PABNA ONLINE: District Broadband and Transit Dependency

Summary

What it says: PABNA ONLINE, a small district ISP in Pabna, Bangladesh, illustrates the microeconomics of local fixed broadband, with a single-homed network, thin margins, and critical dependency on its upstream provider Windstream.
Main topic: Regional ISP economics; Peering and transit
Context: Telecommunications / Company research / Asia-Pacific

PABNA ONLINE: District Broadband, Transit Dependency, and Thin-Margin Economics of Local Internet Trust in Bangladesh

Thesis

PABNA ONLINE should be understood not as a conventional technology company but as a small addressable unit in the hyper-local broadband economy of Bangladesh. Public archives resolve a material record: AS150745, registered in APNIC under the name PABNA ONLINE, with organization identifier ORG-PO5-AP, an APNIC contact address at Hazi Akbar Ali Super Market in Pabna, and an abuse mailbox on the domain pabnaonline.net.

The network visible in routing tables originates a single IPv4 /24 prefix, 103.81.29.0/24, has valid RPKI origin status in public BGP collectors, has no IPv6 origin visibility, and is observed as single-homed via Windstream Communication Limited, AS139009. It is a small "eyeball" access network rather than a backbone operator.

Its economic significance therefore lies not in the absolute scale of its IP holdings but in what those holdings reveal: a local ISP can be both tiny in global routing tables and commercially important in a dense district market where household demand, on-site installation, bill collection, upstream procurement, local reputation, and regulatory licenses determine survival.

The central intelligence value of PABNA ONLINE is that it exposes the microeconomics of the fixed broadband layer in Bangladesh. A district or thana ISP sells low-price monthly connectivity, but it buys from a vertically constrained stack: international internet gateway capacity, national or shared transmission, domestic exchange connectivity, backup power, customer installation, payment collection, and technical labor. Its revenue is retail and local; its cost base is partly wholesale and national. This mismatch gives the customer-facing operator limited pricing power.

It also makes a small provider's gross margin fragile: the business must advertise more speed over time at similar prices, manage churn, absorb installation incentives, and negotiate upstream access from a position of limited traffic scale.

The evidence is not sufficient to write a clean corporate biography. That itself is part of the finding. The name "Pabna Online" appears in at least two adjacent public identities. The APNIC entity and AS150745 is PABNA ONLINE at Hazi Akbar Ali Super Market. Bangladesh regulatory records also list "Pabna Online" and "Pabna Online Network" as distinct Upazila/Thana license holders at Pabna Sadar, with different addresses and license numbers, both expiring in 2027.

A modern, active website under pabnaonline.net.bd presents "Pabna Online Network" or PON as a fiber ISP in Pabna Sadar, claims over 5,000 connected users, advertises customer self-service, online payments, and a plan scale from 20 Mbps to 80 Mbps, and names individual executives. An older website pabnaonline.net presents Pabna Online as an ISP at the Hazi Akbar Ali Super Market address and advertises plans from 15 Mbps to 80 Mbps. The evidence supports name proximity and market overlap; it does not conclusively prove common ownership or legal continuity.

This distinction matters economically because the answer changes depending on whether AS150745 is the retail operational core, a legacy network entity, a predecessor, or a network resource layer alongside a distinct licensed PON retail entity.

What the public archives prove

The registry record is clear. The APNIC Whois record for AS150745 identifies the aut-num as AS150745, as-name PABNAONLINE-AS-AP, description PABNA ONLINE, country Bangladesh, organization ORG-PO5-AP, and maintenance entities under MAINT-PABNAONLINE-BD. The same record provides the incident response entity IRT-PABNAONLINE-BD, the mailbox[email protected], and a contact address at Hazi Akbar Ali Super Market, Pabna. The organization entity lists PABNA ONLINE as an APNIC LIR in Bangladesh. These facts make the canonical infrastructure identity PABNA ONLINE, not merely a directory label.

