ReliableSite

What ReliableSite actually is

ReliableSite is a US dedicated-server specialist that has done one thing since 2006: bare metal. ReliableSite.Net LLC was founded by Radic Davydov, and the company grew out of the New York City metro market into a small multi-site operator running its own hardware, its own network and its own on-site technicians. There is no shared hosting, no budget VPS line, no platform story — just dedicated servers for production workloads, hosting providers, gaming, and increasingly GPU/AI boxes.

The current footprint is six locations: New York City Metro, Miami, Los Angeles, Dallas, Amsterdam and Querétaro, Mexico. The European and Mexican sites are new — configuration pages still label Amsterdam and Querétaro as pre-order while the homepage lists them as deployable — so if those regions matter to you, confirm live stock before planning around them. The US sites are the mature ones, with on-site engineers 24/7/365, spare parts stocked for rapid hardware replacement, and facilities the company describes as Tier 3/Tier 4 SSAE 16 Type II audited (the New York flagship is Tier 4; Miami is a Tier 3 facility rated to withstand Category 5 hurricanes). Our category is /category/dedicated-servers.

Pricing: what you really pay

Published list pricing is refreshingly concrete. The entry point in current stock is the Intel Special at $49/month with Linux: 4 cores / 4 threads (an Intel Core i7 860 or better, benchmarked at 2,700+ by their own table), 32 GB DDR3, a 1 TB SSD, 5 TB of monthly bandwidth, 1 IPv4 plus 1 IPv6, KVM over IP and full root. From there the ladder is regular: a 64GB Intel Special adds DDR4, 2 x 2TB SSDs and a 1 Gbps unmetered port for $50 more per month; the Ryzen line starts around $129/month for a Ryzen 5600X with 64 GB DDR4 and 512 GB NVMe; 8-core Ryzen boxes run $139/month; storage boxes with 32-40 TB of raw HDD are $149/month; and a 20-core Intel Special with 256 GB RAM starts at $199/month. Rapid-deploy configurations advertise a 10-minute setup. The catalog's own filters show the ceiling: core counts up to 192, memory up to 2 TB, and a price slider reaching $5,000 per month for the largest builds — configuration is done through one order form rather than a maze of tiers.

Bandwidth is where the pricing model shows its character. Baselines vary by configuration — the $49 special ships with 5 TB monthly transfer, while many higher configs include 1 Gbps unmetered — and unmetered upgrades are published as line items: 2 Gbps unmetered at +$99.00/mo, 3 Gbps at +$190/mo, 5 Gbps at +$350/mo, 10 Gbps at +$599/mo. The Metal+ membership bundles a 10 Gbps port, a free second IPv4, control-panel tooling (ADMN, Blesta) and hardware discounts. DDoS protection across all IPs is +$0.00, as is KVM over IP. Management is an explicit choice: unmanaged/self-managed is free, one-time migration is +$29.00 setup, and managed dedicated support is +$99.00/mo.

Our one caution on money: ReliableSite does not offer refunds — third-party roundups describe a strict no-refund policy and note that payment disputes are pushed toward cryptocurrency resolution. Treat the first month as a purchase, not a trial, and validate your configuration carefully before committing.

Performance & infrastructure

The network story is the main reason people choose this provider. Every server gets DDoS protection at no charge, covering all IPs with no packet-per-second limit and data-center bandwidth capacity; the free tier is specified to defend against attacks up to 20 Gbps, with bigger protection available as an add-on. The company also offers free IP and ASN announcement, which matters for operators bringing their own address space. The backbone is advertised as 1 Tbps+ capacity with 10 Gbps (and higher) port options. The company publishes straightforward geographic framing for location choice: New York for North American and European reach, Miami as a low-latency gateway to Latin America, Los Angeles for West Coast, Mexican and trans-Pacific traffic — practical guidance when you are picking a first site for a latency-sensitive workload.

The availability commitment is unusually blunt: a 100% uptime SLA covering network, power and cooling, backed by redundant power and cooling at each site. Claims of that shape deserve scrutiny, but the track record holds up in public data — third-party monitoring write-ups cite a customer running 1-minute checks on a Los Angeles server for 1,100+ days at over 99.99% uptime, and community reports of 99.99%+ measured over years. We have not run those measurements ourselves, and we would not quote them as our own; we note them because long-horizon third-party uptime data is rare and worth something.

