Crypto-Friendly Hosting for Solana Validators: a Hetzner Alternative

If you run a Solana validator, RPC node, or trading infrastructure, your hosting provider's terms of service are part of your threat model. Hetzner offers some of the cheapest high-clock dedicated hardware in Europe, but its acceptable-use policy prohibits exactly the workloads Solana operators run, and it has enforced that ban at scale. This guide explains what happened, what "crypto-friendly" actually means, and how to evaluate an alternative built specifically for Solana.

Why Hetzner is a problem for Solana operators

Hetzner's anti-crypto stance is not a gray area or an unenforced clause. It is explicit policy, it covers proof-of-stake and node operation rather than just mining, and it has already taken validators offline.

The policy timeline

As of June 2026, the relevant history is well-documented across multiple outlets:

• May 2021: Hetzner added a crypto-mining ban to its terms after a surge of users renting storage capacity for Chia plotting and farming, citing concerns about premature drive wear. As reported by BleepingComputer and Decrypt at the time, the new language prohibited "the operation of applications for mining cryptocurrencies."
• August 2022: Hetzner clarified, via customer support and reported by CryptoSlate, that the prohibition extends well beyond mining. The quoted statement was that using their products for "any application related to mining, even remotely related, is not permitted. This includes Ethereum. It includes proof of stake and proof of work and related applications. It includes trading. It is true for all of our products, except colocation."

In plain terms: running a validator, running a node, storing blockchain data, and trading are all prohibited on Hetzner's cloud and dedicated servers. The only documented exception is colocation, where you own the hardware. If exact current wording matters for your compliance review, verify it against Hetzner's live AUP before relying on these dates.

The November 2022 enforcement event

This is the part operators should take seriously, because it is not hypothetical. As reported by Decrypt and corroborated by Helius, on November 2, 2022 Hetzner enforced its terms against Solana, and over 1,000 validators went offline within hours. The trigger was account-level enforcement: a validator's account was locked, with restoration conditional on removing all Solana activity.

The network itself stayed up. Solana did not halt consensus, and most affected operators migrated within days, with nearly all delinquent stake back online within a few weeks. Solana Labs co-founder Anatoly Yakovenko and the Solana Foundation publicly urged validators to move to crypto-friendly providers. But "the network survived" is cold comfort if your validator was one of the thousand that suddenly stopped voting, started missing leader slots, and risked losing delegated stake.

How much of Solana was exposed

The widely-cited figure is that roughly 40% of Solana validators were on Hetzner, representing about 20% of total network stake. As of June 2026, it is worth noting this specific pairing traces back to a single August 2022 analysis by Google Cloud engineer Sam Padilla using a monitoring tool he built, and is repeated across nearly all coverage rather than independently confirmed by Foundation telemetry. The same Decrypt analysis estimated that Hetzner, Equinix, and AWS together hosted around 65% of Solana stake at that time.

Why didn't the network halt? Solana consensus stalls when at least one-third of stake (33.4%) goes offline. The roughly 20% removed by the Hetzner enforcement was significant but below that threshold. Some 2022 coverage rounded the halt threshold to "39%," but the canonical figure is one-third of stake for a liveness halt.

What "crypto-friendly hosting" actually means

The term gets used loosely, so here is a concrete definition. A crypto-friendly host is one whose terms of service and acceptable-use policy explicitly permit the workloads you run: validators, RPC nodes, staking, indexing, and trading infrastructure. The test is simple. Are the exact activities Hetzner prohibits allowed in writing, so you are not exposed to sudden account termination?

Cheap hardware that you might lose at any moment is not actually cheap. The real cost of a provider that bans your workload includes emergency migration under time pressure, missed slots and lost rewards during downtime, and the operational risk of running on infrastructure that could be reclaimed if a TOS sweep flags your account. For a validator, that risk is structural, not occasional.

The post-Hetzner landscape

The ecosystem has diversified considerably. As of June 2026, Helius data (around epoch 685) shows Solana validators spread across roughly 135 hosting providers, with Teraswitch (~24% of validators by Helius's count, with the top two providers combining for ~43% of stake) and Latitude.sh (~19%) leading, followed by providers like OVHcloud and Cherry Servers. The Nakamoto coefficient is currently around 19 (down from a 2023 peak near 34).

Regulation is pushing in the same direction. Per Solana Foundation Delegation Program rules, as of June 2026 reporting indicates that starting May 1, 2026 SFDP participants must run on an ASN that holds under 25% of overall network stake, a rule explicitly designed to prevent the single-provider concentration the Hetzner episode exposed. If you are in or planning to join SFDP, confirm the current threshold on an official Solana source, since program rules change. The practical takeaway for operators is the same either way: choosing a provider where validators are explicitly welcome, and where you are not concentrated alongside everyone else, is now both a resilience decision and increasingly a delegation requirement.