BGP visibility narrows operational interpretation. BGP.tools reports AS150745 as active, allocated by APNIC, registered on 1 February 2023, classified as an eyeball network, originating one IPv4 prefix and no IPv6 prefixes. Its sole originated block is 103.81.29.0/24 and the listed upstream is AS139009, Windstream Communication Limited. Hurricane Electric's BGP view independently shows one IPv4 prefix originated, zero IPv6 prefixes originated, one IPv4 peer observed, 256 IPv4 addresses originated, and valid RPKI-originated status for the visible route.

IPinfo similarly characterizes AS150745 as a stub AS, single-homed, not providing transit, with the same 103.81.29.0/24 block and Windstream as the sole upstream and listed peer.

The IP block record connects the routing entity to retail geography. AbuseIPDB's APNIC-derived WHOIS page for 103.81.29.25 identifies the ISP as PABNA ONLINE, usage type as Fixed-line ISP, domain name pabnaonline.net, country Bangladesh and city Pabna in Rajshahi division. Its raw Whois output shows 103.81.29.0–103.81.29.255, netname PABNAONLINE-BD, description PABNA ONLINE, and allocated non-portable status. This means the block is not simply a generic hosting allocation; it is visibly attached to a local fixed broadband access provider.

Bangladesh's official license list adds the missing regulatory layer. The BTRC Upazila/Thana ISP license list shows "Pabna Online" in Pabna Sadar thana at Holding No. 252, Hazi Akbor Ali Super Market, Parbotiganj, Pabna, with license number 14.32.0000.702.47.596.22.144 and validity until 2 March 2027. It separately lists "Pabna Online Network" in Pabna Sadar thana at Holding 1462/1, Shop no-01, Radhanagar Pabna, with license number 14.32.0000.702.47.645.22.183 and validity until 11 April 2027. The same page also lists Pabna Fiber Net in Pabna Sadar, underscoring that the local market has multiple licensed operators.

The result is a two-layer identification. The hard network target is PABNA ONLINE, AS150745, ORG-PO5-AP, 103.81.29.0/24, Pabna. The wider commercial target includes a Pabna Online Network identity with an adjacent name that could be a linked successor, a parallel retail brand, or a distinct competitor using a similar geographic name. For economic analysis, both must be observed, but they cannot be merged without additional legal or ownership evidence.

Name ambiguity and why it matters economically

For a large listed telecom company, name ambiguity is generally a data-quality nuisance. For a local ISP, it is a central economic variable. Customer acquisition is local, and local customers often identify the provider by the shop address, the technician's phone number, the cable route, the help desk, the Facebook page, or the bill payment portal rather than a precise legal corporate name. A broadband customer in Pabna Sadar may experience "Pabna Online" as the installer who answers the phone, the customer ID used for payment, the local fiber termination, and the person who restores service after a line cut.

This is a different kind of brand capital from a national telecom brand. It is narrow, relational, and operational.

The older pabnaonline.net site and the APNIC entity align around the Hazi Akbar Ali Super Market and the pabnaonline.net domain. The website advertises Pabna Online as an internet service provider, lists broadband, FTP and IPTV services, shows plan pricing, gives customer support and account management phone numbers, and places the office on the third floor of Hazi Akbar Ali Super Market on Abdul Hamid Road, Pabna. This address matches the APNIC registry address and the BTRC license listing for Pabna Online.

The newer pabnaonline.net.bd site presents Pabna Online Network, or PON, as a local fiber ISP in Pabna Sadar. It advertises high-speed fiber for households in Pabna Sadar, claims over 5,000 connected users, describes transparent support and billing, lists Pabna Sadar service areas, and provides self-service, fast payment, and mobile payment flows. The ISPAB member directory lists Pabna Online Network as an Upazila/Thana license holder with membership C-622, mobile number 8801703021101, website pabnaonlinenetwork.net, BTRC license number 14.32.0000.702.47.645.22.183, trade license 01515, and no directors on file.

This BTRC license number matches the distinct Pabna Online Network entry in the BTRC list, not the Pabna Online entry.

