Summary
- ExaCloud Factory, S.L. has more than a marketing website: public RIPE, RIPEstat and PeeringDB records show AS212755, RIPE LIR status, visible IPv4 and IPv6 announcements, valid RPKI for observed prefixes, ESpanix connectivity and facility presence in Digital Realty Madrid sites.
- The economic question is not whether the company can describe reliability. It is whether customers will pay enough for local accountability, redundant Madrid infrastructure, support and routing diversity to offset fixed data-center, transit, hardware, staffing, compliance and channel costs.
- Public pricing suggests a ladder that starts low, with FTTO fiber from EUR 49.95 per month, VPS from EUR 29.95, colocation from EUR 69.95 and dedicated servers from EUR 99, while higher-value dedicated circuits, transit and tailored cloud services appear to require configuration or sales contact.
- The public record is thin on customer concentration, churn, gross margin, utilization, capex cycles and service-level payouts. That opacity should make the judgment conditional: ExaCloud looks operationally credible, but the profitability of its reliability promise remains unproven.
Reliability only pays when downtime has a buyer
The first test for ExaCloud Factory, S.L. is not technological sophistication. It is whether the buyer has a business problem expensive enough to justify paying for a higher-grade supplier. Spain has abundant low-cost connectivity, mass-market fiber, hyperscale cloud and large converged telecom groups. In that environment, a small or mid-sized business rarely pays a premium just because a provider says it is reliable.
It pays when a failed office link stops billing, when a hosted application misses orders, when a remote workforce loses access, when a local support call matters more than an anonymous ticket queue, or when a migration from owned hardware to outsourced infrastructure needs somebody accountable nearby.
That is the incentive ExaCloud is trying to monetize. Its own public pages sell a unified promise: data-center services, dedicated servers, IaaS, VPS, VDI, VoIP, cyber-security, professional fiber, dedicated circuits, IP transit, software work and partner resale. The language is not narrow access-ISP language. It is closer to a local technology infrastructure provider that wants to own the reliability layer for business customers. The sales point is continuity, support, flexible scaling and Madrid-hosted infrastructure rather than only cheap bandwidth.
This distinction matters because the cost side is fixed and unforgiving. A provider that sells reliability has to pay for facilities, power, cross-connects, routers, switching, spare equipment, upstream capacity, number-resource management, monitoring, support coverage, contract administration and security compliance before it knows whether each customer will actually use the premium. If the customer sees ExaCloud as interchangeable with low-cost fiber, commodity VPS or a global cloud console, the company is trapped in a price comparison it cannot easily win.
If the customer sees ExaCloud as a local accountable operator that can combine colocation, cloud, connectivity and support, the same infrastructure can carry a higher willingness to pay.
The public evidence points to both sides of that tension. ExaCloud publishes entry prices that look accessible: EUR 49.95 per month for a 600 Mbps professional fiber plan, EUR 59.95 for a 1 Gbps plan, EUR 29.95 for an 8 GB RAM VPS, EUR 69.95 for a 1U colocation offer and EUR 99 for a dedicated server plan. Those prices help customer acquisition and reduce friction, but they also show how hard it is to turn reliability into margin at the low end.
The stronger economic opportunity likely sits above the entry products: dedicated circuits, IP transit, higher-tier servers, rack commitments, managed cloud, cyber-security bundles and customers who buy several services from the same supplier.
The core question, therefore, is not whether ExaCloud can operate a technically real network. The evidence says it can. The question is whether it can convert that network into sticky accounts that value local support, redundancy and accountable service enough to pay for the infrastructure that creates those qualities.
The company boundary is Spanish, but the offer is broader than access
The legal identity is clear. ExaCloud's own legal notice and service contract identify ExaCloud Factory, S.L. as a Madrid-registered company with NIF B88643671, registered in the Madrid Mercantile Registry, with a social address at Miguel Yuste 18, 1.1, 28037 Madrid. The RIPE NCC member page and RIPE Database records match the same address family, Madrid location, country code ES and operational contact at [email protected]. The RIPE organisation entity, ORG-EFS18-RIPE, lists the entity as a Local Internet Registry, with the same registration number, a Spanish country record and creation in July 2020.
That gives the article a firm boundary: ExaCloud Factory, S.L. is the company being assessed, not an ASN, not a prefix, not a route object and not a generic cloud brand. The network records matter because they show how the company supports its offer. They are not the company itself. This distinction is especially important for a provider whose public footprint is richer in operational records than in financial disclosure.
