Toolin.ai
TensorDock

TensorDock

Officially listed

A decentralized GPU cloud marketplace that delivers high-performance compute at affordable prices.

views 53favorites 0Paid
Visit website

TensorDock

TensorDock is a decentralized GPU compute cloud marketplace aiming to become "the Airbnb of cloud computing." By pooling computing resources from hosts worldwide, it delivers high-performance compute at highly affordable prices for AI training, inference, and cloud gaming users.

Core Capabilities

  • Full KVM virtualization: Provides KVM-based virtual machines with complete control of the underlying operating system, supporting custom drivers and even running Windows.
  • Massive global compute: Nodes deployed across more than 20 countries, managing a rich GPU fleet of over 30,000 cards ranging from RTX 4090 to H100.
  • Dynamic seamless scaling: Supports on-demand deployment with no tedious quota applications — via API you can scale from zero to hundreds of GPUs within tens of seconds.
  • Minimal storage billing: Stopped instances are charged only a very low NVMe storage fee, and the platform charges nothing for data ingress or egress.
  • Flexible spot instances: Heavily discounted spot resources save up to 80% on cloud costs for interruptible computing tasks.

Best For Suits AI startups with limited budgets but huge compute needs, university researchers, and cloud developers who need low-level system access. It is also a perfect choice for running cloud gaming, large-scale 3D rendering, and heavy computing tasks at minimal cost.

Unique Advantages Compared with pure container platforms, TensorDock offers a more secure, better-isolated KVM virtual machine experience; against traditional cloud giants, its free bandwidth and minimal quota system dramatically lower costs and entry barriers. This is the compute home for hardcore technical users seeking the ideal balance of performance and price.

Editor's Review TensorDock has successfully broken the entry barrier of premium cloud compute with astonishing prices and a free, flexible infrastructure. Although decentralized underlying resources mean node quality occasionally fluctuates and there is a technical learning curve, its full root access and no-gotcha billing make it an ultimate cost-cutting, efficiency-boosting weapon.

Pricing

### 💰 定价模式：按量付费 **起步价**：按量计费，低至 $0.012/小时 #### 主要方案 - **GPU 实例**：~$0.12 - $2.36 / 小时 - 支持从消费级到企业级 H100 等多种型号 - **CPU 实例**：低至 $0.012 / 小时 - 提供 Intel Xeon 和 AMD EPYC 处理器 - **竞价实例 (Spot)**：享 50%-80% 折扣 - 适合可中断的批量计算任务 #### 试用/其他信息 采用预存资金后按秒后付费机制，不收取数据出入站流量费（Ingress/Egress），停机仅收取极低存储费用。 — Visit website

FAQ

Is TensorDock free?

TensorDock uses on-demand, post-paid billing and offers no free quota. But its pricing is extremely competitive, starting at less than $0.20 per hour — ideal for cost control.

What are TensorDock's main features?

TensorDock mainly provides high-performance KVM-based GPU and CPU cloud servers with full root access, no bandwidth fees, per-second billing, and API-driven automated scaling.

What advantages does TensorDock have over traditional cloud providers?

Compared with traditional cloud giants, TensorDock is typically about 80% cheaper, charges nothing for data ingress/egress traffic, and requires no tedious compute quota applications.

Is TensorDock suitable for cloud gaming?

Very much so. Because it supports a complete Windows operating system and allows installing custom drivers for GPU passthrough, many gamers pair it with tools like Parsec to build low-latency cloud gaming nodes.

What are TensorDock's spot instances?

Spot instances are discounted compute resources offered by TensorDock, priced 50%-80% below on-demand instances. They suit non-critical tasks that can tolerate interruption at any time.

How stable is TensorDock?

Because it aggregates third-party hosts in a decentralized model, some nodes' network and stability may fluctuate. It better fits fault-tolerant R&D and computing environments without extremely strict SLA requirements.