OrbitServers as a Hetzner alternative

OrbitServers is built around crypto and Solana workloads rather than treating them as a liability. It accepts crypto payment, and it is run by the same team behind the OrbitFlare Solana RPC service, so the hardware and network choices are made by people who operate Solana infrastructure themselves. Validators, RPC nodes, staking, and trading bots are first-class supported use cases, not workloads you have to hide.

Hardware matched to Solana roles

Solana is demanding on single-thread performance, memory, and NVMe throughput. The bare-metal lineup is chosen so you can match the machine to the job rather than overpaying for a one-size-fits-all box:

CPU	Cores / clock	Best fit
Ryzen 9 9950X	16c / 32t, up to 5.7GHz boost, 64MB L3	MEV and arbitrage send paths, HFT, sniping and trading bots (highest single-thread clock)
EPYC 4564P	16c / 32t, up to 5.4GHz boost, 64MB L3	High-clock all-rounder for RPC and trading infra
EPYC 4584PX	16c / 32t, up to 5.7GHz boost, 128MB L3 (3D V-Cache)	Cache-sensitive workloads: databases, indexers, some RPC
EPYC 7763	64c / 128t, up to 3.5GHz boost, 256MB L3	High-core duty: full RPC nodes, validators, indexers, dense virtualization

All these chips run DDR5 ECC (to 192GB) except the high-core EPYC 7763, which uses 8-channel DDR4 ECC and scales to 512GB or more. For a full Solana RPC node, plan on roughly 192GB+ RAM, 16+ cores, and 2TB+ NVMe; the recommendation pages for RPC nodes and Solana validators go into the per-role sizing in more detail.

Where the metal lives

Locations matter for Solana because proximity to Jito's block engine affects bundle inclusion and trading outcomes. OrbitServers bare metal is available in Frankfurt, Amsterdam, London, New York, and Salt Lake City, all of which are Jito-connected edges with low latency to Jito. For Solana-specific deployments, there are dedicated Solana RPC hosting pages per location.

A practical networking distinction: the EU locations (Frankfurt and Amsterdam) include a 10Gbps port with unlimited bandwidth, which suits high-throughput RPC and indexing. The US locations (New York and Salt Lake City) are metered with a generous transfer allowance rather than unlimited. Choose based on where your validators, RPC consumers, or trading venues sit, and review the network page for the full picture.

What every dedicated server includes

• Full root access on Linux or administrator access on Windows
• IPMI / KVM out-of-band management, so you keep control even if the OS is unreachable
• A dedicated IPv4 address
• Included DDoS protection, with up to 100Gbps networking available
• Instant deploy on select Frankfurt and Amsterdam configurations; otherwise typically within 24 hours

Out-of-band management is not a luxury for validators. When you need to recover a node, swap a config, or reboot after a failed restart, IPMI/KVM is the difference between a quick fix and a support ticket queue.

Pricing and what fits which operator

Bare metal starts at $299.99/mo. If you are running RPC consumers, dev environments, or smaller trading bots, the VPS range runs from $39.99/mo (Starter, 4 vCPU / 8GB / 50GB NVMe) up to $449.99/mo (Colossus, 48 vCPU / 128GB / 320GB), deploys instantly on payment, and carries a 99.9% uptime SLA. For teams that want to bring their own hardware, the colocation route, which was Hetzner's one carve-out for crypto, is available on custom quote via the colocation page.

• Validators: high-core dedicated metal, 192GB+ RAM, in a location near your delegators. See Solana validators.
• RPC runners: high-clock or cache-heavy CPUs with EU unlimited bandwidth for read-heavy workloads. See RPC nodes.
• Traders and bot operators: the highest single-thread clock on a Jito-connected edge. See Solana trading and trading bots.
• Multi-tenant or staging: high-core servers for virtualization.

A note on the social proof

OrbitServers currently holds a 5.0/5 rating on Trustpilot across 21 reviews. That is a small but real sample from a focused operator base, and we would rather state it plainly than dress it up. On latency, OrbitServers locations are genuinely Jito-connected edges with low latency to Jito, but we deliberately avoid quoting a specific measured millisecond figure as a hard guarantee, because real-world latency depends on your peers, your software, and network conditions on the day.

Bottom line

The Hetzner episode was a clear lesson: hosting on a provider that prohibits your workload is a liability that can materialize in hours, not months. Crypto-friendly hosting is not a marketing label, it is the practical baseline of running where validators, nodes, staking, and trading are explicitly permitted, on hardware sized for the role, near the edges that matter for Solana. If you are migrating off Hetzner or building a new validator from scratch, the next step is to pick the right machine and location for your role.