The name was born the day the metaverse's oldest world closed

On 10 March 2023, Microsoft switched off AltspaceVR, the pioneering social-VR platform it had rescued from collapse six years earlier — the scheduled end, announced that January, of the metaverse's longest-running public world. On the same day, according to Verisign's registry record, somebody registered the domain meteversecloud.com through GoDaddy. Also on the same day, the American Registry for Internet Numbers created an organisation record, ML-1432, for a company calling itself "Meteverse Limited." at suite 1108, 250 Consumers Road, North York, Ontario.

The coincidence is the whole story in miniature. By March 2023 the metaverse was not a rising trend a founder might surf; it was a liability being written off in public. Meta's Reality Labs had just reported a US$13.7 billion operating loss for 2022. Within three weeks Disney would eliminate its metaverse division entirely. Nobody minting a consumer brand that quarter reached for the word — which is precisely what makes this name interesting. A trend name adopted during a hype cycle is a bid for cheap customer acquisition. A trend name adopted at the moment of the trend's death, spelled with one letter changed and attached to a company that publishes no prices, no address and no customer list, is doing a different job.

Consider the timestamp problem as a salesperson would live it. It is mid-2026, and you are pitching content delivery to a procurement team under a name that reads like a 2021 pitch deck misspelled. Every discovery call starts minus one point of credibility; every security questionnaire invites the question of why a global network is named for a fad that predeceased it. If Meteverse were what it appears to be — a small Canadian cloud outfit that bet its brand on the wrong wave — that daily credibility tax, compounded across a funnel, would be the story, and the arithmetic would be grim: naming consultants charge five figures to fix what a founder chose in an afternoon, and rebrands of small hosts routinely cost more in lost search equity and churned bookmarks than the original name earned.

The question we set ourselves was to price exactly that: what a hype-cycle name buys in acquisition while the trend is hot, what it costs in credibility when the trend dies, and what the underlying business of servers, margins and churn looks like once the name stops working. The honest answer, reached early and worth stating plainly, is that the premise does not survive contact with the record. Meteverse is not small, the name was never aimed at customers, and the business underneath is one of the larger content-delivery networks on the public peering map. There is no funnel to tax. What the name buys is not acquisition. It is distance — and distance, for this particular network, has measurable economics.

The paper trail runs through California to Shenzhen

Start where the storefront is thinnest, because the thinness is measurable. On the open web, the entity amounts to one domain, one brochure site and a scattering of third-party references — a boundary worth respecting, and worth testing. The meteversecloud.com website is a four-page brochure: download acceleration for game installers and app packages, video-on-demand caching, streaming over HLS and RTMP, and edge computing "including Openstack-based CVM and Kubernetes container products". No pricing. No legal entity. No copyright line. The about page offers a single address-free email and the claim of being "an exceptional global content distribution acceleration service provider". The company that claims a global acceleration platform serves its own homepage, we observed while researching, through Cloudflare — a competitor's network — from a Toronto point of presence. On the visible web, that is the entire company.

The registries are less shy. ARIN's public transfer log — a primary document that records every change of resource ownership in the North American registry — shows that on 13 April 2023, one month after the name was minted, autonomous system 54994 and three blocks of IPv4 address space (138.113.0.0/16, 157.185.128.0/18 and 168.235.192.0/20) moved from "QUANTIL NETWORKS INC" of the United States to "Meteverse Limited." of Canada in a single set of transfers, all timestamped to the same second.

