An island cloud, not a hyperscale rival Most importantly, cloud.mu is not trying to be “the Mauritian AWS”. Public evidence points to a more realistic and economically plausible business: a locally assembled cloud and hosting platform selling web hosting, VPS, backups, domains, dedicated servers, reseller services, and a thin managed convenience layer to customers who care about three things hyperscale cloud is structurally bad at in small island markets: jurisdictional familiarity, local support, and predictable latency for Mauritian users. Its pitch is explicit: infrastructure “hosted locally in Mauritius”, fast local access, and support from a real local team. Its pricing table is equally explicit: low-cost shared hosting, very cheap Linux VPS, Acronis-based backups, reseller packages, and dedicated servers priced noticeably higher in Mauritius compared to South Africa. That is not a hyperscale profile. It is utility infrastructure for small markets with retail packaging.
That distinction matters because the core commercial question is not whether cloud.mu can beat hyperscalers in cloud computing. It cannot, and it is not set up to try. The real question is whether a Mauritian platform can reliably sell enough “island advantages” to offset the island’s structural disadvantages. Those advantages are real but limited: local data residence under Mauritian law, lower in-country latency, a certain trust premium in a jurisdiction that markets itself as orderly and compliant, and easier handholding for local SMEs and institutions that do not want to buy infrastructure from a distant self-service console. The disadvantages are also real and, over the long term, more severe: a population of roughly 1.25 million, little large-enterprise depth, expensive international bandwidth, exposure to submarine cable geography, high fixed operating costs spread over a small domestic base, and public procurement patterns that tend to pool at telecom or state scale rather than boutique hosting scale.
So the “Island Cloud Problem” is not a technical mystery. It is an economic problem. Local cloud in Mauritius can be commercially useful. It can even be strategically valuable. But it is valuable the way a ferry service is on an island: because it solves a specific local need better than distant alternatives, not because it becomes the ocean itself. The challenge for cloud.mu is to collect enough recurring revenue from those specific local needs before the very things that make it attractive —locality, intimacy, manual support, low-latency proximity— are eroded by cheaper offshore tools, better connectivity, or customers deciding that “good enough” offshore is okay after all.
Based on publicly available evidence, cloud.mu appears commercially real, technically non-trivial, and strategically constrained. It has its own ASN, its own prefixes, public looking-glass infrastructure in Mauritius and Johannesburg, visible peering at the Mauritius Internet Exchange and NAPAfrica Johannesburg, and a locally framed retail catalogue. That is enough to distinguish it from a simple white-label web hosting reseller. But the same public evidence also shows that its “sovereignty” is weaker than the marketing language implies. The platform sits on external transit, external data centre facilities, external control-panel and backup software, and external submarine routes. Even when the servers are local, some of the risk is not. That is the island-cloud contradiction in one sentence.
Who cloud.mu appears to be Corporate identity is more telling than the homepage. cloud.mu’s public contact page names Hosted Ltd as the registered company, with an office at The Dot Building in Moka. Yet its privacy policy still identifies the operator as DataKeepers Ltd, registration number C18158096, and its refund policy page still uses the DataKeepers name and an older address at The Core Building in Ebène. VAT records from the Mauritius Revenue Authority help reconcile the discrepancy: the same registration number, C18158096, appears under Hosted Ltd. The simplest reading is that DataKeepers Ltd has been renamed Hosted Ltd, while some legal and billing surfaces still carry the earlier brand. That is not unusual for a small infrastructure firm, but it is economically meaningful: it suggests a real operating business that has changed skin without fully rebuilding its compliance surfaces—something typical of companies that grow incrementally rather than through a highly managed corporate process.
The network side reinforces that the company is more than a storefront. BGP records show AS328699, registered to DataKeepers Ltd, active since July 2020, with five transit providers and multiple originated IPv4 and IPv6 prefixes. Those same records list the website as cloud.mu, the ASN name as cloud-mu and a Mauritian address in Ebène. The organization also publishes a public looking glass and appears on PeeringDB under the name cloud.mu with the address at The Core Building in Ebène. In other words, whatever the legal name change, the network personality was built as DataKeepers and is being sold to the market as cloud.mu. That gives the company a stronger asset than a domain and a support inbox: it controls IP space and routing policy under its own ASN. In the hosting business, that is a meaningful threshold.
