Nexus Synergy — Engagement-Package Compendium (Main Volume)
Nexus Synergy — Engagement-Package Compendium (Main Volume)
Nexus Synergy — Engagement-Package Compendium (Main Volume)
⚠️ STANDING HONESTY BANNER — read before any number below
Nexus Synergy Ltd does not yet exist. This compendium is written pre-incorporation, pre-revenue, pre-pilot, with €0 cash on hand. No company has been formed; the founder is on a Stamp 1G immigration permission and cannot legally sign, invoice, or be a director of a trading entity until that is resolved (Stamp 4) or a partner/co-founder/brokered contracting vehicle is in place.
- Every buyer named in these 24 packages is a TARGET, not a customer. Nobody has been contacted on the company's behalf. No LOI, MOU, pilot, or contract is signed. The warm intros are potential paths the founder believes are reachable — not commitments. Patrick Walsh's "conditional cheque on a first LOI" is a genuine but conditional signal, not a commitment (confirm in writing before relying on it).
- Every probability in this compendium is a
[PRIOR]— a subjective, pre-pilot estimate used only to prioritise effort. None is a forecast. Each is dated 2026-06-05 and must be re-scored after every real conversation. A Brier calibration score is explicitly not claimed until ≥5 outcomes resolve.- Framing is civil-security / dual-use only. Resilience, common-operating-picture, critical-infrastructure protection, maritime-domain-awareness, search-and-rescue, coordination. Not warfighting.
- The Declined List is a hard filter applied before ranking. Predictive policing, real-time biometric identification, emotion recognition, social scoring, untargeted scraping, and psychometric profiling never enter the board. (An entire candidate cluster — financial-crime / justice, the empty
vol-D— was excluded on these grounds; it is published as Declined precisely so the omission is visible, not silent.)If any figure here is read as a promise of revenue, a signed deal, or a live deployment, it has been misread.
§1 — Methodology: how to read a package
This compendium is a set of 24 full-depth engagement packages, organised into four published cluster
volumes (A civil-protection, B maritime-critical-infrastructure, C airspace-cyber, E European) plus this main
volume and an index. (vol-D financial-crime/justice exists as a directory but is intentionally empty —
the cluster was Declined; see the banner.)
Each package is a 23-block ready-to-go GTM playbook for one target, written to the same template so a reader can jump to the same block in any package and compare like-for-like. The blocks run, in order: §1 buyer & mandate → §2 the pain → §3 the wedge → §4 decline-safety invariant → §5 the schema/DB invariants → §6 funding & procurement reality → §7 competitive/incumbency → §8 the demo lens → §9 architecture fit → §10 data & lawful-basis → §11 the pilot design → §12 success criteria → §13 commercials/pricing tier → §14 risk register → §15 engagement playbook (first-contact → signed contract) → §16 PM / timeline (Gantt + RACI + critical path) → §17 funding / procurement vehicle → §18 the two-stage formula scorecard → §19 commercial terms → §20 the legal-gate note → §21 references → §22 the ASCII/SVG mockup set → §23 the one-page leave-behind.
Three disciplines govern every package:
- Zero fabrication. Numbers are either
[verified](cited mandates, funding amounts, pricing tiers) or[PRIOR](subjective scores). Where an input is genuinely unknown it is written[value: TBD]or flagged[unverified]— never invented. A package's own claims prove nothing about reality; they are inputs to be tested, and several rows carry explicit "this is the most uncertain number in the package" flags. - The Gemini Phase-6.5 grounded gate. Each package's load-bearing factual claims are extracted to
_build/validation/claims-<slug>.jsonand run through an independent grounded-verification gate (a Gemini pass that labels each claim SUPPORTED / CONTRADICTED / UNSUPPORTED / UNVERIFIABLE). The discipline is that a claim a model cannot ground is demoted, not carried as fact. See §8. - The anti-cookie-cutter mockup doctrine. Every screen mockup in a package is generated from a distinct
layout archetype (
map_ops,globe_rmp,command_dashboard,stix_board,graph_explorer,cap_composer,ledger,video_studio,settings_form, …) via_build/archetypes/archetypes.py— never one shared template with the text swapped. The same product surfaces appear across packages because they are the same real product (/synergy-console'sMapConsole,HistoryRail,ChatTerminal-as-Decision-Room), but each buyer's lens is a genuinely different composition.
§2 — The formula (restated in one place)
All scoring uses the two-stage LOI→contract model from strategy-paper/formula-worked-examples.md. It is a
prioritisation instrument, not a forecast engine; every factor score is a dated [PRIOR].