The commercial boundary is broader than a simple regional ISP. ExaCloud's homepage and service navigation present five clusters. The data-center cluster includes colocation and dedicated servers. The cloud cluster includes cloud migration, IaaS, VPS, VDI and VoIP. The cyber-security cluster includes backups, Microsoft 365 backups, antivirus, email security and an ExaCloud Cyber Protect offer. The connectivity cluster includes professional fiber, dedicated circuits and IP transit. A services cluster covers software development, Microsoft 365 licensing and talent outsourcing.
The partner page adds a channel model for other technology companies.
This breadth can be an advantage if ExaCloud is selling continuity to small and mid-sized enterprises. A business that needs a fixed office line, a cloud server, backup, Microsoft 365 protection and emergency support may prefer one accountable provider. Bundling can reduce customer-management costs, cross-sell additional services and make churn harder. The company itself leans into that logic by promising one trusted provider for technology needs and by presenting flexible solutions adjusted to business stages.
But breadth also creates operating risk. A company that sells many services must either build capability across multiple disciplines or coordinate subcontractors and upstream suppliers without losing control of the customer experience. ExaCloud's contract explicitly allows it to subcontract part or all of the services it provides where it considers the third party sufficiently capable. That is normal in infrastructure markets, but it means the public promise of accountability depends on vendor management. The customer sees ExaCloud.
ExaCloud still has to manage data-center operators, upstream networks, hardware suppliers, software vendors and support workflows.
The RIPE member page lists serviced areas that include Andorra, Belgium, Spain, France, the United Kingdom, Italy, Monaco and Portugal. That should be read carefully. It is an official membership-directory service-area field, not proof of material revenue or direct operations in each country. The strongest public operating evidence remains Madrid-centered: the registered address, the Digital Realty Madrid facility references, the Spanish connectivity offer, the ESpanix exchange attachment and the Madrid contact numbers.
ExaCloud may serve wider customer requirements, but the visible reliability proposition is anchored in Spanish and Iberian infrastructure.
A small autonomous system turns marketing into an operating commitment
The most important public evidence is AS212755. The RIPE aut-num entity lists AS212755 with the AS name EXACLOUD, links it to ExaCloud Factory, S.L., and records import and export policies with several major upstream or transit ASNs, including AS1299, AS6939, AS3356, AS2914 and AS3257. It also records a block of peerings at ESpanix, including Sarenet, Packet Clearing House, Voztelecom, Neot, Adamo, Angola Cables and VTS-related entries. The entity was created in August 2020 and was last modified in November 2025.
These records do not prove service quality. They do show that ExaCloud has taken on a real operating role. Running an autonomous system requires routing policy, abuse handling, contact maintenance, route registration, upstream relationships, routing security discipline and incident response. It also turns reliability language into a continuing obligation. If AS212755 is unstable, badly filtered or poorly supported, the customer will feel it.
If it is well maintained, ExaCloud has a lever that pure resellers do not have: it can shape connectivity, peer locally, manage prefixes and provide business customers with a more accountable network path.
RIPEstat strengthens the point. Its AS overview lists the holder as EXACLOUD ExaCloud Factory, S.L. and reports the ASN as announced. Its routing-status data, queried on July 11, 2026, showed the first seen route as 91.233.197.0/24 from AS212755 in November 2020 and the last seen record on July 11, 2026. It reported full visibility from its relevant RIS peers for IPv4 and IPv6, four IPv4 prefixes totaling 1,024 addresses, one IPv6 /29 measured as 524,288 /48s, and 12 observed neighbours. RIPEstat's announced-prefixes call listed 84.54.50.0/24, 86.53.147.0/24, 91.233.197.0/24, 194.55.227.0/24 and 2a0a:b440::/29 over the observed window.
The prefix mix is instructive. RIPE Database inverse records link 91.233.197.0/24 and 2a0a:b440::/29 to ExaCloud's organisation record, while 84.54.50.0/24 and 194.55.227.0/24 appear as ExaCloud-assigned Spanish pools. The 86.53.147.0/24 record is described as Exacloud Factory SL but is marked SUB-ALLOCATED PA, maintained by AS3257-IPAM-MNT and named with a GTT reference. In economic terms, this is a hybrid footprint: some directly associated resources plus at least one supplier-linked sub-allocation. That is common for smaller operators.
It also reinforces that resource records are evidence of operating boundary and dependencies, not a complete identity statement.