There is also an intriguing South African shadow in the public record. AS328699 lists AS328170 DataKeepers (Pty) Ltd as one of its upstream providers, and the AFRINIC WHOIS block for the Mauritian network references David Venter as an administrative contact. LinkedIn shows a Cape Town-based DataKeepers company focused on disaster recovery, cloud backup, virtual servers, and DRaaS, with cloud.mu shown among similar pages. That does not prove ownership. It does, however, strongly suggest some operational or corporate kinship between the Mauritian cloud.mu platform and a South African DataKeepers firm. Economically, that matters because it changes how to interpret cloud.mu. It may not be a purely domestic startup born from Mauritian demand. Instead, it may be a Mauritian market edge of a broader South African infrastructure and backup stack. The public record does not fully resolve the cap table, but it makes the “independent local shop” story look incomplete.
That incompleteness is visible elsewhere. cloud.mu’s website claims it “owns and manages” all its infrastructure, but wording across the broader site is carefully phrased as “securely hosted in local Mauritius data centres,” not “in our own data centre.” Public peering records show cloud.mu present at MIXP via Rogers Capital Data Center, and the company’s PeeringDB and older legal addresses point to office buildings in Ebène rather than a company-owned telecoms facility. The available evidence therefore supports a narrower conclusion than the homepage rhetoric: cloud.mu appears to control its network resources and service stack, while hosting inside third-party Mauritian facilities. That is still significant. It just means the scarce asset is not a building permit or a Tier IV shell. It is a routable, peering-enabled, supportable local hosting platform assembled inside someone else’s physical shell.
That distinction is central to the economics. A firm that owns local IP space, manages routing, maintains peering sessions, and runs public looking-glass infrastructure has more control than a reseller. But a firm that hosts in another operator’s data centre rather than owning the facility also has lower fixed capital burdens and a lower physical moat. That is not a mistake. In Mauritius, it may be the only sensible model at cloud.mu’s apparent scale. The market is too small for “own the whole stack” to be attractive, unless you are a telco, a state facility, or a diversified infrastructure group. cloud.mu’s apparent structure therefore looks like an adaptation to Mauritian economics, not a failure of ambition.
What it sells and where the margin sits The catalogue tells you what kind of business this is. cloud.mu sells shared web hosting, WordPress hosting, Linux VPS, Windows VPS, cloud backup, reseller hosting, SSL certificates, domain registration, Microsoft licences, and bare-metal dedicated servers. The entry points are low: shared hosting from Rs289/month, Linux VPS from Rs299/month and cloud backup from Rs259/month. Windows VPS starts higher, with Windows licensing explicitly built into the price, and dedicated servers are a big-jump product. This is a classic tiered recurring-revenue design: bring in very small customers at low cost, upsell them to reseller or VPS plans, and retain a few heavier users on dedicated servers and backup storage. Economically, it is closer to an ISP-adjacent hosting business than an enterprise cloud platform.
The gross-margin profile of these products can be inferred, though not directly observed. Shared hosting is normally the most profitable layer in a business like this because a single machine can support many low-utilisation customers, and the value is a convenience bundle—mail accounts, DNS, SSL, cPanel, backups, support—rather than raw compute. Domains are thinner: cloud.mu sells.MU domains and others, but the list of registrars accredited by MU-NIC does not include cloud.mu, while it does include Register MU and other registrars. cloud.mu’s own domain pricing for.mu and related local namespaces is visible on its cart pages, strongly suggesting it acts as a reseller or registrars’ channel rather than as an official registrar facing the registry. That means domains are likely lead-generation and account-retention tools, not a rich standalone profit stream.
The price relationship is revealing. Register.mu lists.mu registration at $64 for one year and transfer at $114; cloud.mu’s own pricing page shows.mu registration at Rs2,800, renewal at Rs2,900 and transfer at Rs5,600. Those numbers are close enough to suggest that cloud.mu is not “building margin” on domains in a significant way. Domain services here appear more as an add-on: if the customer buys the domain, it is easier to keep it inside the hosting account, easier to sell additional email or DNS-support services, and harder for them to leave. In small hosting markets, that matters more than the domain’s gross margin itself.