Stage 1 — P(LOI) (horizon 12–18 months). S1 = Σ wᵢ·fᵢ, each fᵢ ∈ {0..5}.
P_LOI = 0.55 / (1 + exp(−1.15·(S1 − 2.6))) [L_max=0.55, S0=2.6, k=1.15]
Stage-1 factor vector and weights: mandate_pull (.22), access_warmth (.22), demonstrability (.18),
decline_safety (.14), white_space (.10), cycle_speed (.08), pillar_fit (.06).
Stage 2 — P(Contract | LOI) (horizon 24–36 months). S2 = Σ vⱼ·gⱼ.
P_raw = 0.70 / (1 + exp(−1.1·(S2 − 2.8))) [M_max=0.70, S2_0=2.8, k2=1.1]
Stage-2 factor vector and weights: contractability (.24), funding_to_pay (.22), procurement_clarity
(.18), incumbency_displacement (.16), time_to_value (.12), reference_leverage (.08).
The legal-gate multiplier G = V · T — the term that dominates the whole portfolio.
V(partner-vehicle availability): 0.10 solo founder on Stamp 1G (the default today) · 0.45 co-founder / Dogpatch-brokered entity in discussion · 0.75 MAHI (or a partner) confirmed as the contracting prime · 1.00 Stamp 4 + incorporated & trading.T = min(1, max(0, (H_conv − t_resolve)/H_conv)), withH_conv = 30 monthsandt_resolve= months until a signature is legally possible.
Joint probability, EV, and the board score:
P(Contract) = P_LOI · G · P_raw
EV = P(Contract) · ACV(tier) · μ_strategic · D(t_cycle)
μ_strategic = (1+r_ref)(1+e_expand) , D = 1/(1.25)^(t_cycle/12)
Score100 = S1 · 20 (for reconciliation with the existing master board)
The single most important structural fact the formula encodes: today, every package's G is crushed by
V = 0.10. That is the gate, not the opportunity. Resolving the vehicle (V: 0.10 → 0.45 → 0.75 → 1.00) is
the highest-leverage action in the entire compendium, and — because one resolution lifts every V at once —
it is also the dominant correlation in the P&L (§7).
§3 — The ranked master board (all 24 packages)
Ranked by EV-to-LOI, operationalised as Score100 descending (Score100 = S1·20 scales with P_LOI, and
near-term action turns on getting to an LOI). Columns: P(Contract) today is at V=0.10 (solo, Stamp 1G);
P(Contract) post-vehicle is at V=1.00 (Stamp 4 / incorporated). Every probability is a [PRIOR].
| # | Package (buyer) | Cluster | Tier | S1 | P_LOI | P(C) today V=0.10 | P(C) post-vehicle V=1.0 | Value | Score100 |
|---|---|---|---|---|---|---|---|---|---|
| 1 | a01 NDFEM / storm-hit LA | civil-protection | NOW | 4.02 | 46% | ~1.0% | ~16% (cond.)¹ | high | 77 |
| 2 | e04 Baltic Sentry — cable protection | european | NOW-EU | 3.80 | 44% | ~0.9% | ~9.3% | high | 76 |
| 3 | e01 Belgium REPMUS / MIK seabed (MAHI) | european | NEXT-EU | 3.76 | 43.5% | ~1.3% | ~12.8%² | high | 75 |
| 4 | e05 Nordic Gulf-of-Finland subsea | european | NEXT-EU | 3.76 | 43.5% | ~0.8% | ~8.2% | med-high | 75 |
| 5 | b03 SFPA fisheries-control | maritime-crit-infra | NEXT | 3.72 | 43.1% | ~0.8% | ~8.3% | high | 74.4 |
| 6 | c01 daa / IAA counter-UAS | airspace-cyber | NOW | 3.72³ | 41% | ~0.5% | ~31% (cond.)¹ | high | 74 |
| 7 | e08 EDF-2026 / EUDIS consortium | european | NOW-EU | 3.70 | 43% | ~1.2% | ~18.1% | very-high | 74 |
| 8 | e03 EDA / PESCO CSIP+UMS | european | NEXT-EU | 3.64 | 42% | ~0.8% | ~8.2% | high | 72.8 |
| 9 | a02 DART / Irish Rail + NTA | civil-protection | NOW | 3.72³ | 41% | ~0.4% | ~7% (cond.)¹ | high | 72 |
| 10 | e06 Netherlands Coastguard autonomy | european | NEXT-EU | 3.62 | 42% | ~1.0% | ~10% | med-high | 72 |
