A working reference compiling what the current data shows, what it does not show, and the staged program proposed in the 22 August memo.
| Purpose | Provide a fact-based starting point for deciding how to validate and monetise the Orbis opportunity. It is not a recommendation. |
|---|---|
| Status | Working reference for review and adjustment. Not an investment recommendation, not a technical feasibility study, and not a valuation. |
| Scope | Figures come from the USGS National Mine Waste Inventory (public domain), the Orbis database, our own diligence reports, and cited public announcements. Nothing here rests on a license that has not been signed. |
| Marking | verified queried from our own data or an issuing authority claimed announced by a company or agency, not reconciled to an executed agreement assumed stated but untested |
The 22 August memo proposes that Orbis may be most useful as the front end of a repeatable project-development process rather than as a database sold on its own: identify candidate resources, secure access to test them, validate the material in ElementUSA's laboratory, develop a recovery pathway, and only then seek funding for a defined project. The sections below set out the data available to support that sequence, and the gaps in it.
The idea in one line
Getting from a pile on a map to a plant that pays takes five steps, in order. Three of them are questions about data. The last two are physical work, and they are the two we already do.
The first three boxes are lead generation. The last two are where money is actually made, and no data company can do them. That asymmetry is the whole strategy: we do not need to win the data business, we need to be the only one who can act on what the data says.
What the data says
The USGS national mine-waste inventory maps 764 waste features covering 585 square kilometers. Ranking by estimated metal content alone would place federal-land and Superfund features near the top, where access is constrained. Land status and liability are therefore recorded alongside size.
Access filter, applied in order
A working target set of 144 confirmed plus 91 to verify, against a universe of 764. Those 144 cover 171 square kilometers, which is 29% of all mapped waste area, so filtering for accessibility does not filter out the big piles.
Who controls the land, and what is listed
303 of 764 features have no recorded land owner. That is the largest single group, larger than confirmed private land. It is not a dead end, it is the research gap, and closing it is the highest-value work available that requires no license, no negotiation, and nobody's permission.
Where and what
By state
By feature type
The shortlist
These are the biggest features that sit on private land, carry no Superfund listing, and are not reclaimed. Ranked by mapped area, which is a proxy for volume and not for value. Adding the composition layer is what turns this from a size ranking into a value ranking.
| Site | State | Feature | Status | Area km2 |
|---|
These are large, well-known copper operations owned by major mining companies. They are the opposite of an overlooked pile nobody has valued. The thesis that waste rights are cheap because owners do not know what they hold is least likely to be true here, and most likely to be true among the 303 features whose ownership is not even recorded.
The gap in our own data
The preceding sections draw on data held outright, which has a defined limit. The commodity field on each pile records what the mine sold. Sierrita reads copper, molybdenum, rhenium. That is the product line, not the contents of the tailings, because nobody ever assayed those piles for anything they were not already selling.
Gallium, germanium, scandium, tellurium and the rare earths are the elements of interest, and no composition data is held for any of them across these 764 features.
What our data answers
Held outright. No license required.
What it cannot answer
Requires an outside dataset not currently held.
That table is ranked by area, which is a proxy for volume and not for value. Ranking by size points at the largest copper operations in America, all owned by majors. Those are the most studied, most expensive, hardest-to-acquire piles in the country, and the exact place where the idea that rights are cheap because owners do not know what they hold is least likely to hold.
A ranking by contained critical-mineral value could reorder that list completely, and surface a mid-size pile sitting at position two hundred that is the better buy. On the present basis, the ranking therefore favors the largest and most established operations.
A university research group has built statistical composition estimates for several hundred US metal mines, covering elements no operator ever measured. Two things make it worth a license rather than an attempt to rebuild it.
First, coverage. The 764 features belong to 172 distinct mines. The research set is materially larger, so it would both supplement the current holdings and extend beyond them.
Second, attribution. A figure attributed to a named university research group carries different weight in a federal or investor conversation than the same figure attributed to a commercial party.
They have already offered a sample limited to a couple of elements or a handful of sites. Take it, run it against our 144, and see whether it reorders them. If the order changes, the license is essential and we will know what to pay for. If it does not, the dataset adds less than we have been assuming, and that is worth knowing for the price of a conversation.
The third dataset in the chain, the one covering filed technical reports, answers whether a known project is real and financeable. That is the right question when underwriting a development-stage asset. It is the wrong question for an unassayed waste pile, because no technical report has ever been written about most of these features. That is precisely why they are still available.
It becomes relevant in Phase 2 and beyond, when a pile we control has to be presented to a funder. That license is not currently held. What was purchased previously was a pilot agent, not a data feed, and the contract has not been read. Nothing in Phase 1 depends on it, and no plan should assume it until someone opens the document.
Evidence that cuts against the obvious answers
Orebis has published pricing at $150, $550 and $1,000 per month, and billing has been live since 15 August. As of today there are zero paying customers. Ten accounts exist and all ten are comped internal. 23 reports have ever been generated, 21 of them by Mike and one each by two colleagues. Lifetime AI spend across the whole product is $71.32.