Three hypotheses follow. The first is separation: PABNA ONLINE and Pabna Online Network are legally distinct operators in Pabna Sadar with similar names. In that case, AS150745 belongs to the Hazi Akbar Ali Super Market operator, while the website and customer portal claims of PON should be read as competitive market evidence, not as PABNA ONLINE's own operational scale. The second is succession or affiliation: PON could be a newer retail brand or a linked entity while the AS150745 entity remains with the older PABNA ONLINE registry identity.

In that case, the economic unit could be larger than AS150745 alone suggests, with the ASN acting as a resource within a larger local customer book. The third is functional separation: one identity could own the network resources and another the retail operations, customer billing, or field deployment. This would be economically plausible in a fragmented local ISP market, but it is unproven by public material.

The unresolved identity question changes the evaluation logic. If PABNA ONLINE and PON are separate, then the AS150745 enterprise is likely a small licensed ISP with a /24 and a legacy web presence, facing both national and local competition. Its bargaining power with Windstream is likely limited. If the identities are linked, PON's claim of over 5,000 connected users, though self-reported, would imply a more significant traffic aggregate, better procurement leverage, and stronger local switching frictions.

If the ASN is a legacy or wholesale support entity behind another retail brand, then BGP visibility may understate customer reach but overstate operational independence.

Operational footprint and market layer

The physical footprint is district-scale. The official license category for both Pabna Online and Pabna Online Network is Upazila/Thana, a geographically constrained ISP category. The BTRC ISP guideline categorizes licenses into Nationwide, Divisional, District, and Upazila/Thana, with the Upazila/Thana license authorizing service in a particular administrative area. This is not a licensing architecture designed for unrestricted national expansion. It is an architecture that formalizes local monopolies, local overlaps, and local density plays.

The BTRC framework also limits what an access ISP can own or do. The guideline stipulates that no person or business entity may build, maintain, or operate ISP systems and services without a license. It also stipulates that the licensee must lease or sub-lease the transmission network from NTTN operators unless NTTN service is unavailable, and that last-mile connectivity is limited to approximately 3 km in metropolitan areas and 6 km in other locations. The same framework permits fixed internet/data services and Wi-Fi-based fixed data services subject to instructions and authorizations.

These provisions matter because they transform the local ISP into a last-mile and customer management specialist rather than a fully integrated national infrastructure company.

In practical terms, PABNA ONLINE's visible network layer is below the IIG and above the household. It is an access provider: it takes upstream connectivity, distributes it locally, bills households or small businesses, and tries to differentiate through support, local content access, domestic exchange performance, and neighborhood reliability. Its public ASN and /24 give it a routable identity, but they do not prove ownership of long-distance fiber, data centers, towers, or national backbone assets.

Under the BTRC framework, the rational presumption is that local transmission and upstream functions depend on licensed wholesale layers and infrastructure-sharing arrangements unless specific contrary evidence emerges.

The retail evidence reinforces this access-layer interpretation. The older PABNA ONLINE website advertises broadband internet, FTP service, and IPTV service, and lists several FTP and IPTV resources. Its plans separate "regular" bandwidth from higher speeds for YouTube, Facebook, movie server, and BDIX. This product design is typical of a local ISP economy in which international bandwidth is expensive relative to domestic or cached traffic, and the customer experience is sold by content-category performance rather than a single undifferentiated Mbps figure.

The newer PON site is more explicitly a fiber proposition for households and small businesses. It claims high-speed fiber service to Pabna Sadar, lists active service areas including Radhanagar, Abdul Barek Road, Maktab More, Pabna Sadar, Shalgaria, and Abdarpur, and describes payment via mobile banking, card, or self-service portal. Its top plan includes static IP availability and a dedicated support line for business and enterprise users. It is not a full enterprise carrier proposition; it is a retail access business adding a small enterprise upsell on top of a residential base.

The network layer: a visible /24, not a moat

The network facts of AS150745 are modest but important. A single IPv4 /24 is the minimum globally routable unit that many networks can practically announce without being filtered by others. It gives PABNA ONLINE 256 IPv4 addresses, enough to run infrastructure, assign some public addresses, deliver static IP products selectively, and maintain a visible routing identity. It is not enough to give every household a public IPv4 address if the customer base is several thousand.

