Wattage Hosting, LLC sells hosted capacity that still depends

Summary

Wattage Hosting, LLC has a real public infrastructure footprint. Its website advertises "Data Center as a Service," bare metal, GPU hosting, private cloud, colocation, dedicated servers, consulting, IPMI access, dedicated public IPv4, managed hardware and 10GbE-plus connectivity; its ARIN RDAP AS record lists AS25893 as active and registered to Wattage Hosting, LLC.
The current network evidence is stronger than a parked name but weaker than a fully disclosed multi-site platform. RIPEstat's AS overview marks AS25893 as announced, RIPEstat routing status reports 39 visible IPv4 prefixes, four visible IPv6 prefixes and full IPv4 and IPv6 RIS visibility on 2026-07-12, while Hurricane Electric's BGP Toolkit reports 43 originated or announced prefixes with 43 valid RPKI origins.
The ownership boundary is mixed. ARIN shows Wattage directly holds 23.159.68.0/24 and 2602:fa49::/36, but public routing on 2026-07-12 did not show the Wattage IPv4 /24 as currently announced, and many current IPv4 prefixes are registered to EGIHosting or Subnet Digital LLC.
The main failure path is dependence concentration. RIPEstat neighbour data, RIPEstat looking-glass paths for 50.118.191.0/24 and looking-glass paths for 2602:fa49::/36 all show AS36236, NetActuate, as the visible neighbour before Wattage. That can be a strong upstream platform, but it means a customer should verify transit diversity, rack location, power path, hardware replacement and data portability rather than treating the advertised capacity as self-proving.

The Wattage-Specific Signal Is The Gap Between The Sales Page And The Route Table

Wattage Hosting is not a generic small cloud entry whose only public clue is a name in a registry. Its own page, reachable at wattagehosting.com, presents a clear market promise: "Data Center as a Service" for private cloud and bare-metal infrastructure, built around custom deployments rather than a menu of commodity virtual machines. The page says Wattage specializes in dedicated bare-metal servers, GPU infrastructure, private cloud environments and custom hosting for startups, small and midsize businesses, developers, AI workloads and IT teams. It also lists colocation and hybrid solutions, infrastructure consulting, custom storage and compute, fully dedicated physical servers, IPMI or iDRAC-style remote management, managed hardware maintenance, dedicated public IPv4 and connectivity that starts at 10GbE on all servers.

That is a very particular claim. It is not merely "we host websites." Wattage is selling capacity that sounds close to the floor: physical servers, remote management controllers, drives, memory, high-speed ports, GPUs and customer-specific configurations. The page even uses language that acknowledges operational reality. In its service section it says things sometimes go bad, that people make mistakes, and that Wattage will tell customers when things go wrong. For a hosting buyer, that tone matters because it frames the product less as magic cloud abstraction and more as equipment and people under stress.

The route table gives the second half of the picture. AS25893 is active, visible and clearly connected to Wattage. ARIN's RDAP entry for AS25893 lists the AS name as WATTAGE-01, the registrant as Wattage Hosting, LLC and the status as active. The same record shows the company address as 1150 W Evelyn Ave, Sunnyvale, California, and lists administrative, abuse, NOC and technical contacts under wattagehosting.com email. RIPEstat's AS overview also reports the holder as "WATTAGE-01 - Wattage Hosting, LLC" and marks the AS as announced during the 2026-07-12 observation time.

So the basic operating case is real. Wattage has a public service page, a live AS, visible IPv4 and IPv6 routing, valid route-origin data for sampled prefixes, and the company's own site resolves inside a route currently announced by AS25893. A customer evaluating Wattage should not treat it as an empty shell. The better question is whether the visible public record supports the kind of resilience implied by bare-metal, private-cloud and colocation language.

That answer is more qualified. The public route table is live, but it is also concentrated. RIPEstat's routing-status view reports 39 IPv4 prefixes, 9,984 IPv4 addresses, four IPv6 prefixes and 4,096 IPv6 /48s, with full visibility among the relevant RIS peers. The same view reports only one observed neighbour. RIPEstat's ASN-neighbours endpoint identifies that neighbour as AS36236. RIPEstat's AS36236 overview names the holder as NetActuate, Inc.

