The Next Limit to Moore’s Law is Underground

This is the final part in our series exploring the broader consequences of lowering the cost per kilogram of access to deep space. Resource extraction is our mission, and the technologies and operational capabilities required to achieve it have applications that extend far beyond mining itself. Read the first post here, the second post here, and the third post here.

For more than three decades, semiconductor progress has been defined by shrinking transistors. Today, a leading-edge chip can contain billions of these ultra-fast transistors, which switch at extraordinary speeds. These transistors are connected by a dense, complex network of metal wires, called interconnects, that route signals and power throughout the chip. Copper has been the industry standard material. 

Now, as engineers push semiconductor designs to atomic scale, copper is approaching a physical limit. At the smallest interconnect layers, chipmakers are preparing for a major materials transition. The leading candidate is an ultra-rare platinum group metal: ruthenium.

Ruthenium is one of the rarest industrial metals on Earth, with global production capacity at only 42 metric tons per year. But if the world’s leading semiconductor foundries shift to this metal, Moore’s Law will no longer be limited by human ingenuity, but Earth’s elemental supply. 

Which means we need to look beyond Earth.

But for most of the history of spaceflight, deep space has been prohibitively expensive to reach. Missions beyond Earth orbit have been rare and government-led, typically requiring bespoke spacecraft costing hundreds of millions, or even billions, of dollars. That cost structure has kept commercial activity in deep space close to zero.

AstroForge is working to change that equation.

We are building low-cost, mass-efficient spacecraft designed to drive down the cost per kilogram of accessing deep space enough that missions once reserved for governments and one-off scientific programs can become repeatable commercial infrastructure. 

When the cost of access falls, new markets become possible. That pattern has repeated across other technological and geographic frontiers, and we believe deep space will be no different. One of the first opportunities is creating a new source of critical materials.

An alternative to copper

Copper has been the workhorse metal for semiconductor interconnects since the late 1990s. But as interconnect dimensions shrink, it is starting to hit some physical bottlenecks that make it increasingly difficult to use, especially as interconnects shrink to sub-2nm nodes.  

There are several issues. Copper must be surrounded by barrier and liner material, like Tantalum Nitride, to prevent it from diffusing into surrounding dielectric material. At 10nm linewidth, those layers occupy more than 50% of available wire space. In addition, microscopic copper generates huge electrical resistance, with interconnects now accounting for roughly 75% of an advanced chip’s internal signal delays. Current density creates yet another problem. Forcing current through tiny copper wires can generate immense internal heating and thermal management challenges, which physically degrade the chip over time. 

Ruthenium has emerged as the replacement at the 1 nm (10Å) nodes. Unlike conventional copper interconnects, ruthenium doesn’t need a protective barrier, which means more of the available interconnect volume can conduct current. It also performs well at atomic scale and has strong electromigration reliability. This enables more power-efficient transistor architectures like CFETs. 

Today, companies including TSMC, Intel, IBM Research and Samsung are developing ruthenium-based interconnects, though researchers recognize that the transition will not happen overnight. 

There’s just one problem: ruthenium is exceptionally rare. Part of the platinum group metals, its natural abundance in the Earth’s crust in extremely low concentrations. It is generally recovered as a byproduct during PGM refining, rather than from mines built specifically to produce ruthenium, and for that reason the supply is incredibly inelastic: mining companies cannot simply mine more Ruthenium without displacing millions of tons of Earth. 

Of the 42 metric tons recovered per year, about 35 metric tons is in South Africa, with smaller amounts in Russia and Zimbabwe. The metal is now included on the U.S. government’s critical minerals list due to the economic consequences of potential supply chain disruptions.  

This is not a problem that can be solved with better software or more advanced lithography. Nor does it mean the world is at an imminent threat of running out of ruthenium. But it does mean that this metal, currently produced at speciality-material scale and refined in primarily geopolitically sensitive countries, could become the constraint that prevents the next fab advancement. 

Lowering the cost of reaching a new resource base

The world needs a fundamentally new supply of this material. The Solar System contains a vast amount of PGMs, including ruthenium, often in concentrations that can exceed those found in terrestrial ores. Iron meteorites, the fragments of shattered planetary cores, suggest some metallic asteroids contain PGMs in concentrations orders of magnitude above the grades of working mines. Today, those resources are effectively irrelevant to the conversation because they are too difficult and expensive to reach. 

AstroForge is building specifically to drive that cost down. We are developing a low-cost, mass-efficient spacecraft that can fit on a rideshare manifest and operate throughout the Solar System, for a fraction of the historical cost per kilogram of mass. Our next mission, DeepSpace-2, will rendezvous with a metallic asteroid, and we aim to return asteroid material before the end of the decade. 

The semiconductor industry has spent decades overcoming successive physical limits to keep computing progress moving forward. If the next constraint is not transistor architecture, but raw material access, the solution must come through a new resource base. That is how we will ensure the next breakthrough in computing will not be limited by Earth’s geology.