DEFENSE / NATSEC
A provenance-controlled accountability layer for AI-enabled mission work
Deterministic replay, authority and policy binding, and signed Decision Receipts that verify offline.
Investor hub · Confidential Pre-seed · 2026
Proof, not capability, is the scarce resource.
The runtime admissibility layer for AI in high-stakes decisions.
Most organizations can tell you whether an AI system produced an output. Almost none can prove that the output was decision-valid — authorized, evidence-linked, and policy-compliant — captured at the moment the decision was made. Summit Cognitive closes that gap. Every governed action ships a Decision Receipt: a cryptographically signed, replayable, independently verifiable record of why the answer was allowed, revised, blocked, or escalated.
The thesis
A log shows what happened. A Decision Receipt proves who authorized a decision — and that it holds up under audit later.
The first wave of AI adoption optimized for productivity. The next wave will be judged by decision accountability. Fluency helped AI enter the workflow; defensibility determines whether it can stay there. Unverified cognition is operational risk — and the risk is not that AI is wrong. The risk is that AI is wrong in ways the operator cannot detect and the auditor cannot reconstruct.
Explanation is not proof. Observability tells you what happened; audit logs tell you who did it. Neither tells you why the system reached this conclusion, with this evidence, under this policy, captured at decision time. Summit operates at the knowledge-graph layer, not the model layer — a new control plane we call Cognitive Security. It is the runtime admissibility layer that AI governance, model safety, data security, and audit firms all need underneath them. A category, not a feature.
This is built, deployed, and in active pilot motion — not a model wrapper, not a UI overlay, not a research project. The product is live in production at decrec.summitcognitive.ai, with a public key and ledger for independent verification. Positioning versus Palantir is complementary, not competitive: Palantir builds the investigator's notebook; Summit builds the notarized record beside it. We do not replace the analysis — we make the analysis admissible.
Stage
Pre-seedPre-revenue
Instrument
SAFEYC post-money, negotiable cap
Target raise
$1.5M1 lead + 1–2 co-investors
Cap table
100% founderNo notes, no outside dilution
Non-dilutive first: a Colorado OEDIT Advanced Industries grant (~$250K) opens July 1, 2026. Valuation cap is to be set with the lead investor and is not stated here.
Why now
A regulatory window and a category vacuum are arriving at the same time.
The recurring CISO question Summit answers: When this fails, how do we know it failed — and can we prove to a regulator that we knew? No one else is shipping runtime admissibility infrastructure for autonomous AI actions. Cognitive Security operates at runtime — every decision, every time, signed before the output leaves the system — unlike AI governance, which operates on review cycles.
Regulatory tailwinds are converging on exactly this requirement: the OMB federal AI governance mandate; the EU AI Act (Articles 13–15 high-risk and runtime obligations) binding August 2, 2026 with penalties to €15M / 3% of turnover; and NIST AI RMF converging with bank model-risk guidance (SR 11-7), FedRAMP, CMMC, and IC Directive 503.
CYBER
Monitoring tells you something went wrong
A Decision Receipt proves what happened: signed, replayable, independently verifiable.
ENTERPRISE AI GOVERNANCE
The control plane for high-risk enterprise AI
Every AI-assisted decision ships a receipt — provenance, authority, policy — that holds up to audit.
THE CATEGORY
First-mover in decision admissibility
Summit owns the vocabulary: Cognitive Security, the Decision Receipt, the runtime admissibility layer. If a decision cannot be replayed, it should not be trusted.
Public traction
Live, in production, and independently verifiable. The figures below are public-level only.
3,000+
Signed Decision Receipts issued in production
Stated conservatively · public ledger
98.6%
Receipt acceptance rate (deny by default)
Capability Statement
99.9%
Replay determinism
Capability Statement
Sub-second
End-to-end evaluation latency
Capture through receipt issuance
99.9%+
Uptime since production launch
Tech-architecture one-pager
Ed25519
Hardware-backed signatures, public-key verification
OPA-gated, deny-by-default policy
52
Canonical packages, five integrated services
Built and maintained by one person
TRL 5
Technology Readiness Level
Decision Receipt one-pager · Federal posture
$2.1M
Federal pipeline across three proposals
SBIR Phase I base award $150K
Federal credentialing is complete: UEI assigned, SAM.gov registered, CAGE code assigned. Provisional patent application No. 64/034,952 was filed April 10, 2026; the non-provisional conversion is due April 10, 2027. Two flagship channels — the Dispatches essay series and the Warrant podcast — are in production. Engineering metrics are audit-time figures (late May to early June 2026); some have since advanced. SPRS / NIST 800-171 self-assessment figures and full pipeline detail are available in the gated data room. Independent review by Galois and Trail of Bits is scoped, not completed — Summit does not claim "independently reviewed by."