That turns the article into a dependency analysis rather than a simple validation. Wattage can plausibly sell bare-metal and private-cloud capacity. Public routing confirms a live edge. But the practical resilience of that capacity depends on specific racks, ports, upstream sessions, address-resource permissions, power distribution, remote-hands access, spare server inventory and human response. Those details are not publicly documented on Wattage's reviewed pages. The result is a Medium network evidence grade: enough evidence to say the network is live, not enough to say the platform is independently resilient.

What The Company And Domain Records Establish

ARIN gives Wattage an unusually clean starting point for a small provider. The ARIN entity record for WHL-79 lists Wattage Hosting, LLC at the same Sunnyvale address as the AS record, with a registration date in May 2023 and a last-change date in January 2026. The AS25893 RDAP record shows registration and last-change entries in May 2023 for the AS itself and validated contact roles for administrative, abuse, NOC and technical functions. All four contact roles point to [email protected] and the same 408 telephone number that the Wattage website publishes.

Those records matter because a hosting customer needs more than a checkout page. A routed network needs a responsible party for abuse, route authorization, contact validation, outage escalation and registry maintenance. ARIN does not verify that a provider has spare servers, working backups or round-the-clock staff, but it does establish the named party that controls the network number and direct number resources. In Wattage's case, the ARIN trail is coherent with the website contact details.

The domain record adds a longer brand timeline. The RDAP record for wattagehosting.com shows the domain registered on 2022-03-16, last changed on 2026-03-17 and set to expire on 2027-03-16. It lists GoDaddy.com, LLC as registrar and DomainControl nameservers. The domain therefore predates the AS25893 registration, which suggests the hosting brand was not created only for the current route table. It also shows that Wattage's web presence depends on ordinary third-party domain services, not only on AS25893.

Public DNS makes that dependency visible. A direct lookup during review resolved wattagehosting.com to 50.118.191.11, and the public website responded with Apache headers. The address is inside 50.118.191.0/24, one of the prefixes currently originated by AS25893. However, ARIN's RDAP record for 50.118.191.11 shows the covering allocation as EGIHosting, not Wattage. Wattage's own public site therefore sits at the intersection of a Wattage-originated route and a partner-held IPv4 allocation. That is not bad; it is common in hosting. It does mean the public web proof and the address ownership proof are different things.

Mail also runs through third-party infrastructure. The Wattage domain's MX records resolve to Google's mail exchangers. That is a sensible operational choice for a small provider because customer contact, abuse handling and sales mail often need to survive problems in the provider's own server estate. It also means the communication path for support is separate from the customer-server path. If AS25893 has a routing problem, Gmail-hosted mail can still receive messages. If domain, mail or form delivery fails, servers may still be up but customers may lose a communication channel.

Wattage's privacy policy is also part of the operating picture, though it is not a resilience guarantee. It says Wattage Hosting, LLC operates wattagehosting.com, describes collection and use of personal information, mentions browser log data and Google Analytics, and states an effective date of 4/10/2020. It also says no internet transmission or storage method is 100 percent secure. That policy is basic rather than infrastructure-specific. It does not answer where customer server data is stored, whether support staff can access consoles, how long provisioning records are retained, or which suppliers process billing and support information.

The main company-record conclusion is therefore narrow but useful. Wattage is identifiable, reachable, and tied to its AS. The company is not publicly proving a large corporate apparatus, but it is also not hiding behind anonymous routing. The missing layer is operational disclosure: public documents do not name facilities, power design, rack providers, transit contracts, backup practice, customer data export, hardware replacement targets or a status archive.

The Visible Network Is Live, Dual-Stack And RPKI-Clean

The strongest evidence for Wattage is the current routing data. RIPEstat's announced-prefixes endpoint for AS25893 showed 43 current prefix entries during the July 12 observation window. RIPEstat routing status summarized that as 39 IPv4 prefixes and four IPv6 prefixes. Hurricane Electric's AS25893 page reported the same 39 IPv4 and four IPv6 counts, all 43 originated prefixes valid under RPKI, zero invalid origins and one observed IPv4 and IPv6 peer.

