Sky Link and the rooftop arithmetic of a Tk500 broadband bill

The bill is cheap; the work behind it is not

Start with the bill a Wari household can understand. In April 2025, The Business Standard reported that millions of Bangladeshi residential customers paying Tk500 a month for a 5Mbps shared broadband connection would receive 10Mbps at the same price after an ISPAB announcement; the same report said BTRC's 2021 rules allowed a dedicated internet capacity to be shared among no more than eight customers (https://www.tbsnews.net/bangladesh/tk500-broadband-connection-speed-doubles-10-mbps-1119976). Earlier reporting on the "One Country, One Rate" tariff placed the original entry package at Tk500 for 5Mbps, with higher tiers around Tk700-800 for 10Mbps and Tk1,100-1,200 for 20Mbps (https://www.dhakatribune.com/business/248823/broadband-internet-price-set-at-tk500-a-month). By April 2026, local press reported that BTRC had approved a still more aggressive tariff for at least one ISP, with 30Mbps at Tk500, 100Mbps at Tk1,000 and 250Mbps at Tk3,000, subject to a 1:8 shared contention ratio (https://viewsbangladesh.com/btrc-sets-new-internet-prices/).

Those numbers are not Sky Link's own tariff card. Sky Link's sparse public site does not publish a normal retail package table. That absence matters. It prevents a clean revenue model from being built directly from the company's own advertised prices. But the national tariff setting still frames Sky Link's commercial world. A small Dhaka access provider selling into this market cannot behave as though broadband is a premium product. The customer benchmark is a few hundred taka a month. The gross margin must therefore be made in the difference between collected subscription cash and the hidden costs of producing a connection: upstream capacity, local transport, router ports, fibre drops, rooftop visits, power, batteries, support calls, payment fees and disconnections.

Sky Link's own website makes one part of that arithmetic unusually plain. The homepage is styled as a "Skylink Streaming Hub" and carries an internet bill notice in Bangla telling customers to pay on time for uninterrupted service, otherwise software will automatically disconnect the connection. It lists WhatsApp and support at 01904447897 and gives the email address skylinkwari@gmail.com (https://skylinkbd.net/). That is not polished investor copy. It is more valuable than that. It shows the operating nerve of a neighbourhood ISP: collections are not an afterthought, and service continuity is tied directly to whether the monthly bill arrives.

The governing argument follows from those two facts. Sky Link is not best understood as a mini version of a national carrier. It is a local cash-conversion business sitting on top of a real but narrow internet-number footprint. Its value is not the romance of "fibre broadband" in the abstract. Its value is whether it can turn a Tk500-to-Tk1,200 household bill into positive contribution after the work of keeping a line alive in Dhaka's old neighbourhoods. If upstream prices rise, if customers delay payment, if a roof cable is cut, if power quality forces extra battery spending, if a customer leaves after three bad evenings of buffering, the margin can vanish before an outside observer sees anything in a route table.

That makes the small clues important. A national operator can lose a few dissatisfied homes and barely notice. A neighbourhood operator cannot. A large backbone provider can absorb an international cable maintenance window through alternate paths and customer communication. A small access network may inherit the customer's anger while the fault sits upstream. A listed company can spread accounting and regulatory cost across millions of users. A Wari ISP must collect from each lane, each staircase and each shopfront with enough discipline to keep its suppliers paid. Sky Link's public evidence is therefore not a story of grand scale. It is a story of thin but real operating evidence and a market where cheap access is a daily test of working capital.

The identity is local, and the records do not all say the same thing

The strongest identity evidence places Sky Link in Wari, Dhaka. ISPAB's member list shows "Sky Link" with membership number A-125, BTRC licence type Upazila/Thana, address 22/1, Tipu Sultan Road, Wari, Dhaka-1204, email smforhadreza@gmail.com and mobile 8801676664373 (https://ispab.org/members/S?page=3). The same ISPAB page separately lists "Sky Link Broad Band Internet" in Uttara and "SKY Link Net" in Comilla, which is a useful warning against treating every similarly named broadband brand as the same company. For this report, the relevant trail is the Wari Sky Link associated with skylinkbd.net and AS147029, not the Chittagong or Comilla names that appear elsewhere in Bangladesh's crowded ISP landscape.