RPKI validation is a positive signal. RIPEstat's RPKI validation reported valid status for the visible ExaCloud-originated IPv4 /24s and the IPv6 /29 when checked against AS212755. That does not guarantee clean operations or customer performance, but it shows a baseline of routing security hygiene. For customers buying reliability, valid RPKI is not a visible product feature. It is part of the hidden work that makes the promise less fragile.
PeeringDB adds a market-facing view. It lists ExaCloud for ASN 212755, with an open peering policy, IPv6 support, global scope, one internet exchange, three facilities, 20-50 Gbps of self-reported traffic and a mostly outbound traffic ratio. It identifies the IRR as RIPE::AS-EXACLOUD. The exchange record shows ESpanix Madrid Lower LAN, a 10 Gbps port, IPv4 and IPv6 exchange addresses, route-server peering and operational status. The facility records show Digital Realty Madrid MAD1-2, MAD3 and MAD4. PeeringDB is user-maintained, so it should be treated as useful operating market evidence rather than audited truth.
Still, its details align with ExaCloud's own infrastructure claims and the RIPE record.
The conclusion from the network record is simple: ExaCloud is not merely reselling a generic website under a Spanish brand. It operates a visible autonomous system with local exchange connectivity and Madrid data-center presence. The harder question is whether the network is large enough, differentiated enough and well-utilized enough to earn returns beyond commodity pricing.
Madrid colocation is the cost engine behind the promise
ExaCloud's reliability story rests heavily on Madrid data-center infrastructure. Its infrastructure page says its systems are hosted in premium Digital Realty facilities, referring to Interxion MAD2, MAD3 and MAD4 in one passage and to Interxion MAD1, MAD2 and MAD3 in another physical-location passage. PeeringDB records list ExaCloud facility presence in Digital Realty Madrid MAD1-2, MAD3 and MAD4. Digital Realty's own Madrid page lists four data centers: MAD1 on Calle Albasanz 71, MAD2 on Calle Albasanz 73, MAD3 on Calle Emilio Munoz 49 and MAD4 on Calle de Alfonso Gomez 4.
The Digital Realty page also reports 408.8 thousand square feet, or 38 thousand square meters, of total Madrid colocation space, more than 250 cloud and network service providers and more than 320 customers.
That location choice is credible. Digital Realty Madrid is a dense interconnection environment. ESpanix describes itself as the largest IP interconnection point in southern Europe, with more than 200 connected customers, more than 2 Tb/s of switched traffic, 260 Tb/s of installed capacity and six Spanish points of presence. ESpanix also states that connectivity services help networks gain resilience, lower IP transit costs and improve latency to the Iberian Internet. ExaCloud's own pages say it is connected to ESpanix and can establish peering sessions rapidly with hundreds of operators.
For a customer, the benefit is straightforward: Madrid-hosted infrastructure, local interconnection and a nearby support channel can reduce the distance between problem and resolution. For ExaCloud, the benefit is more complex. Colocation in a premium data-center ecosystem is both a sales asset and a fixed-cost burden. Space, power, cross-connects and remote hands are not free because a customer's server is idle. Redundant power, UPS feeds, cooling, fire protection, access controls and dark-fiber rings create reliability only by turning capital and recurring facility costs into a product.
ExaCloud's infrastructure page describes redundant power, UPS A and UPS B, generators with up to 72 hours of autonomy, N+1 or higher redundancy language, HVAC environmental controls, fire-detection and suppression systems, CCTV, access control and 24/7 security. Some of these are facility-level claims tied to the underlying data centers, not necessarily assets ExaCloud itself owns. That does not make them irrelevant. A colocation and managed-infrastructure provider sells the customer's access to those facility attributes.
The economic issue is how much of the value ExaCloud can capture after the data-center operator, power suppliers, cross-connect providers and hardware vendors have already been paid.
The colocation pricing illustrates the margin puzzle. ExaCloud advertises 1U colocation from EUR 69.95 per month, a half rack from EUR 999.95 and a private rack from EUR 1,250. The offers include Digital Realty Interxion Madrid, 24/7 access, dual UPS power feeds, meet-me-room interconnections and ExaCloud connectivity options at 1 Gbps, 10 Gbps or 100 Gbps. At the low end, EUR 69.95 is an accessible entry ticket, not a rich contribution to a fully staffed reliability business.
