tzulo

What tzulo actually is

tzulo is a small, privately operated hosting company that has been around since 2003 — founded, in its own words, "in early 2003" as an IT development shop for small and medium businesses before expanding into datacenter services. It sells dedicated servers, colocation and IP transit out of facilities in downtown Chicago and downtown Los Angeles, runs its own BGP network (AS11878) with its own IPv4 and IPv6 blocks, and buys transit from five tier-1 providers. The homepage claims a 70 Gb/s backbone with links held under 50 percent peak capacity. IP transit is available from 13 points of presence across North America.

This is not a hyperscaler and does not pretend to be one. The company employs engineers averaging 10+ years of datacenter experience, sells month-to-month with no setup fees, and answers tickets through its own helpdesk. Our correction file for this listing starts with the launch year: a naive reading of old crawl data dated the company to 2026, which is wrong by more than two decades. Two other DB fields need fixing up front — the entry price is not $80 and the maximum price is certainly not $0 — and we get into the real numbers below.

The product line is straightforward. Dedicated servers range from a single Xeon E3-1270v2 with 16 GB RAM and one 2 TB disk at $125 per month up through dual-socket builds and storage boxes, plus a "Build Your Own" configurator covering single, dual and quad CPU systems, 16 GB to 1 TB of RAM, 1 to 24 HDD or SSD disks, and 1 Gbit or 10 Gbit uplinks. Colocation starts at $50 per month for 1U with 1 amp at 208V, a remote-reboot PDU and 100 Mbit delivered, scaling to half racks and full 47U cabinets. Cloud/VPS plans are advertised from $10 per month but the cloud page currently reads "SOLD OUT" — worth knowing before you plan around it.

The facilities story is credible for a provider this size. Chicago is a Tier III environment (the company cites CoreSite space in the financial district, with an earlier Equinix presence) offering 110V and 208V delivery, UPS and generator service on all circuits, N+1 cooling, multiple fiber paths, a meet-me room with 30+ carriers in-house, and 24/7/365 security with biometric entry. Los Angeles is a private downtown facility with 110V, 208V and three-phase power, over 100 dark fibers to Equinix and CoreSite PoPs, 10+ carriers in-house with no cross-connect fees, and remote-hands support. Neither site will win a "most modern campus" contest, but both are real, staffed, carrier-rich rooms.

Pricing: what you really pay

The published dedicated lineup is fixed and small. E3-1270v2 (4 physical cores / 8 threads, 3.5 GHz, 16 GB RAM, 1x 2 TB, 1 Gbit unmetered) from $125 per month. E5-1620 (4 cores / 8 threads, 3.6 GHz, 32 GB RAM, 2x 2 TB) from $150. Dual Xeon L5630 (8 cores / 16 threads, 24 GB RAM, 2x 2 TB) also from $150. Dual E5-2620 (12 cores / 24 threads, 32 GB RAM, 2x 2 TB) from $300. The storage box — dual L5630, 24 GB RAM, a 120 GB SSD plus 8x 3TB disks and an LSI RAID controller — is $325 per month. All plans include IPv4 and IPv6, free reboots and OS reinstalls, no setup fees, and month-to-month billing. Build-your-own pricing is quoted case by case.

Two things follow from that list. First, the hardware generation is old. The E3-1270v2 is a 2012-era part and the L5630 is older still; at $125 to $325 per month you are paying for the network, the unmetered port and the colo-grade facility, not for modern silicon. Compare that to what the same money buys in current-generation hardware from the providers on our /benchmarks page before you assume this is a bargain. Second, the unmetered 1 Gbit port deserves the asterisk their own terms put on it: "Unmetered service is a shared service and is sold with no guarantees of any kind," and the ToS reserves the right to suspend, cap or rate-limit traffic deemed detrimental to the network. If you need guaranteed commit, ask for a quoted port.

Colocation economics are where small providers can still beat the big ones: 1U at 1 amp / 208V with remote reboot and 100 Mbit for $50 per month, a 21U half rack with a 20A/208V circuit (about 3.3 KW usable) and 30 TB on a 1 Gbit port at $450, and a full 47U cabinet with 1 Gbit and a /25 at $1,100, per their published specials. Emergency hardware support for colo customers is free of charge via the helpdesk.

There is no free tier, no trial, and no money-back guarantee — payments are non-refundable per the current ToS, with a narrow exception for billing overpayments reported within three months. Budget accordingly.