BTRC's official ISP Upazila/Thana licence list as of 18 December 2024 gives an even more detailed local footprint. Around page 110, it lists Sky Link in Kotwali, Sutrapur and Wari, all at 22/1 Tipu Sultan Road, Wari, Dhaka, with licence numbers ending 676, 677 and 678 and a date of 16.02.2022; the same visible rows show 18-Feb-2023 in the adjacent validity or renewal field (https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/29e9f4bf494145f5bfee76bd1a384ddc.pdf). That does not by itself prove that the licences are current in July 2026. It proves that the official list still carried those Wari-area entries in late 2024 and that any buyer or large customer should ask for renewal proof, payment receipts and the current BTRC status before relying on the licence footprint.

The APNIC trail is consistent on the core identity while differing slightly on address formatting. BGP.tools reports AS147029 as "Sky Link", registered on 13 September 2021, active and allocated under APNIC, with the website https://skylinkbd.net/ (https://bgp.tools/as/147029). APNIC whois output for the abuse and administrator records around AS147029 gives skylinkwari@gmail.com, a Dhaka Tipu Sultan Road address and +8801679841459 as the administrator phone number, while noting that the abuse mailbox was validated on 17 March 2026 (https://wq.apnic.net/apnic-bin/whois.pl?form_type=advanced&searchtext=AS3198-AP). IPinfo's AS147029 page likewise names Sky Link, country Bangladesh, registry APNIC, 512 IPv4 addresses and zero IPv6 addresses in its summary (https://ipinfo.io/AS147029).

This is enough to treat Sky Link as more than a name in a listing. There is a regulator trail, an industry-association trail, a live website, a support contact and a number-resource trail. It is not enough to treat Sky Link as a transparent company. The public sources do not identify shareholders, audited financials, a management team, a subscriber count, a current tariff card, a physical network map or a current licence certificate. The BTRC list and the ISPAB page also use different contact emails. That can be ordinary small-operator messiness, but it is still diligence work. In this market, a local ISP's public identity is often assembled from fragments: a licence row, an APNIC contact, a Facebook page, a billing phone, a website, and a neighbourhood reputation.

The distinction matters because these fragments carry different kinds of risk. A customer cares whether the line works and someone answers the phone. A regulator cares whether the operator is licensed in the relevant thana. An upstream provider cares whether bills are paid and routing is clean. A buyer cares whether the customers, licences, number resources, cable plant and brand can be transferred together. Sky Link's evidence connects those things, but it does not fully bind them. The company looks real. The economic package remains unproven.

AS147029 gives Sky Link substance, but not scale

The public network evidence is the most important reason Sky Link is analytically interesting. PeeringDB lists Sky Link under ASN 147029, identifies the network type as Cable/DSL/ISP, gives the company website override as https://skylinkbd.net/, reports 1-5Gbps traffic, a balanced traffic ratio, geographic scope Asia Pacific and a selective peering policy; it also includes a note saying Sky Link is an ISP company in Bangladesh that wants to peer with content providers to improve customer experience (https://www.peeringdb.com/asn/147029). That sentence is a small but revealing piece of market language. It says the operator understands that customer-perceived quality is not only the amount of upstream bandwidth bought, but where the traffic comes from and how close the popular content sits to the customer.

PeeringDB should be read with care. Its record for Sky Link self-reports 50 IPv4 prefixes and 32 IPv6 prefixes, but the live routing views visible in other sources are narrower. BGP.tools shows AS147029 originating three IPv4 routes and zero IPv6 routes, with the visible prefixes 103.173.60.0/24, 103.173.60.0/23 and 103.173.61.0/24 marked RPKI-valid (https://bgp.tools/as/147029). Hurricane Electric's BGP Toolkit also shows three originated and announced IPv4 prefixes, zero IPv6 prefixes, three RPKI-valid originated routes, 512 originated IPv4 addresses and observed peers including Fiber@Home Global and Bangladesh Submarine Cable Company Limited (https://bgp.he.net/AS147029). IPinfo gives the same practical address size: 512 IPv4 addresses, no IPv6 addresses in its summary and an APNIC allocation date in September 2021 (https://ipinfo.io/AS147029).

