# Products and revenue Carbon's commercial plan starts with evidence for a model a customer already owns, then expands into discovery, qualification, and ongoing model support. Customers can retain their existing solvers and engineering tools. The offerings below describe product design and the intended delivery sequence. They are not a list of production services or evidence of sales. See [Project status](../docs/publications/PROJECT_STATUS.md) for the implementation record. ## Start with the customer's model **Carbon Evidence Audit** is the preferred first engagement. The customer supplies a candidate model, its intended engineering use, and access to an adequate reference. Carbon scopes and runs an independent evaluation, then reports supported behavior, failure regions, limitations, and proposed follow-up work. A customer without a well-defined evaluation problem starts with **Challenge Feasibility / Design**. That engagement establishes the inputs and outputs, operating conditions, reference access, rights, and evidence needed for a useful program. Both engagements have finite scope. A negative finding can still satisfy the contract when Carbon delivers the agreed evidence. ## Product families The existing P1–P11 identifiers remain the commercial planning taxonomy. | ID | Product | Customer need and intended deliverable | Revenue design | |---|---|---|---| | P1 | Evidence Audit | Assess an existing model for a defined use; deliver an evidence and failure report | Scoped audit, integration, and evaluation fees | | P2 | Challenge Feasibility / Design | Establish whether a problem supports a defensible evaluation or discovery program | Authoring and feasibility fee | | P3 | Sponsored Discovery | Search for a better method against agreed requirements; deliver research outcomes and evidence | Program, integration, evaluation, and reward-administration fees | | P4 | Model Development | Turn a promising method into a reproducible candidate artifact | Development and integration fees | | P5 | Qualified Model Program | Assess one exact model or system for a bounded engineering use | Qualification, evidence, and deployment fees | | P6 | Model Lifecycle | Revisit evidence after changes to a model, reference, runtime, or operating range | Recurring support, lifecycle, and requalification fees | | P7 | Enterprise Evidence Platform | Manage repeated evaluations across teams and model portfolios | Subscription, usage, private deployment, and support | | P8 | Evidence / Qualification API and OEM | Embed Carbon's evidence workflows in another engineering platform | License, usage, OEM royalty, and support | | P9 | Sponsored Frontier Market | Run recurring research competitions for defined scientific improvements | Platform, authoring, evaluation, and administration fees | | P10 | Model / Method Licensing | Obtain permitted rights to reproduced or qualified work | License or royalty, subject to Carbon's rights | | P11 | Physics Intelligence / Experiment Allocation | Use accumulated evidence to choose better future experiments | Future project, subscription, or API fees after prospective validation | P1 and P2 are the initial commercial focus. P3 is the first planned expansion that depends on competitive research supply. The later offerings require their own implementation, customer evidence, and qualification. ## Sponsored Discovery A sponsor brings a physical-modeling problem and a reference path. Carbon defines the program, establishes the evaluation contract, and organizes research. Evaluators determine whether candidates meet the requirements and improve on the comparison baseline. The commercial deliverable is the agreed research and evidence program. Carbon does not guarantee a frontier advance. A candidate from the competition still needs separate evidence for a customer's deployment. The contract separates Carbon's fees from any sponsor-funded participant reward pool. Participant rewards are not automatically company revenue. ## Expansion within an account | Customer situation | Next potential engagement | |---|---| | Audit identifies important failure regions | Discovery or model development | | Candidate supports further investigation | Qualification for a defined use | | Customer deploys or changes a model | Lifecycle support and requalification | | Teams repeat evaluation across a portfolio | Enterprise software and APIs | | An engineering platform needs the workflow | OEM or embedded API partnership | Customers need not buy each stage. Carbon's aim is to reuse the reference integration and evidence workflow when further work makes sense. ## Pricing and delivery economics Carbon plans to price the engineering outcome and evidence burden. Scope includes expert work, reference calculations, evaluation compute, integration, and any qualified private deployment. Contracts specify finite capacity and change-order terms. The pricing principles remain: - Base fees cover legitimate work even when the evidence is negative. Any success fee is additional. - Customer payments cannot change the scientific standard for the same claim. - Rights, exclusivity, support, and private deployment need explicit scope and authority. - Enterprise procurement can use fiat without a mandatory Alpha purchase. The accounting system retains separate revenue identifiers: `AUTHORING_FEE`, `TRUTH_INTEGRATION_FEE`, `PROGRAM_PLATFORM_FEE`, `EVALUATION_USAGE_FEE`, `SPONSORED_REWARD_ADMIN_FEE`, `MODEL_DEVELOPMENT_FEE`, `QUALIFICATION_FEE`, `DEPLOYMENT_INTEGRATION_FEE`, `SOFTWARE_SUBSCRIPTION`, `ON_PREM_LICENSE`, `MODEL_LICENSE`, `METHOD_LICENSE`, `OEM_ROYALTY`, `SUPPORT_SLA_FEE`, `LIFECYCLE_SUBSCRIPTION`, `REQUALIFICATION_FEE`, and `PHYSICS_INTELLIGENCE_FEE`. Per engagement, Carbon tracks contracted and collected value, recurring value, reward pass-through, direct labor, reference and compute costs, integration, support, gross contribution, custom versus reusable work, network activity, reuse rights, and expansion. These commercial fields do not enter scientific scoring. ## Product maturity Commercial progression remains `CONCEPT` → `DESIGN_SPECIFIED` → `INTERNAL_PILOT` → `PAID_PILOT` → `REPEATABLE_SERVICE` → `PRODUCTIZED` → `ENTERPRISE_QUALIFIED` → `SCALE_READY`. Each promotion requires evidence for that product; a software test or a research result cannot substitute for a customer purchase. Each offering needs a buyer, deliverable, cost model, rights agreement, bounded evidence claim, and a reason to repeat or expand it. [Financial model](Financial_Engine.md) and [Commercial operating model](Commercial_Operating_Model.md) retain the delivery and accounting controls. ## Research services Research Concierge / Copilot remains a candidate workflow offering under `OWNER-EVIDENCE-RESEARCH-01`. `G-PR0` freezes the pilot; `G-PR1..5` govern entitlement and rights, campaign memory, non-official execution, billing and spend, and the commercial decision. Evidence Audit remains the first entry product. Free and paid miners receive official-derived scientific content on equal release terms. Interest or a stated budget does not establish revenue or authorize spending.