Hardware choice is broad — AMD Ryzen and EPYC, Intel Core and Xeon, from quad-core specials up to 192-core and 2 TB RAM configurations, with GPU servers (the GPU AI line), storage-heavy HDD boxes and NVMe general-purpose builds. RAID options, custom OS images (AlmaLinux, Debian, Proxmox or bring your own), Windows licensing and FTP backup space are all configured at order time, and IPMI/KVM-over-IP comes free so you can reach the console even when the network path is broken. Everything is operated in-house: their own hardware, their own staff physically present in the facilities — the reason a 3 a.m. hardware failure gets a human with a spare part rather than a ticket queue.

Where it falls short

Start with the refund policy above; it is the sharpest edge in the customer experience. Second, the no-frills model cuts both ways: this is unmanaged by default, and while managed support exists at +$99/mo, the platform has none of the elasticity of cloud — a server is a server, and scaling means a migration. Teams that need burst capacity, per-second billing or global regions will find the model old-fashioned.

Third, location coverage. The US presence is solid (NYC metro, Miami, LA, Dallas), but Europe is a single new Amsterdam site and Asia-Pacific is served indirectly from Los Angeles. If your users are in APAC or you need multi-region redundancy inside the EU, the portfolio does not yet stretch that far. The database row we inherited claimed New York and Phoenix; Phoenix is not in the current lineup at all, and we have corrected it.

Fourth, configuration and pricing opacity at the margins. Stock is limited and location-dependent (the order forms show quantity counts per site), clearance specials rotate, and some advertised value comparisons (the "$198/mo Value" framing on the Intel Special) are marketing math rather than list prices. Read the configurator totals, not the badges.

Finally, small things: FTP backup space is an add-on rather than an inclusion (off-server backup is offered as an in-house service, but you must opt in), DDR3-based specials are older hardware, and enterprise procurement signals like public compliance documentation are thinner than the big dedicated providers publish.

Who should pick this

We'd pick ReliableSite when the workload wants a real machine and predictable bandwidth: game servers (Minecraft, Counter Strike appear in their own product navigation), media and streaming delivery, forex/trading boxes, hosting resellers (there is a white-label reseller program), private infrastructure, and AI inference boxes where GPU pricing matters. The unmetered port options make it a natural home for bandwidth-heavy delivery, and the 10-minute rapid deploys are genuinely useful when a machine dies and you need hardware now. Because stock is tracked per location and shown as quantity counters on the order form, check availability before finalizing a configuration.

We'd look elsewhere when: you need refunds or contractual flexibility, you need APAC or deep EU presence, you want cloud-style elasticity, or you are a beginner who needs managed-by-default (budget +$99/mo for management, or choose a managed provider). For comparing the wider bare-metal market, see /category/dedicated-servers; our /benchmarks pages cover measured compute where we have test data.

Key takeaways

• Bare-metal-only provider since 2006 (founder/CEO Radic Davydov), with in-house hardware, network and on-site staff at every facility.
• Six locations: New York City Metro, Miami, Los Angeles, Dallas, plus newly added Amsterdam and Querétaro, Mexico.
• 100% uptime SLA on network, power and cooling; third-party long-run monitoring reports 99.99%+ over multi-year windows.
• Free DDoS protection on all IPs (specified to 20 Gbps), free KVM/IPMI, free IP/ASN announcement, 1 IPv6 included on current configs.
• Published pricing: entry Intel Special at $49/month (4 cores, 32 GB, 1 TB SSD, 5 TB transfer); unmetered upgrades from 2 Gbps at +$99/mo to 10 Gbps at +$599/mo; managed support +$99/mo.
• Strict no-refund policy and unmanaged-by-default posture; our overall view lands at 3.7/5.

Sources

• https://www.reliablesite.net/default.aspx
• https://www.reliablesite.net/dedicated-servers/
• https://www.reliablesite.net/dedicated-servers/specials.aspx
• https://www.reliablesite.net/dedicated-servers/quad-core-server/intel-special-32gb
• https://www.reliablesite.net/dedicated-servers/20-core-server/intel-special-256GB
• https://whtop.com/plans/reliablesite.net/140876
• https://whtop.com/news/22540-reliablesite-offering-servers-in-miami-for-low-latency-to-the-americas
• https://hostveritas.net/host/reliablesite
• https://reliablesite.net/hosting-news/reliablesite-dedicated-hosting-backup
• https://gist.github.com/obsjs06/93713b9940f335c71bebaec9322705dd