The reconciliation is straightforward. PeeringDB tells us how Sky Link or its maintainers described the network at one point; BGP.tools, Hurricane Electric and IPinfo show the smaller live footprint visible from public routing and enrichment systems. A serious customer should not treat the PeeringDB prefix counts as audited capacity. A serious customer also should not dismiss the network as a mere reseller. A /23-equivalent IPv4 footprint with valid route-origin authorization is a meaningful base for a small access network. It gives the operator technical identity, abuse accountability and routing visibility. It does not prove high redundancy, enterprise support, large traffic, direct exchange presence or a large subscriber base.

The upstream view sharpens the interpretation. BGP.tools lists two upstreams for AS147029: AS10075 Fiber@Home Global Limited and AS132602 Bangladesh Submarine Cable Company Limited, while also showing other observed peer entries involving those networks and EXABYTE LTD (https://bgp.tools/as/147029). Hurricane Electric's view similarly places Fiber@Home Global and BSCCL among AS147029's observed IPv4 peers (https://bgp.he.net/AS147029). For a neighbourhood ISP, that is a plausible supplier structure. The last-mile brand sells households and small businesses a local connection. Larger national and international infrastructure companies supply reachability, transport, international bandwidth and route diversity.

This dependence is not a weakness by itself. It is the normal structure of fixed broadband in many countries. The question is whether Sky Link can buy or receive enough upstream quality at a price that leaves room under the retail bill. Fiber@Home Global describes itself as the largest private-sector ITC and IIG operator in Bangladesh (https://www.fiberathomeglobal.net/). Fiber@Home Ltd's broader site describes NTTN, SDH/DWDM, IP/MPLS and GPON-based FTTx infrastructure that helps ISPs deliver IP services to retail users (https://www.fiberathome.net/). BSCCL is the state-linked submarine-cable operator and international bandwidth provider; PeeringDB lists AS132602 as Bangladesh Submarine Cables PLC with 24x7 NOC contact information (https://www.peeringdb.com/net/13680). Sky Link's route table therefore points to real upstream institutions, not anonymous transit.

But it also points to bargaining limits. Sky Link does not look, from public sources, like a network with its own direct BDIX port, colocation list or published backbone map. Its PeeringDB page shows no public peering exchange points and no interconnection facilities for AS147029 (https://www.peeringdb.com/asn/147029). BDIX itself says it has more than 130 peering members and current total average traffic around 250GB, and that it exists to keep local traffic local (https://bdix.net/). If Sky Link reaches local content efficiently, it may do so through upstreams, caches, private arrangements or local servers rather than through a public exchange record under its own name. The customer may not care how YouTube, FTP, gaming or local video arrives. The margin cares a great deal.

Local traffic is the hidden subsidy in cheap broadband

BDIX is not a decorative acronym in Bangladesh broadband. It is one of the mechanisms that makes a low retail bill tolerable. BDIX describes itself as the first internet exchange point in Bangladesh, established to provide physical interconnection so members can exchange and route local traffic locally (https://bdix.net/public-peering/). Its parent SDNF page says membership fees are modest compared with the potential saving of upstream transit costs (https://www.sdnf.org.bd/bdix/). PeeringDB's BDIX exchange page describes it as a large not-for-profit, open and neutral internet exchange provider in Bangladesh, with multiple POP locations including Felicity IDC, Coloasia, Colocity, Dhakacolo and NRB Telecom (https://www.peeringdb.com/ix/2516).