Quantil is not an obscure seller. Quantil Inc took assignment of that autonomous system in February 2013, per the registry's archived delegation statistics, and was the American CDN brand of Wangsu Science & Technology — 网宿科技, Shenzhen ChiNext code 300017 — the largest listed specialist content-delivery company in China. The log carries the brand's whole biography in four entries: Quantil Inc passes its resources to Quantil Networks Inc in an August 2017 internal reorganisation; Quantil Networks buys 138.113.0.0/16 — 65,536 addresses of 1990-vintage space — from Travel Tripper LLC, a hotel-booking software firm, in December 2019; and the accumulated estate then leaves the United States for Toronto in April 2023. The connection is not an inference. The archived PeeringDB record for this exact network, captured in March 2021 and unchanged through early 2023, names it "Quantil Networks", gives its alias as "WANGSU", points its website at en.wangsu.com and opens its notes with "Founded in January 2000, Wangsu Science and Technology Co., Ltd (Stock code: 300017)". The same PeeringDB object — created in January 2016, never deleted — today reads "meteversecloud", organisation Meteverse Limited, North York, Ontario. The network did not move. The nameplate did.

Continuity runs through every technical artefact. The routing-policy object Meteverse maintains in the RADB registry, AS54994:AS-ALL, still lists among its members the customer set of "QYNTCL" — Quantil — and AS38107, which belongs to CDNetworks, the Korean CDN that Wangsu acquired from KDDI for roughly US$185 million in 2017. Sampling the address space the network announces today: individual /24 blocks under 163.171, 148.253 and 14.0 resolve in the registries to CDNetworks objects in Vietnam, Thailand and Korea, and 174.35.0.0/16 to CDNetworks Inc in the United States. The network-data firm Netify, which fingerprints traffic platforms, records flatly that "Meteverse Limited manages IP addresses for organisations including Meteverse Limited, CDNETWORKS, and WANGSU". Even the telephone number on Meteverse's ARIN contact records carries a Los Angeles 310 prefix — the dialing code of Quantil's old California operation, attached to a Toronto postal address.

Then there is the strangest artefact of all. Quantil.com is still live, and its text is a near-verbatim twin of meteversecloud.com — the same four products, the same "Cloud CDN Provider" tagline, the same OpenStack-and-Kubernetes phrasing — except the Quantil copy is served, per its response headers, through Wangsu's own delivery network. Two storefronts, one script, one network underneath. The trade press noticed the wind-down years ago: CDN Planet's profile of Quantil records that the brand stopped selling and now redirects buyers to CDNetworks, Wangsu's surviving international marque.

What the Canadian corporate record adds is, tellingly, almost nothing. A search of Corporations Canada's federal register for "meteverse" returns zero results, so the company is not federally incorporated; the Ontario address implies a provincial registration, but Ontario's business registry offers no open query interface — the province will sell a profile report for eight dollars through its transaction system — and the two commercial registry aggregators we tried during research either refused automated access outright or returned nothing under the name. We state the attempt and its outcome because the gap is itself a finding: the only public documents in which Meteverse Limited has a verifiable existence are the number-resource registries. Its Shenzhen parent's disclosure archive is equally silent; a full-text search of Wangsu's announcements on the exchange's disclosure platform for "Meteverse" returns nothing, and the 2025 annual report describes an overseas business spanning "more than 90 countries and regions" without naming the Toronto vehicle that holds its Western address space. A company visible only where visibility is technically compulsory is a design, not an accident.

What five and a half terabits actually looks like

Strip the name away and audit the machine. As of this week, RIPE's routing statistics show the network announcing 1,072 IPv4 prefixes covering 275,712 addresses plus 161 IPv6 prefixes, visible to every one of the 324 full-table feeds that RIPE's route collectors receive — full global reach, no partial propagation. CAIDA's ranking data counts 97 transit providers and 272 peers, the signature of an operation that buys local connectivity in dozens of markets rather than hauling traffic home.

The interconnection footprint is where the scale registers. Meteverse's PeeringDB entries list 110 exchange connections totalling 5,514 gigabits of port capacity across roughly ninety exchange fabrics, plus presence in 74 interconnection facilities across 38 countries. The distribution is the strategy in map form. The largest ports sit in Asia: 800G at Singapore's SGIX, 400G at Equinix Singapore, 400G at BBIX Manila, 300G at Hong Kong's HKIX, 100G each in Tokyo, Mumbai and Riyadh. Latin America gets 200G in São Paulo, 100G each in Rio and Bogotá, 40G in Santiago — and 100G at Venezuela's NAPVE in Caracas. There are 100G at IRAQ-IXP in Baghdad, 40G at OpenIX Beirut, 40G at IXPN Lagos, ports in Yangon and Mandalay, 60G in Bishkek, and 50G split between Moscow and Saint Petersburg. Western Europe, the most contested CDN market on earth, gets a comparative shrug: 20G at AMS-IX Amsterdam, 40G at DE-CIX Frankfurt, 20G at LINX London.