| 11 | e02 EMSA CISE national node | european | NEXT-EU | 3.52 | 41% | ~0.85% | ~8.5% | med-high | 70 |
| 12 | a04 TII roads / tunnels (M50) | civil-protection | NOW | 3.385⁴ | ~40% | ~0.5% (cond.)¹ | ~6% (cond.)¹ | high | 67.7 |
| 13 | a05 DCC / LGMA urban-ops | civil-protection | NEXT | 3.80→3.48⁴ | 44%→band | ~0.65% | ~6.6% | med | 64 |
| 14 | e09 France seabed-infra (DISCOVERY-ONLY) | european | NEXT-EU | 3.18 | 36% | ~0.2% | ~2.3% | high | 63.6 |
| 15 | e10 Spain PESCO HARMSPRO | european | NEXT-EU | 3.14 | 36% | ~0.34% | ~3.4% | high | 63 |
| 16 | a03 OPW / Met Éireann street-flood | civil-protection | NEXT | 3.14⁵ | ~33–44% | (cond.)¹ | (cond.)¹ | high | 62.8⁵ |
| 17 | b01 Naval Service / MAHI MDA | maritime-crit-infra | NEXT | 3.14 | 35.8% | ~0.8% | ~7.9%² | very-high | 62.8⁶ |
| 18 | c03 ComReg telecoms / subsea | airspace-cyber | NEXT | 3.60→band⁴ | ~40% | ~0.6% | ~5.9% | high | 62.2 |
| 19 | b05 Uisce Éireann water OT-security | maritime-crit-infra | NEXT | 3.08 | ~35% | ~0.4% | ~3.8% | high | 61.6 |
| 20 | b04 Dublin Port CER/NIS2 | maritime-crit-infra | NEXT | 3.00 | 33.7% | ~0.6% | ~5.7% | high | 60.0 |
| 21 | b02 Coast Guard SAR (virtual-JRCC) | maritime-crit-infra | NEXT | 2.96 | 33% | ~0.4% | ~7%³ | med | 59.2 |
| 22 | c02 NCSC-IE NIS2 / CSIRT | airspace-cyber | NEXT | 3.54→band⁴ | ~41%→band | ~0.5% | ~4.9% | high | 58.1 |
| 23 | c04 HSE / NAS major-incident | airspace-cyber | NEXT | 3.36→band⁴ | ~30–39% | ~0.4% | ~4.3% | high | 56 |
| 24 | e07 FRONTEX / EUROSUR (gated, decline-policed) | european | LATER-EU | 2.76 | 30% | ~0.2% | ~2.1% | high (strat.) | 55.2 |
Ranking note. This board ranks on EV-to-LOI (Score100 ↓, a proxy for P_LOI·value), which is the right sort for near-term action — get to a first LOI. The EV-to-Contract is the post-vehicle prize: read the "P(C) post-vehicle V=1.0" column for that, and note that the very-high-value rows (e08 EDF, b01 Naval/MAHI) jump materially once the gate is resolved even though their to-LOI rank is mid-board.
Footnotes / reconciliation notes (zero-fabrication trail):
- ¹ Several civil-protection rows (a01, c01, a02, a04, a03) report the V=1.0 column as a conditional P(Contract|LOI) rather than the joint, because their own §18 tables present the gate cascade that way; where a cell is marked "(cond.)" it is the conditional figure as written in that package, not the joint. The joint today-figures (P(C) V=0.10) are stated where the package gives them.
- ² Rows e01 and b01 are MAHI-gate-resolver cases: their realistic vehicle is V=0.75 (MAHI prime), where P(Contract|LOI) reaches ~21.6% (e01) and ~16.6% (b01). MAHI's value is that it lifts the gate (V 0.10→0.75), not Stage 1 — exactly worked-example (D). The V=1.0 joint figures shown are the Stamp-4 ceiling.
- ³ Where a package's §18 states a slightly different S1 in prose than its front-matter Score100 implies (rounding within ±0.05), the front-matter Score100 is carried as the row value.
- ⁴ Rows a04, a05, c02, c03, c04 compute a raw S1·20 above their board anchor; each package applies the board's adversarial haircut (access-warmth realism / cold-access / clearance / procurement-freeze discounts) and carries the board-anchored Score100 for portfolio consistency (raw shown for transparency). a05's raw 76 trims to ~69.6 then defers to the 64 anchor; c02's raw 70.8 defers to 58.1; c03's raw 72 defers to 62.2; c04's raw 67.2 defers to 56; a04 carries 67.7.