This matters for Sky Link even without proof of a direct BDIX port. The economics of a shared Tk500 connection improve sharply when popular traffic is local, cached or exchanged domestically. A household does not buy "international transit". It buys the feeling that a phone, television and laptop can stream, browse and download without evening collapse. If the heavy traffic comes from local caches or domestic peers, the customer experience can be better than the international bandwidth bill suggests. If it must all cross expensive or congested upstream routes, the same retail package becomes harder to defend.

Sky Link's own public surface leans into local traffic. The website is labelled a streaming hub and points to movie servers, IPTV links, FTP and TV server headings, while giving support and billing contacts (https://skylinkbd.net/). Those headings should not be overread as proof of content rights, CDN contracts or specific internal servers. They are still useful market signals. Many Bangladesh access providers sell not just open internet capacity but a local-entertainment experience: domestic FTP, IPTV lists, faster local resources and lower-latency traffic that feels generous to a household even when the international path is constrained.

The supplier stack decides whether that promise is affordable. If Sky Link can keep a meaningful share of bits inside Bangladesh or inside a low-cost upstream ecosystem, the customer's Tk500-to-Tk1,200 bill can support service. If every evening demand spike turns into international transit, phone support and customer refunds, the economics change. A 10Mbps shared package under a 1:8 contention rule is not a guarantee that every household receives 10Mbps at all times. It is a commercial compromise between affordability and statistical sharing. The operator's skill is in making the compromise invisible most of the time.

That skill is partly technical and partly social. Technically, the provider must shape traffic, maintain routers, provision CPE, watch upstream saturation and manage abuse. Socially, it must teach customers what the plan means, answer calls when local video fails, and collect bills without turning every payment dispute into churn. The website's automatic-disconnection notice is therefore not a harsh aside. It is a margin-control device. A provider charging low monthly fees cannot finance unpaid connections for long, particularly if upstream suppliers and power companies expect payment on time.

The same logic explains why the most important Sky Link source is not a glossy plan table. A plan table would tell us what the company hopes to sell. The route table tells us whether it has its own public network identity. The payment notice tells us the business knows the danger of arrears. The BTRC and ISPAB rows tell us where it is meant to operate. Together, they describe a local ISP whose value depends less on headline speed than on the practical ability to convert neighbourhood demand into predictable cash.

The field operation is where the gross margin leaks

A broadband connection in Wari is physical before it is digital. Someone has to run fibre or copper to a building, get permission from a landlord or shopkeeper, climb a stairwell or rooftop, fix an optical terminal or router, replace a damaged drop, answer a phone call, collect or reconcile a payment, and return when the same customer says the connection is slow after sunset. These tasks do not show up in AS147029's route table. They are exactly where a small ISP's profit leaks.

The Bangladesh policy debate points in the same direction. In the same TBS report that covered the Tk500 speed increase, Faiz Ahmed Taiyeb, acting head of the Posts and Telecommunications Division, said all telecom fibre networks in metro areas must be fully underground within three years and that overhead fibre cannot be considered telecom-grade (https://www.tbsnews.net/bangladesh/tk500-broadband-connection-speed-doubles-10-mbps-1119976). For a large carrier, undergrounding is a capital-planning problem. For a thana-level ISP, it can be an existential shift in cost structure. Overhead cable is cheap, quick and vulnerable. Ducted or formally shared infrastructure is cleaner and more resilient, but it requires access rights, coordination and cash.

Power is another quiet cost. Bangladesh power tariffs were raised in 2026, with Dhaka Tribune reporting that commercial and office users would pay Tk15.36 per unit after the adjustment (https://www.dhakatribune.com/bangladesh/power-energy/411666/how-much-does-electricity-tariff-increase-per). That rate is not a precise Sky Link energy bill. It is a useful proxy for the pressure on a small network site, office or support point. A single 100-watt device running all month consumes roughly 72 kWh. At Tk15.36 per kWh, that is about Tk1,106 before other charges. If a small cluster of routers, switches, optical gear, fans and charging equipment sits behind a handful of low-priced customers, electricity and battery replacement can consume the gross contribution from several lines before labour is counted.