The likely margin depends on density, power usage, cross-sell to connectivity, support charges, contract length, remote-hands usage and whether the customer grows from a single unit into a larger commitment.
ExaCloud is therefore making a classic infrastructure-services bet: use a trusted Madrid data-center base to sell higher-margin services layered on top of space, power and bandwidth. The data center creates the platform, but the business is only attractive if customers buy enough managed value, connectivity and continuity to keep the platform utilized.
The product ladder mixes low entry points with bespoke reliability
ExaCloud's public product ladder is designed to let a business start small and move upward. That is commercially sensible. A customer may first buy a VPS, a fiber line or a 1U colocation slot, then later add backup, cyber-security, VoIP, a dedicated circuit, a larger server, a rack, IaaS or transit. The website's recurring phrase is configurability: plans can be adjusted to the customer, resources can be scaled, and support can advise which configuration fits.
The VPS page is the clearest low-friction offer. It lists three plans: Silver at EUR 29.95 per month with 8 GB guaranteed RAM, 4 vCores, 150 GB SSD, one dedicated IPv4, 1 Gbps connectivity and unlimited transfer; Gold at EUR 49.95 with 16 GB RAM, 8 vCores and 300 GB SSD; and Platinum at EUR 69.95 with 24 GB RAM, 12 vCores and 450 GB SSD. The VPS page says a 7-day money-back period applies. These are familiar commodity-cloud entry points. They may fill capacity and create a funnel, but they will not by themselves prove a defensible business if customers compare only CPU, memory and disk.
Dedicated servers sit one step higher. ExaCloud lists Silver from EUR 99 per month, Gold from EUR 190 and Platinum from EUR 290, with HP ProLiant generations from Gen 8 through Gen 10 or Gen 11, redundant power supplies, storage options, IPv4 and IPv6 and connectivity up to 100 Gbps. This is where reliability, hardware renewal and support become more visible. The customer is buying not only a machine but also placement, connectivity, Spanish support and replacement responsibility. ExaCloud's service contract says it manages hardware and replaces necessary hardware elements to keep the service functioning.
That is a real cost, especially if the provider promises responsiveness while maintaining aging and newer generations in the same product family.
Professional fiber is a different entry route. The FTTO page lists EUR 49.95 per month for 600 Mbps symmetric service, EUR 59.95 for 1 Gbps symmetric and EUR 69.95 for 1 Gbps with VoIP. It includes fixed IP, router, free installation and an option to connect locally between the office and ExaCloud data centers when paired with cloud services. The economics depend on access build, wholesale arrangements, customer density, installation costs and support tickets. Low headline price can win small-business attention, but the reliability margin likely appears when the fiber service is part of a wider ExaCloud relationship.
Dedicated circuits and IP transit are more strategic. The dedicated-circuit page emphasizes guaranteed bandwidth, SLAs, 24/7 monitoring, support and professional advice. The transit page targets operators and ISPs, offering global access, continuous connectivity, scalability, low latency, BGP or static routing, multiple BGP sessions on one port, IPv4 and IPv6, link aggregation and blackhole server support for DDoS protection. It lists port categories from Gigabit Ethernet at 100 Mbps to 1 Gbps, 10 Gigabit Ethernet at 500 Mbps to 10 Gbps and 100 Gigabit Ethernet at 10 Gbps to 100 Gbps. Those products fit the AS212755 evidence.
They also carry the most direct exposure to upstream cost and traffic growth.
The product ladder is therefore coherent. It offers cheap enough starting points to attract customers and higher-value products where reliability is more monetizable. The risk is that the public site makes price transparency more visible for entry products than for the high-value reliability products. Without public customer cases, utilization metrics or contract sizes, the outside analyst cannot know whether ExaCloud's revenue mix is weighted toward profitable bundles or toward low-end plans that mainly absorb support and infrastructure cost.
The unit economics hinge on utilization, support load and power discipline
The cost base behind ExaCloud's promise is not mysterious, even if the company's accounts are not public in the materials reviewed. A provider with colocation, cloud, dedicated servers and connectivity must cover facility commitments, power, cross-connects, routers, switches, server inventory, storage, software licences, monitoring, billing systems, staff, spare parts, abuse handling, security processes, RIPE membership and upstream connectivity. The public evidence lets us identify the categories; it does not disclose the unit margins.