VPS and dedicated servers are a different story. They look like the company’s attempt to capture higher average revenue per account without leaving the SME self-service segment. Linux VPS pricing is aggressive enough to win local developers, agencies, and small-business systems that need predictability more than hyperscale elasticity. Windows VPS is priced to make the licence visible rather than hide it, which is typical of providers selling to small businesses that still run Windows-dependent applications but don’t want to navigate Microsoft licensing complexity themselves. There is a Microsoft licensing section for SQL Server Web Edition sold on top of hosted Windows servers. Economically, that is useful because it turns a commodity virtual machine into a higher-margin compliance and convenience bundle. You are not just selling RAM and SSD; you are selling “we handle the annoying Microsoft part”.
Cloud backup has its own logic. cloud.mu markets backup explicitly under the Acronis brand, “unlimited devices” and storage tiers up to 20 TB. In the managed-services economy, backup is attractive because it is sticky, contractual, and emotionally defensive: customers buy it to avoid catastrophe, not to seek experimentation. That typically lowers churn. But the margin is not pure. Acronis itself positions its cloud product as a platform for MSPs, and cloud.mu is visibly packaging that kind of service. So backup margin here likely depends on how effectively cloud.mu arbitrages local trust and support against upstream software and storage costs. If utilisation is low or support is labour-intensive, backup can become a deceptively expensive promise to keep. If managed well, it becomes annuity-like recurring revenue with strong lock-in.
The dedicated-server pages expose the island economics most bluntly. cloud.mu sells the same dedicated server families in Mauritius and South Africa, and the Mauritian location is materially more expensive. For the publicly shown entry-level dedicated product, the Mauritius price is Rs13,499/month versus Rs7,499/month in South Africa. That gap is the article’s thesis in numerical form. The company itself is telling the market that “locality” is a premium product, whether because local infrastructure costs are higher, local capacity is scarcer, or both. If a customer just wants raw compute, South Africa is much cheaper. If the customer pays the Mauritius premium, they are buying something beyond raw compute—typically latency for Mauritian users, local legal location, or local convenience. cloud.mu’s own pricing therefore confirms that its real product is not cloud capacity; it is the locality wrapper around cloud capacity.
That is why the business model looks commercially clever, though constrained. The company does not appear to assume it can monetise “enterprise digital transformation” at the scale of Mauritius Telecom. Instead, it monetises proximity and simplification. It sells self-service hosting with human support, domains with DNS handholding, local backups with a familiar vendor stack, VPS with clear upgrade paths, and resale infrastructure to agencies that want to host clients under their own brand. The reseller pages explicitly market custom nameservers, custom branding, WHMCS integration, and the ability to set your own prices. That is not just technology. It is channel strategy. In a small market, indirect distribution through local agencies and freelancers can matter more than trying to build a large direct sales force.
The network says this is real infrastructure, but not sovereignty in the strong sense The strongest public evidence for cloud.mu is not its marketing copy. It is the routing record. AS328699 originates multiple IPv4 and IPv6 prefixes, appears at MIXP with 10 Gbps links, and also appears at NAPAfrica IX Johannesburg with 10 Gbps connectivity. BGP.tools shows five upstream providers: Kaldera, DataKeepers (Pty) Ltd, Mauritius Telecom, Hurricane Electric, and Vodacom. That combination matters. It demonstrates that cloud.mu is operating a substantial network edge with local peering and external reach, not simply reselling a server rack under someone else’s ASN. In small markets, having your own ASN and routing policy is a hard strategic threshold because it allows the operator to shape traffic paths, be multi-homed for transit, and present credibly to enterprise customers who know enough to ask where the packets actually go.
The Mauritian part of the story is real. At MIXP, cloud.mu appears with two visible port entries at Rogers Capital Data Center. Presence there means that local traffic to other Mauritian networks can often stay local rather than leaking to distant transit. That is exactly how a Mauritian cloud operator creates observable value against offshore hosting: not by inventing superior servers, but by shortening path length and coordination friction for domestic traffic. The site’s claims of “lowest possible latency” for Mauritian users are therefore at least directionally backed by the peering data. If their customers are mostly in Mauritius, local peering can actually be worth money.
