Summary

  • The identity chain is unusually specific for a small hosting provider. Brazilian company data records the active trade name CONECTHOST DATACENTER under CNPJ 39.652.848/0001-30, and LACNIC names the same CNPJ as registrant of AS272524 and an IPv6 /32 allocation.
  • The network is in operation, not merely registered. RIPEstat observed AS272524 announcing one IPv4 /24 and one IPv6 /47 with full visibility among its reporting route collectors on July 15, 2026. The IPv4 route had a valid route-origin authorisation; the IPv6 route returned unknown, not invalid.
  • ConectHost advertises KVM virtual servers, dedicated hardware, automated provisioning and management, daily backups, edge DDoS mitigation, four Brazilian data-centre locations and human support around the clock. These are provider claims; the public record does not establish occupancy at every named facility, tested failover, backup isolation, staffing depth or measured availability.
  • Accountability is the weak edge of an otherwise credible footprint. At review time, website links for terms pointed to missing pages and the advertised public status host returned an error. A production buyer should require the legal counterparty, service scope, support clocks, incident communications, data-copy map, recovery objectives and exit rights in an executed schedule rather than infer them from the brand or landing page.

One tax identifier connects the trade name to the network

Hosting brands can be deceptively difficult to attribute. The name on a website may differ from the legal counterparty, while the network may be operated by a founder, an upstream carrier or another company entirely. CONECTHOST DATACENTER has a better starting position because two independent public records meet on a precise identifier.

The Brazilian company-data record for CNPJ 39.652.848/0001-30 identifies the legal name as LUCAS MONTEIRO DE SOUZA INFORMATICA and the trade name as CONECTHOST DATACENTER. It marks the business active, gives an opening date of November 3, 2020 and records a Sao Paulo address. Its principal activity is a broad telecommunications classification, with computer repair, office-equipment rental and internet-access-room activities also listed. Casa dos Dados republishes Receita Federal data and is not a certified corporate extract, but its CNPJ and trade name are concrete enough to test elsewhere.

That test succeeds in LACNIC's record for AS272524. The regional internet registry names CONECTHOST DATACENTER as registrant, uses 39652848000130 as the organisation handle and records the autonomous system on August 13, 2021. It also publishes administrative and abuse contact information under the ConectHost domain. The corresponding IPv6 record for 2804:8284::/32 assigns the allocation to the same trade name and CNPJ.

The join establishes a legal and technical point of accountability. It does not establish beneficial ownership, current signing authority, headcount or a transfer of every asset shown on the sales site. The legal name is an individual entrepreneur's business name rather than the trade name customers see. An order should therefore state the CNPJ, legal name, trading name, billing entity and authorised signatory together. That small piece of contract hygiene prevents a customer from discovering during an incident that its purchase order, support account and network escalation refer to different identities.

The dates also place a useful limit on the story. The current company opened in 2020 and the ASN was registered in 2021. A website statement that implies a longer market history would need separate evidence about any predecessor operation. Neither a brand's age nor an individual's prior experience should be silently attributed to the present counterparty.

The catalogue spans automation, infrastructure and human intervention

ConectHost's current website presents three main infrastructure products: KVM virtual private servers, higher-bandwidth VPS plans and dedicated Intel Xeon servers. The published VPS plans include root access, solid-state storage and a 1 Gbps connection; dedicated offers advertise exclusive hardware and 10 Gbps connectivity. Prices and configurations make the catalogue testable at a point in time, but they do not reserve capacity or define the exact hardware a later order will receive.

The more interesting proposition is operational. ConectHost says payment by Pix can trigger provisioning in seconds, and that its platform can provision, suspend, manage and bill services programmatically. That can remove repetitive account work for a small platform team. It can also move failure into a new control plane: payment confirmation, account state, identity permissions, inventory, orchestration and billing all have to agree before a machine is usable.

Automation therefore needs evidence at the workflow boundary. A buyer should test account creation, API authentication, idempotent retries, failed payments, quota exhaustion, image selection, credential delivery, rebuilds, suspension and cancellation. It should record the time from a confirmed order to a reachable server, then repeat the exercise under a realistic burst. The relevant metric is not the site's displayed average provisioning time; it is the distribution for the product and location the customer will actually use, including failures that require a person to intervene.