For a small hosting provider, that is meaningful. A one-prefix host can be real, but Wattage's current public table is broader than that. The IPv4 set includes slices of 45.38.0.0/15, 50.117.0.0/16, 50.118.0.0/16, 104.252.0.0/15 and 172.252.0.0/16. The IPv6 set includes 2602:fa49::/36 and more-specific /48s. The website itself, at 50.118.191.11, is reachable inside the same AS. These are not theoretical registrations sitting idle.

Route-origin security is also better than a casual low-end host might show. The RPKI validation result for 2602:fa49::/36 marks AS25893 as valid with a ROA that allows announcements down to /48. The RPKI result for 50.118.191.0/24 is also valid. The RPKI result for 172.252.24.0/24 is valid, as is the result for 45.39.243.0/24. That reduces one class of operational risk: route origin ambiguity.

RPKI is not uptime, though. It tells networks that AS25893 is authorized to originate a prefix under the relevant ROA. It does not tell customers whether a server is powered, whether the top-of-rack switch is redundant, whether a disk shelf has a spare, whether a null route will be lifted quickly, whether a DDoS filter is correctly tuned or whether a support engineer can replace memory at 3 a.m. The route table establishes reachability and route hygiene; it does not establish repair capacity.

The direct resource story is also uneven. ARIN shows Wattage directly holds 2602:fa49::/36, registered in August 2023, and 23.159.68.0/24, registered on 2026-06-01. Both are relevant. The IPv6 /36 is currently visible in routing and has a valid ROA. The IPv4 /24 has a valid RPKI result, but RIPEstat's prefix overview reported it not announced on 2026-07-12, with routing status showing a first and last observation on 2026-06-02 and zero current peers seeing it.

That distinction matters. Wattage has its own ARIN IPv4 space, but the current AS25893 production surface appears to rely heavily on IPv4 space registered to other holders. ARIN shows 50.118.128.0/17 under EGIHosting, 172.252.0.0/16 under Subnet Digital LLC, 45.38.0.0/15 under Subnet Digital LLC and 104.252.0.0/15 under Subnet Digital LLC. Hurricane Electric's prefix list labels many of the announced /24s as EGIHosting and some as Subnet Digital LLC or other descriptions. That is a legitimate model if contracts and ROAs are in place. It is also a dependency model.

If an IPv4 partner changes routing permission, abuse policy, geolocation metadata, allocation terms or customer acceptance rules, Wattage customers may feel the impact even if Wattage's own AS remains active. If a route is valid but a prefix develops reputation problems, customers may see mail rejection, fraud scoring, payment friction, CAPTCHA pressure or blocked access. If a customer assumes "dedicated public IPv4" means Wattage-owned address continuity, the current public record says that assumption needs clarification.

One Visible Upstream Is The Main Network Constraint

The most important number in the public route data is not the prefix count. It is the neighbour count. RIPEstat routing status reports one observed neighbour for AS25893. RIPEstat ASN-neighbours identifies that neighbour as AS36236. Looking-glass paths for 50.118.191.0/24 repeatedly end through AS36236 to AS25893, and looking-glass paths for 2602:fa49::/36 show the same last-hop pattern.

AS36236 is not an obscure network. ARIN's RDAP record for AS36236 lists NetActuate, Inc as the registrant and shows the AS has been registered since 2005. RIPEstat's AS overview marks it announced. PeeringDB's AS36236 profile lists NetActuate, says its general peering policy is open, and reports a large traffic and prefix footprint. NetActuate's own home page advertises global edge infrastructure, BGP anycast, cloud, virtual machines, bare metal and colocation across more than 45 locations with 24/7 expert support. Its colocation page says customers keep control of hardware while NetActuate provides Tier 1 network connectivity, redundant power, physical security and on-site support across those locations.

That can be a positive dependency. A small hosting provider using a larger network platform may get better reachability, DDoS handling, remote-hands access and global routing than it could build alone. It can also let the provider focus on customer hardware configuration and service while a specialist carrier handles the network edge. Wattage's advertised focus on custom bare metal and hardware support fits that model.