Utilization is the first variable. A dedicated server, rack slot or virtualized host has attractive economics only when enough capacity is sold and support incidents are contained. Underutilized racks and servers turn into capital drag. Overutilized systems create performance incidents and churn. ExaCloud's IaaS page promises that customers can scale resources up or down in minutes and pay for what they need. That is attractive for customers, but it pushes the provider to manage pooled capacity carefully. If too much spare capacity is held for bursts, capital sits idle. If too little is held, the scaling promise becomes fragile.
Support load is the second variable. ExaCloud repeatedly markets 24/7 support or 24/365 support across colocation, dedicated servers, transit, IaaS and connectivity. Its contract states that technical and customer support is free through email, chat or phone and available 24/7. Support is a differentiator only if it is staffed, skilled and responsive. It is also expensive. A customer paying EUR 29.95 for a VPS or EUR 49.95 for fiber can erase a large share of monthly gross margin with a single long support interaction.
The business case improves if support is mostly preventive, if higher-value accounts use it for retention, and if entry customers self-serve.
Equipment refresh is the third variable. ExaCloud's dedicated-server page advertises HP ProLiant Gen 8, Gen 9 and Gen 10 or Gen 11 options, and says hardware is constantly renewed or advanced. A mixed hardware estate lets the company segment prices and reuse depreciated servers, but it also creates maintenance complexity. Older gear may support low-price tiers, yet older gear can draw more power per unit of compute, require spare parts and create customer perception risk if marketed against modern hyperscale clouds. Newer gear supports premium workloads but requires upfront capital and disciplined utilization.
Power and data-center cost are the fourth variable. ExaCloud's infrastructure page leans heavily on redundant power, UPS, generators, cooling and certified renewable energy. Those features improve resilience and enterprise credibility, but they are part of an industrial cost structure. If power prices, data-center charges or cross-connect fees rise faster than contract prices, ExaCloud needs either price escalation, efficiency gains or higher-value service bundling.
Its service contract gives the company the right to increase monthly prices with 45 days' notice, and for longer billing periods at the next period where notice requirements are met. That clause protects the provider, but price increases test customer loyalty.
RIPE costs are small but symbolic. The RIPE NCC Charging Scheme 2026 sets an annual contribution of EUR 1,800 per LIR account, with fees for certain independent resources and ASN assignments, and a EUR 1,000 sign-up fee for new members. For ExaCloud, that fee is not decisive. The larger point is that operating as a resource-holder has a recurring compliance and administration burden. Maintaining RIPE entities, abuse contacts, routing records and RPKI is part of the reliability product even when customers never see the invoice line.
The unit-economic judgment is therefore conditional. ExaCloud can make money on reliability if it maintains high utilization, sells multi-product accounts, keeps support efficient, refreshes equipment without overinvesting and uses Madrid interconnection to reduce upstream cost. It will struggle if the customer base is dominated by low-price standalone products, heavy support demand and price-sensitive churn.
Upstream diversity reduces outages but raises the fixed-cost bar
ExaCloud's routing evidence points to a strategy of diversity. The RIPE aut-num entity lists import and export relationships with Arelion's AS1299, Hurricane Electric's AS6939, Lumen's AS3356, NTT's AS2914 and GTT's AS3257, alongside ESpanix peers. RIPEstat's routing-consistency data shows some of these relationships visible in BGP and some recorded in whois but not seen in BGP at the query time. That gap is normal: routing policies can be inactive, backup, filtered, changed or observed differently by collectors. The important point is that ExaCloud's public routing policy is not single-homed.
For a reliability product, upstream diversity has clear benefits. It can reduce dependence on one transit supplier, improve route selection, create bargaining leverage, support maintenance windows and make outages less binary. ESpanix peering can also keep some Iberian traffic local, reduce latency and reduce paid transit consumption. ExaCloud's transit page explicitly sells low latency, route optimization, BGP configuration, multiple sessions and blackholing. Those claims fit the network record.
The tradeoff is cost and complexity. Multiple upstreams, exchange ports and facility presences require routers capable of handling full tables or chosen routes, engineering knowledge, monitoring, configuration management, route filtering, incident response and commercial management. A small operator can buy redundancy, but it must sell enough reliability to pay for redundancy. The customer must value the fact that ExaCloud can reach the internet through more than one path, not just see another Mbps price.
PeeringDB's self-reported 20-50 Gbps traffic range is a useful scale clue. It suggests a network beyond hobby size, but not a mass national carrier. The "mostly outbound" ratio and "Content" info type may reflect hosted workloads, cloud services, servers or content leaving ExaCloud-hosted systems. That would fit a company with dedicated servers, VPS, colocation and transit products.