That is not the port map of a company selling hosting to Canadians. It is the delivery fabric for Chinese games, apps and video services expanding into exactly the markets where Western hyperscalers are thinnest and where a gigabyte delivered locally is worth most: Southeast Asia, the Gulf, Africa, Latin America, Central Asia. The storefront says as much — "game installation packages, application APK" — and the parent's filings complete the picture. Wangsu's 2024 annual report discloses overseas revenue of CNY 2.844 billion, 57.66 per cent of the group's total; the 2025 report shows the group at CNY 4.661 billion in revenue with CNY 2.313 billion — about US$320 million — still earned abroad even after divesting a managed-services unit. The self-declared traffic figure on the PeeringDB record, 20–50 terabits per second, is the operator's own claim and should be read as such; the 5.5 terabits of paid public ports, by contrast, are corroborated by the exchanges' own participant listings and are bought, monthly, with real money.

The address space has its own archaeology, and it argues against reading Meteverse as anything new. The routing history of 138.113.0.0/16 shows the block spending 2017 to 2022 shredded into /24s and announced by dozens of unrelated Asian networks — Thai, Indonesian, Vietnamese, Taiwanese — the classic signature of address space sublet by the month on the lease market while its owner waited for a better use. From May 2021, well before any Canadian entity existed, autonomous system 54994 itself began originating slices of the same block; the network Meteverse now fronts was already consolidating that space under Quantil's number two years before the nameplate changed. Even the PeeringDB object that carries the Meteverse identity today was created in January 2016 — the machine kept its logbook through three names.

One more registry detail completes the asset picture. Meteverse Limited directly holds four ARIN blocks totalling 151,552 IPv4 addresses — the three inherited from Quantil plus one later purchase — while the rest of the 275,712 addresses it announces remain registered to others, chiefly CDNetworks entities across Asia and the United States. It also operates as an address-issuing member of the European registry — individual Meteverse-maintained network objects show up under RIPE with service addresses in Malaysia — and it is still buying. The same transfer log that documents the 2023 re-flagging records that on 15 September 2025, Meteverse received 153.43.0.0/16 — 65,536 addresses of vintage 1991 space — from HP Inc. A dormant brand does not buy a /16 from Hewlett-Packard. A growing network does.

The economics of a name that no longer sells

So what is "Meteverse" for? Take the hypotheses in order of cost.

The charitable one — that this is straightforward hype-branding gone stale — fails on timing and on audience. The name was created in the write-off quarter, not the boom, and the entity does zero consumer marketing: no advertising trail, no sales copy beyond the brochure, no price list to anchor a buyer. Hype names earn their keep in acquisition funnels. This company has no funnel.

The second hypothesis is trademark hygiene: metaversecloud.com belongs to an unrelated virtual-worlds firm, and "Meta" itself was by 2023 a litigious trademark neighbourhood, so the altered vowel may simply be the cheapest name that cleared a conflict search. Plausible as mechanics, but it explains the spelling, not the choice of a spent buzzword by a serious network operator.

The hypothesis that fits the record is that the name's blandness is the product. Between 2021 and 2023, being visibly Chinese became a priced risk in Western telecommunications. The FCC revoked China Telecom Americas' operating authority in October 2021 and China Unicom Americas' in January 2022; Canada, Meteverse's chosen flag, banned Huawei and ZTE from its 5G networks in May 2022. None of these actions touched Wangsu or any CDN. But number resources — address blocks, autonomous systems — are the layer of the internet where nationality is legible to anyone with a terminal window. Re-registering the family's Western assets from "Quantil Networks Inc, US" (a brand the trade press had spent a decade linking to Wangsu) to a fresh Canadian holder with a name that smells of a defunct consumer fad is a legibility choice. A "Wangsu Networks America" invites a Washington hearing; a "Meteverse Limited" invites a shrug. The one letter changed from "metaverse" even defeats the laziest keyword search.