Therefore, the likely access architecture, as with many small ISPs, involves private addressing and carrier-grade NAT for much of the residential base, with public IPv4 sold or reserved for business customers, routers, servers, or special cases. The sources prove the /24; the CGNAT inference follows from arithmetic and market practices, not from a direct public configuration record.

The valid RPKI signal matters commercially. Both BGP.tools and Hurricane Electric show a valid RPKI origin status for the visible 103.81.29.0/24 origination. For a small district ISP, a valid route origin authorization is not a revenue product by itself, but it improves routing hygiene and reduces the risk that upstreams or route filtering networks will reject the announcement. It is a small but real sign that the operator or its upstream support stack has kept modern routing registration in order.

The absence of IPv6 visibility is more telling. AS150745 originates no IPv6 prefixes in the public BGP views examined. In a mature access network, IPv6 can reduce pressure on scarce IPv4 addresses and improve long-term address scaling. In a small ISP, however, IPv6 deployment requires router configuration, customer equipment compatibility, support training, monitoring, and sometimes customer education. The economic problem is that IPv6 saves future address cost but does not obviously increase a household's Tk 500 or Tk 800 monthly bill today.

The absence of visible IPv6 is therefore unsurprising; it is a rational symptom of thin-margin retail broadband where immediate customer acquisition and support costs outweigh long-term protocol modernization.

The single-homed upstream position is the strongest indicator of bargaining fragility. IPinfo classifies AS150745 as a stub AS and a single-homed network that does not provide transit; both BGP.tools and Hurricane Electric show Windstream, AS139009, as the observed upstream or peer. Single-homing reduces complexity and cost: one upstream contract, one operational relationship, one route policy, fewer interconnections, and fewer debugging variables. But it also lowers resilience and weakens negotiation.

If Windstream raises the effective price, experiences congestion, changes terms, or suffers an upstream event, PABNA ONLINE has limited visible route redundancy.

Windstream is structurally larger than PABNA ONLINE. BGP.tools shows AS139009 with hundreds of peers, multiple upstreams, and a wide downstream cone, and its downstream table includes AS150745. PeeringDB lists Windstream as AS139009, IRR as-set AS-WCL-BD, network type NSP, traffic level 1–5 Tbps, geographic scope Asia-Pacific, and notes it is connected with several CDNs and IXs. Windstream's website describes it as a private limited company incorporated under the Bangladesh Companies Act 1994 and an IIG enterprise.

The asymmetry is clear: PABNA ONLINE brings local eyeballs; Windstream brings gateway scale, international paths, CDN and IX relationships, and a broader negotiation surface.

Live directory hints involving PEEREX and Coronet must be read with caution. PABNA ONLINE appears in BGP.tools as a member of several AS-SETs, including as137491:as-customers and as149765:as-coronetiig-bd, but this is not the same as a current observed upstream path. Peerex Networks Ltd., AS137491, is described in public BGP data as PEEREX NETWORKS LTD./IIG with upstreams including BSCCL, Tata, Bharti Airtel, and Hurricane Electric, while Coronet Corporation Limited, AS149765, is a larger infrastructure provider with an AS-SET named AS149765:AS-CORONETIIG-BD and PeeringDB notes describing it as operating an IIG and a nationwide retail ISP.

These are significant ecosystem adjacencies and possible historical, indirect, or routing policy relationships. They do not overturn the current collector evidence that AS150745 is route-visible via Windstream as the sole observed upstream.

Services and customer proposition

PABNA ONLINE's retail proposition is utility-driven rather than brand-driven. The older pabnaonline.net site sells plans starting at 15 Mbps for Tk 500 per month and going up to 80 Mbps for Tk 2,000 per month. Each plan includes a regular bandwidth figure and higher advertised speeds for YouTube, Facebook, movie server, and BDIX traffic. The 15 Mbps plan advertises 30 Mbps YouTube, 25 Mbps Facebook, 25 Mbps movie server, and 25 Mbps BDIX at Tk 500. The 80 Mbps plan advertises 150 Mbps YouTube, 120 Mbps Facebook, 130 Mbps movie server, and 130 Mbps BDIX at Tk 2,000.