It also suggests why traffic engineering matters: outbound-heavy networks need cost-effective upstream and peering strategy because customer revenue may not rise in proportion to bandwidth use if plans include unlimited transfer or flat commitments.
The announced IPv4 footprint is modest. RIPEstat saw four IPv4 /24s, or 1,024 IPv4 addresses, plus a large IPv6 allocation. PeeringDB's profile self-reports 12 IPv4 prefixes, which may reflect route policy, more-specific planning, historical or user-maintained data, and should not be read as a perfect count. Either way, public evidence does not show a huge access network. It shows a focused operator with enough routing footprint to serve hosted, enterprise and transit customers.
That focused footprint can be an advantage if ExaCloud remains disciplined. A regional infrastructure provider does not need to become a national incumbent. It needs enough scale in its chosen facilities and customer base to amortize redundancy. The danger is strategic overreach: selling every service, every country and every customer type while carrying the cost of a reliability network designed for higher-value accounts.
Customer acquisition appears partner-led and price-sensitive
ExaCloud's partner program is one of the most revealing commercial pages. It offers two partner modes. In the "Comercial" model, ExaCloud bills the end customer and pays the partner up to 40% in recurring commissions on the retail price. In the "Up to you" model, the partner can sell the solutions under its own brand, bill its own customers and receive up to a 40% discount on the retail price. The page promises training, marketing resources, technical support and commercial support.
This is a rational channel strategy for a small infrastructure provider. IT consultancies, local integrators, software shops and managed-service providers often control the customer relationship for SMEs. If ExaCloud can become their back-end infrastructure supplier, it can gain distribution without building a large direct sales force. White-label infrastructure also helps ExaCloud fill capacity while the partner handles account ownership.
But channel economics cut both ways. A recurring commission or discount of up to 40% is a large concession unless the retail price is set with channel margin in mind. If a product already has thin gross margin because it includes data-center cost, bandwidth, hardware and support, giving away up to 40% of retail can leave little room for the operator. The model works best when partner customers buy sticky, multi-service bundles with low support friction and when the partner handles enough first-line support to reduce ExaCloud's cost.
It works poorly if partners bring high-maintenance customers at discounted prices while ExaCloud carries the infrastructure burden.
The public site also uses 7-day money-back language across several products. That reduces purchase risk and suggests confidence. It also attracts customers who may test commodity services without long commitment. ExaCloud's contract narrows the refund condition: the customer can request a refund within seven days if unsatisfied, provided use is reasonable and it is the first such request. After cancellation, the contract generally does not promise refunds for amounts already paid. This structure is commercially sensible. It supports acquisition but protects against ongoing churn.
The stronger acquisition story is local accountability. ExaCloud publishes Spanish contact numbers, Madrid address information, support emails and a customer portal. For customers frustrated by large-provider queues, that matters. The company says customers can visit the technical infrastructure and receive advice from specialists. If true in practice, that is a human moat: customers do not pay only for uptime statistics but for the confidence that someone local will answer when the system fails.
The weakness is the absence of public customer evidence. The materials reviewed did not show named customer case studies, revenue concentration, renewal rates, net revenue retention, average contract value or industry vertical mix. For an infrastructure provider, those facts would materially change the analysis. A few demanding customers can validate the reliability proposition, but they can also create concentration risk. Many small customers can diversify revenue, but they can overwhelm support if products are sold too cheaply.
The competitive substitute is not another small ISP; it is cheaper adequacy
ExaCloud's realistic competitors are not only providers that look like ExaCloud. They include every alternative a business can choose instead of paying for local reliability. A small office can buy mass-market fiber from a large operator. A startup can run workloads on a hyperscale cloud. A software agency can resell a global VPS provider. A larger enterprise can buy directly from Digital Realty, a carrier, a systems integrator or a global managed-service provider. In many cases, the substitute is not better reliability. It is adequate reliability at lower perceived cost.
Spain is a tough market for that reason. Public reporting citing CNMC data says Spanish telecom and audiovisual sector revenue grew in 2025, with retail and wholesale telecom scale concentrated in large operators, fixed broadband lines near 19.9 million and fiber representing about 91% of fixed broadband. Another report on CNMC quarterly data said the main operators' revenue share was declining but still substantial, while next-generation access and fiber lines remained very high.
The exact measures differ by report and period, but the direction is clear: Spain is a mature, fiber-rich and highly competitive market where cheap connectivity is widely understood by customers.
