Virtualized compute. Deploy virtual machines through our Web Wizard, CLI, or infrastructure as code, and adjust capacity as your strategies evolve. General-purpose and latency-optimized instance profiles support development, research, and production applications without a hardware procurement cycle.
Bare-metal compute. Run on dedicated, single-tenant servers with no hypervisor and full control over the hardware. Latency-tuned CPU, NIC, and BIOS configurations, together with kernel-bypass networking, help reduce latency and jitter for demanding trading workloads.
Bare metal with FPGA. For ultra-low-latency trading, add programmable hardware acceleration to dedicated servers. FPGA configurations bring latency-critical processing into hardware, supporting strategies where response time and consistency are central to execution.
Storage and connectivity. Pair compute with high-performance block storage for order logs, market data capture, and research, or use local NVMe on bare metal. Both compute models share the Tooq Cloud private network and on-demand provisioning model, connecting your environment to exchange access and Tooq managed services.