The problem is that the public record does not say exactly where the boundary sits. Wattage's website lists Santa Clara, California as the contact location. ARIN lists the company in Sunnyvale. NetActuate's PeeringDB facility list includes many locations, including facilities in San Jose, Fremont, Santa Clara and other markets, but that list describes NetActuate's footprint, not Wattage's deployed racks. Wattage's page does not name a data center, carrier hotel, colocation room, power configuration, cross-connect provider or remote-hands arrangement.

For customers, the distinction is practical. If Wattage owns servers in a NetActuate-connected facility, then hardware replacement depends on Wattage's spare parts, remote access and remote-hands agreement. If Wattage rents dedicated hardware from a supplier, then repair depends on the supplier's stock and ticket queue. If Wattage offers colocation, then customer equipment depends on facility access rules and the customer's own hardware readiness. If Wattage layers private cloud software over bare metal, then recovery depends on storage replication, cluster design and the ability to rebuild nodes when one physical server fails.

A single visible upstream does not mean there is no hidden resilience. It means public BGP evidence does not prove it. Wattage might have standby sessions, private connectivity, out-of-band access or contracts that are not publicly visible. The reviewed sources simply do not show them. A customer should ask whether AS25893 can fail over to a second transit provider, whether the same prefixes are filter-approved elsewhere, whether IPv6 and IPv4 behave the same way during an upstream fault, and whether customer IP assignments can be moved without manual intervention. Those are not abstract questions.

They are the difference between a brief route flap and a long support case.

The absence of a PeeringDB record for AS25893 reinforces the same reading. Many small providers do not maintain a PeeringDB profile, so absence does not prove the network is weak. But it removes one public place where a customer might otherwise see facility claims, peering policy, traffic scale, IX presence, looking-glass information or a NOC page. For Wattage, the public network story is therefore carried by ARIN, RIPEstat, Hurricane Electric and NetActuate path evidence rather than by Wattage's own network-disclosure page.

The Service Offer Is Hardware-Heavy, Which Raises A Different Reliability Test

Wattage's product language is unusual because it leans into physical infrastructure. The website does not emphasize a self-service VPS catalogue. It emphasizes dedicated bare-metal servers, GPU infrastructure, private cloud, colocation, hybrid solutions, infrastructure consulting and custom storage and compute. It says every system comes with SSD or NVMe for the operating system, storage is enterprise-level, the whole system is single tenant, and the team can install Linux distributions or duplicate ready-made images.

That kind of offer can be valuable for customers who have outgrown basic VPS plans but do not want to buy, ship and manage equipment alone. A startup running AI experiments may need GPUs before it can justify a large cloud contract. A developer team may need a single-tenant server for performance isolation. An SMB may want private cloud without hiring its own data-center operations staff. A customer with large storage needs may benefit from a bespoke array rather than paying hyperscale list prices for every terabyte.

The reliability test, however, is hardware-specific. In a pure virtual machine storefront, customers mainly ask about CPU contention, storage performance, snapshot restore and host failure handling. In Wattage's advertised model, they also need to ask about inventory. Which CPUs and GPUs are actually in stock? How quickly can a failed drive be replaced? Are DIMMs, PSUs, NICs, HBAs and fans stocked on site? Who has permission to touch the hardware? Are remote-management controllers on a separate management network? Can a customer obtain serial console access during an OS failure? Are firmware updates scheduled?

Is there a spare chassis for custom builds?

The Wattage page answers some of the customer-experience side. It says managed hardware includes maintenance such as hard-drive replacement and DIMM swaps. It says proof-of-concept work is available. It says IPMI, iDRAC, KVM or other remote management access comes with bare metal. It says configurations reflect the customer's needs. It also says each service includes a dedicated public IPv4 address. Those are positive service claims because they map to real hosted-infrastructure pain points.