The division of responsibility changes by product. Root access leaves guest operating-system hardening, patching, secrets, application monitoring and most recovery decisions with the customer unless an order says otherwise. A dedicated machine reduces compute contention, but it can still share power, switching, transit, remote-hands staff and administrative systems. Daily backups are advertised as a platform feature, yet the public page does not define retention, isolation, encryption-key control, restore granularity or recovery objectives. Those omissions are not proof that the controls are absent.

They are fields that cannot be safely inferred from the word "backup."

This is where a regional provider can either reduce work or merely relocate it. If ConectHost owns provisioning, physical access, network escalation and first-line recovery, the customer may avoid coordinating several suppliers. If those duties remain ambiguous, the customer inherits a supervision task across the same layers, only with less public telemetry than a hyperscale platform normally exposes.

AS272524 is visible, while its two address families tell different stories

Registration proves that an ASN exists; routing observations show whether it is doing work. The RIPEstat announced-prefix record for AS272524 showed two live announcements during the July 1-15 observation window: 87.76.204.0/24 and 2804:8284:1100::/47. Its routing-status view reported the IPv4 route to all 326 reporting IPv4 peers and the IPv6 route to all 322 reporting IPv6 peers at the July 15 observation point.

That is strong evidence of a dual-stack network with broad control-plane reach. It is not a latency test, packet-loss record, capacity measurement or proof that a customer's server was available. A route can remain visible while a hypervisor, storage system, customer VLAN or application has failed. Conversely, a collector snapshot does not reveal every private peering session, backup circuit or physical fibre path.

The resource provenance deserves close reading. LACNIC allocated the IPv6 /32 directly to the same CNPJ that holds AS272524. The visible IPv6 /47 sits inside that allocation. The public record for the IPv4 /24, by contrast, describes an assigned provider-aggregatable block under a larger parent range and names Conect Host LTDA in the contact chain. AS272524 was originating it at the observation point, but the record does not make it an IPv4 allocation held directly under the same CNPJ.

That distinction affects portability and incident handling. A customer receiving addresses from the /24 should ask who can update routing and reverse DNS, whether the space can be renumbered or withdrawn by an upstream, what notice applies to an address change, and how quickly filtering errors can be escalated. For IPv6, it should ask which sub-allocation is assigned, whether addressing survives a product migration and who maintains route authorisation.

Route-origin security is also mixed rather than absent. The captured RIPEstat validation for 87.76.204.0/24 returned valid, with a route-origin authorisation covering AS272524 at a maximum length of /24. The IPv6 validation for 2804:8284:1100::/47 returned unknown with no validating authorisation. Unknown does not mean hijacked or invalid. It means route-origin validation had no applicable authorisation with which to make a positive judgement.

IPinfo's AS272524 page provides a secondary topology view, classifying the network as hosting and listing Hurricane Electric and Golden Link as upstreams, alongside a larger set of visible peers. Such adjacency is useful corroboration for BGP activity. It does not prove commercial terms, purchased capacity, separate building entrances or physical diversity. A buyer relying on redundant connectivity should request a current carrier matrix, port and circuit capacities, route policy, shared-risk analysis and evidence from an actual failover exercise.

Four facility names are a claim to verify per order

The current site says ConectHost operates in four Brazilian locations: Equinix RJ2, Equinix RJ3, a Cirion site in Rio de Janeiro and a Cirion site in Cotia, Sao Paulo. It labels each carrier-neutral, Tier III, N+1 and continuously staffed. Elsewhere, the page describes multiple links, BGP routing, uninterruptible power, generators, controlled access and continuous monitoring.

Named facilities are more useful than a vague promise of "Brazilian infrastructure" because a customer can ask for a site-specific schedule. The reviewed public sources do not independently confirm them. No Equinix or Cirion record examined here confirms ConectHost's current rack, cage, power or cross-connect footprint at all four sites. Nor does the site say which location backs each advertised product, whether a workload spans locations, or whether backups leave the primary hazard domain.

The order should identify the actual facility and service-delivery region, not merely allow the provider to choose "RJ or SP." It should also distinguish a facility's design from the resilience of ConectHost's installed service. A building can have redundant utility and cooling systems while a tenant uses a single power feed, switch, carrier handoff or management path. Facility certification does not automatically attach to every rack configuration or operating procedure inside it.

