P-01
Decision-first models
Forecasting, allocation, and routing models trained on sustainment-specific feature sets — readiness rates, mean time between failure, lift availability — not generic time-series.
The North Star for defense logistics.
Lodestar is an AI-native sustainment platform built for the U.S. Department of Defense — predictive supply, contested-logistics planning, and decision support that compounds across every echelon.
Target buyerDoD / DLA / USTRANSCOM
StageConcept brief · pre-formation
Entry vehicleDIU OTA · Tradewinds · SBIR III
01 /The problem
DoD logistics still runs on systems built for garrison, not contested ops.
Sustainment in the modern operating environment is a data problem the legacy stack cannot solve. ERPs were tuned for steady-state demand; spreadsheets bridge the gaps; analytic platforms surface dashboards rather than decisions. When the warfighter needs a courses-of-action recommendation in minutes, the answer arrives in days — if at all.
- Forecasting horizons are too short for contested-logistics windows.
- Cross-component demand signals (Army / Navy / Air Force / SOCOM / Marines) remain siloed at the data layer.
- Predictive maintenance pilots stall at the integration step — too many systems of record, too few APIs.
- Coalition partner data exchange is manual when it exists at all.
02 /Market anchors
The market is being created right now.
The 2024–2026 cycle marks the inflection from DoD AI pilots to production-scale procurement of sustainment software. Below is a fact-anchored snapshot of the public anchors we benchmark against.
Ceiling values are not-to-exceed and not guaranteed spend. Figures sourced from public DoD press releases, GAO reports, GSA records, SEC filings and reputable industry press as of May 2026.
03 /Our approach
An AI-native stack purpose-built for sustainment.
Lodestar's product thesis is opinionated. We are not a generic data platform with a defense skin — we are a sustainment-decision system from the data model up.
P-02
Action-grade outputs
Every output is a recommended action with quantified confidence and an auditable trace back to the underlying data — built for command approval, not just dashboards.
P-03
Coalition-by-design
Tenant separation, ITAR / EAR partitioning, and data-tagging schema engineered for allied data exchange from day one.
P-04
Open integration
Adapter layer for GCSS-Army, ERP/SAP, MILSTRIP, IUID, and major commercial logistics platforms — not a rip-and-replace.
04 /Why now
Three forces converge in 2025–2026.
- Procurement vehicles matured. DIU prototype-to-production OTAs (JSDT, Jan 2026) and the Tradewinds Solutions Marketplace ($3.26B awards facilitated) provide a non-FAR path that fits AI-native vendors.
- Cloud + foundation models are accredited. AWS GovCloud reached FedRAMP High + IL4/IL5 with Claude and Llama foundation models (May 2025) — the runtime substrate is now compliant.
- Buyer signal is unambiguous. Palantir's $10B Army EA, Govini's $919M SCRIPTS BPA, and the DIU JSDT awards collectively demonstrate that DoD is buying AI-native sustainment software at production scale.
05 /Competitive landscape
We compete on decision quality, not data plumbing.
The market splits between scaled integrators with platform reach and AI-native startups with sharper sustainment models. Our position: AI-native models, integrator-grade integration discipline.
| Company | Bucket | How they differ from Lodestar |
|---|---|---|
| Palantir Foundry | Incumbent platform | Strong data ontology and integration; logistics is one use case among many. Lodestar is sustainment-first. |
| Govini Ark | AI-native scale-up | Supply-chain risk illumination and DIB analytics. Lodestar focuses on action-grade unit-level sustainment decisions. |
| Air Space Intelligence | AI-native startup | PRESCIENCE — multi-domain logistics COA. JSDT awardee. Direct overlap; we differentiate on coalition data model and predictive maintenance depth. |
| Watchtower Labs | AI-native seed | Sustainment planning and simulation. JSDT awardee. Differentiate on production-grade integration and IL5 path. |
| Rune Technologies | AI-native early stage | TyrOS predictive military logistics in contested ops. Differentiate on coalition tenancy and DLA-scale integration. |
| Booz Allen / SAIC | Integrators | Glue layer for DLA and USTRANSCOM. Lodestar is the product layer they would integrate to deliver. |
06 /Compliance posture
Built to clear the buyer-side accreditation gate.
The Lodestar program plan budgets against the explicit accreditation path required by DoD sustainment buyers.
FedRAMP Moderate
Year 1 target
FedRAMP High → DoD IL4 / IL5
Year 2 target via AWS GovCloud reciprocity
CMMC 2.0 Level 2
Self-assessment now; C3PAO certification before Nov 10, 2026 (Phase 2)
ITAR / EAR registration
Pre-revenue
SOC 2 Type II
Year 1 — commercial baseline
07 /DEMO · INTERACTIVE
Try Lodestar — illustrative scenarios
Pick a scenario. The AI generates an analyst-style response on synthetic, illustrative inputs. No live or classified data is ever used.
LIVE · LDS-CONCEPT-2026.05
Select a scenario above to begin.
Illustrative output only. Not classified, not connected to live systems, not a substitute for sustainment-officer judgment.
08 /Where we are
Concept brief — pre-formation, pre-team, pre-revenue.
Lodestar is presented here as a thesis only. There are no signed contracts, no accreditations in hand, no pilot customers, and no formed team. This page exists to invite a focused conversation with operators, capital, and prospective government partners.
- We are seeking conversations with: warfighters and sustainment operators (Army, Navy, Air Force, Marine Corps, SOCOM), former DLA / USTRANSCOM leadership, defense-tech investors, and prime contractors interested in an AI-native sustainment partner.
- We are not seeking: marketing partnerships, generic SaaS resellers, or projects unrelated to DoD sustainment.
Get in touch
Operators, capital partners, and prospective government collaborators welcome. Brief notes please — no pitches in the first message.
Lodestar is a concept-stage brief. No team, contracts, accreditations, customers, or pilots have been claimed or implied. All figures are public, dated, and sourced. This page is not a solicitation, an offer of securities, or marketing of a regulated product.