Disrupting MATERIALS: A Founder's Guide to the AI Opportunity

Every sector has incumbents. Every sector has problems the incumbents can't solve because they're too big, too constrained by legacy, or too invested in the old model. MATERIALS is no different. Here's the AI-shaped opportunity in front of you.

How MATERIALS Works Today

To disrupt an industry, you first need to understand its structure. MATERIALS isn't random. It's organized around specific money flows, customer dynamics, and constraints.

- **Base Metals**: Copper, aluminum, zinc, nickel (infrastructure, EV batteries) - **Precious Metals**: Gold, silver (investment, electronics) - **Bulk Commodities**: Iron ore, coal, aggregate - **Specialty Minerals**: Rare earth elements, lithium, cobalt, manganese (battery/tech-critical) - **Tier 1**: Integrated miners (Rio Tinto, BHP, Glencore) - **Tier 2**: Mid-cap miners, specialists (Li, cobalt) - **Tier 3**: Junior explorers, artisanal mining

The key to understanding the industry is following the money. Who pays whom, when, and for what? What are the margins at each step? Where is value actually being captured? The answers to these questions show you where AI can redirect the economics.

What's Broken: The Multi-Billion Dollar Problem

Every incumbent sector has things that are broken. These aren't small inefficiencies. They're problems that cost the industry billions of dollars annually and that the incumbents can't fix without restructuring themselves.

- **Exploration success rate 1-2%**: Billions spent on projects that don't become mines; seismic interpretation and drilling still largely manual - **Commodity price forecasting dismal**: Analyst forecasts miss actuals by 30-50%; mining capex decisions made on flawed price assumptions - **Stranded assets risk**: $2T+ capital deployed; climate policy/decarbonization threatening coal/high-carbon assets - **ESG compliance gaps**: Water usage, tailings management, emissions reporting inconsistent; ESG-sensitive capital fleeing sector

These problems aren't invisible to the incumbents. They see them every day. They can't fix them because the fix requires dismantling part of their business model. That's where you come in.

Three Specific Startup Plays

Play 1: The Problem-Solving Play

Take the single most expensive problem in the industry. Build an AI solution specifically for that problem. Make it 60-70% as good as the incumbent can do, at 20-30% of the price. Don't try to be better at everything. Be better at that one thing. TAM: Usually $500M-$2B for a properly positioned single-problem solution in a major sector.

Play 2: The Unbundling Play

Most incumbents bundle multiple services or capabilities together. Customers have to buy the whole bundle. Unbundle one piece. Build an AI-native solution for that piece that's 3-5x better than what customers get inside the bundle. Let customers use your solution instead of the bundled version. TAM: Usually $200M-$1B depending on which part you unbundle.

Play 3: The New-Customer Play

Incumbents serve existing customers with existing willingness-to-pay. There are usually customers outside the market entirely—too small to serve profitably, too different to address, in geographies the incumbents don't operate in. Build an AI solution for those customers. It won't appeal to the incumbent's current base. It will appeal to entirely new customers. TAM: Usually $100M-$500M as you expand from greenfield customers.

What Makes It Hard

Every sector has moats. In MATERIALS, the moats are typically: regulatory constraints, capital requirements, network effects, or incumbent distribution. Understanding which moats protect the incumbents in your target space is critical.

MATERIALS (Mining, Metals, Minerals) ### Industry Structure - **Base Metals**: Copper, aluminum, zinc, nickel (infrastructure, EV batteries) - **Precious Metals**: Gold, silver (investment, electronics) - **Bulk Commodities**: Iron ore, coal, aggregate - **Specialty Minerals**: Rare earth elements, lithium, cobalt, manganese (battery/tech-critical) - **Tier 1**: Integrated miners (Rio Tinto, BHP, Glencore) - **Tier 2**: Mid-cap miners, specialists (Li, cobalt) - **Tier 3**: Junior explorers, artisanal mining ### Geopolitical Concentration Risk - **China dominance**: 98% of gallium, 95% of magnesium, 70%+ of rare earths, 60%+ of lithium processing - **Supply chain vulnerability**: Single-country concentration risk for EV supply chain ### Top 3 AI Disruption Vectors 1.

The winning move isn't to attack the moat directly. It's to find a customer segment where the moat doesn't matter. Build for them first. Scale from there.

Founder's Playbook: 12 Months to Product-Market Fit

Months 1-2: Deep Industry Immersion

Spend 8 weeks talking to customers and operators in the sector. Your goal is not to validate your idea. It's to understand if you're solving the right problem, for the right segment, in the right way. Talk to 20-30 operators. Ask them about their biggest pain points. Ask them how much they'd pay to solve it. Ask them who else would care about that solution. You'll either get clearer or you'll pivot. Both are good outcomes at this stage.

Months 3-5: MVP Development

Don't build a perfect product. Build a minimally viable product that solves 60-70% of the problem for a specific customer segment. Focus on the part that requires AI, not the parts that are just engineering. By end of month 5, you should have something that a customer in your target segment will use (even if they're doing it as a favor, not because they're paying).

Months 6-8: Pilot and Iteration

Get your MVP in front of 3-5 pilot customers. They should represent your target segment. Run pilots for 4-8 weeks. Measure: Did the product deliver the promised value? Would they pay for it? What would they change? By end of month 8, you should have strong conviction that this solves a real problem and that customers want it.

Months 9-12: Path to Monetization

Now build the things that make you sustainable: pricing model, sales process, CS/support structure, compliance setup. By end of month 12, you should have 2-3 paying customers, a repeatable sales process, and a clear path to $1M ARR. You don't need to be there yet. You need to see it clearly.

The 10-Year Vision

If you execute well on one AI problem in MATERIALS, you'll earn the right to solve the next one. The biggest founders in this space don't stop at one product. They build platforms. MATERIALS in 2036 will look fundamentally different from MATERIALS in 2026. The companies that build the new infrastructure will be massive. Plan for that from day one, but execute on one problem first.