Every founder I meet building the next generation of AI infrastructure, whether it’s data centers, robots, drones, electric aircraft, autonomous fleets, or the power systems behind all of it, eventually hits the same wall. They are trying to build the future on top of materials that were never designed to carry it.

That is why I led Initialized’s investment in Arcturus, which is emerging from stealth today with $8M in seed funding, joined by Toyota Ventures, Breakthrough Energy Discovery, 1517, and Wireframe Ventures.

EVERY LAYER OF THE AI STACK BOTTOMS OUT AT MATERIALS

It is easy to look at the AI buildout and see software. Easier still to see GPUs. But underneath the chips, models, and agents is something far more boring and far more limiting: copper and aluminum. They are in every motor, transformer, transmission line, bus bar, heat sink, and winding the modern world runs on. And those metals – the literal substrate of the electrified, AI-powered world we are trying to build – have not fundamentally changed since the 1800s.

We are asking 19th-century materials to carry a 21st-century load. They cannot. The grid needs to roughly 10x over the next decade to keep pace with AI compute and electrification. The U.S. already loses about 15% of electricity to transmission resistance, roughly $100B wasted every year ($260B globally). Hyperscalers are building their own nuclear plants because they no longer trust the grid. Data centers are pouring billions into cooling because the metals carrying their power throw off too much heat. EVs and electric aircraft are weight-constrained. Drones and humanoid robots are runtime-constrained. Every one of these gets treated as a separate engineering problem: thermal, packaging, power density, grid resilience. They are the same problem underneath: the material in the middle is the bottleneck.

You cannot software your way out of a materials limit. You cannot build the AI economy on top of materials that were not designed for it.

WHAT ARCTURUS IS BUILDING

Arcturus is developing a new class of metals: copper and aluminum infused with carbon nanomaterials like graphene and carbon nanotubes, using a proprietary laser fabrication process. The output is a drop-in replacement for the wire and conductors already in motors, heat sinks, and infrastructure today: same gauges, same form factors, no system redesign required. Early prototypes have demonstrated step-change improvements across electrical conductivity, thermal conductivity, and mechanical strength, all at once. Most alternatives trade one property for another. None are drop-ins.

In practice that means motors that push more power without overheating, heat sinks that pull more thermal load out of dense GPU racks, drone motors with longer flight times, and eventually transmission lines that carry far more current with far less waste.

The field has been trying to make carbon-nanomaterial wire for decades. Science was never the blocker. Manufacturing was. Getting uniform dispersion at the nanoscale, preserving the carbon-metal interface through processing, holding properties at production length: every team that has tried has hit a wall. 

Arcturus’s founder, Amir Mashal, has spent more than a decade at the rare intersection of disciplines needed to crack it: electrical engineering, materials science, and carbon nanomaterials. Most nanomaterials researchers cannot manufacture, and most manufacturers do not work with nanomaterials. Most electrical engineers do not know either. Amir is fluent in all three, first at UW–Madison and then as a Principal Investigator at HRL Laboratories running aerospace and automotive R&D programs for GM and Boeing. His laser infusion process is the only approach we have seen that improves conductivity, thermal performance, and mechanical strength simultaneously while remaining a drop-in.

WHY I BELIEVE IN AMIR

Amir is obsessed, strategically clear, and willing to do the unsexy work. He built the first prototype laser system in his garage, fabricated the wire that produced Arcturus’s performance data, filed a PCT patent on the process, and ran technical validation conversations with Boeing, GM, Google, ARPA-E, and EPRI. Here’s what I came up with.

He was selected as a 2024 1517 Flux Capacitor Fellow, a 2025 Breakthrough Energy Discovery Fellow, and a National Science Foundation Graduate Research Fellow. Arcturus was also a 2024 ARPA-E Vision OPEN finalist. Each of these recognitions involves months of independent technical diligence. Arcturus holds two patents with three more pending. The garage was a starting point, not where the work has stayed.

He also understands how most materials companies die: on the wrong beachhead. They muscle into regulated, price-sensitive markets like utility-scale grid infrastructure on day one and run out of capital before they ever ship a contract. Amir has chosen the opposite path. Start in unregulated, performance-constrained, price-insensitive markets where customers will pay for better materials today: drone and robotics motors first, then high-performance heat sinks for AI data centers, then EV motors, electric aircraft, and bus bars. Grid infrastructure comes last, after the technology is proven and the manufacturing is mature. Materials companies are won by getting material out of the lab and into real products.

A MATERIALS PLATFORM, NOT A SINGLE PRODUCT

The closest analogy I have for what Arcturus can become is Applied Materials in its earliest days: one materials-process advantage that solved a semiconductor bottleneck, expanded into a sprawling portfolio that came to underwrite an entire industry. The laser infusion process is Arcturus’s equivalent wedge, a manufacturing capability nobody else has demonstrated that opens directly into multiple downstream markets.

Learn more about Arcturus at arcturus.io.