Interview

Mariana Minerals is building the fully autonomous copper mine: 220 employees, two projects, and a plan for 10 mines in 10 years

May 1, 2026 with Turner Caldwell

Key Points

  • Mariana Minerals, founded in 2024 by Tesla's former minerals chief Turner Caldwell, is building autonomous copper and lithium operations by integrating disconnected mine data into unified world models that train reinforcement learning agents to direct haul trucks, drill rigs, and other equipment.
  • The company estimates that optimizing existing mine logistics and equipment utilization alone could double productivity by cutting idle time, which accounts for up to half of diesel consumption on heavy haul machinery.
  • Mariana operates 220 people across two active projects and plans to expand into nickel, cobalt, uranium, graphite, and rare earths across 10 projects in 10 years, betting that falling ore grades and deeper deposits will make autonomous mining increasingly critical.
Mariana Minerals is building the fully autonomous copper mine: 220 employees, two projects, and a plan for 10 mines in 10 years

Mariana Minerals is trying to build the fully autonomous copper mine. Turner Caldwell, the company's founder and CEO, started it in 2024 after nearly a decade at Tesla, where he ran the minerals and metals team and built Tesla's lithium refinery in Corpus Christi. That background is doing real work here — the company sits at the intersection of critical minerals supply and autonomous systems, two areas where Caldwell has direct operating experience.

What the company is building

The core problem in mining is coordination. Geologists, mine planners, operations crews, and refinery teams all work from disconnected data sources — ore compositions arrive at the refinery on a piece of paper. Mariana's approach starts by pulling all of those data feeds into a single integrated architecture, then building world models of the mine and refinery, and using those to train reinforcement learning agents that eventually direct autonomous equipment: haul trucks, drill rigs, loaders, dozers, graders, water trucks. Boston Dynamics Spot robots handle undercarriage inspection and thermal sensing in areas that are too hazardous for routine human access.

The goal is full closed-loop autonomy between the world models, the RL agents, and the physical assets.

We just restarted a copper mine that we acquired at the end of last year with a major focus on autonomy across the entire stack... I think the first phase is gonna double productivity through just better orchestration and better logistics management... We raised a seed round in August. We're at 220 people now across the company. We have two projects.

Two-phase efficiency thesis

Caldwell frames the efficiency gains in two stages. The first is straightforward optimization of the existing architecture — better logistics, better equipment utilization, and cutting idle time. Up to half of diesel consumption on heavy haul machinery comes from idle time in the pit. He estimates this phase alone could double productivity.

The second phase is more structural. Today's mines are designed around human labor, which has pushed the industry toward ever-larger equipment to reduce labor intensity per ton moved. Autonomous systems flip that logic. Smaller swarm mining equipment becomes viable, roads get narrower, and mines can be more selective about what ore they extract. That selectivity improves recovery rates across the whole pit.

Current scale

The company is 220 people, raised a seed round in August 2024, and is running two projects simultaneously. The first is a copper mine — described as a first large-scale commercial mine — that Mariana acquired and restarted at the end of 2024. The second is a lithium refinery in East Texas extracting lithium from oil and gas wastewater. Caldwell started with lithium, not copper, deliberately timing entry when lithium prices were near a trough. The same logic applied to copper — Mariana moved on the asset before the recent price run, when labor and equipment were cheaper and lead times were shorter.

Commodity diversification

Copper prices have roughly doubled since the company was founded, which makes the near-term economics favorable, but Caldwell is deliberate about not building a single-commodity business. The plan is to expand into nickel, cobalt, uranium, graphite, and rare earths across 10 projects in 10 years. The structural case for copper remains strong regardless of short-term price moves: ore grades are falling, deposits are getting deeper, and mineralogies are getting more complex — all of which will make copper harder and more expensive to produce over time.

The 220-person headcount, two active projects, and a 10-mine target inside a company that is barely a year old is an aggressive pace. Whether the autonomy stack matures fast enough to justify that expansion rate is the question the next few years will answer.

Every deal, every interview. 5 minutes.

TBPN Digest delivers summaries of the latest fundraises, interviews and tech news from TBPN, every weekday.