Skip to content
a16za16z

The U.S. Can’t Build AI Without These Materials

It can take more than 15 years to permit and build a new mine in the United States - yet nearly every modern technology we rely on, from smartphones to fighter jets to AI data centers, depends on a steady supply of critical minerals. In this episode, Erik Torenberg is joined in the studio by Turner Caldwell, founder of Mariana Minerals, along with American Dynamism general partner Erin Price-Wright and partner Ryan McEntush. Turner spent nearly a decade at Tesla, working his way upstream from factory design to battery materials and mining. Now, he’s building a new kind of mining and refining company - vertically integrated and software-first- designed to meet the demands of our industrial future. We get into why the industry is so broken, what it actually takes to turn rocks into usable materials, and how the U.S. can rebuild its capacity to mine, refine, and manufacture the things that matter most. Timecodes: 00:00 Introduction 00:53 The Importance of Critical Minerals 01:54 The Mining Process Explained 04:05 Challenges in the Mining Industry 05:46 Career Paths in Mining 06:40 Personal Journey and Insights 11:16 Tesla's Vertical Integration 12:47 Geopolitical & Market Dynamics 16:24 Technological Innovations in Mining 20:47 Industry Challenges & Opportunities 28:19 Mariana's Product & Thesis in Construction and Mining 29:30 Leveraging Technology in Mining 32:02 Automating Chemical Processing & Optimizing Refining Operations 35:25 Challenges in Scaling & Commissioning Refineries 36:55 Deciding What to Build & Where to Partner 38:05 Supply Chain & Commercial Deployment of New Technologies 40:20 Venture Capital & the Mining Industry 42:59 Critical Minerals & Their Importance 47:45 Permitting & Regulatory Challenges in the US 53:46 International Strategy & Future Goals 54:44 Conclusion: Rebuilding Infrastructure & Capability Resources: Find Turner on X :https://x.com/tbc415 Find Erin on X: https://x.com/espricewright Find Ryan on X: https://x.com/rmcentush Stay Updated: Let us know what you think: https://ratethispodcast.com/a16z Find a16z on Twitter: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Subscribe on your favorite podcast app: https://a16z.simplecast.com/ Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures.

Turner CaldwellguestErik Torenberghost
Jul 23, 202556mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 1:55

    Why critical minerals are the hidden foundation of modern life (and AI)

    The hosts frame critical minerals as the physical prerequisite for everything from consumer electronics to defense—and increasingly, AI infrastructure. The conversation sets the stakes: refining and mining are usually invisible to consumers, but now sit at the center of geopolitics and hard tech innovation.

    • Critical minerals show up in everyday devices (phones, headphones, screens, laptops)
    • AI growth increases demand for electrification and infrastructure metals
    • Mining/refining are background processes that need to be understood publicly
    • The topic sits at the intersection of geopolitical urgency and technology
  2. 1:55 – 4:06

    From exploration to engineered materials: how rocks become batteries and magnets

    Turner walks through the full value chain: exploration, permitting, mining, beneficiation, concentration, refining, chemical conversion, and finally engineered materials for batteries and magnets. The key theme is complexity: each stage adds time, specialization, and cost before a material reaches an end product.

    • End-to-end chain: exploration → permitting → mining → ore/waste separation → concentration → refining → specialty chemicals → engineered materials
    • Battery pathway: cathode/anode materials → cell → module → pack → EV or storage
    • Magnet pathway: refined rare earths → blending/slurry → casting/sintering → precision machining → motors
    • Intermediate products move globally before becoming high-purity metals
  3. 4:06 – 5:45

    Why every mine’s processing flow sheet is bespoke—and why that’s a problem

    Processing plants are designed around the specific chemistry and impurity profile of each ore body, making refineries hard to standardize. Turner explains how changing ore characteristics over time forces operators to manage variability, and why flexibility in circuit design is a major opportunity.

    • Flow sheets are custom-built for each deposit’s grade and impurities
    • A ‘library’ of unit operations gets stitched together by human experts
    • Ore properties and impurity concentrations change as the mine advances
    • Need for more flexible circuits that can adapt to heterogenous feedstock
  4. 5:45 – 6:45

    The human stack behind mining: disciplines, roles, and labor constraints

    The discussion maps the wide range of specialized roles required to build and run mines and refineries—from geology through engineering to site operations and back office. They emphasize that labor shortages now span both trades and advanced engineering, becoming a key bottleneck for scaling supply.