The labour calculation is just as unforgiving. A technician visit that takes 90 minutes across travel, access and testing may cost more in wage time, transport and opportunity cost than the operator earns from that customer's monthly fee. Yet the visit cannot always be avoided. A neighbourhood ISP competes on intimacy. If a customer cannot reach a national call centre, the local provider's advantage is that someone nearby can come. If the local provider stops coming, the customer wonders why it is paying at all.

This is why support language matters. Sky Link's site lists phone support, WhatsApp support and email support, and the public Facebook page positions the brand in Wari, Dhaka with a consumer sales message around fast internet and the same skylinkbd.net domain (https://www.facebook.com/skylink2016/). These are unofficial market signals, not audited operational metrics. They nevertheless show that the business is sold through direct customer contact. The economic cost of that model is that a technical issue becomes a human obligation quickly. A call unanswered at 9pm is not just bad service; it can become a missed bill next month.

The attraction of the model is that local trust can beat national scale in narrow pockets. A household may tolerate less formal paperwork if the installer is known, the support number answers and local video works. A small business may prefer the provider that can send a person to the shop rather than a national brand that logs a ticket. Sky Link's challenge is to keep that trust cheap enough to produce margin. Rooftop cable, small batteries, routers, poles, messenger payments and support labour are not glamorous assets. They are the operating surface on which the business lives.

The failure scenario: upstream stress turns into household churn

The most plausible failure scenario is not a spectacular collapse of AS147029. It is an upstream shock that becomes a customer-retention problem. Imagine a week in which one international path is under maintenance, one upstream tightens credit or capacity terms, and evening demand rises because local customers are streaming video. Bangladesh has recent examples of this kind of pressure. The Business Standard reported that internet services across Bangladesh could slow between 9 and 13 April 2026 because of maintenance on a major international submarine cable, with backup traffic routed through other capacity (https://www.tbsnews.net/bangladesh/get-ready-slower-internet-9-13-april-1406306). Internet Society's 2024 analysis of a SEA-ME-WE-5 outage said Bangladesh was experiencing slower connectivity to international hosted services and was helped by Indian terrestrial cables and local content caches (https://pulse.internetsociety.org/en/blog/2024/04/bangladesh-coping-with-submarine-cable-outage-thanks-to-indian-terrestrial-cables-local-content-caches/).

Now put that stress onto Sky Link's economics. The customer does not call BSCCL, Fiber@Home Global, a submarine cable consortium or an exchange. The customer calls Sky Link. If local content still works but overseas video calls are poor, the complaint is still made to the retail provider. If the operator buys more capacity for a few days, the extra cost may not be recoverable from Tk500, Tk800 or Tk1,200 household bills. If it does not buy more, some customers decide the provider is weak. If a technician is sent to a rooftop for a problem that is really upstream, the visit burns labour without fixing the underlying constraint. If enough customers delay payment after a bad week, the automatic-disconnection notice becomes a churn trigger rather than a collections tool.

The upstream market also carries financial risk. The Daily Star reported in February 2025 that 29 IIG operators still owed about Tk205 crore to BTRC, with dues including regular payments and Social Obligation Fund contributions, while Coronet Corporation Limited had cleared a much smaller Tk3 lakh due (https://www.thedailystar.net/business/economy/news/29-iig-operators-still-owe-tk-205cr-btrc-3825121). That article is not evidence that Sky Link's visible upstreams were in default, and it should not be used that way. It does show that Bangladesh's middle layer can be financially tense. A local ISP living under that layer inherits pricing, credit and capacity risk even when the customer sees only the retail name.

This failure scenario changes the economics rather than merely listing risks. The gross margin of cheap broadband depends on a high share of normal days. On normal days, the same installer can activate new customers, local caches keep video tolerable, bills arrive, and upstream capacity is statistically shared. During a stressed week, the same cost base becomes defensive: support calls rise, technicians revisit old customers, WhatsApp fills with complaints, collections weaken, and some households compare alternatives. A small operator can survive one such week. Repeated shocks turn a cheap plan into a negative-margin product.