For ExaCloud, that means the winning pitch cannot be "we also have fiber" or "we also have cloud." It must be "we carry the operational downside for you." The buyer has to believe ExaCloud can integrate the office connection, hosted workloads, backup, security, transit and support better than separate cheap suppliers. The company's product breadth makes that possible. Its local data-center and AS evidence makes it credible. But the customer must experience fewer failures, faster repairs or simpler accountability.
Hyperscale cloud competition is different. Global cloud providers win on breadth, automation, ecosystems and perceived resilience. ExaCloud cannot out-platform them. It can compete where customers want Madrid-hosted infrastructure, predictable support, simpler contracts, local language, physical colocation, fixed IPs, direct connectivity and hybrid arrangements. The IaaS page's language about customer control over data and local infrastructure points toward that niche. It is not a global-cloud replacement. It is a local continuity and sovereignty-adjacent offer for customers that value proximity.
Large telecom operators are another substitute. They can bundle mobile, fixed, TV, cloud, security and enterprise services, often with procurement reach and national field operations. ExaCloud's defense is focus. A smaller provider can be more flexible, faster to configure, more transparent and more personal. The dedicated-circuit and transit pages emphasize personalized advice, monitoring and BGP flexibility. That is the correct strategic comparison: not scale against scale, but accountability against bureaucracy.
The risk is that customers often say they value reliability but buy on price until an outage occurs. ExaCloud's public prices show it understands price sensitivity. The question is whether it can avoid teaching customers to view reliability as a cheap add-on rather than a paid service.
Compliance and service terms shift some risk, but not the burden of trust
Infrastructure services operate under legal and compliance pressure. ExaCloud's contract refers to Spanish law, data-protection duties, public-administration communication where necessary, customer responsibilities and Madrid courts. The legal notice identifies Spanish law and the LSSI framework for online services. The company security policy sets objectives for confidentiality, integrity and availability, legal compliance, continuity planning, employee awareness, incident management and periodic security review.
It says ExaCloud operates and maintains an information-security management system aligned with the policy, conducts risk analysis and reviews security policy and risk assessment regularly, generally annually through internal audit or management review.
Those statements matter because reliability today is not only uptime. It includes incident handling, data-protection process, abuse response, continuity planning and customer confidence that the provider knows its legal environment. EU and Spanish rules raise the bar. The NIS2 Directive expanded cybersecurity obligations across sectors and emphasized risk-management and incident-reporting measures for economically significant entities in covered sectors. GDPR governs personal-data handling. Spain's General Telecommunications Law sets the national electronic-communications framework.
These frameworks do not automatically tell us whether every ExaCloud service falls into every obligation. They do show that a provider selling connectivity, cloud, hosting and security in Spain operates in a compliance-heavy environment.
ExaCloud's contract also manages risk in ways customers should read carefully. It allows service interruptions for improvements, maintenance and restructuring, with a commitment to communicate as far in advance as possible and minimize impact. It limits responsibility for customer software and customer-hosted content. It says ExaCloud's responsibility for limited public evidence service is limited to proportional return of fees for the period when the customer did not receive satisfactory service. It reserves rights to block IPs or services that may compromise network security or integrity.
It can suspend or cancel abusive or unlawful use. It can subcontract services. These are normal protections, but they show how reliability risk is shared rather than eliminated.
The backup clause is particularly useful. For IaaS, Cloud VPS and Cloud VDI, the contract says ExaCloud includes a basic weekly backup service, but describes it as limited and recommends additional backup systems. This is good risk language. It also reveals the commercial opportunity: the basic included backup protects the offer, while paid backup and cyber-security services can monetize a customer's need for stronger continuity.
For customers, the real question is not whether ExaCloud has legal clauses. It is whether the company has operational discipline behind them. A security policy that mentions continuity and incident management is a positive signal. A valid RPKI footprint is another. A public NOC contact is another. But trust is earned through performance under stress, not through document headings. The outside evidence can support credibility, not certify it.
Unofficial signals support a real network, not a proven moat
Unofficial and user-maintained market signals should be used carefully. PeeringDB, routing collectors and BGP databases are not audited financial statements. They are useful because operators maintain or reveal interconnection data, and because public routing systems observe what networks announce. For ExaCloud, these signals align well with the official RIPE record and the company's own site. That alignment reduces identity risk.