The page does not answer the capacity side. It does not publish server counts, GPU models, regions, power density, rack count, cross-connect diversity, storage replication, backup retention, deployment time by configuration, hardware replacement objectives, remote-hands coverage, a maintenance calendar, a public status page or incident notes. The result is not a negative finding. It is an underwriting boundary. A customer can see what Wattage wants to sell, but cannot independently measure how much capacity is installed, how much is spare, or how quickly it can be repaired.

"Installed capacity" and "usable capacity" are different. A rack can contain many servers, but if power is capped, cooling is tight, a single switch aggregates too much load, addresses are constrained, or support staff are unavailable, the customer-visible capacity is smaller. A provider can advertise 10GbE ports, but usable throughput depends on oversubscription, upstream commit, DDoS filtering, traffic mix and acceptable-use rules. A server can have IPMI, but recovery still depends on management-network reachability and whether the customer has credentials during an incident.

For Wattage, the public evidence supports an active route table and a hardware-oriented service proposition. It does not support a claim of abundant spare capacity. That is why the proper buyer posture is verification, not dismissal. A low-profile bare-metal provider can be exactly right for a customer who wants hands-on help. It can also become fragile if a custom build depends on a single rack, single upstream, single address supplier, single switch or single engineer.

Address Locality, Data Locality And The US-Centric Footprint

The assignment classifies Wattage's service area as US, and the public evidence supports that as the safest label. The company address in ARIN is Sunnyvale, the website contact section says Santa Clara, and the phone number uses the Silicon Valley 408 area code. The public site embeds a map query for Santa Clara, California. The domain uses a US registrar, the mail exchangers are Google-hosted, and AS25893 is an ARIN number.

None of that proves where every server sits. A company can be based in California and host machines in Nevada, Arizona, Texas, Virginia or another market. A routed path through NetActuate may enter through many places, and NetActuate's public footprint is global. Public BGP path observation is not a facility locator. IP geolocation is not a contract. ARIN addresses are administrative, not rack coordinates. That is why the article treats data locality as a question customers must pin down before placing regulated or latency-sensitive workloads.

The direct resource mix also matters for locality. The Wattage IPv6 allocation is direct and registered to the company. The directly held Wattage IPv4 /24 is not currently visible in the reviewed RIPEstat snapshot. The visible IPv4 set uses many EGIHosting and Subnet Digital allocations. Those address holders have their own administrative records and possibly their own geolocation histories. A customer using Wattage for region-sensitive access should not assume that the company contact city, IP registry city and server city are the same.

Locality has several layers. Physical locality asks where the server is powered and cooled. Network locality asks where routes enter and leave the network. Data locality asks where customer data, backups, logs, support tickets, management snapshots and billing records are stored. Jurisdictional locality asks which legal entity contracts with the customer and which suppliers may touch the service. Address locality asks what geolocation and reputation systems think about the assigned IP. Wattage's public site answers the branding and contact layer; it does not fully answer the others.

The privacy policy illustrates the gap. It identifies Wattage Hosting, LLC as the operator of the site and talks about personal information, log data, Google Analytics, cookies and communications. It does not describe customer server data, backup location, support access to consoles, retention after cancellation, data export methods or deletion timing. That is common for a simple marketing-site policy, but customers placing sensitive workloads should request a service-specific agreement rather than relying on the website policy alone.

The data-locality issue is especially important for AI and GPU workloads, which Wattage explicitly names. AI training, fine-tuning, inference and data processing can involve proprietary data, customer records, model weights or regulated datasets. If a customer buys GPU infrastructure from a small provider, it should know whether data persists on local NVMe drives, network storage, snapshots or backups; whether drives are wiped after replacement; whether failed drives leave the facility; and whether remote hands can access storage devices. Public Wattage materials do not settle those questions.

The Practical Failure Paths

The most likely failure paths for a Wattage customer come from the same places that make the offer attractive: custom hardware, partner address space and a larger upstream platform.

The first path is rack or facility failure. If a server loses power, a top-of-rack switch fails, a remote-management network breaks or a facility limits access, the customer needs a repair path. Wattage's public page says managed hardware includes maintenance, but it does not name the facility, publish remote-hands coverage or state hardware replacement targets. A rack incident could therefore become a coordination exercise among Wattage, any facility operator, any remote-hands provider and the customer.