The more subtle risk is reputation. A national outage can be explained in the news, but customers rarely allocate blame accurately. In a dense Dhaka neighbourhood, one frustrated shop owner tells another. If a competitor offers a promotion, free installation, a better router or a "real IP" upsell, the switching cost may be low. Sky Link's local advantage is proximity. Its local vulnerability is that failure is also proximate.

Customer loss is a substitution story, not just a complaint story

For a retail broadband provider, churn does not always come from anger. It often comes from substitution. A household upgrades to a larger national ISP because it wants app-based billing, a recognizable brand or a bundle. A student leans more on mobile data because exams and classes demand mobility. A shop moves to a provider that includes a static IP, faster BDIX access, a better router or easier digital payment. A small office decides that if internet is essential, it should pay for a dedicated business connection rather than a shared household line. The customer does not need to prove Sky Link is bad. The customer only needs a better fit.

This is why Sky Link's lack of a public package table is commercially relevant. A transparent retail page can reduce friction for new customers and make the offer comparable. Sky Link's current site tells users how to pay and reach support, but it does not show a clean set of plans, installation terms, contention, upload speeds, fair-use terms, static IP pricing or service boundaries (https://skylinkbd.net/). That may not hurt if sales are done by phone and neighbourhood referral. It does matter for customers who compare providers before calling.

The wider market is crowded and increasingly formal. Bangladesh's fixed-line internet base remains far smaller than mobile internet, but it is large enough to support many local operators. BSS reported from BTRC data that by March 2026 fixed-line internet subscribers, including ISP and PSTN users, remained at about 1 crore 47 lakh 50 thousand during the quarter (https://www.bssnews.net/news-flash/385567). TBS reported that around 65% of 1.4 crore fixed broadband connections would benefit from the 2025 speed enhancement (https://www.tbsnews.net/bangladesh/tk500-broadband-connection-speed-doubles-10-mbps-1119976). A large pool of shared residential users is attractive, but it also means customers have learned to treat broadband as a replaceable household utility.

The substitution risk is different for a larger customer. A small business, school, clinic or office would not leave merely because a rival plan is cheaper. It would leave if support is not documented, if outages affect payment terminals or cloud software, if a static IP or business-grade router is unavailable, if the provider cannot prove route diversity, or if the contract is too informal for procurement. Sky Link's public evidence does not show an enterprise-service page, named corporate customers or SLA terms. That does not mean such services do not exist. It means a serious buyer would have to request private evidence.

The customer-loss mechanics also cut the other way. Local providers can be sticky because the customer tie is personal. The installer knows the building. The support number is not remote. A technician can identify whether the problem is inside the flat, on the roof, in the local splitter, at the upstream handoff or in a power backup. That familiarity has economic value only if it saves more customers than it costs in labour. Sky Link's margin is therefore a discipline problem: keep enough personal service to defend the base, but not so much bespoke attention that every Tk500 connection becomes a managed service.

Competition is governed by regulation as much as brand

Sky Link competes in a market where the regulator shapes both retail price and operating form. The "One Country, One Rate" framework reduced the scope for simple geographic price discrimination. The more recent talk of stronger performance indicators, fibre undergrounding and licensing reform pushes in the direction of more formal operations. That can help consumers and cleaner operators. It can also squeeze small providers that built their economics around informal overhead fibre, local cash collection and low-documentation service.

The BTRC broadband connectivity report, published through official government storage, described Bangladesh's October 2024 ISP and PSTN user base at 13.74 million and set future targets that include higher minimum national broadband rates, wider FTTH coverage and tariff affordability goals (https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/2553c9a48743467faaa8b420c2e6ecb5.pdf). The same policy direction that raises the market's quality floor can undermine operators that lack capital. More underground fibre, stronger QoS metrics, better route hygiene and transparent tariffs are good for the sector. They are not free.