PeeringDB's record shows a network named ExaCloud, ASN 212755, created in December 2020 and updated in September 2025, with public contact records for NOC, abuse, technical and sales roles. The exchange and facility entries align with ESpanix and Digital Realty Madrid. RIPEstat sees the ASN announced. RIPE Database route objects exist for the main observed prefixes. RPKI validation is valid for the visible prefixes. These signals do not prove customer satisfaction, but they make it hard to dismiss ExaCloud as merely a shell around borrowed infrastructure.
At the same time, the signals expose scale limits. Four visible IPv4 /24s and one IPv6 /29 are enough for a serious local cloud and connectivity operator, but they are not evidence of national access scale. PeeringDB's 20-50 Gbps traffic range is meaningful but modest in the context of Spain's largest operators, exchanges and data-center ecosystems. The company appears to be a focused infrastructure and connectivity provider, not an incumbent-style network.
The website itself has some inconsistencies that should temper confidence. The infrastructure page refers to different sets of Digital Realty Madrid facilities in different passages. Some site copy claims more than 18 years of experience, while the RIPE organisation record starts in 2020 and the legal company record reviewed identifies ExaCloud Factory, S.L. as the current legal counterparty. That may reflect predecessor operations, brand history, team experience or site-copy carryover; public evidence reviewed here does not resolve it.
The safest reading is to treat the legal and network records as hard evidence and the experience claim as marketing unless independently confirmed.
Another market signal is sparse customer visibility. There are social links and a partner program, but the public pages reviewed did not show a large set of named enterprise references. That absence is not unusual for SME infrastructure providers. Many customers do not want their hosting, network or security vendors publicized. But for investment or strategic judgment, it leaves a major gap. Reliability businesses are proven by renewal behavior, outage history, support quality and willingness to pay after the first contract period. Those are not visible in route records.
The appropriate conclusion from unofficial signals is therefore balanced. They support operational credibility and a coherent Madrid network story. They do not establish a moat, pricing power or sustainable value creation.
What would change the judgment
The bullish case for ExaCloud is clear. The company has a Spanish legal identity, a public RIPE LIR record, AS212755, valid visible RPKI, ESpanix peering, Digital Realty Madrid facility presence, a broad services ladder, local support language, configurable products and a partner channel. If ExaCloud can turn those assets into multi-product business accounts, it can sell something more valuable than raw bandwidth or commodity compute: a local reliability relationship. That relationship could support better retention, cross-sell and higher gross profit per customer than standalone low-end VPS or fiber plans.
The bearish case is also clear. The public price ladder starts low, channel discounts can be large, hardware and support obligations are real, data-center cost is fixed, upstream diversity is not free, and public evidence is thin on customers and financial performance. A provider can look technically credible while still failing to capture enough margin. Reliability is expensive to produce and easy for customers to undervalue until something breaks.
Several facts would materially change the judgment. First, revenue mix: the share of revenue from recurring colocation, dedicated circuits, transit, managed cloud, cyber-security and bundled SME accounts versus low-end VPS and standalone fiber would show whether ExaCloud is monetizing reliability or mainly selling access. Second, gross margin by product: bandwidth-heavy transit, power-heavy colocation, support-heavy VPS and high-touch managed services have different economics. Third, utilization: rack occupancy, server utilization, router capacity and power commitments would show whether fixed infrastructure is productive.
Fourth, customer concentration: a few large accounts could validate quality but create renewal risk. Fifth, churn and renewal: reliability businesses should show stickiness after incidents and price changes. Sixth, support metrics: response time, ticket volume per customer and escalation costs would reveal whether 24/7 support is a moat or a margin drain.
Other evidence would matter too. Public case studies with identifiable customers, audited service-level performance, outage history, DDoS handling examples, contract lengths, cross-sell rates, partner productivity, capex plans, facility commitments and upstream cost trends would all sharpen the view. So would confirmation of the company's operating history before the 2020 RIPE record, if the "more than 18 years" claim reflects a predecessor business or team history rather than brand copy.
For now, the defensible judgment is conditional but not dismissive. ExaCloud Factory, S.L. has the public signals of a real Madrid-based infrastructure and connectivity operator. It owns enough network responsibility to make reliability a credible product, and its service ladder gives customers reasons to consolidate around it. The unknown is whether customers pay enough for that responsibility. If they do, ExaCloud's local accountability and routing control can be a business. If they do not, the company is left carrying the cost of reliability in a market that often rewards cheap adequacy.