The second path is upstream failure. Public route observations point to NetActuate as the one visible neighbour. If the Wattage-NetActuate session fails, if filtering blocks one of the partner-held prefixes, if a DDoS response null-routes customer space, or if a larger carrier path into NetActuate changes, AS25893 customers may lose reachability even when their servers are healthy. NetActuate is a serious network, but the public Wattage record does not prove a second active upstream.

The third path is address-resource failure. Wattage promises dedicated public IPv4, yet most current visible IPv4 prefixes in AS25893 are registered to EGIHosting or Subnet Digital allocations rather than to Wattage. Valid ROAs make those routes cleaner, but contracts and address reputation still matter. If a partner-held block is withdrawn, remapped or damaged by abuse, customer workloads may need new addresses. That can break allowlists, DNS, TLS pinning, mail reputation, VPN tunnels and customer-facing APIs.

The fourth path is hardware stock. Custom bare metal and GPUs are not instantly replaceable unless the provider has spare equipment, supplier access or hot standby capacity. A virtual machine can sometimes move across a cluster; a custom physical server may require a matching chassis, disk controller, GPU, memory type or network card. Wattage's page suggests custom work and proof-of-concept support, but it does not publish inventory depth. Customers should therefore treat bespoke configurations as powerful but less portable than standardized nodes until Wattage states otherwise.

The fifth path is support and billing continuity. Wattage's site uses a Google form for contact intake, Google mail exchangers for email and DomainControl nameservers for the domain. These are normal outsourced services. They also mean a service incident can involve multiple channels: server route, management login, form submission, email, phone, billing and DNS. If a customer cannot reach the control or support path during an outage, the published "white-glove" promise becomes harder to use.

The sixth path is migration. Public materials do not publish a data export process, image-download method, IP portability commitment, snapshot transfer guide or cancellation migration window. For a single-tenant bare-metal customer, exit may require rebuilding onto another provider, changing IP addresses, moving data over the network and retesting every dependency. That can be expensive even when the monthly server price is reasonable. The safest customer practice is to design migration before the first production deployment.

Who Is Affected If Wattage Fails

The affected users are visible in Wattage's own marketing. The site names startups, SMBs, developers, AI workloads and IT teams. Those groups do not all carry the same risk. A developer using Wattage for a test system may tolerate a long repair window. A startup running customer-facing AI inference may not. An SMB with a single private-cloud environment can be more exposed than an enterprise that uses Wattage as one provider among many. A team using colocation may own the hardware but still depend on Wattage and its upstream path for reachability.

The failure impact also depends on whether Wattage is primary or auxiliary. If Wattage hosts a staging environment, a rendering queue, a GPU training node or a monitoring probe, the business impact may be contained. If Wattage hosts production databases, control panels, customer APIs, authentication components or storage, the same failure can become a company-wide outage. Public evidence does not show a Wattage-managed backup product or multi-site recovery service, so customers should not assume those protections exist by default.

Dedicated public IPv4 makes the impact stickier. If a customer receives addresses from a partner-held Wattage-originated /24 and later has to move, the customer may not be able to take those addresses with it. DNS changes can be quick, but allowlists, customer firewall rules, reverse DNS, mail reputation and VPN configurations can lag. That is why address portability matters more in bare-metal hosting than buyers often expect.

Remote management can reduce the pain, but only if it is reachable when needed. IPMI, iDRAC and KVM access are valuable because customers can reinstall operating systems, view consoles and recover from some boot problems. They also introduce their own security and availability demands. They should be isolated, patched, access-controlled and monitored. Wattage's page says such access is available; customers should confirm how it is protected and whether it remains reachable during a public-network incident.

The support promise is perhaps the hardest to measure from outside. Wattage says it offers responsive service, knowledgeable professionals and real people who care about customer success. That may be true and may be a major reason to buy from the company. But public sources do not show a status page, an incident history, published support objectives or an escalation chart. The customer has to test support before relying on it: ask presales technical questions, request a sample maintenance notice, ask how urgent tickets are handled, and confirm who can act during a rack or route failure.