The ownership and supplier question remains open. Coronet Corporation Limited is relevant market context because it is a real Bangladesh IIG and internet-services company, with a PeeringDB profile for AS149765 that describes high traffic scale and IIG/nationwide ISP activity (https://www.peeringdb.com/net/32178). But the public routing sources reviewed here do not show Coronet as a current visible upstream for AS147029. Public BGP views instead point to Fiber@Home Global and BSCCL (https://bgp.tools/as/147029; https://bgp.he.net/AS147029). Coronet's presence in the wider Bangladesh ISP ecosystem therefore does not prove control, ownership or current transit for Sky Link.

This distinction is important in Bangladesh because the middle-mile and retail layers are intertwined in public perception. Customers often know the local ISP name but not the upstream. Regulators know the licence categories. Route observers know the AS paths. Suppliers know who pays the bill. A credible economic assessment has to keep those views separate. Sky Link can be a local access provider with its own APNIC identity while depending on larger upstreams. It can appear beside Coronet in a market clue without being controlled by Coronet. It can sell local internet service without having public evidence of direct exchange membership.

The best competitive position for Sky Link would be modest but defensible: a tight Wari footprint, clean licence renewals, a stable AS147029 route table, enough upstream diversity to handle maintenance windows, quick local support, transparent billing and clear package terms. The worst position would be the same footprint without documentation: stale licence fields, no published tariff, no proof of cable rights, no current supplier contracts, no subscriber ledger and no explanation of who controls the number resources. The difference between those two positions may not be visible to a retail customer. It would be very visible to a lender, acquirer or regulator.

What capital would underwrite, and what it would reject

A buyer would not pay much for the name "Sky Link" alone. It would pay for active paying subscribers by location, low churn, documented collection history, current BTRC licences in Kotwali, Sutrapur and Wari, control over AS147029 and 103.173.60.0/23, valid RPKI authority, current APNIC contact control, supplier contracts with upstream providers, equipment inventory, building access rights, pole or duct permissions, support records and a credible path to comply with any underground-fibre requirement. The public evidence supports only the starting claim that these are plausible assets. It does not prove their size or transferability.

A lender would be more cautious. It would not treat 512 IPv4 addresses as collateral without knowing whether they are transferable, encumbered, assigned to customers or tied to APNIC membership obligations. It would not treat a licence list row as current authority without a renewal certificate. It would not treat PeeringDB's 1-5Gbps traffic band as audited revenue. It would ask for monthly bandwidth invoices, electricity bills, support payroll, customer receipts, arrears ageing, router logs, fibre-repair spending and complaint rates. It would also ask what happens if Fiber@Home Global or BSCCL changes price, payment terms or route policy.

A large customer would ask different questions. It would want a service boundary. Is the line shared or dedicated? What is the upload speed? Is there backup upstream? Is there a static IP? How quickly does support arrive? Are there written terms for outage credit? Does the provider have permission to run cable into the building? Can the provider show that route-origin authorization and abuse contacts are under control? Can the customer speak to a technical person rather than only a sales number? Sky Link's public site answers only a fraction of those questions.

A regulator would care about the public side of the same facts: licence status, customer-protection terms, tariff compliance, undergrounding or infrastructure-sharing rules, lawful interception and abuse handling, revenue reporting, and whether a small ISP is using number resources and local infrastructure responsibly. The APNIC abuse contact validated in March 2026 is a positive sign (https://wq.apnic.net/apnic-bin/whois.pl?form_type=advanced&searchtext=AS3198-AP). The old-looking licence validity fields in the BTRC list are the counterweight (https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/29e9f4bf494145f5bfee76bd1a384ddc.pdf). The company would benefit from making current licence and package details easier for the public to verify.

The one fact that would most change the judgement is a verified monthly operating pack tying together active subscribers, average revenue per user, churn, upstream Mbps purchased, actual upstream cost per Mbps, support visits, power spend, licence status and arrears. A close second is a current BTRC renewal confirmation for the Wari-area licences. With those facts, Sky Link could be valued as a small but operating local ISP. Without them, it remains a real network and local brand whose economic weight is uncertain.

