Y Combinator’s September 2026 Demo Day put a spotlight on startups working in areas that once sounded like science fiction: floating nuclear reactors, brain-cell computing, humanoid robots, and heavy-lift machines for space. The deals also show where investors are placing big bets, from clean energy and AI infrastructure to defense technology and automation.
Bluecore Energy drew one of the largest headlines after raising a $50 million seed round led by Silverton Partners. Slauson & Co., Harlem Capital, Black Angel Group, and HartBeat Ventures also invested. The company is developing compact, water-cooled small modular reactors designed to operate aboard floating barges, with the Port of Long Beach as its first target.
The idea connects two major trends: the push for zero-emission energy and the search for new power sources for industrial sites and data centers. Bluecore is working with the Nuclear Regulatory Commission and the U.S. Coast Guard as it pursues certification. Kofi Asante described the company’s goal in simple terms: “Our focus is simple. Create and deliver zero-emission energy as safely and quickly as possible.”
Nuclear power and clean technology attract fresh capital
Bluecore’s funding arrives during a strong period for nuclear startups. Nuclear fission companies pulled in roughly $2 billion in venture funding in 2025, and investors have continued writing enormous checks to nuclear startups in 2026. Across the wider cleantech, electric vehicle, and sustainability market, startups raised about $15 billion in the first half of 2026.
Second-quarter funding reached its highest quarterly level since 2024. That momentum gives companies like Bluecore room to tackle long certification timelines, complex hardware, and the challenge of proving that a new energy system can work outside a laboratory.
Atomarine is pursuing a different version of the same floating-power idea. The startup, whose co-founders come from MIT, plans to launch a gas-powered pilot by 2028 before transitioning to floating nuclear power ships in 2032. Atomarine claims to have secured over $4 billion in customer interest through letters of intent.
Those plans show how startups are trying to build a bridge between current technology and a more ambitious future. Atomarine’s gas-powered pilot comes first, followed by the planned move to nuclear power ships. The company’s customer interest suggests that buyers are already considering the concept, even though the nuclear phase is still years away.
New hardware for AI, robots, and defense
Several other startups are attacking the infrastructure behind AI. Dipole Labs built an optical switch that skips the conversion from light to electricity, with the goal of improving energy efficiency in AI data centers. The company’s work focuses on the network hardware that moves information, rather than on the AI models themselves.
Lamb Labs is taking a different approach to AI chips. The company is building custom inference chips called “Model Processing Units,” or MPUs, with hardcoded AI model weights. Its co-founders come from Imperial College London and Oxford, and the design aims to place specific model information directly into the hardware.
Parasma is exploring an even more unusual computing path: using human brain cells as an energy-efficient alternative to AI hardware. Together, these startups point to a race that extends beyond larger models. Companies are also looking for new ways to cut power use and build machines that handle AI workloads with less energy.
Isengard Industries is already generating $10 million in revenue and aims to mass-produce jet-powered attack and counter-drones within allied countries. That puts the company at the intersection of manufacturing and defense technology, where startups are working to produce both offensive and defensive unmanned systems.
Robots leave the lab and move into daily work
Praxis Robotics is building robots for practical tasks and has captured video data in more than 150 environments. The startup is working with publicly traded companies and has recorded almost half a million in sales. Its humanoid robot costs around $1,600 and can help with home tasks such as cleaning and folding clothes.
That price places the robot in a different conversation from large industrial machines. A device built for household work must handle varied spaces and changing tasks, which makes the real-world video collected by Praxis an important part of its approach. The company’s work with publicly traded companies also gives it a path beyond the home.
Waddle Labs is focused on the software layer that controls robots. It uses a layer of large language model agents to write control code, allowing users to give robots natural language commands while the system generates code on its own. Waddle claims developers can tell robots what to do in about 20 minutes using its API.
That approach could make robot programming more accessible. Instead of writing every instruction by hand, developers describe a task and let the system create the control code. The promise depends on how well robots follow those instructions, but Waddle’s goal is clear: shorten the path from an idea to a working machine.
From solar installation to Mars exploration
Cosmic Robotics is building heavy-lift robots and has $25 million in contracts through 2027. The company is installing solar panels across the United States while also aiming to begin an exploratory mission to Mars by 2028.
That gives Cosmic two very different operating targets: work on Earth today and space exploration later. SpaceX is listed as a competitor in Mars exploration, placing Cosmic in a field where ambitious missions must share attention with practical commercial work.
Taken together, these startups reveal how broad the current deep-tech funding wave has become. Investors are backing floating power systems, specialized chips, optical networks, autonomous robots, drone manufacturing, biological computing, and machines built for space. The common thread is not one product category, but a willingness to fund difficult hardware and software projects that aim to change how energy, work, and exploration happen.
Based on