What Would Upgrade The Evidence

Wattage does not need to publish every supplier contract to earn more trust. It would only need to publish enough operational detail for customers to separate marketing capacity from recoverable capacity. The first upgrade would be a network page that explains how AS25893 is operated: active upstreams, whether NetActuate is the only production transit path, whether a second path is configured but not visible, whether IPv4 and IPv6 have the same failover posture, and whether the directly held 23.159.68.0/24 will become a customer-serving block. A simple looking-glass, route-status page or BGP community guide would also help customers distinguish routine route changes from incidents.

The second upgrade would be a facility and hardware statement. Wattage can preserve security and commercial privacy while still naming the metro, facility class, remote-hands model, power redundancy assumption and hardware-replacement process. If the company sells custom GPU or storage builds, customers need to know whether parts are stocked, ordered on demand, swapped by Wattage staff, handled through a colocation provider or handled through a dedicated-server supplier. A statement that separates owned hardware, leased hardware and customer-owned colocated hardware would make the risk much easier to price.

The third upgrade would be support and status evidence. A public status page, even a minimal one, would show whether Wattage treats incidents as customer-visible operational facts. Maintenance notices would show how repair windows are scheduled. A support policy would say which channels are watched, how urgent cases are escalated, how abuse complaints are handled, and whether the same team can act on routing, hardware, billing and remote-console problems. The existing contact details are useful, but they do not show response depth.

The fourth upgrade would be portability language. Dedicated servers and private clouds are sticky by design. Customers need to know whether they can export images, download backups, retain access after cancellation, preserve IP addresses, request reverse DNS, move data over a private link, or schedule a migration window. Wattage's public materials do not need to promise seamless exit to every customer. They do need to define what is and is not portable before a high-dependency workload is installed.

None of these upgrades would change the fact that Wattage is a smaller provider than the large networks around it. They would change what can be verified from outside. Today, the public record proves a live route, a specific bare-metal thesis and a real administrative holder. It does not prove recovery depth. More disclosure would let customers decide whether Wattage is best used as a custom infrastructure partner, a secondary capacity provider, a test-and-build host, or a primary production platform.

The Best Reading Of The Evidence

Wattage Hosting should be read as a real, hardware-oriented US hosting provider with a visible route table and a clear public service thesis. It is not just a directory name. The evidence supports current AS25893 reachability, direct Wattage IPv6 resources, a newly allocated Wattage IPv4 /24, valid RPKI for sampled routes, a public website hosted inside the AS, contact details that match ARIN, and a service page specific to bare metal, private cloud, GPUs and colocation.

It should not be read as a fully proven resilient platform. The visible network has one observed neighbour. The current IPv4 production surface depends heavily on address blocks registered to EGIHosting or Subnet Digital. The directly held Wattage IPv4 /24 was not visible in the current RIPEstat snapshot. Wattage has no public PeeringDB entry. Reviewed public pages do not disclose facility locations, rack count, power redundancy, active upstream diversity, hardware inventory, status history, support objectives, backup practice or customer migration commitments.

Those two statements can coexist. Small infrastructure providers often sit between the customer and a larger network or data-center platform. That model can work well when the small provider is strong at hardware selection, customer support and custom deployment, while the larger upstream supplies reachability and facility services. It becomes risky when customers treat a sales page as if it were proof of spare capacity, tested recovery and independent network diversity.

The practical grade is Medium. Wattage has enough public evidence to support an article about an active hosted-capacity provider. The grade is capped because the route and address evidence shows concentrated dependencies, and the public service documentation leaves the most important failure-handling questions unanswered. Customers should verify multi-site capacity, active transit diversity, spare hardware, support escalation, billing continuity, backup scope and data portability before using Wattage as a sole production platform.

In one sentence: Wattage sells something materially more specific than commodity web hosting, but the thing it sells is still made of racks, ports, address permissions, upstream sessions, power and people. The public record proves the route is live; it does not prove how quickly the company can keep a customer's workload alive when one of those layers fails.