This pricing design is economically rational in the Bangladesh fixed broadband market. Customers perceive video, social media, gaming, domestic content, and streaming as the service. The operator's cost for each category differs. Domestic traffic and cached content can often be delivered at a lower marginal cost than uncached international traffic if the ISP has favorable upstream, NIX, BDIX, or cache access. A provider can therefore advertise higher category-specific speeds while keeping the "regular" internet figure lower.

The product is a cost-discrimination device: it matches customer-visible performance to cheaper traffic classes while protecting the operator from unlimited, expensive international usage.

The PON plan scale, if treated as linked or at least market-adjacent evidence, shows speed inflation and price compression. Its site advertises all plans as unlimited with no FUP, starting at 20 Mbps for Tk 500 and increasing to 26 Mbps for Tk 600, 42 Mbps for Tk 800, 55 Mbps for Tk 1,000, 65 Mbps for Tk 1,200, and 80 Mbps for Tk 1,500. The top plan is positioned for business and enterprise users and includes static IP availability and a dedicated support line. Compared to the older PABNA ONLINE scale, the new scale gives more regular Mbps at the same or lower monthly price, particularly at mid and upper tiers.

The feature stack extends beyond Mbps. PON advertises free installation, a 10 Mbps upgrade bonus for three months when existing customers upgrade, a Tk 100 referral discount, fast payment by customer ID, mobile banking, and card, SMS bill reminders, account self-service, usage monitoring, billing history, and plan changes. These features create two kinds of value: they reduce collection friction for the operator, and they make the customer relationship more legible. A self-service portal and a customer ID do not eliminate churn, but they turn an informal neighborhood service into a managed recurring-revenue system.

The FTP and IPTV features point to the local-content economics of small ISPs. PABNA ONLINE's website lists FTP servers and IPTV links, while the BTRC ISP guideline stipulates that ISP licensees are permitted to provide IPTV services subject to the conditions of the Ministry of Information and that the commission may determine tariffs when necessary. The business interpretation is that local ISPs try to increase perceived value with video, FTP libraries, and domestic content that can be cheaper to serve than international bandwidth.

The regulatory interpretation is that such services sit inside a controlled licensing and content regime, not an unconstrained retail bundle.

Business model and revenue logic

The simplest revenue model is monthly access ARPU multiplied by active subscribers, with installation, reconnection, static IP, small business support, and possibly device fees as supplements. At the older PABNA ONLINE plan prices, regular bandwidth costs the customer about Tk 33 per advertised Mbps at 15 Mbps/Tk 500, Tk 32 per Mbps at 25 Mbps/Tk 800, Tk 25 per Mbps at 40 Mbps/Tk 1,000, Tk 24 per Mbps at 50 Mbps/Tk 1,200, Tk 25 per Mbps at 60 Mbps/Tk 1,500, and Tk 25 per Mbps at 80 Mbps/Tk 2,000.

At the newer PON pricing, the scale compresses towards roughly Tk 25 per Mbps at 20 Mbps/Tk 500, Tk 23 per Mbps at 26 Mbps/Tk 600, Tk 19 per Mbps at 42 Mbps/Tk 800, Tk 18 per Mbps at 55 Mbps/Tk 1,000, Tk 18.5 per Mbps at 65 Mbps/Tk 1,200, and Tk 18.75 per Mbps at 80 Mbps/Tk 1,500. The direction is unmistakable: customers receive more advertised speed per taka over time.

This compression does not mean the operator's cost per delivered Mbps also falls. International transit, domestic transport, NTTN lease, customer support, payment fees, electricity, backup power, fiber repairs, and technical labor all shape unit economics. Some costs vary with traffic, some with subscriber count, some with route length, and some with outage events. A local ISP can survive if oversubscription, caching, and domestic traffic keep peak bandwidth costs below monthly ARPU. It struggles when traffic growth, speed promises, and service expectations increase faster than gross margin.