We should be precise about what the record does and does not support. The transfers, the continuity of routing objects, the shared website copy and the CDNetworks address space are documented facts. The intent behind the naming is our inference from those facts, and no filing states it; an alternative reading — an arm's-length management buyout of Quantil's assets by parties who happened to keep serving Wangsu's network — cannot be fully excluded without the Ontario shareholder register. But the buyout reading has to explain why the "independent" buyer's routing database entries are maintained by the same operations desk that lists CDNetworks as a member, why its contact phone keeps Quantil's Los Angeles prefix, and why its twin website runs on Wangsu's own delivery platform. Occam favours the flag of convenience.

The counterfactuals sharpen the point. Companies that wore the metaverse label in front of paying customers spent 2023 and 2024 quietly taking it off: consumer platforms renamed products, funds rebadged themselves around artificial intelligence, and the word migrated from earnings-call openers to footnotes. Meteverse did the opposite of all of them — it put the word on in 2023 and has left it there for three years, through two annual cycles in which its parent's own reports stopped mentioning the concept. A name that survives untouched while the trend it references becomes an embarrassment is a name nobody inside the firm considers customer-facing. The maintenance decision is itself disclosure.

Here, then, is the inverted brand economics this case actually teaches. In consumer markets a trend name is an asset that decays: it buys clicks in the boom and costs credibility after. In wholesale infrastructure, brand value at the point of sale is approximately zero — game publishers buying delivery in Jakarta procure on route tables, test transfers and price per gigabyte, and would not blink at a vendor named after last cycle's fad. What has value is the absence of associations. The obsolete name is not a depreciating asset; it is a cheap, self-renewing insulation layer. Its datedness — the quality that would kill a consumer brand — is irrelevant to the only audiences that matter: procurement engineers who never weigh names, and regulators who scan for a different word entirely.

Counting the cost of staying invisible

The arithmetic of this business divides into three lines a reader can check: what the identity cost, what the visible network costs to keep connected, and what the address assets are worth. Every price below comes from a published tariff, a registry log or a dated marketplace series; where we infer, we say so.

The identity first. ARIN's published fee schedule prices a resource transfer at US$500 payable by the requesting organisation, and organisation creation at US$50. The entire April 2023 re-flagging of an autonomous system and three address blocks into a new Canadian holder therefore cost on the order of US$550 in registry fees, plus an Ontario incorporation and a virtual-office lease that the province and landlords price in the low hundreds of dollars a year. Set that against a single line of the network's monthly interconnection bill, below: the corporate identity is the cheapest component this company owns — cheaper than one month of one exchange port — and that asymmetry is the clearest evidence that the name was never meant to carry commercial weight. It is an anchored pair worth restating: roughly US$550, one time, to become Canadian at the registry; about US$7,700, every month, list price, to keep just its Singapore public ports lit.

Now the connectivity line, built strictly from exchanges that publish their tariffs. Singapore's SGIX publishes a port-price table: the 800 gigabits Meteverse holds there — the table prices eight 100G ports or two 400G ports identically — lists at S$9,900 a month, roughly US$7,700. Hong Kong's HKIX publishes its charge table — version 1.6, updated January 2025 — at HK$31,200 per 100G port per month, so the 300G there lists near HK$93,600, about US$12,000, before the exchange's aggregation discounts. AMS-IX publishes €3,240 per 100G per month on a twelve-month term; Meteverse's modest 20G in Amsterdam runs about €1,300. For the remaining ~4,400 gigabits spread across ninety-odd fabrics from Baghdad to Bogotá, no unified tariff exists; bracketing them between the cheapest and dearest published per-gigabit rates in this network's own portfolio — SGIX's ~US$9.7 and AMS-IX's ~US$38 per gigabit-month — yields US$43,000 to US$167,000 a month. Summing: the public-peering port bill alone plausibly runs US$64,000 to US$188,000 a month at list, call it US$0.8–2.3 million a year, and that is the bracket's arithmetic, not a measured invoice. It excludes what the tariffs themselves exclude: cross-connects and rack space in 74 facilities, local loops, and transit contracts with 97 providers, none of which publish prices and all of which scale the true network cost to a multiple of the port line.

