Confidential · internal strategy and diligence

The machine-data fabric for physical AI.

PeltBeam connects cameras, sensors, machines and edge compute across private outdoor sites—without trenching and without pretending to be a data-center AI fabric.

NVIDIA makes the models. Palantir turns them into decisions. Starlink delivers internet access. PeltBeam is the intelligent infrastructure connecting the physical world to AI — we ship answers, not just packets.

This version leads with the executive decisions, separates what is credible now from what remains gated, and moves the detailed explorer behind progressive disclosure.

Strong narrow thesisProof-stage evidencePartner-led scale path
Governance: the workbook’s v5 model is a candidate for sales planning. TAM/SAM v4 remains the formal registry until adopted. “$0.53B SAM” is presented here as an annual invoiceable serviceable revenue pool to avoid definition confusion.

01 · Executive dashboard

What management needs to decide

The architecture is compelling. The company-building question is whether PeltBeam can close the deployed proof stack, convert three repeatable lighthouse patterns, and add enough software to scale through partners.

Numbers below preserve evidence marks and distinguish model output from management interpretation.
Decision 01

Close proof before broadening claims

M1a, M1b, MAC-1 and DESENSE-1 are the immediate value-creation sequence. External language should remain bounded by the weakest open gate.

Decision 02

Use a two-engine commercialization plan

Run a reference factory for fast OTA evidence while pursuing industrial revenue through integrators and OEMs rather than founder-led direct selling alone.

Decision 03

Make software an attach layer, not a premature platform

Ship transport-plus-edge packages first. Build runtime, APIs, observability and lifecycle tools only where customers and partners pay for them.

02 · System and proof

Show the architecture—and the evidence boundary

The strongest presentation is one picture for the customer workflow, one ledger for what is credible now, and one gate roadmap for what turns into sellable proof next.

Credible now

Not yet safe to claim

Validation-gate registry

GateCapabilityStatusCommercial unlockTiming

03 · Commercial portfolio

Separate proof wedges, revenue beachheads and growth options

This eliminates the common mistake of ranking a horizontal workload such as sensor-to-edge-GPU transport in the same list as vertical markets without showing how they relate.

Opportunity landscape

Commercial attractiveness vs evidence readiness

Upper-right opportunities deserve the first lighthouse and channel capacity. High-attractiveness / low-evidence opportunities remain gated regardless of their strategic score.

Top right-to-win opportunities

100-point model; use as prioritization, not forecast

04 · Application explorer

55 applications, shown by commercial readiness

The original strategic status is preserved, but the presentation now elevates gating: a high-scoring market cannot appear “GO” when its critical technical dependency remains open.

05 · AI data communications

Position PeltBeam in the missing middle

PeltBeam is the transport and local-execution layer between physical sensors and edge compute. It is not a GPU interconnect and should never be benchmarked against NVLink, InfiniBand or 800GbE fabrics.

06 · Market and revenue model

Use four numbers—each with a different job

The visual structure below prevents TAM, annual serviceable revenue and PeltBeam revenue from being presented as interchangeable quantities.

Board interpretation: the 2030 model output is an execution stretch, not a conservative base. The 2035 output requires roughly 37.5% of the modeled annual serviceable pool and should be treated as category-leadership upside.

Broad category context

Top-down counted TAM vs prior v4 pool

Annual invoiceable serviceable pool

Low / base / high scenario

PeltBeam revenue model

Model output with management interpretation

2030 bottom-up slices

Largest annual serviceable clusters

07 · Competitive positioning

Win the intersection, not every category

The strongest competitive story is a job-to-be-done comparison. Fiber, private 5G, Wi-Fi, tactical MANET, microwave and LEO each win somewhere; PeltBeam fills the missing middle for private outdoor machine data.

Technology classDeterministic serviceFull-duplex multi-hopPer-node edge computeRapid no-trench economicsWhere it wins / PeltBeam posture

Counter-moves and strategic response

08 · Company-building roadmap

Proof first. Software attach second. Scale third.

Each phase has an explicit commercial unlock. The roadmap is designed to avoid building a generalized edge platform before the transport product earns repeatable customer pull.

0–90 days

Lock truth and test plans

Freeze claim governance, correct the gate registry, finalize M1a/M1b/MAC-1/DESENSE-1 protocols and choose three instrumented lighthouse patterns.

3–6 months

Generate proof that sells

Run perimeter, depot-offload and reference-factory deployments with buyer-visible KPIs, logs and deployment economics.

6–12 months

Package the attach layer

Release observability, workload manifests, signed model/update flow, northbound APIs v1 and repeatable vertical SKU documentation.

12–24 months

Scale through partners

Convert references into multi-site expansions, add two integrator/OEM channels and enter ports after the relevant deployed proof closes.

09 · Method and sources

Keep diligence one click away

The explorer is designed for progressive disclosure: executive conclusions first; assumptions, sources and application detail available without crowding the main story.

Source ledger
Full feature ledger
Explicit no-go markets