The data room
Access-gated. Below is the index of what's inside — request access to view the contents.
Confidential financials, the cap table, the valuation, and term-sheet specifics are not published on this page. They live in a redaction-by-default data room released to qualified investors with an active diligence engagement. The cards describe each section without exposing its contents.
-
Executive summary & master deck
Company overview and positioning, plus the master pitch deck — retargeted per investor. Variants run from a 6-slide cold-intro teaser to an ~11-slide standard deck to a 40-slide comprehensive deck.
Request access → -
Corporate & legal
Certificate of incorporation (Delaware C-Corp), bylaws and board consents, founder stock, IP-assignment posture, and active federal registrations (EIN, UEI, CAGE, SAM).
Request access → -
Intellectual property
Provisional patent No. 64/034,952, the continuation and international PCT strategy, and the deep-IP / category-creation surface. Patent-sensitive implementation detail is available under NDA only.
Request access → -
Product & technical architecture
The investor-safe four-layer governed decision pipeline, deployment and security architecture, the live Decision Receipt endpoint, and architecture-diagram appendices.
Request access → -
Compliance, trust & security
SPRS / NIST 800-171 self-assessment with remediation trajectory, NIST AI RMF mapping, the security & trust one-pager, and the SOC 2 / FedRAMP / CMMC roadmap (designed-for-alignment, not yet certified).
Request access → -
Financial model & cap table
The 18-month model (base and downside cases), burn and runway narrative, and the current cap table. Released only to investors in active diligence — figures are not published here.
Request access → -
Go-to-market & traction
Live proof metrics, the anonymized pipeline / relationship map, ICP and pricing appendix, and the pilot, design-partner, and federal capture motion. Pipeline numbers are deliberately ranged, never customer-named.
Request access → -
Team & advisors
Founder bio (cleared variants), the three-role hiring plan, the advisor profile matrix, and the advisory-agreement template.
Request access → -
Federal credentials & pipeline
Capability statement, opportunity register, active submissions, federal readiness packet, and the OTA / FAR / CSO posture.
Request access → -
Investment & term summary
The pre-seed SAFE term summary (illustrative), use of funds, and funded milestones. The term summary is a discussion template — not an offer, and not legal or financial advice.
Request access → -
Content & narrative
Press kit and fact sheet, plus the category-authority program: the forthcoming book Admissible Reality, the Dispatches essay cadence, the paper program, and the Warrant podcast.
Request access → -
Detailed pipeline review
The full investor pipeline and active-diligence relationship map. Available on request to investors with an active diligence engagement.
Request access →
Investor FAQ
What exactly does Summit sell?+
A governed decision platform that cryptographically proves every autonomous action was authorized, evidence-linked, and policy-compliant before execution. The product turns an AI-assisted decision into a governed artifact — the Decision Receipt — that shows why the answer was allowed, revised, blocked, or escalated. We do not replace the analysis; we make the analysis admissible.
How is this different from observability, audit logs, or AI governance?+
Explanation is not proof. Observability tells you what happened and audit logs tell you who did it — but neither tells you why the system reached this conclusion, with this evidence, under this policy, captured at decision time. Summit is not an AI governance company. It is the runtime admissibility layer that AI governance, model safety, data security, and audit firms all need underneath them.
Where does Summit sit relative to Palantir?+
Complementary, not competitive. Palantir builds the investigator's notebook; Summit builds the notarized record beside it — the Amicus Curiae to their investigator. The Decision Receipt wraps any analytic output — from Palantir, any LLM, or any agentic system — and makes it replayable, auditable, and defensible.
What stage and what are you raising?+
Pre-seed and pre-revenue. We are raising a $1.5M pre-seed SAFE (YC post-money standard, with a negotiable cap), designed for one lead plus one to two defense-tech co-investors. The cap table is 100% founder-owned: no convertible notes, no advisor equity issued yet, no employee option pool issued yet. A non-dilutive Colorado OEDIT Advanced Industries grant (~$250K) opens July 1, 2026.
What does the round fund?+
Use of proceeds: the multi-tenant and public-demo path; the first two hires; SOC 2 Type I; the Wave-1 non-provisional patent filings; and pilot business development — a 12–18 month runway. Funded milestones include at least one signed pilot, general availability in Q4 2026, patent counsel engaged with filings, and a federal Phase II / transition motion.
