GeoIntel · Drilling Intelligence

This is a thesis case study. GeoIntel is pre-deployment — architecture locked, rollout phased, first-target cohort identified. The work documented here is the venture architecture, not realized outcomes. The case study is updated as each phase gate clears.

Every few decades, an industry arrives at a strategic inflection point — the moment, to use Andy Grove's framing, when the fundamentals of the business shift so completely that the operating assumptions that built the last era cannot carry the company into the next one. The inflection points are not announced. They arrive as pressure on the operators who already work in the industry. The operators who name the inflection, and re-architect against it, compound. The ones who don't, don't.

Geothermal drilling is inside one of those moments.

The Department of Energy has committed serious capital to the sector under the Enhanced Geothermal Systems program and its adjacent funding vehicles. Oil and gas operators are actively migrating talent, rig time, and capital into geothermal programs — because the skills transfer cleanly, because the permitting environment is meaningfully easier than the federal oil and gas leasing environment has become, and because the demand signal from utilities, data centers, and industrial heat customers is suddenly real. Enhanced geothermal is no longer the subsidized long-shot technology it was five years ago. It is a category with a functioning commercial thesis and a compressed timeline.

And the entire procurement side of the sector is running on spreadsheets.

You cannot orchestrate what you cannot see. And nobody in the geothermal supply chain can see anything.

GeoIntel is a venture thesis built against that gap.

I. The Problem

Modern drilling programs — in any vertical, oil, gas, geothermal, or water — are procurement-intensive businesses. A single program touches hundreds of line items across dozens of suppliers: casing, mud, directional services, bits, cementing, perforating, logging, wellhead equipment, surface iron, downhole tools, completions fluids, wireline, coiled tubing, frac spreads. Every line item has a price band, a lead time, a supplier list, and a regulatory envelope. The program's economics live inside those procurement decisions. A drilling program that spends ten percent more than it needs to on procurement is a drilling program that loses its margin before the first bit ever hits the ground.

Oil and gas figured this out thirty years ago. The supermajors operate against structured procurement systems — master service agreements, category management, supplier intelligence platforms, benchmarking dashboards, e-sourcing systems, strategic sourcing teams that treat procurement as a discipline rather than a function. A mid-cap operator in the Permian today has more procurement infrastructure than the entire geothermal sector combined.

Geothermal does not have that infrastructure. The operators inside the sector are excellent engineers — many of them came out of oil and gas specifically because they saw the transition coming — but the sector grew too fast for its tooling to catch up. The old procurement model worked the way procurement has always worked in a small industry: a phone call to a trusted supplier, a verbal quote, a purchase order issued against a budget line that might or might not reflect current market pricing. When you are drilling three wells a year, that works. When you are drilling thirty, it does not.

The sector's center of gravity is now moving from pilot projects to production programs. The procurement infrastructure has to move with it, or the cost curve that makes the whole commercial thesis work will not hold.

This is what inflection looks like from the inside. The operators who are running geothermal programs today are not asking whether the procurement problem is real. They know it is real. They are asking a narrower question: who is going to solve it, how, and when? Because every month the procurement side remains opaque is a month of cost that cannot be recovered, schedule risk that cannot be quantified, and capital efficiency that cannot be benchmarked against anything.

The pattern here is not new. It has played out before, in other sectors, at other inflection points. Every time, it follows roughly the same shape.

The sector grows faster than its tooling. A procurement gap emerges between what the mature sectors have and what the new sector has. The operators inside the new sector feel the gap but cannot justify building the tooling themselves — because the tooling requires horizontal data (every supplier, every lead time, every price band) that no single operator has. A platform thesis forms. The first platform into the canonical data position captures disproportionate value, because in a procurement intelligence business, the moat is the data layer, and the data layer compounds with use.

Bloomberg did this in finance in the 1980s. The financial markets had grown more complex than the tooling could handle, traders were running Rolodexes and calling brokers for quotes, and Michael Bloomberg walked in with a terminal that aggregated bond pricing across every dealer in New York. The terminal was not magic. It was a canonical data layer wrapped in a user interface that made the data actionable. Inside a decade, the terminal was the default procurement surface for the sector. Inside two decades, it was the default sense-making surface for the entire financial industry, across every adjacent vertical Bloomberg eventually absorbed.

