A use-of-funds estimate, plus four ownership and licence terms. $400K for a licence and an MVP, $200K for an internal LLM, $600K for 30% of Protosentient. Nearly all the terms were said on the 8/3 call. Three are new: the $600K price, a 10-day option window, and a licence covering two segments with nuclear excluded.
Two absences. No line-item budget, though one was offered on the call. No investor: ElementUSA is the named funding party for all $1.2M. Their document closes by stating it is not a binding offer.
| Piece | Terms as written |
|---|---|
| Phase 1 months 1 to 4 $400K | Perpetual licence to EIQ for critical materials and waste recovery. White-label build, MVP at month four, three resources dedicated. Payments: $200K at signing, $100K at day 45 against an unspecified "interim milestone," $100K on MVP delivery. EIQ owned 70% ElementUSA / 30% Protosentient, with a 20% anti-dilution floor on their 30%. |
| Phase 2, optional months 5 to 8 $200K | Internal LLM trained and delivered at month eight. Inference, hosting and operating costs excluded, and fall to EIQ. |
| Platform option $600K | 30% of Protosentient. The same $600K funds the Exchange/Marketplace build, delivered within twelve months. Window opens on execution, closes ten days after MVP delivery. After exercise: EIQ 70/30 ElementUSA, Protosentient 30 / 70. The Exchange passes to EIQ on the Phase 1 terms, so within the two segments only. |
Seven questions a position has to answer. Each structure below answers all seven differently.
| Structure | What we would put forward | What it costs us |
|---|---|---|
| A. Their plan, four terms changed | Accept the licence, EIQ 70/30 and the option. Require exclusivity written in, an MVP definition and acceptance test we set, payments weighted to delivery, and a 90-day option window. | Fastest start, least friction, keeps the relationship intact. We remain a licensee, not an owner. |
| B. Same, but take the option now | Fold the $600K into the initial commitment so the 30% is secured at $2.0M before any outside round can reprice it, and the 10-day window stops mattering. | $1.0M out instead of $400K, committed before the MVP exists. Removes the timing trap. |
| C. Licence only, no equity | Buy Phase 1 as a straight software licence and build. No EIQ entity, no Protosentient stake, no cap table. | Simplest and cheapest to paper. Gives up all platform upside and the marketplace. |
| D. Assignment in-field | The two segments' IP assigned to EIQ rather than licensed, with their 30% of EIQ as the consideration for it. | Strongest ownership position. Hardest ask, and Rob argued against it on the call. |
Nobody at ElementUSA has assessed the platform's defensibility. Silica-X has $200K of unfunded patent work and an unpaid IP attorney. Their plan carries no remedy if the MVP slips or fails. Rob reported an SRNL benchmark, 150 documents in under 10 minutes against 30 minutes and 60 failures for the comparison system, replicated three times by the lab; unconfirmed in writing.
On the option's price: $600K for 30% implies $2.0M post-money, $1.4M pre. A prior internal brief put a defensible range at $1.0M to $2.5M, not retested since June, against a category median pre-money nearer $7.7M and falling. So the price is low; the open question is whether the 30% attaches to the IP, which is question 3 on page 2. Rob said on the call that a round "in this space" runs "10 to 12", speaking generally and without pricing Protosentient. He has contacted Andreessen and Lightspeed, and takes Protosentient to Nvidia within two weeks.
Their plan is a licence and a build: three resources dedicated through MVP delivery, a payment schedule, a delivery date. It does not describe them operating, staffing, supporting or leading EIQ, and it contains no cost of running the company.
| Line | Our own deck's year-one allocation | Funded by their plan |
|---|---|---|
| Engineering and platform | 55% of ~$4M = ~$2.2M | $600K (Phase 1 + Phase 2) |
| Go-to-market | 30% = ~$1.2M | Nothing |
| Operations | 15% = ~$600K | Nothing |
| Platform equity | Not in the deck | $600K option |
So the year-one figure is not $1.2M. On our own deck's percentages, roughly $1.8M of go-to-market and operations sits outside their plan entirely. Phil put year-one G&A at "easily another million" on the same call, which points the same way.
Four things their plan leaves to us. Running it: Phil said an outside CEO should be recruited, and not from our existing leadership. Not costed anywhere. Knowledge capture: Rob named it the real delivery risk, extracting knowledge from engineers who have day jobs, which is our engineers' time. Selling it: Rob said "we'll give you government, we're tired of government." Everything after month four: no support terms, no SLA, no maintenance and no continuing engineering commitment, while they hold 30% of EIQ.
On their role, they are contractors rather than officers. "Three resources dedicated to the engagement" is vendor language, and they run Silica-X. Three consequences. Contractor work product belongs to the contractor unless the agreement assigns it. Nothing commits their time exclusively or bars adjacent work. And our own deck lists Rob as Chief Architect and Phil as Chief AI Officer, which reads as officers to anyone doing diligence; either the titles get backed by agreements or they get softened.
| Term | Origin |
|---|---|
| EIQ 70 / 30, and 30% of Protosentient to us | Rob proposed both on the 8/3 call |
| 20% anti-dilution floor on their 30% of EIQ | Phil raised it on the call, understood to be workable. Never formally accepted |
| A licence rather than an IP assignment | Rob argued for it at length on the call. Never formally accepted |
| $600K for the 30% | New. On the call the stake was contributed, not bought |
| Option window closes 10 days after MVP | New. Not discussed |
| Two segments only, nuclear excluded | New. Rob had said "all things minerals, all things nuclear" |
On the call the stake was consideration, not a purchase. Rob: "we contribute IP, that's where we get our equity. You contribute capital, that's where you get your equity." No price was attached. In their new document only their side is contributed: their 30% of EIQ is for "platform and licence contribution," and our 30% of the platform is a purchase.
| June proposal | 20% of the Wastepoint parent, contributed, for capital, the customer and commercialisation |
| 8/3 call | 30% of Protosentient, contributed |
| 8/3 document | 30% of Protosentient, $600,000 cash |
| Nvidia meeting | Within two weeks of 2026-08-03, per Rob. A tier-one valuation would sit against the $2.0M their option implies. |
| The option window | Opens on execution, closes ten days after MVP delivery. Silica-X sets the delivery date, so earlier delivery shortens it. |
| Counsel | Not engaged. Nobody has read the licence or the option. |
| Our own deck | Already with Ernest, describing a $4M seed raise. Their plan has no investor in it. |
| Blue Bear | Rob has not approached them; what he knows is secondhand from our side. He said he would want them representing Silica-X on the larger Protosentient round, and proposed giving them tag-along and drag-along rights plus first call. |