No firm conclusion about demand should be drawn from one week of availability. As matters stand, any option assuming subscription revenue has no supporting evidence yet.
Diligence reports held on comparable projects contain 65 findings about government funding across 16 reports. Every one describes money going to a physical asset. DoD took $400M of preferred equity in one producer and became its largest shareholder. One developer closed a $996M DOE loan. Another has an $800M EXIM facility still in due diligence.
Two patterns worth stating plainly: a closed federal loan is not money, because equity conditions precedent can block every draw, and in each case the private equity gap, not the federal piece, was the binding constraint.
On 18 August 2026 the Department of Energy selected nine projects for $162 million to recover critical minerals from mine waste and industrial feedstocks, targeting scandium, copper, antimony and rare earths. The program is Mines and Metals Capacity Expansion, run by the National Energy Technology Laboratory, and it follows a roughly $1 billion announcement in August 2025.
The stated eligibility criteria are based on technology readiness level rather than on resource size or data holdings. One track funds bench scale, moving TRL 4 to 5 up to 7. The other funds pilot scale, moving TRL 6 to 7 up to 7 to 8. What qualifies an applicant is a working process.
ElementUSA has described four years and more than $30 million of process development plus the Cedar Park laboratory. On the stated eligibility criteria, the applicant for this program would be the operating company on its process rather than the database.
Alcoa was one of the five pilot-scale selections. Alcoa is a bauxite-residue party, and our diligence file separately records another federally backed party pursuing gallium recovery at the same site.
Every investment vehicle found in this sector prices against production. A $1 billion critical minerals fund launched by a specialist manager with the IFC deploys equity, credit and royalty into construction, production and expansion. The closest analogue to a government-backed platform investor holds a portfolio of producing and late-stage development assets, funded by $105 million of federal equity, with a stated focus on reaching commercial production before investing further.
No vehicle was found that finances an information product. The search covered the vehicles named above and is not exhaustive.
On timing: a tailings recovery company has raised a $43 million Series B from two automotive strategics and is producing rare earths at 200 tonnes a year, and several other seed-stage companies were funded through late 2025 and early 2026 on waste-feedstock approaches. The one-to-two-year window is an assumption, and on this evidence the recovery-technology side of it is already contested. No funded example of the targeting layer itself was found. That may indicate an unserved gap, or that buyers do not pay for it separately. The available evidence does not distinguish between those readings.
The precedent
ElementUSA is advancing a critical-minerals project at Gramercy using bauxite residue as feedstock, aiming at gallium, scandium, rare earths and a list of co-products. It is the closest available template for what has to exist before public money arrives.
| Support | Amount | Status |
|---|---|---|
| Defense Production Act Title III, DoD | $29.9M | verified officially announced, gallium and scandium demonstration |
| DOE award with Colorado School of Mines | $67M | claimed company-announced, terms to confirm |
| Louisiana performance-based grant | $6M | claimed utility and infrastructure, performance terms |
| State programs and tax-exempt bonds | not stated | assumed anticipated participation, confirm availability |
| Louisiana investment decision | ~$850M | claimed state announcement, staged path to ~1Mt per year |
The two federal awards sum to $96.9 million, which matches the figure independently recorded in our own diligence reports from a separate source. Two independent sources give the same total. The awards do not fund the commercial facility; they support validation, demonstration and de-risking ahead of a larger financing.
In the Gramercy case, government support followed the assembly of a defined resource, secured rights, technical work, a scale-up plan and identified end markets. Separately, the current federal program for mine-waste recovery gates eligibility on technology readiness level, meaning a working process at bench or pilot scale. On both the Gramercy sequence and the stated program criteria, the qualifying asset appears to be the process and the defined project rather than the database.
Colorado School of Mines is already a partner on the $67M DOE award, so an academic validation relationship exists institutionally rather than needing to be created.
The program
The memo proposes testing whether Orbis can originate one commercially credible project, rather than working the whole database or assigning it a value before anything is validated. Every gate has a named owner, a budget, a deliverable and a written go or no-go.
Diversified deliberately, not ranked on pile size alone. Criteria include likely mineral relevance, volume and continuity, clarity of land, waste and mineral rights, whether samples can be taken legally and safely, infrastructure, environmental context, fit with our own laboratory strengths, multi-product potential, and fit with current federal priorities.
Note the constraint: we hold no composition data, so likelihood of useful grades cannot be scored today. Diversification on commodity, feature type, land status and geography is available now. Grade likelihood is where outside data or an assay enters.
Verify ownership, regulatory status, historical production, prior assays, waste volume, infrastructure and owner receptivity before anyone travels. Advance only where ownership is traceable, sampling looks feasible and no fatal flaw appears.
Before meaningful spend or any disclosure of a favorable thesis. Full acquisition is not required. Sampling permission, confidentiality, an exclusive evaluation period, a right of first negotiation, an option, or a pre-agreed commercial framework. The agreement must also cover ownership and permitted use of results, samples, process IP and data.
