# FlexDedi > FlexDedi is a cloud dedicated server provider: single-tenant AMD EPYC bare-metal servers, plus NVIDIA L40S and RTX PRO 6000 Blackwell GPU servers, that you order and manage like cloud instances. Every server has unmetered traffic on a 10 Gbps to 2 × 100 Gbps port, always-on DDoS protection and IPMI access, and is deployed in one of 10 data centers in North America, Europe and Asia-Pacific. Prices from $299 per month, no setup fees, monthly billing, paid in Bitcoin. Every page below also exists as Markdown: add .md to its URL (the home page is https://flexdedi.com/index.md). Everything in one file: https://flexdedi.com/llms-full.txt ## Key facts - **What:** cloud dedicated servers. Single-tenant bare metal (no hypervisor, no other tenants) that you order and control like a cloud instance. - **Hardware:** AMD EPYC 4005 and 9005 (Zen 5) processors with 16 to 192 cores, 128 GB to 1.5 TB of DDR5 ECC memory and NVMe Gen4 or Gen5 storage; GPU servers with NVIDIA L40S (48 GB) or 2 × RTX PRO 6000 Blackwell (96 GB each); a 264 TB storage server. - **Prices:** 10 configurations from $299 to $2,999 per month, in US dollars, excluding taxes. No setup fees, no contract, monthly billing, cancel at the end of any month. - **Network:** unmetered traffic on every server (no monthly cap, no overage or egress fees), ports from 10 Gbps to 2 × 100 Gbps, always-on DDoS protection, 1 IPv4 address and a /64 IPv6 block, free private VLANs between servers in a location. - **Locations:** 10 data centers: New York, Ashburn, Dallas, Los Angeles, Amsterdam, Frankfurt, London, Singapore, Tokyo, Sydney. - **Delivery:** Core and Pro configurations are pre-racked and ready in minutes; GPU, Max and Storage servers are delivered the same day. - **Payment:** a prepaid balance topped up in Bitcoin (BTC on the Bitcoin network), from $100 to $25,000 per top-up. Servers renew every month from the balance; no card or bank account is needed. - **Management:** control panel, REST API and Terraform provider, IPMI and HTML5 KVM console, one-click images for Ubuntu, Debian, AlmaLinux, Rocky Linux, Windows Server and Proxmox VE, or any custom ISO. - **Support and SLA:** 24/7 support by engineers through tickets in the client area; 99.99% monthly network availability SLA with service credits; failed hardware replaced within 4 hours of diagnosis. - **Servers are unmanaged:** customers get full root access and run their own software stack. ## Servers (USD per month, excluding taxes) - [Core 16](https://flexdedi.com/dedicated-servers/core-16): AMD EPYC 4545P (16 cores / 32 threads), 128 GB DDR5 ECC, 2 × 1.92 TB NVMe Gen4, 10 Gbps unmetered, $299/month, ready in minutes (pre-racked and tested). Good for: Web apps, APIs, game servers, CI runners. - [Core 16X](https://flexdedi.com/dedicated-servers/core-16x): AMD EPYC 4585PX (16 cores / 32 threads), 192 GB DDR5 ECC, 2 × 3.84 TB NVMe Gen4, 10 Gbps unmetered, $399/month, ready in minutes (pre-racked and tested). Good for: Game servers, trading, latency-critical apps. - [Storage 264](https://flexdedi.com/dedicated-servers/storage-264): AMD EPYC 4565P (16 cores / 32 threads), 128 GB DDR5 ECC, 12 × 22 TB HDD + 3.84 TB NVMe, 25 Gbps unmetered, $699/month, delivered the same day. Good for: Backups, object storage, media libraries. - [Pro 32](https://flexdedi.com/dedicated-servers/pro-32): AMD EPYC 9355P (32 cores / 64 threads), 384 GB DDR5-6000 ECC, 2 × 3.84 TB NVMe Gen5, 25 Gbps unmetered, $749/month, ready in minutes (pre-racked and tested). Good for: Databases, SaaS back ends, virtualization. - [GPU L40S](https://flexdedi.com/dedicated-servers/gpu-l40s): AMD EPYC 9355P (32 cores / 64 threads), 256 GB DDR5-6000 ECC, 2 × 3.84 TB NVMe Gen5, 1 × NVIDIA L40S 48 GB, 25 Gbps unmetered, $979/month, delivered the same day. Good for: AI inference, rendering, video transcoding. - [Pro 48](https://flexdedi.com/dedicated-servers/pro-48): AMD EPYC 9455P (48 cores / 96 threads), 576 GB DDR5-6000 ECC, 2 × 7.68 TB NVMe Gen5, 25 Gbps unmetered, $999/month, ready in minutes (pre-racked and tested). Good for: Kubernetes nodes, analytics, large databases. - [Pro 