But the same network record deflates any grander sovereignty narrative. cloud.mu also peers in Johannesburg and depends on South African and international upstream providers. Mauritius Telecom’s own commercial materials highlight the island’s submarine cable ecosystem—SAFE, LION/LION2, MARS, and T3—and the T3 cable itself is documented as connecting Mauritius to South Africa, with Liquid Telecom as a landing partner in Amanzimtoti. In other words, Mauritius does not live in splendid network isolation. It sits on a submarine geography whose resilience is better than before, but is still regionally entangled. cloud.mu’s Johannesburg presence and South African upstreams make operational sense for performance and reach. They also mean the platform’s local-sovereignty promise is only partly about route independence. If the South African side of the system falters, Mauritian service quality can still suffer.
This is not theoretical. cloud.mu’s own announcements disclose a “network connectivity issue” on 29 August 2024 caused by an outage with one of its upstream providers, which intermittently affected international connectivity to its Mauritius data centre for about two and a half hours, while services remained operational but less reachable. The more dramatic anecdote occurred in May 2026, when South African providers were hit by sustained DDoS attacks. MyBroadband reported that Datakeepers also suffered disruptions; local LinkedIn chatter complained that “cloud.mu was down” for hours and linked the issue to attacks in South Africa; and Mauritian press coverage from l’express framed the outage as affecting services hosted on cloud.mu. Some of that evidence is informal and should be treated as such. Taken together, though, it shifts the commercial interpretation. The whole point of buying “local cloud” is supposed to be insulation from external complexity. The public outage trail shows that locality reduced some risks but did not eliminate exposure to regional infrastructure.
The sponsorship of Ubuntu and other Linux mirrors is a subtler piece of network evidence. SysAdmin Journal documents that cloud.mu sponsored server and bandwidth for Mauritius-hosted mirrors for Ubuntu, Fedora, AlmaLinux, and openSUSE, and that the Ubuntu country mirror for Mauritius points to a cloud.mu subdomain. This matters for two reasons. First, it strongly suggests cloud.mu has enough spare bandwidth, enough operational competence, and enough local peering relevance to support a national-utility caching and distribution function. Second, it reinforces the company’s role as a domestic traffic localiser. Mirrors don’t just create goodwill; they reduce international dependence for software updates. In that limited sense, cloud.mu does create real digital-infrastructure value for the island. The catch is that mirror sponsorship is also a reminder that this is a bandwidth-and-caching business as much as a cloud business. That is strategically useful, but it is not the same as being a broad enterprise-computing platform.
So what does the network evidence actually prove? It proves that cloud.mu is a real operator with its own network identity, visible peering, routable allocations, and practical integration with the local Internet. It proves that “local hosting” is not just a slogan glued onto an offshore reseller account. It does not prove ownership of a Mauritian data centre building, immunity to regional outages, or sovereign independence in the stronger geopolitical sense that some of the data-sovereignty marketing implies. Locality here is a quality-of-service and commercial-trust variable. It is not autarky.
Mauritius as a market, regulator, and buyer Mauritius is large enough to sustain local digital infrastructure, but small enough that the addressable market is always the central strategic constraint. The country has roughly 1.25 million people and a diversified economy with ICT and financial services among the important sectors. That is enough to produce a meaningful long tail of websites, SMEs, agencies, schools, NGOs, and local applications. It is not enough to generate hyperscale-grade utilisation from domestic demand alone. For a local hosting platform, that means the operating game is keeping boxes full enough, support efficient enough, and churn low enough that modest recurring revenues can cover a stubborn base of fixed costs in power, colocation, transit, software licensing, staff, and spares.
Here is where Mauritius’ regulatory and policy environment cuts both ways. The country’s Data Protection Act 2017 does not create a broad legal obligation to keep all personal data in Mauritius. Section 36, as summarised in official and quasi-official materials, permits the transfer of personal data abroad where the controller or processor provides appropriate safeguards, where the data subject gives explicit consent, or where other specified grounds apply. The National Data Strategy speaks more forcefully about data sovereignty and notes that data classification frameworks should be harmonised with the Government Cloud or data localisation policies. But the public-policy posture is more nuanced than a simple “local hosting required” rule. Economically, that means cloud.mu’s sovereignty pitch is usually a soft advantage, not a hard mandate: it simplifies comfort, due diligence, and local accountability, but it does not automatically exclude offshore alternatives for private-sector customers.