Two refinements keep the bracket honest. First, list prices overstate what a buyer of this size pays: HKIX's own table publishes discounts of HK$3,120 a month per 100G port for bundled links and again from the third port onward, exchanges compete for content networks precisely because they attract the eyeball networks that pay full freight, and a 110-port customer negotiates. Second, the bracket prices only ports. The tariffs are explicit that local loops, cross-connects and rack space are excluded, and none of the 74 listed facilities publishes a rate card; every one of those buildings adds a monthly line we can name but not number. Both refinements push in opposite directions, which is why we publish the bracket rather than a point estimate and mark the whole line as tariff-anchored arithmetic, not an observed invoice.

Finally the balance sheet in the walls. Meteverse's 151,552 directly registered IPv4 addresses are marketable assets with a dated public price series. IPv4.Global's marketplace reporting shows /16 blocks falling from nearly US$50 per address in mid-2024 to just over US$24 by March 2025, with the firm's 2025 full-year review recording large-block prices at decade lows amid more than 40 million addresses transferred. At the March 2025 print the holding replaces at roughly US$3.6 million; at the softer late-2025 large-block levels, nearer US$2.3–2.7 million. The September 2025 purchase of HP's /16 has no disclosed price — the transaction, parties and date are registry fact; the value is our single-sourced estimate from that same marketplace series, and we flag it as such — but at prevailing large-block rates it represents a US$1.0–1.6 million cash commitment. Keeping it all registered costs, per ARIN's schedule, US$8,820 a year in the "Large" category. The shape of the ledger is unmistakable: registry identity costs hundreds, registry assets are worth millions, connectivity costs millions over any planning horizon — and none of it depends on what the company is called.

Where the revenue lives, and where it doesn't

Whose money flows through this? Entity-level revenue is invisible: no Canadian filing is public, no price list exists, and the parent's segment reporting stops at "overseas". So we triangulate, and we present both methods as inference.

Method one works down from the filed segment. Wangsu's audited overseas revenue was CNY 2.844 billion in 2024 and CNY 2.313 billion in 2025 — about US$320 million — with the decline attributed to divesting the Cloudsway managed-services unit, not to delivery. That revenue is booked through the group's selling entities, principally CDNetworks out of Singapore and Seoul; nothing suggests the Toronto holder invoices customers at all. Method two works up from capacity: 5,514 gigabits of public ports, run at the 20–40 per cent average utilisation typical of content networks (an operating assumption, not a measurement), implies 1.1–2.2 terabits per second of average public-peering egress; at wholesale content-delivery rates of one to three US dollars per megabit-month — a market band we assert from practice, not from a filed tariff, and weight accordingly — the delivery layer that Meteverse's ports serve would gross very roughly US$13–80 million a year. The two methods reconcile only one way: the traffic Meteverse carries is a meaningful but minority slice of a parent overseas book that also includes security services and managed cloud, and the Canadian entity itself is best modelled not as a business with customers but as a cost-and-asset vehicle — a holder of addresses and port contracts whose "revenue" is an intercompany settlement we cannot observe.