- ⁵ a03 (OPW): front-matter Score100 = 62.8 is carried here; its board anchor is 60.6 (the §18 body treats the flood pilot as a build, not a demo-and-deploy, hence demonstrability=2). Within tolerance.
- ⁶ b01 (Naval/MAHI): front-matter Score100 = 62.8; board anchor 66.4 (within ~4 pts). Value band very-high (national-centre + EDF scale) — the row's prize is in the post-vehicle column, not the rank.
- EU rows e01–e10 had
board_anchor.score100 = "[PRIOR-derive]"in the spec; their Score100 is the Stage-1-derived value computed in each §18 and is flagged as derived, not pre-anchored. - e05, e09, e07 are research-flagged / low-confidence in the spec (
confidence: unverified/ DISCOVERY-ONLY); they are scored (named buyer + procurement route established) but carry an explicit re-score-on-first-contact flag. No row is null.
§4 — Cross-package synthesis
The cluster narrative. The 24 targets resolve into four strategic stories:
- Civil-protection (A: a01–a05). The warmest, fastest-to-discovery cluster, anchored by the live Storm-Éowyn-review recommendations and the CER/NIS2 resilience mandate. One storm-replay demo unlocks five Irish targets (NDFEM, DART, TII, DCC, OPW) — see §5. This is where the first LOI is most realistically manufactured, even though the contract is gated like everything else.
- Maritime / critical-infrastructure (B: b01–b05). Higher strategic value, slower cycle. The MAHI founder-to-founder relationship is the unlock: it does not raise demand, it resolves the legal gate by acting as the contracting prime (V 0.10→0.75). The Naval/MDA row (b01) is the bridge into the whole European maritime cluster.
- Airspace-cyber (C: c01–c04). Mixed: daa counter-UAS (c01) is a genuine NOW with the cleanest decline-safe hook on the board (object-tracking, explicitly not passenger biometrics); NCSC/HSE/ComReg are mandate-strong but access-cold and clearance/accreditation-gated.
- European (E: e01–e10). The strategic depth tier and the non-dilutive engine. Driven by the acute EU cable / seabed forcing functions (Baltic Sentry, e04, is the single most-funded, most-time-sensitive mandate in the portfolio) and the EDF-2026 / EUDIS consortium play (e08), which is the one row with a 100%-funded EU grant line on a clock rather than a customer budget.
The Stamp-1G legal-gate cap binds the whole portfolio. Read down the "P(C) today V=0.10" column: every single row is ~0.2%–1.3%. That is not 24 independent weak opportunities — it is one common cause (the founder cannot sign) multiplied across 24 rows. The binding constraint is the vehicle, not the buyers' appetite.
The MAHI gate-resolver + EU-consortium wedge. MAHI BV (Belgium) appears as the gate-resolver in b01, e01,
e02, e03, e05, e06, e08 — wherever it can act as the contracting prime, V jumps to 0.75 and the maritime/EU
rows become genuinely pursuable. The EU consortium play (EDF-2026 / EUDIS, e08; close 29 Sep 2026) is a
Belgian-MAHI + Irish-Nexus team contributing the COP decision-layer module to a CSIP/seabed consortium — a
non-dilutive route that also is the vehicle resolution for the maritime cluster.
§5 — The shared storm-replay COP demo
Status: build-spec, not yet built. One demo on the live /synergy-console map replays a real past Irish
storm (Éowyn 24 Jan 2025 primary; Chandra 7–8 Feb 2026 coastal-DART variant) from open/recorded data, then
re-skins per buyer via a saved-view lens. The fusion, map, action loop, and audit ledger are identical across
lenses — which is exactly the product thesis. One build → five discovery conversations.
| Master rank | Buyer | The lens this demo shows them |
|---|---|---|
| #1 | NDFEM / a storm-hit LA | responder COP above public-alerting tools; Éowyn-review Recs 3.11.6 (reach vulnerable people) + 3.11.7 (operational comms) |
| #2 | DART / Irish Rail + NTA | coastal-line closure decision loop (Storm Chandra Seapoint–Blackrock); a low-cost decision-support overlay, not a TMS replacement |
| #6 | TII | M50 / national-roads incident coordination across the Interagency Incident Coordination Group |
| #8 | Dublin City Council / LGMA | Local Critical Entities Group situational awareness + vulnerable-person outreach |
| #11 | OPW + Met Éireann + LAs | street-level flood operating picture closing the "county-level warning told people nothing about their street" gap |
(Master-rank numbers above are the source-card ranks the demo file uses; they map to a01, a02, a04, a05, a03 in this compendium.)
