> For the complete documentation index, see [llms.txt](https://tensorspace.gitbook.io/tensorspace-whitepaper/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://tensorspace.gitbook.io/tensorspace-whitepaper/product-utilities-features/tensorunit.md).

# TensorUnit

At the heart of TensorSpace lies the TensorUnit, a robust infrastructure not bound by peer-to-peer limitations. This is the On-Demand Computing Power for your computational needs. TensorUnit offers the following: 

* Flexible Resource Pool: TensorSpace eliminates the need to purchase expensive AI hardware. Instead, users gain access to a vast array of processing units (PUs) owned and managed by TensorSpace. This model ensures consistent performance and streamlined maintenance.
* Diverse PUs: TensorUnit offers PUs with varying capacities and specializations (GPUs, TPUs, etc.) to match diverse AI workloads.
* Hourly Rentals: Pay only for what you need. Rent PUs for periods as short as one hour, maximizing cost-effectiveness and adaptability for projects of all sizes.
* USDT Pricing: Rental fees are transparently priced in USDT, providing a stable and familiar payment method for web3 users.