This step addresses how material can be tested without acquiring the pile, and before results change an owner's expectations.
A defensible sampling plan with locations, mass, representativeness and chain of custody, then characterization at Cedar Park. Elemental content and mineralogical form, particle size and liberation, variability across the waste body, preliminary recovery response, and early reagent, energy, water and residue requirements.
Preliminary flowsheet and a screening-level project case for each. Colorado School of Mines where independent characterization, validation or credibility materially strengthens the work.
Not on assay results alone. The package needs verified feedstock volume and access rights, assay and mineralogy with variability, preliminary recovery results, expected products and likely customers, screening-level capital and operating ranges, environmental and permitting considerations, and a costed work plan with milestones.
The pitch is a defined project, never the size of the database. Federal and state money for characterization, engineering, pilot and demonstration. Strategic investors for offtake and product qualification. Private equity later, for development and construction. Resource owners contributing access for royalty or participation.
| Period | Activity | Decision |
|---|---|---|
| Weeks 1-3 | Confirm criteria, rank 8-12 sites, begin ownership and technical diligence | Select candidates for owner contact |
| Weeks 3-7 | Owner outreach, access agreements, sampling plans | Authorise sampling on protected sites |
| Weeks 6-12 | Representative sampling and laboratory characterization | Reduce to technically promising candidates |
| Weeks 10-16+ | Recovery testing, flowsheets, commercial screening | Choose one or two lead projects |
assumed The 90 to 120 day timing depends on owner responsiveness, sampling permissions, travel, laboratory capacity and how difficult the materials turn out to be. None of those has been tested.
Risks
| Risk | Why it matters | Control |
|---|---|---|
| Database overreach | Accessibility and pile size do not establish value | Use Orbis for screening only; require physical validation before any economic claim |
| Sampling bias | A favorable grab sample may not represent the waste body | Qualified sampling plans, multiple locations, documented chain of custody |
| Rights leakage | Good results raise owner expectations and attract competitors | Confidentiality and evaluation rights before spend or disclosure |
| Technical recovery | A mineral can be present but locked in a form too costly to recover | Characterize mineralogy and test recovery before advancing |
| Environmental liability | Historic waste can carry remediation and permitting exposure | Define responsibility contractually; site-specific diligence. 194 of 764 features are Superfund-listed |
| Premature funding approach | Assays alone will not support a credible request | Add flowsheet, rights, scale-up plan, economics and end markets first |
| Promotional positioning | Overstatement weakens credibility with agencies and investors | Separate verified facts, company claims, assumptions and open questions, as this page does |
| IP ambiguity | Disputes over Orbis, site data and new process IP | Agree ownership, licensing and economics before the pilot begins |
Decisions requested
Who owns Orbis today, and what role, economics and authority would Michael hold within ElementUSA? No record reviewed answers this. The memo asks for it to be confirmed before the program begins.
The opportunity nobody priced
Colorado School of Mines runs a federally funded international program on critical minerals from mine waste. They are currently writing the standard procedure the government will use to decide which waste sites deserve public money, and they have said they want commercial operators to tell them what actually drives a real investment decision.
A scoring rule decides where funding goes. If it weights tonnage and element count, money flows to the biggest piles. If it weights recoverability with proven technology, existing infrastructure, permitted sites and a demonstrated flowsheet, that describes ElementUSA and very few others. Same dollars, different destination.
This is not lobbying for a contract. It is helping write the test we will later be graded on, at the invitation of the people writing it. It costs hours, needs no data license, and builds the federal relationship Phase 2 depends on.
Honest assessment
Open questions
If it includes the site-level results, every exclusive arrangement is worth far less within a year, and the whole timeline compresses. If it publishes the method only, the asset survives. One question settles it and it changes the value of every option on this page.
Owner: Mike · Blocking: the value of Phase 2 · Cost to answer: one conversation
The researcher believes it is hers. That is a reasonable reading of her own grant, not a determination. The university technology transfer office decides this, and every structure that could pay her depends on the answer.
Owner: Mike and Ellis · Blocking: step 1.3 and all of Phase 2
We have assumed data and API access three separate times. What was purchased appears to have been a pilot agent, not a data license. The contract has been sitting unread since 12 August. Nothing in Phase 1 depends on it, but our credibility with partners does.
Owner: Chris · Blocking: nothing in Phase 1 · Cost to answer: read one document
The weakest point in the plan. Every step before 1.5 is analysis, and analysis does not secure a pile. We have not scoped what an option or lease costs, who negotiates it, or which balance sheet it sits on.
Owner: unassigned · Blocking: step 1.5, which is the entire point
If this capability serves ElementUSA, the returns show up as better projects and there is nothing to raise against. If it is a standalone business, it needs outside revenue, which means selling access, which undoes the strategy above. Both are defensible. Doing neither deliberately is not.
Owner: Mike, Rick and Ellis · Blocking: everything downstream