The five-beat narrative (≈6 minutes):
- T-12h — warning. Met Éireann red/orange warnings paint onto the map; the COP pre-stages assets.
- T-0 — impact. ESB outage polygons bloom; tide/surge breaches a coastal cell; road + rail closures appear; a telecoms-degradation layer flags areas that lost both power and comms.
- Triage. The system surfaces the highest-priority cluster automatically — the area that lost power and comms and contains a concentration of registered vulnerable residents: the intersection no single agency's tool shows today.
- Human-gated decision. The operator reviews the recommendation in the Decision Room
(
ChatTerminalright-dock), sees the dissent/considerations, and signs off one action: open a Community Support Centre + dispatch a generator + issue a targeted local alert. - Write-back + audit. The action is logged with a full PROV-O attribution trace on the
HistoryRail; the operator can scrub back and replay exactly who knew what, when, and who authorised what — the auditable record the Éowyn review asked for.
Decline-safety, stated up front in every showing: operational coordination + resilience only; the vulnerable-person layer uses consented, already-held LA registers / aggregate census geography (surfacing who registered for help, never inferring who "might" be vulnerable); everything human-gated and fully audited. This is the line that keeps the whole cluster off the Declined List.
§6 — Funding / consortium map
The Irish non-dilutive stack (phase → instrument → corrected eligibility gate, from €0 cash):
Phase 0 Incorporation + constitution → founder cash; €25k s.137 bond IF no EEA-resident director [gate: Stamp 4]
Phase 1 Technical validation → Innovation Voucher €5k (CeADAR/UCD) [solo-accessible]
Phase 2 Free-discovery / feasibility → HPSU Feasibility Grant ≤€30k [HPSU path]
Phase 3 Early pilot capital → PSSF ≤€100k convertible [± NDRC €100k SAFE] [convertible = only intentional dilution]
Phase 4 Maritime beachhead pilot → Defence Innovation Challenge €3.2m pool [ACADEMIC-ROUTED — via a university PI, NOT solo]
Phase 5 Maritime/dual-use dev at scale → EUDIS / EDF-2026 (close 29 Sep 2026) + €120k voucher [CONSORTIUM-ONLY — needs MAHI]
Phase 6 Scale equity → Innovative HPSU Fund ≤€1.2m [post-HPSU-designation]
Phase 7 Growth → EIC Accelerator / DTIF / Horizon Europe
The corrected eligibility gates are load-bearing (and were corrected because they materially change the plan): the Research-Ireland Defence Innovation Challenge is academic-routed — it must go through a university PI (e.g. CeADAR/UCD), not the solo founder; and EUDIS / EDF are consortium-only — they need MAHI (or another partner) and a ≥3-Member-State team. Award probability is 0 until the gate resolves; this is why the Year-1 grant fan is honestly thin.
The EU consortium plays. EDF-2026 / EUDIS (e08, close 29 Sep 2026) with a Belgian-MAHI + Irish-Nexus team contributing the COP decision-layer to a CSIP/seabed-protection consortium; the EDA/PESCO CSIP+UMS enablers (e03) where Ireland already participates (neutrality-compatible); and the EU Cable Security Toolbox / CEF Digital lines behind the Baltic and Nordic cable rows (e04, e05).
The partner-vehicle resolution of the Stamp-1G gate. Three paths, in increasing strength: a Dogpatch-brokered entity (V→0.45); MAHI as confirmed contracting prime (V→0.75, the realistic resolver for the maritime/EU cluster); or full Stamp 4 + incorporation (V→1.0). An EEA-resident director (Patrick Walsh / Manuel Loureiro) or the €25k s.137 bond unblocks incorporation independently of the founder's own Stamp-4 timeline.
§7 — Portfolio $0-start Monte-Carlo P&L
Method (from strategy-paper/pnl-monte-carlo.md): the P&L is built bottom-up from the 24 named pipeline
targets, never top-down from a market %. A correlated Monte-Carlo (Gaussian copula, ≥10,000 seeded
iterations) samples, per iteration: draw S1, S2 from triangular/PERT three-point distributions off the factor
scores ±σ; compute P(Contract) = P_LOI · G · P_raw; draw a Bernoulli win; if won, draw close-month and tier.