GeoIntel is the same pattern applied to a different sector at a different inflection point. The timing window is open now, in a way it will not be again in five years — because the first platform into the canonical position in this category will hold the position, and the sector will not support two canonical platforms.

II. The Approach

GeoIntel is a three-phase intelligence platform, plus a fourth architectural commitment that governs how the platform compounds into a category rather than staying trapped inside a single vertical.

Each phase is independently valuable. Each phase also builds the foundation for the next one. The architecture refuses the trap of trying to deliver all three at once — because the canonical data layer has to exist before any analytics sit on top of it, and the analytics have to prove themselves before the procurement orchestration layer has anywhere credible to land.

01 — Phase 1 · The canonical data layer

The first phase is infrastructure. GeoIntel consolidates supplier, equipment, lead-time, and regulatory data into a single normalized source of truth for the sector.

This is the moat. Nothing downstream works without it.

The sector's data is not missing. It is scattered — in supplier PDFs, in procurement team spreadsheets, in permit filings, in conference slide decks, in vendor quotes that get forwarded to one person and then forgotten by the entire organization. Every operator in the sector knows some piece of the picture. Nobody has the whole picture. The sector has built its procurement function around the assumption that the whole picture is impossible to see.

GeoIntel's job in Phase 1 is to disprove that assumption. The platform pulls data from every available surface — public filings, supplier-provided data sheets, FERC and state regulatory databases, industry association publications, opt-in operator data contributions, scraped public quote archives. The data is normalized against a standardized taxonomy of drilling line items, supplier capabilities, and regulatory jurisdictions. The data is kept current through a combination of automated refresh, supplier-contributed updates, and operator-flagged corrections.

The first operators on the platform become the first readers of a picture of their own supply chain they have never been able to see clearly. The surprise, for most of them, is not what the data says about their competitors. It is what the data says about their own procurement decisions over the last eighteen months. The gap between what they paid and what the sector paid, for the same line items in the same time windows, is almost always larger than the operator expected.

That surprise is the platform's first commercial wedge. An operator who sees a six-figure procurement gap against the sector benchmark, in a single report, does not need to be sold on the value of continued access.

02 — Phase 2 · Enrichment and benchmarking

Once the canonical layer exists, the question the platform answers shifts. Operators stop asking what is the price of X and start asking what is the historical price band of X, what is the current availability window, and what are my peers paying right now.

Phase 2 is the enrichment and benchmarking layer.

Historical price bands become a service, not a guess. For every line item an operator's drilling program touches, the platform shows the 30-day, 90-day, and 12-month price distribution across the sector — and where the operator's own procurement sits inside that distribution. Supplier availability becomes visible. The operator's procurement team moves from reactive pricing to informed pricing. The procurement decision becomes a decision made against data, not against the first quote that came in.

There is a second-order effect that Phase 2 produces, which is more important than the benchmarking itself. Once the sector's pricing becomes visible, the sector's suppliers respond to it. Suppliers that were quoting opportunistically — knowing that the operator had no way to verify — lose that advantage. Pricing across the sector compresses toward a narrower, more defensible band. The operators win. The disciplined suppliers win. The suppliers who were harvesting informational asymmetry lose. This is how a market matures — and it is almost always a platform that forces the maturation, because the operators inside the market cannot coordinate to force it themselves.

03 — Phase 3 · Forecasting and orchestration

The third phase is where the platform becomes indispensable. Procurement shifts from informed to forecast-driven.

Rig programs see supply risk before it crystallizes. An operator planning a six-well program for the following quarter can model supplier lead times, availability windows, and regulatory envelopes against the program's actual drilling schedule — and surface conflicts before they become execution problems. Alternate supplier paths are pre-modeled. Regulatory shifts are propagated across every program the platform sees. The platform does the work that, today, the procurement team does manually in a spreadsheet every Monday morning and that still misses the risks nobody thought to look for.

The platform becomes the default procurement surface for mid-cap geothermal operators. And the intelligence layer that any driller checks before signing a material order.

04 — The adjacent-market bridge

The final architectural commitment is the one that turns a platform into a category.