64](https://flexdedi.com/dedicated-servers/pro-64): AMD EPYC 9555P (64 cores / 128 threads), 768 GB DDR5-6000 ECC, 4 × 3.84 TB NVMe Gen5, 2 × 25 Gbps unmetered, $1,349/month, ready in minutes (pre-racked and tested). Good for: Private cloud, CI farms, high-traffic platforms. - [Max 96](https://flexdedi.com/dedicated-servers/max-96): AMD EPYC 9655P (96 cores / 192 threads), 1,152 GB DDR5-6000 ECC, 4 × 7.68 TB NVMe Gen5, 2 × 25 Gbps unmetered, $1,899/month, delivered the same day. Good for: HPC, big data, dense virtualization. - [GPU RTX PRO](https://flexdedi.com/dedicated-servers/gpu-rtx-pro): AMD EPYC 9455P (48 cores / 96 threads), 576 GB DDR5-6000 ECC, 2 × 7.68 TB NVMe Gen5, 2 × NVIDIA RTX PRO 6000 Blackwell 96 GB, 25 Gbps unmetered, $2,199/month, delivered the same day. Good for: LLM inference and fine-tuning, 3D, simulation. - [Max 192](https://flexdedi.com/dedicated-servers/max-192): 2 × AMD EPYC 9655 (192 cores / 384 threads), 1.5 TB DDR5-6000 ECC, 4 × 7.68 TB NVMe Gen5, 2 × 100 Gbps unmetered, $2,999/month, delivered the same day. Good for: In-memory analytics, consolidation, HPC. ## Which server for which workload - [AI inference & ML](https://flexdedi.com/solutions/ai-inference): start with GPU L40S ($979/mo). One NVIDIA L40S with 48 GB of VRAM and FP8 support handles 7B to 32B-class models with long contexts, or a 4-bit 70B with short ones. At $979/mo it is the entry point for dedicated GPU inference. - [Databases](https://flexdedi.com/solutions/databases): start with Pro 32 ($749/mo). 32 Zen 5 cores, 384 GB of RAM and 2 × 3.84 TB NVMe Gen5 in a mirror: a solid primary for PostgreSQL or MySQL datasets up to about 3 TB. Order two, a primary and a standby, at $749/mo each. - [Game servers](https://flexdedi.com/solutions/game-servers): start with Core 16X ($399/mo). The pick for simulation-heavy games: 5.7 GHz boost, 128 MB of 3D V-Cache and 192 GB of RAM for many instances per server. $399/mo in all 10 locations. - [Streaming & media](https://flexdedi.com/solutions/streaming): start with Pro 48 ($999/mo). 48 cores and 576 GB of RAM for transcoding ladders, packaging and an in-memory cache of hot segments, with 2 × 7.68 TB NVMe Gen5 and a 25 Gbps port. Ready in minutes in 8 locations. - [SaaS & web apps](https://flexdedi.com/solutions/saas): start with Pro 32 ($749/mo). The workhorse for SaaS back ends: 32 cores, 384 GB of RAM and NVMe Gen5 for a primary database or a dense container host. $749/mo, ready in minutes in 8 locations. - [Private cloud](https://flexdedi.com/solutions/private-cloud): start with Pro 64 ($1,349/mo). A balanced cluster node: 64 cores, 768 GB of RAM and four NVMe Gen5 drives for local storage or Ceph OSDs, with 2 × 25 Gbps ports. Ready in minutes at $1,349/mo. - [CI/CD & dev](https://flexdedi.com/solutions/ci-cd): start with Pro 48 ($999/mo). 96 threads and 576 GB of RAM run about a dozen 8-thread jobs in parallel with room for tmpfs build directories, plus 2 × 7.68 TB NVMe Gen5 for caches. $999/mo. - [Big data & analytics](https://flexdedi.com/solutions/analytics): start with Max 96 ($1,899/mo). 96 cores and 1,152 GB of RAM on a single socket, so there is no cross-socket NUMA traffic to tune around. A strong single-node ClickHouse or Spark host at $1,899/mo. ## Locations - [New York](https://flexdedi.com/locations/new-york), US East: 4 of 10 configurations - [Ashburn](https://flexdedi.com/locations/ashburn), US East: 10 of 10 configurations - [Dallas](https://flexdedi.com/locations/dallas), US Central: 6 of 10 configurations - [Los Angeles](https://flexdedi.com/locations/los-angeles), US West: 7 of 10 configurations - [Amsterdam](https://flexdedi.com/locations/amsterdam), Netherlands: 10 of 10 configurations - [Frankfurt](https://flexdedi.com/locations/frankfurt), Germany: 10 of 10 configurations - [London](https://flexdedi.com/locations/london), United Kingdom: 4 of 10 configurations - [Singapore](https://flexdedi.com/locations/singapore), Singapore: 5 of 10 configurations - [Tokyo](https://flexdedi.com/locations/tokyo), Japan: 2 of 10 configurations - [Sydney](https://flexdedi.com/locations/sydney), Australia: 2 of 10 configurations ## Buying, billing