The dominant correlation is ρ(legal_gate, ALL) = +0.9. One Stamp-4 / vehicle resolution lifts every
target's V at once — a single common cause, not independent per-deal risk. Secondary structural correlations:
civil-protection cluster +0.6 (one storm-replay demo + one Patrick relationship co-move NDFEM/DART/TII/DCC/OPW);
maritime cluster +0.5 (shared MAHI vehicle + Defence intro); all others 0.0. The Cholesky factorisation induces
correlation only through the shared latent, so an "all-win-at-P90" tail is impossible by construction — a
deliberate guard against false precision.
Two output distributions, NEVER summed. (1) Booked-ARR — P10/P50/P90 fans for Years 1–3, recognised at
the Contract stage only; a grant-funded pilot is never ARR. (2) Grant income — a separate correlated
Monte-Carlo over the non-dilutive stack, each instrument = Bernoulli(award) × amount × timing, with award-prob
= 0 until its corrected eligibility gate (§6) resolves. The document renders them as two side-by-side fan charts
under an explicit "ARR and grant income are never summed" banner.
Three scenarios (publish all three; never a point runway):
- Conservative — only Innovation Voucher (€5k) + HPSU Feasibility (€30k) land; no buyer pilot in Year 1; §137 bond taken.
- Base — + PSSF (€100k convertible) lands ~Q3; one Tier-1 LOI converts via a partner vehicle in Year 2.
- Upside — + Innovative HPSU Fund (€1.2m) post-designation; a MAHI EUDIS-consortium bid funds the maritime dev track; 2–3 Tier-1 contracts in Years 2–3.
Runway = cash_on_hand / net_monthly_burn, both as distributions, from €0.
Numbers TBD where genuinely unknown — not fabricated. Verified inputs: pricing Tier-1 €120–240k, Tier-3
€2.4–4.8m; funding amounts as in §6; EDF-2026 close 29 Sep 2026. Genuinely still [TBD] (do not invent):
Tier-2 ACV midpoint; per-tenant compute & sovereign-inference cost-per-call; conformity-file cost; founder
cash-comp; headcount ramp; gross margin; net revenue retention. Live numbers belong in financial-model.xlsx
(to be built) and the gtm-genesis ledger — no ARR is asserted here. (Comparators such as NHS-FDP £330m/7yr
or Quantexa-HMRC £175m/10yr are context only, never this company's revenue; the Adverant/NexusROS deck numbers
are explicitly not carried into this model.)
§8 — Validation summary
Every package's load-bearing claims are extracted to _build/validation/claims-<slug>.json (one file per
package — all 24 are present) and run through the Gemini Phase-6.5 grounded gate, which labels each claim
SUPPORTED / CONTRADICTED / UNSUPPORTED / UNVERIFIABLE. The discipline is the point: a claim the gate cannot
ground is demoted from fact, not silently carried; a CONTRADICTED claim is a hard stop for that claim. This is
the operational form of the zero-fabrication rule — a package's own AS-WRITTEN assertion proves nothing until it
survives independent grounding.
[gate results: see
_build/validation/gate-summary.tsv]
(The summary table is regenerated by _build/run-gate.sh; per-package rows populate as the gate completes. The
header schema is slug · SUPPORTED · CONTRADICTED · UNSUPPORTED · UNVERIFIABLE · ERROR · total.)
Closing note — honesty, and the window
This compendium is a prioritised plan against 24 named targets, written from €0 with no entity, no signed
relationship, and a founder who cannot yet legally contract. Held honestly, that is its strength: the binding
constraint is named (the legal gate), the highest-leverage action is named (resolve the vehicle), the first
realistic LOI is named (a storm-hit LA via the storm-replay demo), and the non-dilutive runway is mapped with
its eligibility gates corrected. Every number is a [PRIOR] to be re-scored on first contact, and a whole
candidate cluster sits on the published Declined List rather than in the pipeline.
The window is real and dated. The EU AI Act (prohibited-practice bans live; high-risk obligations phasing in) makes conformity-by-construction a wedge, not a tax. NIS2 / CER designations land 17 Jul 2026, creating funded resilience-spend lines across the civil-protection and critical-infrastructure clusters. The EDF-2026 call closes 29 Sep 2026 — a hard date for the consortium play. And Ireland holds the EU Council Presidency in H2 2026, putting sovereign-EU decision-intelligence on the agenda at exactly the moment this portfolio is ready to move. The gate is the constraint; the window is the reason to resolve it now.