Mariana Minerals raises $310M Series B led by Khosla Ventures

Something interesting is happening at the intersection of AI hype and industrial reality. According to Fortune, Mariana Minerals has closed a $310 million Series B led by Khosla Ventures, with participation from Andreessen Horowitz, Breakthrough Energy Ventures, Greenoaks, StepStone Group, BHP Ventures, Mitsubishi Corporation, and Greycroft. That brings total capital raised to $400 million at a $1.5 billion valuation, and it puts a two-year-old startup in the awkward and interesting position of being a software company that is also, literally, a copper mine.

The pitch from CEO and cofounder Turner Caldwell is that we're entering what he calls a "metals-driven economy" — the data centers, chips, grid, and EVs everyone talks about all trace back to a small set of critical minerals, and copper is the one nobody has a good substitute for. Caldwell spent nine years on factory design and construction at Tesla before starting Mariana in 2024 with cofounders Baker Tilney and Juan Lozano. That background matters because the company's core claim is a factory-style playbook applied to mining: acquire an idled site, layer in autonomous software, and restart it faster than a conventional operator could.

Copper One, in Utah, is the proof point they're leaning on. Mariana bought the previously idled mine in 2025 and, per the Fortune reporting, restarted it within four months, targeting 50,000 metric tons of refined copper annually. Lithium One in Texas broke ground the same year and is expected to reach commercial production in 2027. The stated goal, in Caldwell's words, is to "reduce the cost of these core inputs to the modern economy over time."

The honest caveat is that the piece is largely a fundraise announcement, and what it does not give you is verified production numbers, an independent read on unit economics for the autonomous software, or a clear picture of how much of that 50,000-ton target Copper One is actually hitting today. Autonomous mining also has a long history of overpromising. Permitting and construction risk on the 2027 lithium timeline is real. Take the specifics as reported, not settled.

The forward-looking part is who is on the cap table. BHP Ventures and Mitsubishi are strategic checks from the incumbent metals world, and the price signal Caldwell keeps hitting, that downstream customers face cost pressures when copper gets expensive, is exactly the sort of thing hyperscalers, EV OEMs, and defense buyers will pay to hedge. If a software-first operator can meaningfully bend the supply curve on even one critical mineral, this stops being a curiosity and starts being a template.