Performance & infrastructure

We have not benchmarked tzulo machines, and we would not extrapolate from spec sheets this old anyway. What is verifiable: each server ships on an unmetered 1 Gbit port (10 Gbit available on custom builds), the network runs on Juniper and Cisco gear with diverse fiber paths, and the company states its links peak at no more than 50 percent utilization. For streaming and audio/video workloads they explicitly market a "streaming optimized network," and their long history on hosting forums includes media-hosting customers.

Uptime is the honest mess it usually is at this tier. The company page says the Chicago facility has maintained a 99.99 percent global uptime average — an average, not a commitment. Their ToS is unusually blunt: "Tzulo does not offer a Service Level Agreement. All Services... are provided on a best-effort basis, and no uptime, latency, or other performance commitment is made or implied." They even publish an explanation at tzulo.com/legal/sla for why they do not sell one. Whether you admire that or not, it is the opposite of the fine-print SLA theater bigger hosts practice: you know exactly what you are buying.

Redundancy exists at the facility layer — UPS and generator on all circuits, N+1 cooling in Chicago, A/B power options, multiple fiber paths — and they maintain SLAs with their own upstreams for power, cooling and transit. Those do not flow down to you. Power delivery supports 110V/208V in Chicago and adds three-phase in Los Angeles, which matters for high-density colo builds.

Where it falls short

Start with the two absences that define the offering: no backups and no SLA. The FAQ states it plainly — tzulo does not maintain any backups of customer servers, and all clients are responsible for their own data, restoration and migration. Combine that with unmanaged service (they will do reboots, reinstalls and connectivity restoration, nothing at the OS or application layer) and the operating model is clear: you bring the sysadmin.

The hardware list reads like a time capsule. Selling a Xeon L5630 in a current lineup means either you value legacy compatibility or you have not refreshed stock lately; for modern workloads — container hosts, CI runners, anything vector-heavy — the per-core performance will disappoint next to current Ryzen and Xeon Scalable offerings at similar prices.

DDoS protection is not documented anywhere we could find. Their terms explicitly reserve the right to rate-limit traffic during attacks, which is network self-defense, not customer protection. If your workload is attack-exposed, plan on your own mitigation in front of tzulo's port.

Scale and polish are limited. Cloud servers are sold out, the website is dated, support is ticket-only with response targets (0-4 hours for high-priority connectivity, 3-6 hours for OS issues, up to 24 for low priority — targets, not guarantees), and there is no self-serve automation surface comparable to the API-driven clouds. Payments are non-refundable.

Who should pick this

Our view: tzulo fits a narrow, real audience. US colocation customers who want cheap 1U and cabinet space with remote hands and no cross-connect fees in Chicago or downtown LA will find the pricing hard to beat. Media and streaming operators who need an unmetered 1 Gbit port on a network built for that traffic, and who already run their own infrastructure team, are the other natural fit. Small businesses around Chicago that want a named local provider with a phone-book-sized helpdesk rather than a ticket queue in three time zones also come to mind.

We would not point here for anything needing an SLA, bundled backups, managed service, or current-generation hardware. For those needs, the mainstream dedicated and cloud providers give you more accountability per dollar. And if you are simply hunting raw price-performance, run the numbers against our /benchmarks comparisons first — at $125 per month for a 2012-era quad-core, tzulo is competing on network and facility, not on compute value.

Key takeaways

• Operating since 2003: dedicated servers, colocation and IP transit from Chicago (Tier III, CoreSite) and downtown Los Angeles, on its own ASN 11878.
• Real pricing: dedicated from $125/mo (E3-1270v2, 16 GB, 2 TB, 1 Gbit unmetered) to $325/mo storage builds; colo from $50/mo per 1U; month-to-month, no setup fees.
• No SLA, by explicit published policy — best-effort service with an explanation on their site; no contractual uptime credits.
• No backups, unmanaged by default: you own your data and your OS. Emergency hardware help for colo is free.
• Hardware is several generations old and the unmetered port is shared with no bandwidth guarantees per their ToS; no documented DDoS mitigation.
• Best fit: US colo buyers and streaming workloads wanting unmetered transit and a small, reachable team — not SLA-driven production — see more in /category/dedicated-servers.

Sources

• https://www.tzulo.com/
• https://www.tzulo.com/company.html
• https://www.tzulo.com/dedicated.html
• https://www.tzulo.com/colocation.html
• https://www.tzulo.com/faq.html
• https://www.tzulo.com/tos.html
• https://www.tzulo.com/cloud.html
• https://www.whtop.com/review/tzulo.com