Frequently asked questions

NPDC is short for Net-Positive Data Centers — a distributed-compute architecture for community-governed AI infrastructure. Instead of one megacampus drawing 100+ MW from a single grid connection and a single watershed, NPDC operates as a federated mesh of 100–500 kW MicroDC nodes. Each node is co-deployed with a host-community council that holds a binding vote on whether the node continues to operate. The architecture is anchored in three research threads: Stanford's Intelligence Per Watt programme, Microsoft's BitNet 1-bit LLM framework, and Dr Luci Attala's anthropology of water.

Net-Positive Score = (community benefits / community costs) − 1.0. Costs include water consumed, power consumed, land occupied, and utility-rate impact on households. Benefits include jobs (FTE local employment), tax revenue, waste heat delivered as thermal value, water recycled and returned to municipal supply, and subsidised compute access for schools, libraries, and non-profits. The deployment target is a score above 0.2 (a 20% net-positive return); a sustained negative score triggers the reversibility clause and the host-community council can decommission the node at NPDC's expense.

The host-community council holds a contractually binding seat in NPDC's tri-partite governance model. Three seats — investors, host community, epistemic stakeholders — must reach consensus on deployment and continuation decisions; no single seat overrides the others. The veto is operationalised in three ways: a co-designed net-positive metric set agreed at deployment, a public real-time dashboard exposing water, power, heat, and revenue-share data at the node, and the reversibility clause that gives the council legal authority to trigger decommissioning at NPDC's expense if the metrics are not met after twelve months. Trust is not asserted; it is engineered into the deployment contract.

NPDC is seeking one forward-thinking host community to deploy the first 100 kW MicroDC reference node in a twelve-month pilot. Requirements: a host-community council willing to co-define net-positive metrics, local government open to consortium co-design, existing grid capacity for a 100 kW draw, and a real use case for waste heat (district heating, greenhouse agriculture, or industrial process). What the community gets in return: co-defined metrics, binding veto at the twelve-month review, a real-time public dashboard, revenue-share on compute sales, waste heat to the council-nominated use case, and decommissioning at NPDC expense if net-positive targets are not met.

A host-community council — or a councillor authorised to enquire on the council's behalf — writes to the consortium via the contact form with "host-community application" selected, or directly to consortium@netpositivedatacenters.org. The first round of conversation is exploratory: NPDC explains the reference architecture, the council explains the local context, and both sides decide whether the site meets the pilot requirements. The exploratory conversation carries no commitment; the binding seat is signed only at deployment.

NPDC is convened by a consortium of operators, host-community advisors, and epistemic-stakeholder institutions. The founding seats are listed on the team page. Some seats are confirmed but the named occupant is not yet public; those entries read To be confirmed. NPDC does not claim institutional partnership unless the partnership is formally confirmed, and lists only confirmed partnerships publicly. Naming will catch up to the architecture as the v1 whitepaper publishes.

The NPDC v1 technical whitepaper is in preparation. Target publication: Q3 2026. The v0.1 outline is complete; drafting is underway against the four-pillar architecture and the convergence research. The page that documents the architecture must be the v1 document, not the working outline, so NPDC does not surface the draft. Leave an email on the whitepaper page and NPDC will write to you the day v1 publishes.

Epistemic governance is the seven-dimensional knowledge-justice layer that allocates priority compute on NPDC infrastructure. The seven dimensions are: Historical, Indigenous, Cross-cultural, Scientific, Artistic, Marginalised, and Future-generational. Compute allocation is treated as a knowledge-justice question, not an engineering optimisation — priority queues route to Global South researchers, indigenous institutions, and long-term climate modelling instead of routing strictly to highest-bidder commercial inference. Stanford's Intelligence Per Watt finding (77% of inference runs on local hardware) is what makes this allocation tractable in a way it never was on hyperscale.

Microsoft BitNet has proven that 1-bit LLMs run on commodity CPUs at 96% less energy than full-precision peers, with 0.4 GB memory, outperforming LLaMA 3.2 1B on standard benchmarks. The GPU dependency that made hyperscale data centres water-hungry — GPU cooling infrastructure, NVIDIA supply chain, massive power draw — is no longer a precondition for competitive inference. A 100 kW NPDC node running BitNet-class models on CPUs serves the 77% practical-AI tier without needing a reservoir. The water argument the consortium hears most often — that AI inherently requires extractive water use — is settled by Microsoft's published research.

NPDC is one project in the wider Cogniosynthesis research programme. The seven-dimensional epistemic-governance layer in NPDC is a direct application of the Cogniosynthesis framework's knowledge-justice principles. Other Cogniosynthesis projects — CognioNews, CognioFem, Bio-Debt, Ministry of Digital Obscurity — sit in adjacent registers (news synthesis, structural-women-and-gender analysis, biological-debt ledgering, satirical archival commentary). NPDC is the infrastructure layer. See cogniosynthesisportal.uk for the ecosystem.