is set to launch in Q4 2026.

Forge Automation runs software enabled factories that deliver custom metal parts in four days or less. Instead of mechanical engineers having to wait two to three weeks for their low volume CNC parts, they can simply upload a CAD model on our platform and we will ship it out in four days or less. We aim to give mechanical engineers the "Amazon experience" for custom metal parts.

Ultralight is a GPU-accelerated toolkit to embed modern HTML in games and native apps. Available for C and C++ on Windows, macOS, Linux, Xbox, PlayStation, ARM64 devices, and more. Developed in collaboration with leading game studios, built GPU-first on a new, lightweight fork of WebKit.

We make personal AI robots - general-purpose physical machines that anyone can teach with code, language and demonstrations. We believe building open platforms like the first PCs this is the only way to bring the robotics future that we all want, where everyone can participate in the robotics revolution.

Finding new cures requires seeing what and where the abnormal molecules are in a diseased tissue, but current tools see less than 1% of those molecules. Biocartesian combines microscopy and new chemistries to see 50X more, offering unprecedented insights into disease biology and new therapies.

Sensei helps robotics companies scale and outsource their training data collection. Our hardware platform enables the collection of human-demonstration data at a tenth of the cost and twice the speed of current teleop approaches. Our software platform acts like Scale AI for robotics data: a large network of paid human operators use our low-cost collection platform to fulfill data-generation requests.

We’re building the next generation of warehouse robotics. In the US today, retailers spend approximately $10B per year paying human laborers to pick up and move cardboard boxes in warehouses. Existing solutions for automating this are expensive and difficult to install, which is why manual operation is still so prevalent. Our solution is different. We make swarms of small mobile robots that install into existing warehouses to provide a low-cost and robust automation solution for case picking and mixed-SKU palletization. Our novel technology allows these robots to be installed and operate at significantly lower cost than existing solutions while being both flexible and robust.

We build underwater data centers to cut power usage by up to 30%, operating GPUs cheaper and more sustainably. Our 1 MW capsule is being tested underwater in the SF Bay

Zetta builds energy-efficient chips ("XPUs") for AI to accelerate discovery. Our XPUs are highly programmable, letting them optimize the dataflow for each model and making them faster and far more efficient than anything on the market today, saving data centers billions in energy and cooling costs.

Sorcerer is deploying a global network of persistent airborne sensors to fuel advanced weather forecasting. Our high-altitude balloons collect 1000x more data than existing systems and are already used daily by meteorologists to track extreme weather across the US and Central America.

1849 bio designs microbes enabling cheap metal extraction allowing miners to unlock value from low quality copper and gold ores. Surprisingly, the mining industry is one of the largest scale users of biotech in the world with biomining processes accounting for ~1% of global copper production. Biomining is ultra-low cost, running around ~$1/ton of ore vs ~$7/ton for conventional processes. Unfortunately, while biomining is cheap, it can’t be applied to over 80% of copper ores, leaving vast resources without profitable extraction methods. An estimated ~$800B of copper sit today in waste materials and stockpiles with negative unit economics. While a great deal of effort has been spent on optimizing microbial metal extraction processes, very little effort has been spent on optimizing the microbes themselves. To change that, we’re creating new biotech tools and platforms applied directly to the types of biology most relevant to miners. This enables us to develop new microbes and tackle some of the most difficult problems in biomining, unlocking billions in value from unprofitable resources while being more environmentally friendly than conventional processes. We’re world class microbial engineers. We met while doing our PhDs in synthetic biology, where we spent our time applying and developing the most advanced bioengineering technologies to engineer living cells. 

Starcloud is building data centers in space, initially to provide GPU compute to other satellites, and later to address the rapidly growing demand for energy caused by the deployment of AI. Falling launch costs give Starcloud access to abundant energy, radiative cooling, and the ability to rapidly scale in space. In November 2025, just 21 months after founding, Starcloud launched its first satellite, Starcloud-1 to space. This has an Nvidia H100, which is 100x more powerful GPU than has ever been operated in space before. With this, they became the first entity to train an LLM in space and the first to run a version of Gemini in space. Starcloud will soon also run high-powered inference on Capella SAR data on orbit for the first time. Starcloud is launching its second satellite in October 2026, which will have 100x the power generation of the first and generate more cash than it costs to build and launch. Starcloud expects that within 10 years, most new data centers will be being built in space for the energy. In March 2026, Starcloud raised a $170M Series A led by Benchmark, becoming the fastest unicorn in YC history, just 17 months after demo day.

Conductor Quantum is building quantum superintelligence: AI that operates quantum computers to make scientific discoveries beyond human reach. Quantum computers let us understand the world at its most fundamental level, the path to new drugs, new materials, and discoveries no human can reach alone. A quantum computer is the perfect simulator of nature: it encodes the logic of reality into a programmable machine, atom by atom, electron by electron. Give AI that simulator and you open the door to discovery. The bottleneck is operating the hardware. Today, engineers spend days or weeks by hand to bring a chip to operating conditions for just two qubits. A qubit is the information-carrying unit of a quantum computer, the equivalent of a bit in a classical one, and a useful machine needs billions. Removing the human from that loop is only half the problem. Every command an AI sends to a quantum computer must be optimised for the specific hardware it runs on. We build the AI that operates quantum computers and tunes every machine it runs on. That is the path to quantum superintelligence.