Why a Cooling Startup Just Hit a $1.64 Billion Valuation
At $1.64 billion, Frore Systems just became the first thermal management company to reach unicorn status on the strength of functional cooling components alone—and its core products contain microstructures that no CNC machine can produce at scale. San Jose-based Frore Systems secured a $143 million Series D funding round (Company PR / SEC Filing, March 16, 2026), bringing total capital raised to $340 million. The round will scale production of the company's LiquidJet and AirJet cooling platforms for AI data center and edge markets. That this milestone belongs to a thermal company—not a GPU or memory vendor—signals that heat dissipation has become a first-order constraint on AI infrastructure, on par with compute density itself. Parallel material breakthroughs in high-strength aluminum alloys and the deployment of hydrogen-ready turbine components reinforce a broader shift: additive manufacturing is becoming the primary production method for components where internal geometry, not just external shape, determines performance.
What Frore's Technology Actually Is—and What It Isn't
On March 16, 2026, Frore Systems announced the closure of its $143 million Series D round (Company PR / Regulatory Filing). The company produces two core products: AirJet, the world's first solid-state active air-cooling chip, which uses ultrasonic MEMS (Micro-Electro-Mechanical Systems) membranes to generate high-velocity pulsating air jets; and LiquidJet, a direct-to-chip liquid coldplate for data centers that uses 3D short-loop jetchannel microstructures to handle up to 1,400W per GPU—a thermal load class exemplified by NVIDIA's Blackwell Ultra architecture (Tom's Hardware, 2026).
A note on manufacturing process: Frore's products are not produced via additive manufacturing in the conventional AM sense. LiquidJet's micrometer-scale jetchannel architectures are fabricated using semiconductor manufacturing techniques adapted to metal wafers—a process closer to MEMS foundry work than to laser powder bed fusion or directed energy deposition. AirJet similarly relies on semiconductor fab processes, with production operations in Taiwan. This distinction matters: Frore represents a convergence of semiconductor process precision with thermal hardware, not an advance in metal AM per se. What connects it to the broader AM story is the geometric principle—internal microstructures that are physically impossible to manufacture by conventional subtractive means at production volume—and the funding signal it sends about where the functional complexity of industrial components is heading.
The funding aims to expand manufacturing capacity in California and Taiwan to meet demand from AI server infrastructure and high-performance edge devices, where thermal design power targets are outpacing the capabilities of passive vapor chambers and traditional cold plates.
The Geometry Problem That Unites Cooling and Additive
The core reason Frore's jetchannel technology is relevant to the AM industry is structural: it represents the same fundamental manufacturing challenge that drives adoption of LPBF and DED in aerospace and energy—the need to produce internal geometries at scale that subtractive processes cannot reach.
Traditional cooling solutions rely on machining or stamping that cannot maintain the surface-area-to-volume ratios required for the 1,000W–1,400W thermal loads of modern AI accelerators. LiquidJet's jetchannel diameters operate in the micrometer range, versus the millimeter-level channels achievable with conventional CNC. By fabricating these structures through a semiconductor-derived process, Frore achieves coolant delivery precisely mapped to the hot-spot distribution of individual GPUs—something impossible with standard coldplate manufacturing.
This builds on prior work in the AM heat exchanger space—notably Conflux Technology (established 2017) and GE Research (2019–2022)—which demonstrated that LPBF-produced internal channels could outperform machined equivalents in aerospace applications. Frore's distinction is scale and substrate: targeting high-volume semiconductor packaging rather than low-volume aerospace hardware, and using semiconductor fab rather than AM. Both approaches share the geometric logic; the manufacturing method differs by application.
Parallel Signals: Materials and Energy Infrastructure
Complementing this hardware shift is a materials breakthrough that widens the frontier for AM-produced thermal components. Researchers published results in 2026 on a high-strength aluminum alloy for LPBF exhibiting significantly improved strength and thermal stability at elevated temperatures, identified through machine-learning-guided composition screening of large elemental combination spaces. By targeting nanometer-scale precipitate formation, these approaches solve a long-standing trade-off between printability and high-temperature performance in aluminum AM.
In energy infrastructure, Siemens Energy has deepened its operations in Finspång, Sweden, using LPBF to redesign burner and combustor components for hydrogen-fueled turbines. These designs require intricate fuel-air mixing pathways and internal cooling channels achievable only through additive means, to prevent flashback in hydrogen combustion (Analyst Estimate / Industry Case Study). Consistent with trends in computational engineering reported in our previous coverage of LEAP 71's aerospike work, these energy components represent AM shifting from a prototyping method to a primary production route for mission-critical infrastructure.
Desktop Metal has introduced its PureSinter Furnace, a high-purity thermal processing system for sintered metal parts (Company PR). This addresses atmospheric contamination of reactive alloys during debinding and sintering—a secondary bottleneck for scaling binder jetting and other bound-metal AM processes where high-purity copper or specialized tool steels are required for thermal components.
Where the Value Is Migrating
The Frore funding round is most useful as a market signal rather than an AM milestone directly. It confirms that investors now place unicorn-level value on the ability to manufacture complex internal geometries at production volume—regardless of the specific process used to achieve them. For the AM industry, the implication is that the competitive frontier has moved inward: from the external shape of a part to its internal function.
We project that the next 24 months will see acceleration in thermal-first designs, where the cooling architecture is the primary constraint around which the rest of the component is engineered. This maturation reaches a threshold where AM—and adjacent precision fabrication methods—compete not on price-per-part against CNC, but on physical possibility: they are the only manufacturing technologies capable of meeting the thermal requirements of the AI era.
Counter-Signal & Risks: The growth of LiquidJet and similar functional platforms assumes sustained high yield rates at production volumes not yet demonstrated outside pilot runs. A critical failure mode is channel clogging in sub-millimeter geometries under real-world coolant conditions where purity cannot be perfectly controlled. For AM-produced thermal components more broadly, the reliance on specialized high-strength aluminum alloys requires stabilization of the metal powder supply chain, which faces volatility in raw materials like scandium and zirconium. If yield rates do not stabilize by Q4 2026, the industry may retreat toward hybrid solutions combining traditional cold plates with simplified AM inserts.