    • Mining requires geologists, geophysicists, mining and process engineers, chemists, metallurgists, and more
    • Site operations depend on diverse on-the-ground skilled workers
    • Back-office functions are essential to operating complex projects
    • Labor pool contraction is becoming one of the industry’s biggest challenges
  5. 6:45 – 11:15

    Turner’s path from Tesla upstream: cost, scale, and incentive misalignment

    Turner describes moving from factory design to battery manufacturing to upstream materials, driven by tracing what actually drives costs. He contrasts manufacturing economics (scale lowers cost) with commodity economics (more demand can raise prices), highlighting a structural mismatch between miners and industrial buyers.

    • Career progression: factory construction → cell manufacturing → cathode/refining exposure
    • Following cost upstream reveals metals as a primary cost driver
    • Commodities often price upward with demand due to constrained supply
    • Mismatch between mining incentives and manufacturing expectations
  6. 11:15 – 12:46

    Tesla’s vertical integration lessons: incentives, speed, and risk transfer

    They use Tesla as a case study for why companies vertically integrate when suppliers won’t invest, innovate, or scale fast enough. Turner emphasizes vertical integration as a deliberate choice to reshape incentives—while also absorbing partners’ risks into the company’s own operating risk.

    • Tesla integrated early because required parts didn’t exist or weren’t prioritized
    • Vertical integration resolves supplier misalignment on innovation and scale
    • Insourcing shifts risk from partners to the integrating company
    • Requires confidence in managing expanded technical and execution risk
  7. 12:46 – 17:55

    On-the-ground mining realities and why automation adoption lags

    Turner shares what mine sites feel like in practice and explains why automation often gets deprioritized: long unprofitable phases, capital fatigue, and remote operations. The labor scarcity trend is changing the equation, but incumbents still move slowly due to operational risk and conservative culture.

    • Mine sites can be calmer and more methodical than outsiders expect
    • Automation is limited not because machinery doesn’t exist, but due to economics and deployment friction
    • Years can pass before reaching ore (waste removal/shafts), reducing appetite for extra capex
    • Remote locations plus shrinking labor availability increase urgency for automation
  8. 17:55 – 24:04

    Why incumbents struggle with tech: outsourcing, bureaucracy, pilots, and slow cycles

    They describe how major mining companies approach digital efforts via consultants and ‘innovation arms,’ but adoption is limited. Risk is evaluated at the micro-change level (where each change threatens downtime), leading to endless pilots and few commercial deployments because new plant build cycles are so infrequent.

    • Majors often outsource digital work to consultants; recommendations may not be implemented
    • Plants are high-stakes, so counterintuitive model outputs are culturally hard to trust
    • Small changes can risk multimillion-dollar downtime, discouraging experimentation
    • Pilots are common, but commercial adoption is slow due to multi-year project cadence
  9. 24:04 – 25:38

    Geopolitics and China’s advantage: policy support plus a massive skilled labor engine

    The conversation shifts to why Chinese players scaled so quickly: early strategic recognition, supportive policy, and a deep, experienced talent pool. Turner highlights construction and commissioning manpower as a decisive factor that enables faster iteration and ramp-up.

    • Top-down policy enabled domestic and international resource build-out
    • China’s advantage is also a large, experienced technical labor pool
    • Example: 13,000 people mobilized for refinery construction/commissioning
    • Western projects struggle to mobilize even a fraction of that workforce
  10. 25:38 – 28:19

    Industry structure: juniors, majors, ‘orphan’ deposits, and Mariana’s entry point

    Turner explains how juniors explore and majors deploy capital, leaving many subscale deposits stranded despite containing viable metal. Mariana’s thesis is to unlock these ‘orphan’ assets by building and operating more efficiently, then scaling up over time.

    • Juniors focus on exploration and flipping resources; majors want multi-billion-dollar scale
    • Many deposits are too small for majors and become ‘orphans’
    • Mariana targets subscale assets where metal exists but execution economics fail for incumbents
    • Strategy: build a platform, prove efficiency, then scale project size
  11. 28:19 – 31:54

    What Mariana is building: a vertically integrated, software-first minerals operator

    Turner defines Mariana as a project developer and operator, not a point-solution SaaS company. The company’s software focus targets two bottlenecks: construction execution (reducing churn and latency) and plant operations (moving toward autonomous optimization).

    • Mariana focuses on detailed engineering, permitting, construction, commissioning, and operations
    • Goal: enable a small team to do what traditionally requires huge organizations
    • Capital Project OS aims to reduce three-week data lags and run projects like manufacturing
    • PlantOS aims to reduce human-in-the-loop control in complex refineries
  12. 31:54 – 35:39

    PlantOS and reinforcement learning: optimizing refineries with 1,000+ variables

    They describe refineries as ‘big robots’ with sensors and actuators, but still heavily reliant on human operators and static operating strategies (like blending feedstock to reduce variability). Turner argues RL can handle multivariable, high-latency, interconnected circuits to reduce energy/reagents and improve recovery—the biggest cost lever.

