Our thesis
Agents are beginning to become economic actors.
See evidence: are agents transactingPayment rails move money, but that alone does not establish whether the purchased work satisfied the agreement.
See evidence: what agents buyAs agents buy more data, research, code, and completed outcomes, fulfillment becomes something that must be judged.
See evidence: what agents buyA fulfillment layer needs explicit acceptance criteria, evidence, verification, and an earned-value result.
See evidence: the market gapThat result should be portable and independent of whichever wallet, marketplace, or payment rail moves the money.
See evidence: what it unlocksIf other systems consume verified outcomes, they can eventually inform payment, routing, reputation, pricing, and other economic systems.
See evidence: what it unlocksSpoolis is the bet that fulfillment verification becomes this missing layer.
See evidence: how large it could becomeEvidence
As agents move from buying simple API calls to buying data, research, code, and completed tasks, a payment status can no longer confirm the work was delivered to specification. Fulfillment verification becomes a distinct infrastructure layer. That is the Spoolis thesis.
Strengthening. The latest available stored x402 transaction and volume signals increased by at least 5% on average. Derived Aug 16, 2026.
Excludes dust bursts: facilitator-day rows averaging under $0.001 per transaction.
Source: Dune Analytics · as of Aug 13, 2026
HTTP 402 · coverage shown in stored provenance
Source: Dune Analytics · as of Aug 13, 2026
Built for high-frequency microtransactions.
Dashboard totals · network coverage not established
Source: MPPscan by Merit Systems · as of Aug 16, 2026
Built for higher-value utility purchases.
Sources use different coverage and should not be compared as complete protocol totals.
Machine commerce is measurable, but this page will not call its scale meaningful until the real series supports that conclusion.
If this activity does not become economically meaningful, nothing downstream matters.
Tracking. Tracking until measured purchasing or spend composition is available. Derived Aug 16, 2026.
e.g. Exa · Brave · OpenRouter
Payment status confirms the call succeeded.
e.g. Apollo · Firecrawl · 2Captcha
Was it complete, valid, and accurate?
emerging · few named services yet
Did it meet the criteria, and what was earned?
Share of listed services in the catalog snapshot, not transactions or spend.
Source: x402 Bazaar discovery · as of Aug 16, 2026
Spend composition remains unmeasured. The transition is important enough to track and too important to invent.
Verification becomes valuable only when agents increasingly buy things whose fulfillment must be judged.
Tracking. Tracking because attestation and service-listing growth do not directly measure the fulfillment-verification gap. Derived Aug 16, 2026.
Machine-payment infrastructure is rapidly becoming more capable and interoperable while fulfillment verification remains fragmented.
The gap is both the opportunity and the risk. If another standard absorbs this function, our thesis weakens.
Tracking. Tracking because the visible calculator is illustrative and does not explain a directional market signal. Derived Aug 16, 2026.
Illustrative workloads · thesis-economics-examples-v1 · as of 2026-08-14
Spoolis price card verification_intro_v1, effective 2026-08-11
Verification: 100 deterministic checks
Cost: $0.003000
Verification can be economically negligible for some machine-native work today.
If verification costs scale with work value, this never becomes infrastructure. If they stay tiny, it can.
Available condition states are stored derived reads. Unavailable or unexplained states remain tracking.
The evidence for the thesis strengthened this month.
Derived Aug 16, 2026Outside forecast figures remain unavailable until each primary publication, definition, horizon, publication date, and verification date is recorded in the data. The scenario below is illustrative.
Forecasts are hidden pending primary-source verification. A forecast will render only after its primary publication, definition, horizon, publication date, and verification date are recorded in the data.
All inputs and bounds are illustrative · thesis-scenario-v1 · as of 2026-08-14
Illustrative scenario, not a forecast. No input is presented as an observed current value.
The opportunity is highly sensitive to how much agent commerce becomes outcome-based, not merely how large agent commerce becomes.
Our most important market variable is not total agent GMV. It is the amount of economic activity whose fulfillment is non-trivial to establish.
These citations render from the same configuration used by ingestion, so the page and jobs stay aligned.
GET https://api.dune.com/api/v1/query/{query_id}/results
daily refreshPOST https://base.easscan.org/graphql
weekly refreshGET https://www.mppscan.com/
daily refreshGET https://openrouter.ai/api/v1/models
daily refreshGET https://api.owlracle.info/v4/base/history
daily refreshGET https://api.cdp.coinbase.com/platform/v2/x402/discovery/resources
daily refreshA verified outcome can become useful far beyond the transaction that created it. If software can reliably know what was promised, what actually happened, and what was earned, other systems can act on that result.
Potential consumers
Portable primitive · illustrative receipt · as of Aug. 14, 2026
Could enable
The wedge is fulfillment verification. The larger opportunity is making economic outcomes programmable.
If outcomes become structured, signed, portable, and widely consumed, they could become inputs to reputation, routing, pricing, marketplaces, and other economic systems.L1 Availability. Did it respond. A commodity signal anyone can observe.
L2 Criterion satisfaction. Did the delivered work satisfy the agreed criteria under the recorded verification method. This is what the signed receipt captures today. Read how the fulfillment layer works.
L3 Downstream economic success. Did the work create the value the buyer hoped for. Today this is implicit human judgment. Whether it becomes a recorded, portable signal is an open question we track, never something we can extract.