GeoIntel is not a geothermal-only platform. The same canonical data layer, with modest extensions, spans adjacent verticals — core drilling, water well, deep geothermal, shallow geothermal, lithium brine. The verticals share suppliers. They share equipment categories. They share regulatory surfaces. They share the underlying problem of a scattered data landscape and a reactive procurement function.

The strategic logic of the adjacent-market bridge is compounding. Each vertical the platform covers sharpens the benchmarks in the other verticals. A supplier's price discipline in water well informs the expected price for comparable line items in geothermal. An availability signal from core drilling propagates through the platform's forecasting layer for every adjacent vertical. The data layer gets more valuable the more of the market it covers — which is the defining property of a winner-take-most information platform.

This is how Bloomberg became Bloomberg. They did not set out to cover the entire financial industry. They set out to cover bond pricing. The canonical data layer they built for bonds became the foundation for equities, then derivatives, then commodities, then every adjacent asset class. Each expansion made the previous coverage more valuable — because the data inside one asset class was never fully independent of the data in the others.

GeoIntel's architectural commitment is to build for the same pattern. The first vertical is geothermal, because the inflection point is happening now. The adjacent verticals are pre-mapped. The platform is designed from day one to absorb them, rather than to be re-architected to support them later.

III. The Results

GeoIntel is a venture thesis. The results are architectural, not operational — and this case study will be updated at each phase gate.

Architectural Milestone	Status
Canonical data layer — scope and spec	Locked
Enrichment & benchmarking — roadmap	Staged
Forecasting & orchestration — thesis	Locked
Adjacent-market bridge — architecture	Pre-mapped
First-target cohort — mid-cap drillers	Identified
Deployed competitors holding the canonical position	0
Sector inflection window	Open · estimated 24–36 months

The sector's windows are open in a way they will not be again.

DOE capital is flowing. Talent is migrating from oil and gas. Regulatory pressure is pulling toward faster permitting, not slower. The procurement infrastructure gap is visible to every operator who has tried to build a rig program in the last eighteen months. And the commercial thesis for geothermal is finally standing up under independent scrutiny — utilities are signing PPAs at prices that support positive project economics, data center buyers are signing direct offtake agreements, and industrial heat customers are finally treating geothermal as a real alternative rather than a subsidized experiment.

What has been missing is the platform. GeoIntel is the platform.

The case study updates as the platform moves into Phase 1 build and the first operators onboard. The next update will name the first operator cohort and the first benchmark coverage report.

IV. Why This Pattern Compounds

The reason platforms like this compound is structural. It has nothing to do with execution excellence, though execution matters. It has to do with the shape of the information economy inside a maturing sector.

Every operator in a scattered-data sector is paying a hidden tax. The tax takes the form of procurement decisions made against incomplete information, schedule risks that were not forecastable, supplier pricing that went unverified. The tax is not visible on any operator's income statement — because nobody knows what the alternative looked like. The tax is visible only once the platform exists.

The moment the platform exists, the tax becomes visible, and the operator who is paying it rationally subscribes to the platform to stop paying it. Every additional operator who subscribes makes the platform's canonical data layer more valuable. Every additional supplier who contributes data expands the coverage. Every additional vertical the platform covers multiplies the cross-benchmark value.

This is what a winner-take-most information platform looks like in the early innings. It looks like a slow compounding loop that suddenly accelerates once the canonical position is established — because every operator who does not subscribe is visibly disadvantaged against every operator who does.

Hamilton Helmer, writing about the durability of strategic advantage in 7 Powers, called this shape Counter-Positioning plus Cornered Resource. The platform is positioned against a status quo the incumbents cannot adopt without cannibalizing their own spreadsheet-based procurement consulting practices. And the platform owns a cornered resource — the canonical data layer — that compounds with each new contribution. Those two powers, operating together, are the shape of a category-defining business.

The sector has not had a platform of this shape before. It has now.

V. Sources

1. Internal NEXT venture thesis, GeoIntel volume series, 2026.
2. U.S. Department of Energy, Geothermal Technologies Office — strategy and funding documentation. energy.gov/eere/geothermal
3. Enhanced Geothermal Systems program — DOE roadmap and cost-reduction targets.