ElementIQ: Decision Brief for the Ellis CallOn Phil's Capital Deployment Plan of 2026-08-03. Ellis is a direct recipient of that email, so this is a decision, not a briefing.

Internal. ElementUSA only. Not for Silica-X.
Prepared by Mike, 2026-08-03
Page 1 of 2: decisions and sequence
Download PDF
Bottom line

The deal changed shape. It is no longer a JV formation, it is a software license with a build attached, plus a separate equity option, and there is no investor in it. ElementUSA is named the funding party for all $1.2M: $400K for a license and an MVP, $200K for an internal LLM, $600K for 30% of Protosentient. Downside is capped at $400K with no dilution of ElementUSA, which is a better risk profile than the $4M raise we were carrying this morning.

My recommendation: do not sign Phase 1 as written, and do not walk. Counter on four terms (exclusivity, an MVP definition we set, payments weighted to delivery, and a 90-day option window), fund Phase 1 internally if those land, and treat the $600K option as the real prize because it is priced at a $2M valuation while Rob is telling VCs the number is 10 to 12.

The sequence, and what not to do out of order

Today
Acknowledge only

Chris replies to Phil: received, reviewing, questions this week. Ellis does the same. No substance.

This week
Counsel first

Two hours with a startup and VC lawyer on the license and the option. Before any answer goes back.

Then
Questions to Silica-X

One written list, framed as clarifications. Exclusivity first. Their answers tell us if this is a deal.

Then
Decide Phase 1

$400K is an ElementUSA capital call, not a JV negotiation. Chris and Ellis, with counsel's read.

Separately
Blue Bear on the option

Narrow ask now: is $600K for 30% a fair price, and would you fund it. Not a structure question.

Do not take this to Blue Bear before counsel, and do not answer Phil substantively before counsel. Our own deck with Ernest already describes a different deal (page 2), and the last term nobody challenged on a call was in writing six hours later.

Four decisions we need from you and Chris

1

Do we want this at all, given we already have a working alternative?

Chris said on 7/31 that nothing on the near-term path requires outside money, and Orebis already has the data, a working chat layer and no investment needed. Honest options: sign Phase 1, counter, keep building in-house and revisit, or walk. This sets our leverage on everything else and it has never been asked out loud.

Recommend: counter, do not walk. Their verification layer and CPU efficiency are real and we cannot build them. But say out loud that we have an alternative, because we negotiate differently once we know that. Decided
2

Do we counter before signing, and on which terms?

Four terms carry the risk: exclusivity in the two segments (never stated in the document), an MVP definition and acceptance test we set (defined nowhere), payments weighted to delivery (today $300K of $400K lands before delivery), and a 90-day option window (today 10 days, and they control the trigger).

Recommend: all four, as clarifying questions rather than a counter-offer. Nothing else needs to move for us to say yes. Decided
3

Who funds the $600K option, and do we ask Blue Bear now or later?

$600K for 30% implies a $2M post-money valuation, the bottom of the $1.0M to $2.5M range our own research called defensible. Rob says a round here is "10 to 12" and that he sees Nvidia within two weeks. Both cannot be true. If Rob is right the option is deep in the money; if he is inflating, we have our answer on valuation generally.

Recommend: ask Rob directly what valuation he is quoting tier-one VCs, and take a narrow question to Blue Bear (is this price fair, would you fund it). Do not reopen structure with them. Decided
4

How do you want to handle my position?

I am engaged hourly by ElementUSA and I am advising on a transaction where ElementUSA pays Silica-X. On the 8/3 call Phil raised, unprompted, whether I go on an EIQ salary. I have not pursued it and I am raising it now rather than after terms are set. Any compensation for me should come through ElementUSA, never from Silica-X's side of the table, and if you want me to recuse from any part of the Silica-X terms, say so.

Decided

Four red flags in the document

  1. They control when our option expires. It closes 10 days after MVP delivery, they decide when that is, and MVP is defined nowhere. Note the direction: delivering early shortens our window. If Rob has a tier-one conversation going, letting our cheap option lapse is rational for them.
  2. $300K of the $400K is paid before delivery. $200K at signing, $100K at day 45 against an "interim milestone" that is never specified. No remedy, refund or termination right if they miss, from a counterparty with unpaid IP counsel and $200K of unfunded patent work.
  3. The license is never called exclusive. It says perpetual, not exclusive. Nothing stops Protosentient licensing the same two segments to a competitor or working them itself.
  4. We are the only party who can be diluted. We fund 100% of EIQ and hold 70%. Their 30% carries a 20% anti-dilution floor. We have no protection in either entity.