That nuance is important because local-cloud narratives often overstate legal lock-in. In Mauritius, the better framing is that local hosting can make regulated or cautious customers feel safer, can help keep data under a familiar legal regime by default, and can reduce the compliance conversation around cross-border transfers. That has commercial value. But it is not the same as a sovereign monopoly. A bank, law firm, insurer, clinic, or media company may still decide that an offshore provider plus contract language and governance controls is sufficient. In a small market, that means cloud.mu’s proposition lives or dies on organisational preference, not only on law. Preferences are weaker moats than mandates.
The public sector is even less open than the sovereignty story might suggest. Mauritius already has a substantial state-hosted cloud position through the Government Online Centre, which official assessments describe as the central government data centre and the “Government Cloud” that provides hosting and IaaS for public systems such as e-procurement. Government strategy documents continue to assume that digital public services are hosted on the Government Cloud, and the Public Sector Investment Programme includes explicit funding lines for the GOC Tier IV Data Centre, the GOC Data Centre, and related hosting infrastructure. This makes the state both a cloud promoter and a vertically integrated customer of its own cloud. For an operator like cloud.mu, that narrows rather than widens the obvious path to public demand.
The procurement record confirms the ceiling. In the 2023 procurement for hosted data centre services for a disaster recovery site for the Government Online Centre, the two visible bidders were EMTEL Ltd and Mauritius Telecom Ltd, with EMTEL’s first-choice quote materially below Mauritius Telecom’s. cloud.mu was not visible in that bid opening. One must be careful: absence from a bid opening does not prove inability. But combined with the public sector’s GOC architecture and the tendency of large tenders to pull in telco and infrastructure incumbents, it suggests that cloud.mu’s realistic government-adjacent role is probably not “host the state.” It is more likely to catch smaller institutions, quasi-public niches, subcontracting opportunities, or overflow needs that do not require operator-scale comfort. In island markets with thick procurement, that distinction is an important commercial boundary.
The competitive map inside Mauritius also constrains the strategic space. Mauritius Telecom markets my.t Cloud, private cloud, colocation, backup as a service, and two data centres, including a Tier IV site in Rose Belle. Kaldera markets cloud, hosting, backups, managed infrastructure, and says it has three data centres. Mauritius Computing Services has long marketed itself as a cloud and hosting provider in the country, with historical press positioning it among the early movers. These are not imaginary competitors. They are the incumbent categories that capture large-account trust: telco, ISP, long-established enterprise IT integrator. cloud.mu therefore appears to occupy the space between them and purely offshore self-service. That can be a viable niche. It is also exactly the kind of niche that gets crowded whenever incumbent providers decide they want the same SME and mid-market portfolios more aggressively.
Where the moat lives and where it leaks cloud.mu’s moat, such as it is, does not start with “Mauritius” alone. Many firms can put a rack in Mauritius. The moat starts with the combination. Public evidence suggests that cloud.mu combines visible local network presence, a retail-grade self-service catalogue, human support, a domain-and-DNS wrapper, low-friction onboarding, and community embedment. That combination is harder to replicate than any single feature. The company sponsors local developer and software-craft events, appears in community publications as a hosting sponsor, and shows up in practical tools like country Linux mirrors. It also appears in client-side footers such as Capetech, which explicitly says it is “Powered by cloud.mu”. That is not the same as enterprise reference, but it is exactly how a small infrastructure brand wins mindshare in a constrained domestic market: by becoming the default recommendation among developers, agencies, and small-business operators who want someone local to call.
There is also a trust moat in being visibly Mauritian without being purely bureaucratic. Big offshore clouds intimidate some buyers because they require self-service competence and governance maturity. Big local incumbents intimidate other buyers because they look expensive, slow, or oversized. cloud.mu positions in the middle: local enough to answer the phone, structured enough to publish incident reports and operate a visible network, and cheap enough to feel approachable. The Digital Marketing Specialist job posting on LinkedIn is revealing in that respect. It frames the company as being interested in driving organic and paid search growth “for our group and select clients.” That sounds like a business trying to behave not just as a hosting provider, but as part of the local digital go-to-market stack. In small markets, that adjacency can matter as much as the technology. If you help the client launch, not just host, you become harder to displace.