and policies - [Deploy a server](https://flexdedi.com/order): configure, create an account and pay in one form - [Pricing](https://flexdedi.com/pricing): every monthly price, price per core and per GB, billing FAQ - [Pay with Bitcoin](https://flexdedi.com/bitcoin-dedicated-servers): how the prepaid BTC balance works - [Fleets and enterprise](https://flexdedi.com/enterprise): volume pricing from 5 servers - [Cloud migration](https://flexdedi.com/migration): savings estimate and migration help - [Network](https://flexdedi.com/network): unmetered ports, DDoS protection, latency between locations - [Support](https://flexdedi.com/support): 24/7 tickets by engineers - [Acceptable Use Policy](https://flexdedi.com/legal/acceptable-use): The rules for using FlexDedi servers and network, covering prohibited content, spam, network abuse, fair use of unmetered ports and how abuse is handled. - [Cookie Policy](https://flexdedi.com/legal/cookies): No analytics, advertising or tracking cookies. Only strictly necessary ones: Cloudflare's security cookies and the client area's session cookie. - [Privacy Policy](https://flexdedi.com/legal/privacy): How FlexDedi collects, uses, shares and protects personal data from its website, forms and services, how long it is kept and how to exercise your rights. - [Service Level Agreement](https://flexdedi.com/legal/sla): FlexDedi commits to 99.99% monthly network availability and to hardware replacement within 4 hours of diagnosis, with service credits, exclusions and claims. - [Terms of Service](https://flexdedi.com/legal/terms): The contract for FlexDedi dedicated servers, covering accounts, orders, monthly USD billing, unmetered traffic, your responsibilities and liability. ## Platform - [Instant provisioning](https://flexdedi.com/platform/provisioning): Servers marked Ready in minutes are already racked and burn-in tested. Pay, and your server usually arrives within 15 minutes with your operating system installed. - [Control panel](https://flexdedi.com/platform/control-panel): One place for day-to-day operations on your servers, from a power cycle to a full reinstall. The client area is live with your servers and 24/7 support tickets; these actions roll out in it next, each one also exposed through the REST API. - [API & Terraform](https://flexdedi.com/platform/api): Everything you can do in the control panel, you can do over HTTPS. Call the REST API directly, or declare servers in Terraform and review every change before it is applied. - [OS library](https://flexdedi.com/platform/os-library): Choose an operating system when you order or reinstall, and your server comes up with it installed. For anything else, boot your own ISO through the HTML5 KVM console. - [DDoS protection](https://flexdedi.com/platform/ddos-protection): Every IP address on every FlexDedi server sits behind always-on filtering at the edge of our network. It is included in the price, with nothing to enable or configure. - [Private networking](https://flexdedi.com/platform/private-networking): Put replication, storage and cluster traffic on a private VLAN between your servers in the same location. It is free, and internal traffic is not billed. - [Out-of-band access](https://flexdedi.com/platform/out-of-band): Every server comes with IPMI and an HTML5 KVM console that work independently of the operating system. Fix a broken network configuration, watch a boot or install a custom ISO from your browser. - [Monitoring & alerts](https://flexdedi.com/platform/monitoring): The control panel shows each server's bandwidth and hardware health, and email alerts tell you when something needs attention. Pair it with your own monitoring inside the OS for full coverage. ## Documentation - [Configure IPv6 on your server](https://flexdedi.com/docs/ipv6-networking): Assign an address from your /64 IPv6 block with Netplan on Ubuntu 24.04, ifupdown on Debian or PowerShell on Windows Server, then test it and set reverse DNS. - [Connect servers over a private VLAN](https://flexdedi.com/docs/private-networking): Link your FlexDedi servers in one location over a free private VLAN, with a