Public evidence register

The core identity evidence is the Sky Link website, which supports the Wari support and payment-discipline reading through its customer notice and contact details (https://skylinkbd.net/). ISPAB's member page supports the Wari Sky Link association, membership number A-125 and Upazila/Thana licence type (https://ispab.org/members/S?page=3). BTRC's December 2024 ISP licence PDF supports the Kotwali, Sutrapur and Wari Sky Link licence rows at 22/1 Tipu Sultan Road (https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/29e9f4bf494145f5bfee76bd1a384ddc.pdf).

The network evidence is PeeringDB for AS147029, which supports the Cable/DSL/ISP classification, self-reported traffic band, website and peering note (https://www.peeringdb.com/asn/147029); BGP.tools for AS147029, which supports the active APNIC-allocated status, visible upstreams and live prefixes (https://bgp.tools/as/147029); Hurricane Electric's AS147029 page, which supports the three originated IPv4 prefixes, RPKI-valid status and observed peers (https://bgp.he.net/AS147029); IPinfo, which supports the 512 IPv4 address summary and zero IPv6 summary (https://ipinfo.io/AS147029); and APNIC whois output for the Sky Link abuse and administrator contacts (https://wq.apnic.net/apnic-bin/whois.pl?form_type=advanced&searchtext=AS3198-AP).

The market-economics evidence is The Business Standard's April 2025 account of Tk500 shared broadband speed doubling and Bangladesh's 1:8 contention rule (https://www.tbsnews.net/bangladesh/tk500-broadband-connection-speed-doubles-10-mbps-1119976), Dhaka Tribune's 2021 "One Country, One Rate" tariff report (https://www.dhakatribune.com/business/248823/broadband-internet-price-set-at-tk500-a-month), and Views Bangladesh's 2026 report on newly approved broadband tariffs with 30Mbps at Tk500 for one provider (https://viewsbangladesh.com/btrc-sets-new-internet-prices/). BDIX's public pages support the local-peering and transit-saving context (https://bdix.net/, https://bdix.net/public-peering/, https://www.sdnf.org.bd/bdix/). Fiber@Home Global and Fiber@Home official pages support upstream and national transport context (https://www.fiberathomeglobal.net/, https://www.fiberathome.net/). BSCCL and PeeringDB sources support submarine-cable and IIG context (https://bsccl.com.bd/, https://www.peeringdb.com/net/13680). TBS, Internet Society and The Daily Star support the outage, cable-maintenance and IIG financial-pressure scenarios (https://www.tbsnews.net/bangladesh/get-ready-slower-internet-9-13-april-1406306, https://pulse.internetsociety.org/en/blog/2024/04/bangladesh-coping-with-submarine-cable-outage-thanks-to-indian-terrestrial-cables-local-content-caches/, https://www.thedailystar.net/business/economy/news/29-iig-operators-still-owe-tk-205cr-btrc-3825121).

The judgement

Sky Link deserves to be tracked because the public record supports a real local Bangladesh ISP identity: Wari licence rows, ISPAB membership, a customer support surface, AS147029, APNIC contact control, valid IPv4 route-origin evidence and visible upstream relationships. The company also deserves a restrained judgement because the public record does not support claims of large scale, direct exchange presence, strong enterprise service, transparent current tariffs, current licence renewal, audited finances or confirmed ownership.

The economics are less mysterious than the company is. Cheap broadband in Bangladesh is a narrow-margin craft. The customer sees a low bill and expects video, games, classes, calls and downloads to work. The operator sees upstream invoices, local caches, cable faults, power bills, router ports, payment delays and churn. Sky Link's advantage, if it has one, is that it can be close to the customer and visible in the routing system at the same time. Its weakness is that closeness is expensive and routing visibility is not the same as scale.

The most honest conclusion is that Sky Link is a credible micro-ISP candidate, not a proven infrastructure prize. It may be valuable inside a small geographic pocket if its active customer base, licence renewals and supplier contracts are sound. It may be fragile if the same evidence turns out to be stale, informal or dependent on narrow upstream goodwill. The Tk500 bill tells the story. There is room for profit only if the operator does many small things right every month, and almost no room if it has to keep doing them twice.