The confirmed links, demo conversations, and sample queries are scoped to the
~134 Training & Simulation sector PEs — the programs managed by
PEO STRI, NAWCTSD, AFLCMC Simulators, PM TRASYS, and DHA medical simulation.
The underlying database is not T&S-only. All 1,740 RDT&E PEs,
41M contracts, and 154k narrative sections are queryable for any program across
any appropriation. Asking about a PE outside the T&S sector will return
funding history, R-2 narratives, and FPDS contract search — it just will not
have confirmed-link intelligence, because the link-verification pipeline has
only been run for T&S so far.
Additional sectors follow the same pipeline. C2, ISR, Aviation
systems, Shipbuilding, Cyber, and others can be activated by running the ensemble
scorer and AI-assisted verification pass against their PE cohort. The architecture
is sector-agnostic; T&S is the first production cut.
PE↔contract links — confidence tiers
Every dollar in RDT&E flows through a Program Element (PE). This tool links PEs
to FPDS contract awards using a four-method ensemble scorer followed by AI-assisted
verification, with every high-confidence link cross-checked against an expert-reviewed
sample. The result is a graded confidence score for each PE↔contract pair.
Tier
Gate
Meaning
Measured precision
Use
Confirmed
≥ 0.8
Contract description matches R-2 narrative and cost items; verified with high confidence
91.8% real (95% CI: 85–96%)
Customer-facing claims, proposal intelligence
Likely
0.6–0.8
Plausible match; not fully verified
~51% real
Internal analysis only; label explicitly
Ensemble-only
< 0.6
Unverified; weak signal
~14% real
Do not surface
Gold-set validation: Precision figures come from a stratified expert review
of 545 links across 62 sector PEs (July 2026). "Real" = gold confirmed + likely.
The tool defaults to the ≥ 0.8 gate for all answers. If an answer loosens to 0.6–0.8,
it will say so explicitly.
The confirmed set covers 5,409 links across 485 PEs, spanning
$812.4B in total obligated dollars across linked contracts.
What the dollar totals don't mean
The $812B total obligated value of confirmed-linked contracts is not the
training & simulation market size. Many links connect RDT&E program
elements to platform procurement contracts (ships, aircraft, vehicles) whose descriptions
reference the training systems being developed. A $1.3B AMRAAM production lot linked
to the PE that funds its simulator is captured in that total.
The RDT&E training & simulation sector runs approximately $9–10.5B/yr
in recent budget cycles. That is the relevant market sizing number for BD work.
Known data gaps
Air Force RDT&E — PB2025 and PB2027 missing.
The AF never posted FY2025 RDT&E J-books publicly; PB2027 books may not have been
posted at pipeline-build time. All other appropriations (Army, Navy, USMC, DARPA, OSD,
Defense Health) are complete for both cycles.
Do not interpret absence of an AF 3600/3620F PE in PB2025 or PB2027 as
evidence of termination or restructure. The tool will say "data not available
for this cycle" and recommend verification via SAF/FM or the Defense Comptroller portal.
Procurement (P-1) funding is raw-only — not yet harmonized into
the PE funding model. O&M is deferred to a future release.
OTA contracts (Other Transaction Authority) do not appear in FPDS
and are not in this database. Some competitive awards are invisible here.
IDIQ task orders below the simplified acquisition threshold
(~$250k) are not universally reported and may undercount activity on large MAC vehicles.
How to read an answer
Every dollar figure states cycle + column type + FY
(e.g., "PB2027 BY1 request for FY2027: $261.2M"). BY1 = request;
CY = executing year estimate; PY = prior-year actuals.
Every link claim states the confidence band and cites the evidence
(e.g., "confirmed at ≥0.8 — contract description matches J-book line").
No citation = analysis, not database fact.
Cross-cycle comparisons name both cycles and column types.
Thin cycles (PB2025, PB2027 at ~730 PEs vs. ~1,050 in adjacent cycles) are flagged.
Anomalies are explained mechanically first (CR, OCO migration,
BA-9 restructure) before inferring a program decision.