    • Refineries have many coupled unit operations with recycle streams and long delays
    • Feedstock changes constantly due to ore body variability; industry blends inputs to cope
    • RL can target global optima faster than humans, potentially mirroring DeepMind’s data center gains
    • Improving recovery reduces the amount of metal that must be mined upstream
  13. 35:39 – 36:55

    Commissioning speed as a strategic weapon: months vs years to reach spec

    Turner contrasts commissioning timelines: some Western refineries take years to reach stable throughput and product specs, while Chinese operators may do it in months. He links delays to system latency and human-driven tuning, making autonomous control and better integration central to Mariana’s approach.

    • Some recently built refineries still aren’t fully commissioned years later
    • Benchmark: ~6 months commissioning in China vs 2–4 years for many Western projects
    • Circuit latency (24–48 hours for effects to propagate) slows tuning and ramp
    • RL-driven operations aim to cut time-to-spec and accelerate cost-down curves
  14. 36:55 – 40:20

    Build vs partner: starting with proven unit ops, then accelerating new tech deployment

    Mariana’s near-term plan is to integrate commercially proven unit operations and win via superior system-level operation—also aligning with what project finance will underwrite. They discuss industrial supply chain weaknesses (e.g., long lead times for pumps/tanks) and Mariana’s intent to become a deployment partner for novel processing technologies.

    • Initial focus: proven unit operations + better integration/operations to capture uplift
    • Project finance prefers de-risked unit operations over first-of-a-kind chemistries
    • Non-obvious bottlenecks: broken industrial supply base and extreme equipment lead times
    • Mariana wants to be the customer/partner that accelerates commercialization of new processes
  15. 40:20 – 42:57

    VC view: why mining needs end-to-end ownership (not point solutions)

    Ryan explains a16z’s long-standing interest in mining as a massive, tech-underpenetrated market. The core investing conclusion: startups selling point tools struggle against incumbents’ adoption friction, so a vertical operator that controls the full lifecycle is better positioned to capture the value of technology improvements.

    • Mining is one of the largest markets with low tech adoption
    • Selling point solutions is hard due to decentralized incumbents and low risk appetite
    • Geopolitics (China’s scale) reinforces urgency for a new approach
    • Owning the full atom-to-product lifecycle enables capturing efficiency gains
  16. 42:57 – 47:45

    Which minerals matter most: ‘big metals,’ batteries, rare earths, and commodity cycles

    Turner prioritizes high-volume metals (aluminum, copper, iron, zinc) as the biggest mass-flow challenge, alongside battery and defense-linked materials (lithium, nickel, magnesium, uranium, rare earths). He emphasizes that timing matters—build at commodity troughs—and highlights lithium and copper as key near-term focuses.

    • Biggest growth by mass: aluminum, copper, iron, zinc
    • Battery chain metals: lithium (needs major capacity growth), nickel, manganese
    • Defense and energy security: aluminum, magnesium, uranium, rare earths
    • Strategy includes diversification and building at cycle lows; lithium is cited as trough-positioned
  17. 47:45 – 53:46

    US permitting bottlenecks—and policy levers to mobilize capital and capability

    They cover why the US struggles to bring mines online: exploration permitting thresholds, slow and opaque environmental review workflows, and broader NEPA burdens. Turner proposes demand-side support (offtakes/price floors), smarter use of government capital, and efficiency improvements—potentially aided by language models—to speed review and execution.

    • Exploration permitting (e.g., >5 acres on federal land) can throttle discovery activity
    • US resource estimates are limited by lack of exploration, not just geology
    • Permitting reviews lack transparency and can be streamlined with better workflows/tools
    • Top policy asks: demand-side support (offtakes/price floors), thoughtful gov participation in capital stack, reduce regulatory friction created by funding-linked requirements
  18. 53:46 – 56:13

    Beyond US borders: scaling internationally and the 10-project, 10-year goal

    Turner frames the US as the starting point for building a repeatable platform, not the final boundary. The long-term success metric is restoring the capability to build and operate complex minerals infrastructure quickly, cost-effectively, and responsibly—reducing national anxiety about supply security.

    • Strategy begins in the US but anticipates international expansion (and possibly seabed resources)
    • Goal: build 10 projects in 10 years with increasing scale
    • Core mission: rebuild the ability to construct and operate complex minerals plants
    • Success looks like reduced fear about securing critical minerals due to regained industrial capability

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.