Who buys this?+
Three priority categories: federal mission systems (AFWERX, SBIR, direct-to-CDAO via the Tradewinds Solutions Marketplace); regulated enterprise (banks under SR 11-7, healthcare under HIPAA and the 21st Century Cures Act, insurance under state AI-governance mandates); and design partners on the pilot waitlist. The base-case milestone is three to five design partners converting to paid contracts within six months of funding.
How big is the opportunity?+
The AI-governance market today is estimated at $4.4B (a third-party figure, cited as the adjacent market — admissibility is a category Summit is creating, not just a slice of governance). Exit thesis is execution-dependent over a 3–5 year horizon: acquisition by a platform or defense prime, or an IPO on a federal contract base over a 7–10 year horizon. Comparable framing and ranges are in the data room.
Why is this defensible?+
First-mover in decision admissibility — no one else is shipping runtime admissibility infrastructure for autonomous AI actions. Summit owns the vocabulary for the category. The IP surface is anchored by provisional patent No. 64/034,952 (filed April 10, 2026; non-provisional due April 10, 2027), with continuation and international PCT strategy funded by this round. The patent is pending, not granted.
Is the product real, or a deck?+
Real and in production. Live at decrec.summitcognitive.ai with a public key and ledger for independent verification. More than 3,000 signed Decision Receipts issued, 98.6% acceptance (deny by default), 99.9% replay determinism, Ed25519 signatures. The codebase is 52 canonical packages and five integrated services. It is not a model wrapper, not a UI overlay, not a research project.
What's the governance at the SAFE stage?+
No board today; investor information rights only, no board seat at the SAFE stage. MFN on later SAFEs in the same round; pro-rata and standard information rights are offered to the lead and major investors.
Team
Brian C. Long
Founder & CEO · Technical lead / PI
Cyber, media, and intelligence work across more than two decades, on institutional decision systems and cognitive infrastructure. Sole founder by design at this stage — the entire codebase (52 canonical packages, five integrated services, the Decision Receipt architecture, demo hub, and trust portal) was built and is maintained by one person. Author of the forthcoming Admissible Reality. (Distinct from the founder of Attentive / Adaptive Security.) Working to secure a speaking slot at Web Summit, Lisbon (November 9–12, 2026).
Planned first three post-funding hires: a federal capture lead (Col. Steve Lawlor, month-1 retainer), a security engineer (month 2), and a design-partner success hire (month 3); team math at month 18 is six people plus a network of credentialed advisors. Advisor network cited includes Steve (Col.) Lawlor (federal capture), Domingo Guerra, Davey Rathgeber (computational linguistics), Goodman (data science), Cantrell (design thinking), Silverman (legal), and Kevin Long — who runs Slipstream Solutions, a separate company with a D2P2 SBIR relationship at the Bespin Software Factory (a partnered prime path).
Request access
The full data room — deck, financial model, cap table, technical architecture, IP posture, and term summary — is released to qualified investors with an active diligence engagement. Leave your email and we'll be in touch to verify and grant access. You can also write directly to the founder.
For accredited / qualified investors. We respond personally; this is not an automated subscription.
Request received — thank you. We'll verify and follow up personally.
Couldn't reach the request service — please email hello@summitcognitive.ai to request access.
Important disclosures
Confidential · Not for forwarding
This page and any materials referenced are confidential and intended only for the recipient. They are not for distribution or forwarding without permission.
Not an offer. This page is for informational purposes only. It is not an offer to sell or a solicitation of an offer to buy any securities, and it is not legal, tax, accounting, or investment advice. Any such offer would be made only to accredited or otherwise qualified investors, through definitive offering documents, in compliance with applicable securities laws. Any term summary referenced is an illustrative discussion template — non-binding, with no security offered or sold by it.
Forward-looking statements. Statements about the market, product roadmap, pipeline, milestones, exit paths, and financial projections are forward-looking — they are plans and model-based estimates, not commitments, guarantees, or results. Actual outcomes may differ materially. Market-size and funding figures are third-party estimates cited for context. Engineering metrics are audit-time figures (late May to early June 2026) and some have since changed. The company is pre-revenue with no paying customer today.
Compliance & review posture. Compliance frameworks are designed-for-alignment, not yet certified (SOC 2 not certified; FedRAMP scoped; CMMC planned). Independent review by Galois and Trail of Bits is scoped but not completed; nothing here should be read as "independently reviewed by." The provisional patent is pending, not granted.
Summit Cognitive, Inc. is a Delaware C-Corporation headquartered in Alma, Colorado (Park County), founded April 2026. Contact: hello@summitcognitive.ai.