But the moat leaks in four places.
The first leak is the law. As noted earlier, Mauritius’ data protection regime supports careful cross-border transfer rather than absolute local retention. cloud.mu can sell peace of mind, but it cannot rely on a sweeping legal obligation forcing businesses to use Mauritian infrastructure. When customers are sophisticated, the sovereignty premium becomes a negotiation variable rather than a regulatory necessity.
The second leak is utilisation. A small country produces small aggregate demand. Shared hosting can work because of oversubscription, but the heavier parts of the stack—backup repositories, virtual infrastructure, and especially dedicated servers—need enough paying use to support local power, space, inventory, and support. cloud.mu’s own Mauritius–South Africa price gap is evidence that utilisation and cost are harder locally. If a customer is price-sensitive and does not need Mauritian residence, the offshore or regional option becomes hard to resist, even on cloud.mu’s own menu.
The third leak is connectivity dependence. Local hosting helps domestic latency, but the island still depends on submarine cables and regional routing. cloud.mu’s Johannesburg footprint, South African upstreams, and the public outage trail mean that some of the operational risk sits below the “hosted in Mauritius” label. That matters especially for customers who buy local hosting for resilience rather than for latency alone. If the sales pitch is “stay local and stay safe,” customers will eventually ask what happened when South African infrastructure was under DDoS stress. The public record’s answer is not fatal, but neither is it reassuring.
The fourth leak is enterprise procurement. Larger buyers care about certifications, vendor risk, disaster-recovery architecture, contractual frameworks, and institutional permanence. Mauritius Telecom and Kaldera can wrap cloud and hosting inside wider connectivity, managed security, and data centre propositions. Government has its own cloud. The big integrators and telcos already occupy the “trusted infrastructure” seat for many serious procurements. cloud.mu may be able to win at the margins, but the public evidence does not yet show it occupying the commanding heights of Mauritian enterprise-infrastructure demand. Visible customer signals point more towards communities, websites, SMEs, agencies, and selected enterprise workloads than towards a crushing presence in high-end regulated-infrastructure contracts.
So where exactly is the durable value? It is in being the credible local default for the middle market: not the state, not the hyperscale-native multinational, but the Mauritian organisation that wants a website, email, VPS, backup, or application hosting under a familiar jurisdiction with a local support path and decent domestic performance. That is a genuine business. It is also a business that creates strategic value beyond its own revenue line. The Linux mirrors, local IX presence, and locally reachable hosting infrastructure reduce frictions in the domestic digital economy. cloud.mu therefore has strategic value to Mauritius even if it never becomes a large company. The trouble is that strategic value to a country is not the same as large economic rents for the operator. Many strategically useful infrastructure firms remain only modestly profitable because they solve collective problems in markets too small to overpay them.
What the public record still cannot answer For all the visible evidence, there are crucial things the public record does not resolve.
It does not resolve the exact ownership structure. The renaming trail from DataKeepers Ltd to Hosted Ltd is strongly supported by public records, and the DataKeepers South African connection is suggestive. But the public sources reviewed here do not clearly show whether cloud.mu is a subsidiary, sister company, franchise, closely allied operator, or simply a Mauritian business using a South African founder network or partner. That matters because ownership affects available capital, procurement credibility, and resilience options. If cloud.mu is backed by a broader group, its commercial durability is stronger than the website alone suggests. If it is primarily a lean local operator, then concentration risk is higher.
It does not resolve customer concentration. Public customer signals show community sponsorship, some visible hosted sites, media reliance in at least one press account, and a reverse DNS footprint that suggests a broader tenant base. But none of that tells us whether the firm has one giant anchor customer, fifty mid-sized contract customers, or thousands of small accounts. That is perhaps the most important missing commercial fact. A local host with diversified small accounts can survive churn better than one exposed to a few large logos. Conversely, a few larger recurring backup or dedicated-server contracts could explain how a small-market operator can afford visible network investments. The public evidence ledger is not rich enough to answer this with confidence.