For critical deployments, evidence should include the rack and power design, A/B feed availability, generator and fuel assumptions, cooling path, physical-access process, cross-connect inventory, maintenance windows and remote-hands authority. If geographic recovery is part of the purchase, the schedule should name the second site, the replication method, the recovery-point and recovery-time objectives, and the test that demonstrates applications can run there without relying on the failed location's credentials or control plane.

Brazilian hosting is not yet a complete data-locality map

ConectHost's products are explicitly marketed as running in Rio de Janeiro and Sao Paulo. The visible routes are Brazilian, and the company's infrastructure page names Brazilian facilities. For teams seeking local latency, billing in reais or a domestic operational contact, that is a meaningful proposition.

It does not answer every sovereignty question. Workload data can appear in primary disks, snapshots, backup repositories, monitoring platforms, ticket attachments, billing systems, email, fraud checks and vendor diagnostics. A server in Rio can be managed through an account system elsewhere. A backup can remain in Brazil but share credentials or a physical failure domain with production. Public route geolocation cannot settle any of those questions.

A buyer should request a copy-by-copy map covering location, operator, subprocessor, purpose, retention, encryption and deletion. It should ask where platform metadata and support material live, who can exercise privileged access, and whether remote support can move diagnostics across borders. The contract should also explain what happens during migration, abuse investigation, disaster recovery and service exit, when normal placement rules are most likely to be tested.

Locality has a labour dimension as well. A Portuguese-speaking technical team in the same market can be valuable when a customer needs a console action, route escalation or rapid commercial decision. That advantage depends on authority and depth, not merely proximity. The useful questions are how many escalation layers exist, which skills are staffed overnight, who may change BGP or hypervisor state, and how responsibility transfers when the first responder cannot restore service.

Public support promises need clocks and a working evidence surface

ConectHost promises human support 24 hours a day through tickets and WhatsApp, including holidays. Its site also advertises 99.9% uptime, while a status-style panel on the landing page displays 99.98%. Neither figure is accompanied there by a measurement window, service boundary, exclusion list or incident history. An older product page narrows the 99.9% cloud commitment to platform hardware and network, excluding failures caused by customer changes, but the public terms link associated with that page returned 404 during this review.

The advertised status host also returned a gateway error to the review request. Either failure could be temporary, and neither proves that customer support or hosted systems were down. Together they show why assurance cannot depend on a link merely existing. A status service is useful only if customers can reach it during a provider incident, understand which components are affected and retrieve enough history to compare reported availability with their own monitoring.

An executed service schedule should define severity, acknowledgement time, engineer engagement, workaround and restoration separately. It should state when each clock starts and stops, what information the customer must provide, which maintenance and external events are excluded, how updates are delivered, and what remedy follows a miss. For dedicated equipment, it should set remote-hands response and approval limits. For virtual infrastructure, it should identify who can move or rebuild a guest. For backups, it should require an authorised restore request and a measured completion result.

The customer should keep independent probes outside AS272524 and measure reachability, application success and latency from relevant user regions. It should retain tickets, incident updates, post-incident reports and restore-test results. Those records turn a friendly local support proposition into an operating history that can guide renewal, capacity planning and escalation.

Buy the evidence path, not the confidence of the name

CONECTHOST DATACENTER clears an important first threshold. The public trade name resolves to an active Brazilian counterparty and the same CNPJ appears in an authoritative autonomous-system record. AS272524 is visibly carrying dual-stack routes, with valid route-origin authorisation for its current IPv4 announcement. The product catalogue is specific enough to describe real customer workflows rather than a purely aspirational brand.

The remaining gaps are manageable, but they are not cosmetic. The IPv4 resource has a different provenance from the directly allocated IPv6 space. Facility presence and redundancy are self-described. Backup placement and recoverability are not publicly specified. Uptime figures lack a visible measurement history, while terms and status links did not provide a dependable public evidence surface at review time. Support is promised continuously without public clocks or staffing detail.

For a buyer, the next step is not to reject the provider or to accept the data-centre label at face value. It is to convert each claim into an artefact: CNPJ and signing authority in the order; facility and resource assignment in the service schedule; carrier paths and route controls in the network design; every data copy in a locality map; support and recovery in timed tests; and independent monitoring in operation. ConectHost has enough public substance to justify that diligence. Operating assurance begins when the answers survive a failed route, a lost host and a restore request, not when the sales page says the system is ready.