ElementIQ and Silica-XAgenda and state of play for the Chris and Ellis call

Internal. ElementUSA only. Not for Silica-X.
2026-08-03, after the call with Phil and Rob
Download PDF
Bottom line

Their own team says the build needs about $2 million, not the $4 million we have been carrying. No structure is agreed, and the two of them brought competing proposals. Rob has already started showing the platform to tier-one VCs, so the window to set terms is short.

What we want: one company, or two with the money landing where the equity is. Control protections for them instead of a bigger percentage. No raise above $2M until a plan justifies more.
If we only cover three things: items 1, 2 and 3 in the left column. The money, where it lands, and the dilution floor. Everything else can wait a week.
1

The money is half what we thoughtCover

Phil put development at $1M over 12 months and said his and Rob's salaries are inside that number. Add about $1M of first-year G&A and the honest ask is $2M. Rob went further: the tech is built, MVP ready today, and the money is for infrastructure, APIs and patents, "not coding."

Do not carry $4M into the Blue Bear meeting. Their own team will contradict it. And at $2M, ask whether we want a raise at all.
2

Where the money lands decides everythingCover

Both proposals split this into a minerals company and a platform company (item 4). Phil's routing default was "if it's easier, do it all EIQ." Take it seriously, because nearly every dollar they described builds the platform: finishing the tech, infrastructure, API work, the patents, their salaries, the marketplace. The only ElementIQ-specific asset the money builds is the mineral knowledge base, and that comes out of our own engineers' heads.

So under Rob's split we raise the money, take the dilution, and pay to build something we own 30% of. The rule to hold: the IP they already built bought their stake. Money spent from here buys equity for whoever spent it.

Settle it without a fight: ask for one more column on the budget they already owe us, showing which entity's asset each dollar creates. Accounting question, not an accusation.
3

The 20% dilution floorCoverDecision

Phil wants Silica-X held at 20% no matter how much gets raised, and says you told him there are ways to do that. Here is the cost: if a round dilutes everyone by half, they stop at 20% and we absorb the share they did not give up. He asked for the floor and for equal dilution in the same breath, and both cannot be true.

Nobody challenged it on the call, so right now it looks accepted. Challenge it before you design around it. Decided
4

Two structures, neither agreed

ElementIQ (minerals)ProtoSentient (platform)
PhilOne company holding both. 50/50, dilute equally, with a 20% floor under them
RobUs 70, them 30Us 30, them 70

Rob calls the platform the bigger company, "reached beyond, into the AI economy." If he is right, his split hands us more of the smaller company and less of the bigger one. For scale: on 2026-07-03 Phil's ask was 10% earning toward 20%. On this call it was 50%.

5

Blue Bear: who is in the roomDecision

Recommend going in with three or four structures and asking them to poke holes, rather than with a number, and without Phil and Rob in the first meeting. Chris's reputation is on the line and we need an unfiltered read. They can join the second.

They are not neutral. Rob has pre-positioned them: he wants them as his own representation on the bigger ProtoSentient round and offered tag-along rights and first call. One input, not a ruling.

Decided
6

Counsel, and who we negotiate withDecision

Two hours with a startup and VC lawyer before Blue Bear, not three LLCs after it. Four non-experts designed a cross-funded two-entity structure live on a call.

Rob named his own motive: "once you lose control, your days are numbered." So the 50/50 and the floor are about control, not price, and control is far cheaper for us to give than equity. Offer board seats, consent rights on named actions and an exit floor. Take the percentage back.

Negotiate with Rob in the room. Phil is who signs. Decided

Proof we already have and are not using

SRNL ingested 150 documents both ways. Their system: 30 minutes, 60 failures, 12 stuck in queue. Silica-X: under 10 minutes, 149 clean, one corrupt file, replicated three times by the lab, on CPUs. Arm's-length, quantified, stronger than the Cory Trivelpiece story, and in no deck of ours.

Leverage and risk

For us: they do not want the minerals and government lane, Rob volunteered it. And they are cash-constrained, with unpaid IP counsel and a platform built on two six-year-old laptops, which argues for taking time.

Against us: Rob takes ProtoSentient to Nvidia within two weeks and has already contacted Andreessen and Lightspeed, so a tier-one could price the platform before we set terms.

Still unanswered, none of it urgent today

IP assigned or licensed? Rob argues a license is better for us since they carry prosecution and litigation cost. Partly true, and it turns on terms: perpetual, irrevocable, royalty-free, survives their insolvency, covers future improvements.

The patents, which are the moat. Of the 13, how many are issued versus pending? Claims are still being refined, their IP attorney works free accruing debt, and $200K of IP work is unfunded.

Employees or vendor? Employees of the new company, or an outside shop that gets paid and also holds equity.

Two more: the three universities and five national labs "coming on board" are signed, piloting, or just interested? And they want Blue Bear to recruit an outside CEO, explicitly not Chris and not you.

Next steps

Budget request, plus the "which entity's asset" columnMikeThis week
Open items to Silica-X in writing, so silence stops looking like consentMikeThis week
Engage a startup and VC lawyerChris and EllisBefore Blue Bear
Decide whether to lock ElementIQ terms before Rob's Nvidia meetingChris and EllisInside two weeks