It does not resolve facility architecture in enough detail. We can see use of local data centres, IX presence via Rogers Capital Data Center, and a Mauritius/South Africa location split on dedicated-server offers. We cannot fully see the rack count, redundancy design, storage architecture, power contracts, or whether South Africa is merely an alternative sales location or a genuine failover backbone for some Mauritian services. In an island market, those distinctions are not footnotes. They can determine whether “local cloud” means “primary in Mauritius and secondary abroad”, “primary in third-party colocation with regional spill”, or simply “some products are local, others are not”.
It does not resolve the support economics. cloud.mu heavily emphasises support—phone, live chat, email, real people—which is often exactly what the local market wants. But support is labour, and labour is margin. Without headcounts, ticket volumes, or renewal rates, you cannot tell whether support is the reason cloud.mu wins business or the reason cloud.mu may struggle to scale profitably. The existence of a job posting for a marketing role suggests active demand generation, but there is no public evidence here showing whether customer acquisition cost is low because of word-of-mouth, or high because the local market is expensive to educate and price-sensitive once approached.
Finally, the public record does not resolve whether cloud.mu can move upmarket without losing its economics. The website’s dedicated servers, backup tiers, and Microsoft licensing offer the ingredients of a move from “web hosting provider” toward “serious SME infrastructure provider”. But Mauritius is full of firms that can handle websites and not many that can clear enterprise procurement hurdles. The decisive evidence would be certifications, audited uptime or facility standards, evidence of disaster-recovery exercises, named enterprise customers, and repeated public-sector or regulated-industry contracts. That evidence is thin in the material reviewed here. The sober conclusion, therefore, is not that cloud.mu lacks strategic value. It is that its strategic value is currently easier to evidence than its ability to harvest large enterprise rents from that value.
Evidence record Source name URL Source type What it supports What it doesn’t prove Why it matters economically Quote cloud.mu homepagehttps://cloud.mu/Official company page | Central positioning around local hosting, local support, full product family, and “own and manage” infrastructure rhetoric | Does not prove facility ownership or financial performance | This is the company’s demand-side narrative and reveals the premium it tries to charge for locality cloud.mu contact pagehttps://cloud.mu/contact/Official company page | Current public operator name as Hosted Ltd and office address in Moka | Does not prove legal history or equity ownership | Establishes current operating identity and local commercial presence cloud.mu privacy policyhttps://cloud.mu/legal/privacy-policy/Official legal page | Operator identified as DataKeepers Ltd with registration number C18158096 | Does not prove why the brand changed or when | Shows legal lag and supports the rebranding/change-of-name interpretation Mauritius Revenue Authority VAT listhttps://www.mra.mu/download/ListofVATRegPersons.pdfOfficial government record | Registration number C18158096 appears as Hosted Ltd | Does not prove all historical names or actual ownership | Helps reconcile legal identity; the name-change evidence is important for vendor due diligence and continuity BGP.tools AS328699https://bgp.tools/as/328699Network record / measurement | ASN, prefixes, upstream providers, Mauritian address, network age, and DataKeepers branding | Does not prove customer numbers, network utilisation, or facility ownership | Confirms cloud.mu is a real network operator, not just a reseller bgp.he.net AS328699https://bgp.he.net/AS328699Network record / measurement | Company website, looking glass, IX participation, and originated prefixes | Does not prove service quality or commercial success | Validates the infrastructure layer and multi-homing/peering posture MIXP PeeringDB recordhttps://www.peeringdb.com/ix/1508IX / Peering record | cloud.mu presence at MIXP, with ports at Rogers Capital Data Center | Does not prove exclusive colocation or complete architecture | Local peering is a real low-latency advantage and a solid operating asset in Mauritius cloud.mu looking glasshttps://lg.cloud.mu/Official network tool | Public routing/diagnostic surface in Mauritius and Johannesburg | Does not prove uptime or capacity utilisation | A public looking glass is a marker of serious network operations and transparency Domain pricing pageshttps://my.cloud.mu/cart.php?a=add&domain=registerandhttps://register.mu/cart.php?a=add&domain=registerOfficial company / registrar pages |.MU pricing at cloud.mu and Register.mu, showing likely reseller-style domain economics | Do not prove