The 9 buzziest startups from Y Combinator’s latest Demo Day, according to VCs

The Y Combinator batch stands out for its concentration of deep tech startups, a departure from the accelerator’s historical focus on software and consumer applications. This shift aligns with broader industry trends where hardware, particularly in AI, energy, and robotics, is increasingly seen as a critical bottleneck. For engineers, the startups’ technical approaches offer a preview of where infrastructure may evolve: floating data centers to bypass land-use restrictions, optical networking to eliminate power-hungry conversions, and custom chips to sidestep memory bandwidth limits.

The startups’ value propositions hinge on solving specific, measurable inefficiencies. Dipole Labs’ optical switches, for example, target the energy and latency costs of converting data between electrical and optical signals in AI clusters. Lamb Labs’ hardcoded inference chips aim to eliminate the memory-bandwidth bottleneck that plagues traditional GPUs. These solutions are not incremental; they propose rearchitecting core components of data center infrastructure. However, adoption will depend on whether the performance gains justify the cost of integrating proprietary hardware into existing systems.

Defense and robotics startups in the batch highlight a growing intersection between commercial and military applications. Isengard Industries’ jet-powered drones and Cosmic Robotics’ heavy-lifting robots are positioned as dual-use technologies, with revenue already tied to defense contracts. For engineers, this signals a potential expansion of career opportunities in hardware development outside traditional tech hubs. The focus on locally producible hardware also suggests a shift toward supply chain resilience, a priority for both defense and commercial sectors.

The startups’ ambitious timelines and revenue claims, such as Atomarine’s $4 billion in letters of intent or Nori’s $1,600 humanoid robot, reflect the high-risk, high-reward nature of deep tech investing. While these projections may attract VC capital, they also underscore the technical and regulatory hurdles ahead. Floating nuclear data centers, for instance, will face scrutiny over safety and environmental impact, while autonomous robots must prove reliability in unstructured environments. Engineers evaluating these projects should weigh the technical feasibility against the market’s appetite for disruption.