Under the term sheet, OpenWorld will serve as advisor, structuring agent and services provider for the program, and its OpenWorld Enterprise platform will finance, coordinate and monitor the network.Key Highlights
Up to $1 billion in tokenized financing: OpenWorld structures and administers the tokenized financing program on OpenWorld Enterprise. Tesla Powerwall batteries, solar panels and AI computing equipment serve as collateral within the structure.
Exclusive Tesla Powerwall inventory: Meridian holds exclusive rights to a reserved inventory of Tesla Powerwall systems that qualified for federal clean-energy tax credits before those credits were discontinued. [The inventory covers more than 160,000 Powerwalls, enough to support approximately 80,000 homes.]
Initial deployment of 10,000 homes: [GPU-equipped computers at participating homes, with larger installations at local schools, hospitals and commercial campuses.]
AI inference without new data centers: AI providers pay to run trained models on the network with no new centralized data-center sites, no additional grid interconnections and no water-based cooling. The program is intended to supply capacity to one or more leading AI inference providers.
Hosts get paid: Homeowners receive monthly hosting payments and lower-cost electricity, with household energy needs taking priority over computing workloads. Commercial hosts receive a share of site revenue and backup power.
Timeline: Test installations scheduled to begin in the fourth quarter of 2026; first commercial site targeted for the first quarter of 2027.
Builds on the BeyondAI transaction: Follows last week's proposed transaction with BeyondAI, which would transform BeyondAI’s industrial AI intellectual property rights into financeable collateral on OpenWorld’s tokenization platform.
Solving the Power Problem in AI Infrastructure
Access to power is a growing constraint on AI infrastructure. New data centers face lengthy grid interconnection timelines, local opposition and significant water demands. The planned network of home-hosted, distributed AI inference computing would place GPU-equipped computers at participating homes and larger installations at local schools, hospitals and commercial campuses, each powered by on-site solar generation and Tesla Powerwall storage. AI providers would pay to use this computing capacity for inference, the process of running trained AI models, without requiring new centralized data-center sites, additional grid connections or water-based cooling.
Participating homeowners would receive monthly hosting payments and lower-cost electricity, with household energy needs taking priority over computing workloads. Commercial hosts would receive a share of site revenue and backup power.
Click Here