ElementIQ: Supporting DetailQuestions to ask, what changed, and what is already inconsistent

Internal. ElementUSA only. Not for Silica-X.
Page 2 of 2

Questions for Silica-X, in this order

  1. Is the license exclusive in critical materials and waste recovery? If not, what stops a competitor licensing the same segments?
  2. Define both segments in writing. Does red mud count? Tailings? Nuclear waste glass? Nuclear is not a named segment even though it is the credential you lean on hardest.
  3. What is the MVP, in acceptance-test terms, and who signs off? We propose: it ingests an agreed set of our documents and answers an agreed question set with sourced citations, verified by us.
  4. What happens if MVP slips or fails? There is no remedy, refund or termination right in the document.
  5. Does the $600K go into Protosentient as capital, or to Silica-X as consideration? If it is consideration, Protosentient receives nothing and cannot fund the Exchange build the document promises.
  6. Can Protosentient use our mineral knowledge in the other 70% of markets? That data is the moat, and the document is silent.
  7. Is the license assignable if EIQ is ever sold, and does it survive Silica-X insolvency? Without both, EIQ may not be sellable.
  8. Employees or vendor? Are the three resources EIQ employees or a Silica-X contract? If a contract: benchmark the rate, recuse Phil from approving his own invoices, assign IP on everything built.
  9. Of the patents, how many are issued versus pending, and is the $200K of IP work a use of these proceeds or a pre-close condition?

Questions for Blue Bear, narrower than last week

The structure question is largely settled by Phil's document, so do not reopen it. Ask three things: is $600K for 30% of an application-layer AI platform a fair price at this stage, would you fund or syndicate that $600K, and does a licensee rather than an IP owner change how you would value EIQ. That last one is the question we cannot answer ourselves.

And keep them out of the first Silica-X exchange. Rob has already pre-positioned them: he wants Blue Bear as his own representation on the bigger Protosentient round and offered them tag-along rights and first call. They are an input, not a referee.

What changed without us agreeing

It is a license, not an assignment. Rob argued this on the call, nobody accepted it, and it is now the structure. EIQ will own a license, not IP.

The 20% floor became a line item. Unchallenged at 10:00, in writing by 14:15.

Our platform stake was repriced from contributed to purchased. On the call Rob said "you contribute capital, that's where you get your equity." It now costs $600K.

The marketplace moved behind that paywall. The $600K is what funds the Exchange build, so the Spotify thesis only happens if we exercise.

What is genuinely good about the offer

Downside capped at $400K with zero dilution of ElementUSA. The license covers any improvements serving our segments, in perpetuity, which is the future-updates parity clause off our own June list. Rob is also right that a licensee does not pay to prosecute or litigate the patents.

And asking worked: development fell from Phil's on-call "$1M finishing the technology" to $400K for the MVP, and we finally have a headcount, three resources through delivery at about $33K per person-month including materials.

Two inconsistencies to fix before anyone talks to Blue Bear

Ellis has already sent a 10-page deck to Ernest describing a VC raise. Phil's plan has no investor in it, a different structure and a different number. Whoever speaks to Blue Bear next needs to know which story we are telling.

Our own collateral is out of sync. The teaser deck carries ARR of 0.6 / 3.5 / 12.5 and margins of 70 / 78 / 82, and the Google diligence deck still shows 1.5 to 2M and 5 to 6M on its roadmap slide. Fix or pull that slide.

Timing, and who it favors

Delay helps us on Phase 1 and hurts us on the option. They are cash-constrained, so waiting improves our terms on the build. But Rob takes Protosentient to Nvidia within two weeks, and a tier-one pricing the platform either makes the option far more valuable or gets it withdrawn.

Nothing here is binding. Final terms are subject to definitive agreements, and the document lists exactly one open item, LLM operating costs. There are at least nine, above.

Owners and timing

Acknowledge receipt to Phil, no substance, and confirm Ellis does the sameChrisToday
Engage a startup and VC lawyer on the license and the optionChris and EllisThis week
Send the nine questions to Silica-X, exclusivity firstMikeAfter counsel
Ask Rob what valuation he is quoting tier-one VCsMikeBefore Nvidia
Decide whether $400K Phase 1 is funded internallyEllis and ChrisInside two weeks
Reconcile the Ernest deck and the diligence-deck roadmap slideMikeBefore Blue Bear