Netplan example, MTU checks, firewall rules and uses such as database replication. - [Connect to your server with SSH](https://flexdedi.com/docs/ssh-access): Set up SSH keys from Linux, macOS or Windows, save hosts in ~/.ssh/config, disable password logins without locking yourself out and recover access by KVM. - [Getting started with your FlexDedi server](https://flexdedi.com/docs/getting-started): Deploy your first dedicated server: pick a configuration and a city, pay from your balance, find your IP and credentials, then log in and secure it. - [Reinstall the OS or boot rescue mode](https://flexdedi.com/docs/reinstall-rescue): Back up your data, reinstall a one-click OS image, or boot the rescue system to mount your disks, fix fstab and GRUB, and reset a password through chroot. - [Use IPMI and the HTML5 KVM console](https://flexdedi.com/docs/ipmi-kvm): Power-cycle your server, boot an ISO, reach the BIOS and repair a broken system with IPMI and the HTML5 KVM console, and keep out-of-band access secure. ## Guides - [Bare metal vs VPS vs cloud: which one do you need?](https://flexdedi.com/guides/bare-metal-vs-vps-vs-cloud): Bare metal, VPS or public cloud? Compare isolation, noisy neighbors, performance consistency, elasticity, egress pricing and real cost at steady load. - [Hosting game servers on dedicated hardware](https://flexdedi.com/guides/game-server-hosting-performance): Why clock speed and cache decide game server performance, how to estimate instances per server, and how to run game servers on Linux with DDoS protection. - [How to choose a dedicated server in 2026](https://flexdedi.com/guides/how-to-choose-a-dedicated-server): A practical guide to choosing a dedicated server in 2026, from CPU and memory channels to NVMe, RAID, network ports, location, delivery and budget. - [Install Proxmox VE on a dedicated server](https://flexdedi.com/guides/proxmox-on-dedicated-server): Install Proxmox VE 9 on a dedicated server from the one-click image or the ISO, then set up repositories, ZFS or LVM-thin, networking, firewall and backups. - [Secure a new Linux dedicated server in 15 minutes](https://flexdedi.com/guides/secure-a-new-linux-server): A 15-minute hardening baseline for Ubuntu 24.04 and Debian servers covering updates, a sudo user, SSH keys, firewall, fail2ban, monitoring and backups. - [Sizing a GPU server for LLM inference](https://flexdedi.com/guides/llm-inference-gpu-sizing): Work out how much VRAM an LLM needs from weights, KV cache, batch and context, with formulas, vLLM commands and what fits on L40S and RTX PRO 6000 GPUs. - [Software RAID 1 on NVMe with mdadm](https://flexdedi.com/guides/software-raid-nvme-mdadm): Check, create, monitor and repair mdadm RAID 1 and RAID 10 on NVMe servers, replace a failed drive step by step, and keep both disks bootable with UEFI. ## Blog - [AMD EPYC 9005 'Turin' on dedicated servers: what changes](https://flexdedi.com/blog/amd-epyc-9005-turin-dedicated-servers): What AMD EPYC 9005 Turin brings to dedicated servers, from Zen 5 cores and 12 DDR5 channels to PCIe 5.0 NVMe, and how to choose between single and dual socket. - [Introducing FlexDedi: dedicated servers, deployed like cloud](https://flexdedi.com/blog/introducing-flexdedi): FlexDedi launches single-tenant AMD EPYC servers from $299 to $2,999 a month, with unmetered traffic, 10 locations and delivery in minutes for many models. - [Unmetered bandwidth explained (and why we include it)](https://flexdedi.com/blog/unmetered-bandwidth-explained): Metered, unmetered, 95th percentile and per-GB egress compared, what an unmetered 10, 25 or 100 Gbps port means in practice, and how it changes cost planning. - [Why our servers have 384, 576 or 1,152 GB of RAM](https://flexdedi.com/blog/ddr5-memory-channels-server-sizing): How DDR5 memory channels set our RAM sizes, why one DIMM per channel matters, how bandwidth trades against capacity, and how to size RAM for databases and VMs. ## Optional - [Full content of the site in one Markdown file](https://flexdedi.com/llms-full.txt) - [API reference](https://flexdedi.com/docs/api) - [Changelog](https://flexdedi.com/changelog)