That structure is unremarkable in multinational network groups and rational here. A holder in the ARIN service region can receive ARIN transfers, sign North American exchange and facility contracts under local law, and keep the asset register a jurisdiction away from both Shenzhen and Washington; whether it books a cost-plus service margin or holds the assets at nominal value is a transfer-pricing choice invisible from outside. What the outside can see is the group's willingness to spend on this layer across a decade: roughly US$185 million for CDNetworks in 2017, a /16 from a hotel-booking company in 2019, a /16 from HP in 2025, and port capacity added on six continents in between. Where the money finally lives is Shenzhen; the 21st Century Business Herald's April 2026 analysis of Wangsu's results notes the group earned CNY 799 million in net profit on that shrinking top line, even as its founder and former chairman sold down hundreds of millions of yuan in stock across 2025 and early 2026 — the founder roughly CNY 130 million in June 2025, the former chairman nearly CNY 590 million across two disclosed rounds with a third block trade in March 2026. Insiders harvesting while the overseas machine keeps buying /16s is not a contradiction either; it is a listed parent monetising a mature domestic franchise while capital flows quietly to the frontier that still grows.

The sound of no customers complaining

For a network of this size, the most striking market signal is silence. The hosting bazaars where operators live and die in public — LowEndTalk, WebHostingTalk, the relevant subforums — contain, as far as our searches could find, not a single customer thread about "Meteverse" or "meteversecloud". No reviews, no outage complaints, no "is this legit" posts. For a retail host, silence like that means no customers; for a wholesale carrier whose buyers are a handful of games and video platforms under contract, it means the customers do not shop in public. Either way it confirms that the brand does no selling. The absence of any job posting under the Meteverse name anywhere we looked points the same direction: the people who run this network are employed by someone else.

Customers do leave fingerprints in one place, though: the routing-policy object. The membership list of Meteverse's registered routing set — the machine-readable declaration of whose routes it will carry — names, alongside the Quantil and CDNetworks legacy sets, a short roster of downstream networks: a Hong Kong games studio, two Indonesian providers, another Hong Kong trading entity, and Zenlayer, the Los Angeles edge operator. CAIDA's topology inference agrees, counting five customer networks beneath the system against 272 settlement-free peers. Five visible wholesale customers under a network with ports on six continents is not a contradiction; it is the expected shape of a captive carrier whose real client list — the game publishers and video platforms — buys delivery contracts from the group's selling arms and never touches the routing layer under its own name. The one named customer that is neither Chinese nor an affiliate, Zenlayer, is itself a wholesale edge platform: infrastructure selling to infrastructure, three layers removed from anyone who would ever read the word "Meteverse" on an invoice.

The consumer-facing residue rounds out the picture at the margins. Website-technology trackers maintain a hosting category for "Meteverse Cloud", implying its address ranges do appear under third-party sites their crawlers visit; a consumer trust scanner rates the domain legitimate and safe — a verdict generated, with some irony, from signals like domain age and traffic rather than from any knowledge of what the company is. The signals that do exist beyond that are reputational residue in the address space. IP-intelligence services disagree about the same network in instructive ways: ipinfo tags at least some Meteverse addresses in Toronto as VPN exits, while ip2location marks others clean; anti-spam databases such as CleanTalk maintain blacklist pages for the ASN recording scattered abuse from a handful of individual addresses — pages that, in a small archaeological joke, still list the network's owner as Quantil Networks. This mottled pattern is what recycled and multi-tenant space looks like — blocks that spent 2017–2022 sublet across dozens of Asian networks, as the routing history of 138.113.0.0/16 shows, carry scar tissue in the reputation databases long after consolidation. One further trace worth watching: since late January 2026, a /24 from the former HP block has been announced by Zenlayer, a Los Angeles-based edge provider that also appears in Meteverse's routing-policy object — consistent with capacity or address leasing between the two, which would be a modest secondary monetisation of the IPv4 estate. What would settle these readings: a sustained multi-source pattern of proxy flags (indicating the space is being rented to anonymisation services at scale) versus their gradual disappearance (indicating cleanup for first-party CDN use). Watching which way the flags trend over the next year is a better telescope into this company's commercial choices than anything it publishes.

