Data spaces
Offer data or permitted queries. Define who may use them, for which purpose and under which conditions. Keep source custody explicit.
European AI & Data Spaces
In development · testing in progress
Connect data,
AI and compute.
Control the rules.
Share the value.
Build and operate shared AI services across organisations and infrastructure—with controlled data access, verifiable execution and transparent allocation of costs and revenue.
ONE SHARED SERVICEFOUR INDEPENDENT ROLES
Customer
Requests & accepts
Own bookService operator
Connects & delivers
Own bookAI executor
Runs the model
Own bookData owner
Controls access
Own bookdaSPE.AI
Connect → execute → verify → account
Agreed rulesMeasured contributionsShared value
Your infrastructure. Your records.
A shared account everyone can check.
DATA OWNERSAI BUILDERSTELECOMS & CLOUD PROVIDERSENTERPRISE CUSTOMERS
The partner opportunity
Your contribution.
A bigger business.
Bring your data, models, applications or compute. Together, we turn them into services customers can buy—with clear control, measurable contribution and an agreed share of the revenue.
A data owner can supply a permitted query. A model provider can turn it into an answer. A telecom can bring the customer, network and infrastructure. daSPE.AI connects the work and the account between them.
01 / The platform
From separate capabilities
to a shared service.
Data custody & verification ↓
AI services
Combine models, applications and data into a service. Connect the request, model version, execution and accepted result.
Cloud–edge execution
Connect work across company infrastructure and selected providers. Bring an existing application through its API and a company edge.
Shared commercial operations
Measure delivery, allocate costs and revenue under agreed rules, and net reciprocal obligations. Return a verifiable account to each company.
Governance, evidence & Chamber
Each company keeps its own book. Shared rules govern participation; the Chamber handles reviews and disputes with human decisions.
02 / For telecoms and infrastructure providers
Turn capacity and reach
into a shared AI business.
Bring your enterprise customers, connectivity, cloud or edge capacity. Combine them with partners’ models and data, with an agreed account of who delivered what.
Illustrative partner offer
A private AI assistant for enterprise customers.
A telecom brings the customer and managed service. An application partner supplies the assistant, an AI provider runs the model, and the enterprise authorises access to its knowledge.
Telecom + enterprise
The telecom packages the service; the enterprise defines the task and accepts delivery.
Enterprise knowledge
Permitted retrieval from approved sources, under the enterprise’s access rules.
Application + model provider
A chosen application uses a model on approved GPU capacity.
daSPE.AI
Records accepted work and allocates the agreed shares across participating providers.
What we agree in the pilotOne customer task · approved data access · execution location · service price · provider shares
These are proposed partner services. We are developing and testing the shared workflow, from permitted data access to execution and contribution accounting. Explore the Co-Tru Agent Market workflow ↗
03 / One understandable business process
Agree the service.
Follow the work. Share the value.
- 01
Connect
Each partner connects one application, data service or resource through its edge.
- 02
Agree
Define access, purpose, service levels, price and contribution shares together.
- 03
Deliver
Each party records its part. The customer accepts the result; measured usage feeds the account.
- 04
Reconcile
Check the records, allocate value and return each partner’s account. Refer exceptions to the Chamber.
Business continues after the first receipt; final cryptographic proofs arrive asynchronously. During onboarding, we agree who operates each server edge and where the source records are retained.
Customer payment and internal accounting
The intended commercial model is simple: the customer pays the daSPE.AI operator in the agreed currency. Internal contributions, costs and obligations are calculated in µCLU. External invoicing and payment follow the agreed commercial terms.
Own the shared service.
Choose the infrastructure.
AWS and Databricks already offer substantial cloud, data, AI and governance capabilities. Our focus is the joint business between independent companies: their contributions, evidence, obligations and agreed shares.
Bring the systems your organisation already uses. We scope their connection into the shared workflow during onboarding.
04 / Built around European needs
Control that can
be demonstrated.
Make data access, AI execution and responsibility visible. Map each pilot’s technical controls to the obligations that apply to its service and participants.
Trace the AI work.
Connect model versions, actions and evidence. Identify accountable people and retain the decisions needed for human oversight and review.
Official framework ↗Keep meaningful choice.
Document access conditions, interfaces and export formats. Include provider switching and data custody in the pilot’s integration plan.
Official framework ↗Build European capability.
Make operators, execution locations and technology dependencies explicit. Work with European infrastructure and service providers.
Cloud and AI Development Act ↗Technical controls support applicable obligations; legal compliance depends on the deployment and responsible parties. CADA is a legislative proposal, not an awarded certification. Legislative status ↗
05 / Engine architecture & verification
Each contribution recorded.
Each account checked.
Company books connect the business activity. Co-Tru Ledger Technology anchors the evidence. The accounting engine reconciles accepted work and calculates the agreed shares.
Keep the evidence with its owner.
The customer, service operator, AI executor and data owner each record their part in their own book. A shared task reference connects the records for reconciliation.
Source records remain in each participant’s controlled storage. The shared ledger receives 32-byte commitments; retained originals and proof paths allow the records to be checked later.
Calculate the shares. Compare the results.
Python and GnuCOBOL calculate allocation independently using the agreed shares and integer µCLU amounts. Their outputs must match before the allocation is applied.
Reconciliation connects accepted work, measured usage and the applicable terms. Differences remain visible for review under the Chamber’s rules.
One net position per participant.
Companies can buy and supply services in the same period. The engine offsets reciprocal obligations and calculates a net position for each participant.
Gross obligations, bilateral offsets and multilateral net positions remain visible in the account. The reduction depends on the actual pattern of trade.
Keep work moving. Verify the final record.
The business workflow continues after the first receipt. Subsequent confirmations and the final cryptographic proof arrive asynchronously through the existing evidence path.
Closed accounts stay unchanged. Later eligible records enter a linked supplementary account; reviews and disputes follow the Chamber’s rules and human decisions.
Company recordsReconciliationAllocation & nettingVerifiable returns
06 / Founding partner invitation
Bring one capability.
Build one service together.
We are preparing the first partner pilots around one real service at a time. Data, AI, application, infrastructure or a customer need—tell us what you bring.
A practical first engagement.
Open the partner connection guide ↗
- Choose one service.Name the customer problem, your contribution and a technical contact.
- Agree the boundaries.Define data access, execution location, price, shares and service rules.
- Connect and demonstrate.Connect the API and edge, run the workflow and inspect each party’s account.
Speak directly with Dražen Kapusta
drazen.kapusta@cotrugli.org ↗