gross margin or registry-side contract terms | Shows that domains are probably a retention and bundling tool, not a deep profit stream MU-NIC registrars listhttps://www.nic.mu/registrars/Official registry record | Accredited.MU registrars include Register MU but not cloud.mu | Does not prove that cloud.mu does not have a reseller agreement or indirect access | Important for understanding whether cloud.mu controls a scarce registrar permission or merely channels it Product pricing pageshttps://my.cloud.mu/store/web-hosting,https://my.cloud.mu/store/linux-vps,https://my.cloud.mu/store/cloud-backup,https://cloud.mu/dedicated-servers/Official company pages | Retail pricing scale for shared hosting, VPS, backups, and dedicated servers Mauritius vs South Africa | Do not prove realised selling prices, discounts, or margins | The public price book reveals the business model, segment focus, and the premium charged for Mauritian locality ICTA commercial licenseeshttps://www.icta.mu/licences-issued/Official regulator page | ICTA’s public telecoms licence list is populated by telcos/ISPs; searches for Hosted Ltd and DataKeepers yield no matches on the page text | Does not definitively prove the company has no other approval or agreement | Suggests the moat is not an obvious telecom regulatory franchise, but an operational market position Mauritius government cloud recordshttps://ppo.govmu.org/Documents/Publications/Maps/MAPS_Assessment_Mauritius_e-proc_vol1.pdfandhttps://mof.govmu.org/Documents/budget_2025_2026/2025_26_PSIP.pdfOfficial government report/budget | GOC operates the Government Cloud; the state continues to invest in GOC data centre infrastructure | Do not prove complete exclusion of private providers from all state workloads | Shows that the state is itself an infrastructure incumbent, limiting the addressable market for private local cloud DR hosting bid openinghttps://cpb.govmu.org/Documents/pub_openings/N7_CPB_24_2022-Hosted%20Services%20of%20a%20data%20centre.pdfOfficial procurement record | The visible bidders for GOC disaster-recovery hosting were EMTEL and Mauritius Telecom | Does not prove cloud.mu was unwilling or unable to bid | Indicates where sovereign-style contracts gravitate: towards telco-scale actors Mauritius data transfer regimehttps://www.fscmauritius.org/media/105843/the-data-protection-act-2017.pdfandhttps://dataprotection.govmu.org/Documents/NDS_final.pdfLaw / official policy PDF | Cross-border transfer is permitted with safeguards/consent; sovereignty is a policy theme, not a sweeping localisation mandate | Does not prove how each regulator or sector interprets edge cases | Shows that cloud.mu sells soft compliance and comfort, not an absolute legal necessity for most private buyers Outage trail and regional dependencehttps://my.cloud.mu/announcements/view/2024-08,https://mybroadband.co.za/news/cloud-hosting/648362-south-african-infrastructure-providers-wiped-off-the-internet-by-sustained-ddos-attacks.html,https://lexpress.mu/s/une-cyberattaque-massive-en-afrique-du-sud-558137Official status page / press / local press | cloud.mu disclosed upstream-related outage; South African DDoS reports and local press linked regional attacks to impact on Mauritian services | Does not prove the exact internal failure path for each interruption | Commercially crucial because it tests the robustness of the thesis that “local insulates you”
What would change the island-cloud bet Three facts would materially change the commercial view.
The first would be solid evidence that cloud.mu has durable enterprise anchors—for example, named long-term contracts in finance, health, media, education, or regulated industry; credible third-party certifications; or repeated wins in quasi-public procurements. That would shift the company from “real local infrastructure merchant” to “institutional platform,” and make the margin story more attractive.
The second would be solid evidence on ownership and capital backing. If public records eventually showed that Hosted Ltd / cloud.mu is backed by a larger South African or regional infrastructure group with meaningful balance-sheet support, some of the fragility implicit in the small-market economics would weaken. If the opposite were true—if it is effectively a very lean local operator—the concentration and resilience risks would look sharper.
The third would be a better view on how much of the locality premium customers will continue to pay as connectivity keeps improving. cloud.mu’s own Mauritius–South Africa price split says the premium exists today. The long-term question is whether Mauritian customers continue to buy it for sovereignty, latency, and support—or whether they gradually decide that the island cloud should be local only at the edge, while the real workloads live elsewhere. That is the real Island Cloud Problem. The public evidence shows that cloud.mu has built a serious answer to it. The public evidence does not yet show that the answer scales indefinitely.