Rivals, price wars and the flag problem

The competitive logic of the whole construction starts at home. China's domestic CDN market has been through a brutal price war — the business press headline on Wangsu's mid-2025 results was, literally, "squeezed by the price war, transformation accelerates" — and the group's CDN-and-edge segment shrank 15.6 per cent in 2025, its only declining line. Overseas delivery, at nearly half of group revenue, is the margin refuge; the security and value-added line grew 18.65 per cent but delivery abroad remains the volume engine. Meteverse's port map is therefore best read as Wangsu's answer to a strategic squeeze: follow Chinese game and app publishers into the Global South, where the delivery economics reward exactly this shape of network. In Frankfurt or Ashburn, transit is nearly free and a marginal CDN differentiates on nothing; in Baghdad, Caracas or Yangon, international transit is scarce and expensive, a local exchange port turns a transcontinental haul into a metro hop, and the first content network in the building earns real margin on every gigabyte a game launch pushes through it. Akamai and CloudFront price those markets high; Cloudflare's self-serve model fits poorly with hand-held enterprise contracts; local incumbents lack global backbones. Meteverse's true competitors are the other Chinese-origin edge networks running the same play — Tencent's and Alibaba's international edges, EdgeNext, Kaopu-style distributed clouds — all of which face the same flag problem, and several of which have made the same choice: Singapore holdings, Hong Kong holdings, and now, in this case, an Ontario one.

That flag problem is the risk section, and it cuts both ways. Nothing in the public record shows Meteverse, Quantil, CDNetworks or Wangsu subject to any Western enforcement action, and CDNs sit outside the licensed-carrier perimeter where the FCC's Section 214 revocations bit. But the perimeter moves. The precedents of 2021–2022 established that Chinese-owned network operators can lose Western operating authority on national-security grounds without any finding of misconduct; a future extension of such logic to content networks, or a Canadian analogue under its telecommunications-security amendments, would make the Toronto wrapper's thinness suddenly material — regulators, unlike customers, do eventually read transfer logs. The network's deep positions in Moscow, Caracas, Yangon and Baghdad, rational as pure delivery economics, add sanctions-adjacency risk that Western-headquartered rivals simply decline to hold. And the group's own disclosure that overseas expansion is its growth story means any forced retreat from ARIN-region assets — including a fire-sale of 151,552 addresses into a falling IPv4 market — would land on the parent's most profitable segment. Customer switching costs run shallow in wholesale delivery; a games publisher can re-point traffic in weeks, which is why the moat here is not lock-in but price and presence — and why the quiet efficiency of the whole machine matters more to Shenzhen than any name ever will.

What would change this judgement

Our judgement is that meteversecloud is the deliberately forgettable Canadian face of Wangsu's international delivery network: a real, growing, capital-intensive operation whose brand economics are inverted — the name is worth nothing to customers and several hundred dollars of registry fees to its owners, and its obsolescence is a feature. Specific documents would move us. An Ontario corporate profile or shareholder register naming Wangsu, CDNetworks or their officers would convert our attribution from registry-inference to corporate fact; one naming unrelated principals would force the management-buyout reading and a rewrite of the intent analysis. Wangsu's next annual report naming the entity among consolidated subsidiaries would do the same job from the Shenzhen end. A divestiture entry in the ARIN transfer log — Meteverse selling blocks rather than buying — would signal retreat, as would decay in the PeeringDB port inventory, which has so far only grown; port additions in Western Europe would instead suggest a push upmarket. Any FCC, ISED or Team-Telecom-style proceeding touching Chinese-owned content networks would reprice the whole wrapper. Movements in the IPv4 market cut the asset math both ways: a recovery in large-block prices would rebuild the balance sheet inside the walls, while a continued slide toward the marginal value of leasing would make the 2025 HP purchase look like buying at the falling knife — the lease-versus-own arithmetic on 151,552 addresses is sensitive to a few dollars per address either way. And if the storefront one day publishes a price list, hires under its own name, or answers its contact address with a legal entity and a phone number that matches its continent, we would happily retire the camouflage thesis — that is what a company that wants customers looks like, and this one, so far, does not.

Evidence register