Frequently asked questions

The planned founder membership includes 2 years of access after the platform launches. This covers private AI chat, coding assistance, document analysis, research tools, productivity tools, a member dashboard, roadmap voting rights, transparent budget reports, and a discounted renewal option after the founder period ends. Usage is subject to fair-use limits — this is not unlimited access.

No. Founder membership is not an investment. It does not provide equity, ownership, dividends, profit sharing, or any financial return. You are paying for platform access and community participation only. If you are considering this as a financial investment, please do not join — this project is not appropriate for that purpose.

No. Members do not receive company equity or ownership rights of any kind. The membership is for platform access and community participation only. There are no shares, tokens, certificates, or ownership documents issued to members.

No. The project does not offer profit sharing, revenue sharing, dividends, or financial returns of any kind. The cooperative is structured to provide platform access and member services, not to generate financial profit for members.

No. Usage will be subject to fair-use limits. This is necessary to keep the platform stable and fair for all members. Shared GPU infrastructure has finite compute capacity. If usage were unlimited, a small number of heavy users could consume all available resources and degrade the service for everyone else. Fair-use limits include monthly usage allowances, daily abuse protection, file upload limits, context limits, and queue protection.

Fair-use limits protect the cooperative so a small number of heavy users cannot consume the entire shared infrastructure. A private GPU cluster has finite compute. Without limits, automated scripts or batch jobs run by a few members could make the platform unusable for hundreds of others. Limits may include monthly usage allowances, daily request caps, file upload limits, maximum context lengths, and request queuing. Members who genuinely need more capacity may have the option to purchase additional usage at cost.

No. Training a frontier model from zero would require hundreds of millions of dollars and large research teams. $200,000 is not enough for that. The realistic plan is to use strong open-weight models, apply fine-tuning where it adds value, build RAG pipelines for document retrieval, implement tool use and agent workflows, and use commercial API fallback where the cooperative's own capacity is insufficient. The plan is grounded in what is actually achievable with this budget.

The final model stack will be selected during testing and planning. The plan is to evaluate strong open-weight models such as those in the Llama, Mistral, and Qwen families, specialized coding models, document-focused fine-tuned models, and API fallback options. The model selection will be based on performance benchmarks, licensing terms, memory requirements, and suitability for the cooperative use cases. Final model choices will be published before launch.

4× H100-class infrastructure is a strong target for running large open-weight models at useful inference speeds for private AI use cases including chat, coding assistance, document analysis, and research workflows. The target provides meaningful capacity for a 1000-member cooperative while remaining within the $200,000 budget target. The final hardware choice depends on pricing, availability, hosting costs, and the funding result at the time of purchase.

H200-class hardware may be considered if pricing and availability make it a better value than H100s at the time of purchase. The project should not overpromise H200s. The public target remains 4× H100 or better. If H200s become a realistic option within budget, the cooperative will publish the comparison and make the decision transparently.

GPU prices fluctuate significantly. If GPU prices change, the cooperative may adjust the hardware plan, rent or lease GPUs first instead of buying, purchase fewer GPUs initially and expand later, use API fallback while waiting for better pricing, or open an expansion vote after publishing an updated budget to the community. No hardware purchase will happen without publishing the updated plan.

The cooperative will not blindly spend funds if final costs are higher than planned. If GPU prices, hosting costs, or other expenses exceed the $200,000 target, the cooperative may choose a phased launch using API fallback first, lease or rent GPU capacity temporarily, purchase fewer GPUs initially and vote on expansion later, delay the hardware purchase until pricing improves, or open optional expansion funding with full member approval. No major change to the plan will happen without publishing the updated budget and getting community agreement.

If the project does not reach the required funding threshold during the reservation phase, it may continue as a smaller beta with rented or API infrastructure, delay the hardware purchase, reduce the initial scope, or refund reservations according to the published refund policy. Reservations will be refundable if the minimum funding target is not reached. The cooperative will not proceed to hardware purchase until the funding and member threshold is confirmed.

The final refund policy will be published before payments open. The planned approach is that reservations are refundable up to a defined deadline, and if the minimum funding target is not reached, all reservations are refunded. No payment is collected during the waitlist phase — there is nothing to refund at this stage.

No. The project is currently in idea validation. No payment is open and no hardware has been purchased. The current step is building a community, collecting waitlist signups, and refining the plan. Payment will only open after the full plan, budget, technical architecture, refund policy, and legal structure have been completed and published.

The 2-year founder access period begins after the member platform officially launches, not when someone joins the waitlist or makes a reservation. If someone joins the waitlist today, their clock does not start until the cooperative platform is live and access is granted. The expected timeline will be published when the reservation phase opens.

Yes. Businesses may join the waitlist and participate in the project. However, business usage often requires higher limits than individual use. Business accounts or team plans may require separate pricing after launch. Members with business use cases should note this in the optional message field when joining the waitlist so the cooperative can plan accordingly.

API access may be considered later, but it is not guaranteed in the first founder launch. API access would require separate limits and pricing because API calls can consume infrastructure significantly faster than browser-based usage. API plans would be evaluated based on demand, infrastructure capacity, and fair-use considerations. If offered, they would likely be a separate tier beyond the founder membership.

Before accepting full founder payments, the project will publish a clear budget showing expected GPU costs with current vendor pricing, server and hosting costs, storage, legal and accounting, software development estimates, operating reserve, and fallback options. After launch, monthly budget and infrastructure reports will be shared with all members. The goal is full transparency before anyone is asked to pay.

Members may vote through the member dashboard or community polls after the platform is active. Voting topics may include feature priorities, model choices, infrastructure expansion, usage policy updates, and major cooperative decisions. The governance model will be defined during the planning phase and published before the reservation phase opens. Founding members have voting rights as part of their membership.