World Of Provider

From an empty garage to a working provider

World Of Provider starts with $5,000, an empty garage, and a paused clock. The opening build uses most of that cash on a server rack plus power, cooling, networking, and storage, leaving a small reserve for the GPUs that make the first model possible.

The floor is an isometric grid, but the important decisions live in linked capacity bars. A rack without enough electrical or cooling headroom cannot become a dependable business, and storage and network limits matter once customer work begins to arrive.

Fit the model before taking the contract

Each rack exposes GPU slots, total VRAM, power draw, temperature, utilization, and its running model. In the tutorial, two 12 GB Vexa G1 Ember cards leave enough room for the 6 GB Orrin 3B model and the first contract from four indie-developer customers.

Accepting that contract is only the start. Running the clock creates requests, the rack processes them through its queue, and completed jobs move the balance and revenue figures while upkeep keeps draining cash. The first three jobs made the loop clear: capacity becomes service, service becomes revenue, and revenue has to outrun operating costs.

Read the whole facility, not one rack

The useful habit is to check the top-bar totals before buying more compute. Extra GPUs increase power and heat, while larger contracts can pressure network and storage capacity; the Manage view brings those bottlenecks together with revenue, costs, and net profit.