Posted March 4, 2019

I’ve shared a lot about a future enabled by mobile — covering how mobile changes product design, app monetization, and business models to social communication and more. But what we don’t talk about enough in the software industry is the hardware — in particular the hard, physical, raw materials like cobalt — that enable mobile to eat the world, for better or worse. That’s why I’m excited today to share our investment in a very different kind of company for both a16z, and for me.

Much has been written about the raw materials and components in smartphones and other devices, from drones to electric vehicles (what others have called the “peace dividend of the smartphone wars”). But materials — and especially mining — has been an impenetrable area for software. Yet that’s not as true of other sectors, like the oil and gas sector, which is much more advanced in using software. Software for materials mining and especially cobalt hasn’t advanced much at all, and certainly not in a way that startups can help… until now.

The story of cobalt

Cobalt — first discovered by a Swedish chemist in 1739 due to its ability to turn glass blue — is a transition metal. As suggested by its name, it was first used to color glass and ceramics in everything from pottery to tombs. Given no major obvious application beyond these, cobalt was traditionally extracted as a byproduct of copper and nickel mining… never really a focus of mining itself.

That changed, however, as we began to manufacture electrical devices on mass scale for a massive base of people using portable devices that require recharging — such as laptops, mobile phones, laptops, electric bikes, and most recently, electric vehicles. Your smartphone probably has about $1 worth of cobalt; your laptop about $10 worth; and if you drive an electric vehicle, the battery has roughly $1,000-$1,500 worth of cobalt in it.

But this is just the beginning. As electric vehicles become more ubiquitous and more widely available to more people thanks to the same factors that drove costs down for every other electronic device that came before, there is even more demand for cobalt than before. Because cobalt is a superior material for the battery cathodes that power all those devices.

The physics that helped material scientists fully understand why cobalt is a superior metal for batteries was only established in the past 15-20 years. What they found is that cobalt beat other transition metal candidates such as nickel and magnesium on charging rate, energy density, and cycle life — the very chemical properties we need for the batteries that power our modern technologies.

All of sudden, this once neglected element is suddenly in very high demand worldwide.

Software eats mining

Many industry analysts estimate annual EV sales to exceed 10M by 2025. There’s a supply crisis looming, and we’re already seeing it play out in the commodity markets.

Over 98% of the 110 kilotons of globally produced cobalt a couple years ago was a byproduct of copper and nickel mining. That byproduct supply was sufficient to satisfy historical global demand. But as the production of electric vehicle batteries rapidly accelerates, refining existing sources will not be enough.

The irony of all this is that cobalt is not a rare metal. In fact, it’s more common in the earth’s crust than even lead or tin, and as common as nickel. Our inability to efficiently find cobalt is largely due to the current concentration in supply (68% is mined in the Democratic Republic of the Congo, a conflict region suffering from political instability and extremely dangerous mining conditions). And unlike in the oil & gas industry, there simply wasn’t enough demand to warrant technology development that helps identify where and how to find more of it.

Mineral exploration today is a largely manual process where skilled geologists use various heuristics to identify patterns in maps and subsurface data. While expert judgement is essential to this, the manual nature of the process makes searches slow (and inefficient), seemingly random (not systematic), and very blunt (not able to draw on subtle patterns).

Enter KoBold Metals. By building a digital prospecting engine — full stack, from scratch — using computer vision, machine learning, and sophisticated data analysis not currently available to the industry, KoBold’s software combines previously unavailable, dark data with conventional geochemical, geophysical, and geological data to identify prospects in models that can only get better over time, as with other data network effects. There is also growing research literature establishing these methodologies for mineral prospectivity, but KoBold Metals — like other startups that are more able to take ideas from academia to industry — is the first to bring this technology to market. More importantly, such sophisticated software-defined prospecting techniques can be applied to the entire universe of mineral exploration… and beyond.

What excites me about this — much like the online to offline/ offline to online trends I’ve previously covered in depth — is that such startups represent a software overlay over the physical world, and in the most tangible, raw way possible. We’ve already seen software start to eat the world in previously hardware-only — or otherwise entrenched industries — such as transportation, finance, education, and more. Now it’s for mining.

That’s why I’m excited to announce Andreessen Horowitz’s investment in KoBold Metals and especially its founding team. KoBold Metals brings together all the industry and domain know-how — in natural resources, the mining industry, geology, geochemistry, geophysics — required to attack this problem. But the team also brings a unique technological focus and depth not typically available to the industry, even if the players tried building the software themselves. CEO Kurt House has traveled deeply through the idea maze here, having founded multiple ventures at the intersection of technology and natural resources; he also teaches about entrepreneurship in energy as an adjunct professor at Stanford. COO Jeff Jurinak is a 35-year veteran of exploration, who previously served as the Chief Reservoir Engineer at ConocoPhillips. CTO Josh Goldman brings experience in both big data and machine learning as well as corporate strategy after multiple years at McKinsey’s Natural Resource Practice, and after earning a PhD in atomic physics from Harvard.

I’m also personally excited about this investment as someone who came to the venture capital industry 8 years ago: Centuries ago, queens (and kings) funded expeditions to explore new lands and resources that later created significant business opportunities… not to mention changed the shape of the world forever. This investment is a throwback to that old world, wrought new. Extracting cobalt in a better, faster, safer way could mean greater access to electronics for us all.