Hardware and AI agents draw serious capital this week. Four rounds totaling over $500M closed in days, with sovereign funds, chip giants, and top-tier VCs all moving toward physical AI and deep hardware. Chip giants are now writing checks. Nvidia's name appears in two of this week's major rounds, and that pattern is worth examining closely. As SiliconANGLE reported, Nvidia and Samsung co-invested in Nous Research's $90 million Series B, led by Robot Ventures with Y Combinator also participating [[1]]. The round values Nous - the developer of the Hermes AI agent - at $1.5 billion. Separately, as TechCrunch and Unite.AI both reported, Nvidia joined Qualcomm Ventures, Samsung, and Microsoft's M12 fund in Mecka AI's $60 million Series B, led by Sequoia [[3]] [[5]]. When semiconductor companies invest in software and data startups rather than simply selling to them, it signals that they are trying to lock in the application layer early. For founders raising in AI infrastructure or physical AI, a strategic check from a chip manufacturer now carries a different weight than it did two years ago - it can function as both capital and a distribution signal to other investors. Robot data infrastructure attracts institutional conviction. Mecka AI's round is notable not just for its size but for its breadth. As Unite.AI reported, the $60 million Series B drew new investors including Nvidia, Qualcomm Ventures, Samsung, and M12, alongside existing backers Kindred, Framework Ventures, and Neo [[5]]. The company collects and analyzes human motion data to train humanoid and other robots - a category that sits beneath the more visible robotics hardware plays but arguably enables them. The $500 million valuation reported by TechCrunch suggests the market is pricing robot data infrastructure comparably to earlier enterprise SaaS rounds, despite the field being younger and more capital-intensive [[3]]. For founders building data pipelines, labeling systems, or simulation environments for physical AI, this round provides a useful valuation anchor when talking to investors about comparable companies. Stealth and patience paid off for atomic Machines. Atomic Machines emerged from six years of stealth this week with $250 million already raised and a product - the PrimeSwitch PS-150 relay - already shipping to early-access data center customers, as Startup Fortune reported [[2]]. The company, led by Jeff Holden, who built Amazon Prime and later served as Uber's first chief product officer, is working on what it describes as a compiler for matter: manufacturing at molecular scale. The structure of this raise carries a lesson. Six years of stealth and $250 million in capital before a public announcement is not a path most founders can or should follow, but the outcome illustrates how investors at the frontier of deep hardware are willing to wait for proof of shipping product before the world knows a company exists. If you are raising for hardware or manufacturing, investors will increasingly want to see something physical in a customer's hands, not just a prototype in a lab. Quantum computing reaches a Series A milestone. Universal Quantum closed what it describes as the largest Series A ever raised by a quantum computing company - over $100 million - backed by a syndicate that includes sovereign funds, institutional investors, and quantum-specialist funds, as the company announced directly [[4]]. The UK and Germany-based company is focused on trapped-ion systems designed for real-world scale. The sovereign fund participation is a detail founders in deep tech should register. Government-linked capital is increasingly filling the gap at the pre-commercial stage of hardware and science-based startups, where commercial venture capital has historically been reluctant to commit. If your startup sits in quantum, advanced materials, or similarly long-horizon categories, a sovereign or national innovation fund may be a more realistic lead investor than a traditional VC at the Series A stage. This week, if you are raising. * If you are raising in AI agents or robot infrastructure, identify whether Nvidia, Qualcomm Ventures, or Samsung have a strategic investment mandate that aligns with your category - strategic checks from chip companies are now influencing how other institutional investors read a cap table. * Hardware and deep tech founders should treat a shipping product in a customer's facility as the minimum credible proof point before approaching institutional investors, regardless of how early your round is. * If your startup has a long path to commercial revenue, research whether a sovereign wealth fund or national innovation fund operates in your sector, as this week's Universal Quantum round shows they are actively leading rounds that traditional VCs may not. Sources. ai agents robotics deep tech startup funding

Atomic Machines emerges with $250M to build micro-machines from code. Jeff Holden's Atomic Machines spent six years in stealth. Its first device, PrimeSwitch, is a relay for AI data centres that the company says opens in 50 microseconds. October 7, 2026 - 1:15 pm Atomic Machines has come out of six years in stealth with $250M raised. It says its system builds working micro-machines straight from code. The Bay Area start-up announced the launch on Wednesday. Its first product is a power switch for AI data centres. Founder and chief executive Jeff Holden joined Amazon as its tenth engineer and built Amazon Prime. He then became Uber's first chief product officer. There, he set up its self-driving, AI labs and air-taxi units, according to the company. "3D printing made shapes programmable: any geometry, straight from a file. We're making machines programmable," Holden said. What the Matter Compiler does. The company calls its system the Matter Compiler, after a machine in Neal Stephenson's novel The Diamond Age. It says code alone decides which device the system builds. The machines it makes have moving parts, several materials and features down to single-digit microns. No moulds, masks or other fixed tooling are needed. The system measures and corrects every step in real time, and its AI runs its own experiments on the hardware, Atomic Machines said. Targets include gears the size of a grain of sand and robots small enough to work inside the body. Another is cooling built into AI chips. Conventional manufacturing hits its limits at a few tens of microns, the company said. Below that, chipmaking tools take over. Each new device made with them, known as MEMS, needs years of custom process work. Sweden's Silex Microsystems runs a MEMS foundry business of that kind. Holden explains why he started the company in an essay on its website. He compares the system to the ribosome, which builds proteins from a genetic code. The first product. The first device is PrimeSwitch PS-150, a relay that opens in 50 microseconds. The company says that is about 1,000 times faster than a conventional contactor. It carries 150 amps of continuous current. The package measures 9.5mm across and 3mm thick. The switch is meant for AI data centres that are moving to 800-volt direct current. At those voltages, a fault can release destructive energy in microseconds, the company said. Mechanical protection takes milliseconds to open. PrimeSwitch is shipping to early-access customers for evaluation. Atomic Machines did not name them. It will show the switch in public for the first time at the Open Compute Project Global Summit in San Jose, from 12 to 15 October. Data centre power is drawing other start-ups too. Emerald AI raised $150m in August for software that eases the load on grids. Who is backing it. Investors include OneIM, Gigafund, XTX Ventures, KAS Venture Partners, Valor Equity Partners and Sozo Ventures. The Regents of the University of California also invested. The company did not give a valuation or say how the $250M was split across rounds. "If vibe coding is how software gets made now, this is the start of vibe manufacturing," said Luke Nosek, managing partner at Gigafund. In vibe coding, people describe software in plain language and AI writes it. The trend has already produced start-up unicorns.