Installation economics are the hidden balance sheet. A free installation offer can win customers, but the operator still bears the drop fiber, connectors, ONU/router coordination, technician time, travel, pole or building access, and activation support. If the customer stays two years, the acquisition cost amortizes. If the customer churns after a few months, the economics break. That is why referral credits, SMS reminders, self-service portals, and local support matter. They are not peripheral marketing features; they are mechanisms to reduce churn and collection leakage.

Customer switching cost is neither zero nor very high. A household can change broadband providers if competitors are physically present, but switching requires scheduling another installer, possibly replacing or reconfiguring the router, settling unpaid bills, changing the payment routine, and trusting a new technician. In apartment buildings or dense neighborhoods, cable access and relationships with local technicians may matter more than formal contracts. The operator's most valuable asset may be the already-wired customer premises, the already-configured router, the already-known bill ID, and the already-saved local support number.

Small business upsell is attractive but limited. The top PON plan includes static IP availability and a dedicated support line. For shops, small offices, schools, clinics, CCTV users, or freelancers, a stable connection and static IP can justify a premium. But the address resource constraint matters: a /24 cannot support unlimited static IPv4 assignments. Without IPv6 or upstream-provided address pools, the operator must ration public IPv4 addresses, price them, or reserve them for higher-value customers. In this sense, address resource visibility is not merely a technical artifact; it constrains product design.

Procurement leverage and upstream negotiation

PABNA ONLINE's upstream bargaining position appears weak at the route-visible level. A single-homed /24 access network has limited leverage against a Tbps-scale IIG or IP transit provider. It can threaten to move traffic, but migration requires new interconnection, routing changes, commercial negotiation, possible service disruption, and support capacity. If the operator lacks redundant upstreams, switching is a one-off event rather than continuous negotiation sustained by live load balancing.

Windstream's own position is not without risk. The Bangladesh IIG market has been under regulatory and financial pressure. A February 2025 Daily Star report stated that 29 IIG operators still collectively owed about Tk 205 crore to the BTRC, with dues including regular payments and Social Obligation Fund contributions; the same report named Windstream Communication, Peerex Networks, and others among operators with outstanding amounts, while Coronet was reported to have settled Tk 3 lakh dues. This article is a press report, not a court filing for PABNA ONLINE, but it shows that upstream financial pressure can propagate downward.

If IIGs face dues, bandwidth sale caps, disputes, or license stress, the local ISPs that depend on them face service and price risk.

The Internet Society's 2023 analysis of Bangladesh internet resilience highlights the structural bottleneck: although Bangladesh has many IIGs and high retail ISP diversity, international transit diversity was not uniformly distributed, and the requirement that ISPs buy bandwidth from registered IIGs created a bottleneck. The same analysis noted that small transit providers such as Windstream saw more prolonged connectivity impacts around a major infrastructure incident, while large transit providers rerouted more quickly.

This supports the economic interpretation of single-homing: cheaper transit and lower operational complexity can be rational for hyper-local ISPs, but redundancy costs money and the savings create fragility.

Coronet and Peerex are important to the ecosystem even if they are not PABNA's current observed upstream. Peerex's public BGP profile shows a Bangladesh network labeled IIG with upstreams including BSCCL, Tata, Bharti Airtel, and Hurricane Electric, while Coronet's PeeringDB profile describes an IIG and nationwide retail ISP with 1–5 Tbps traffic levels, extensive CDN and exchange connectivity, and an IRR as-set AS149765:AS-CORONETIIG-BD. PABNA ONLINE's membership in Coronet's and Peerex's AS-SETs may reflect routing policy propagation, historical transit, customer set inclusion, indirect wholesale arrangements, or broad IRR hygiene.

It should not be overinterpreted as evidence of current direct transit.

The broader lesson is that supplier power is stratified. At the international frontier, IIGs and their upstreams control routes and bandwidth economics. At the national layer, NTTN and transport providers control leased fiber and transmission. At the local layer, pole access, building access, power reliability, and municipal permissions affect service delivery. The last-mile ISP is close to the customer but dependent on every upstream layer.

Regulatory constraints and margin pressure

BTRC regulation shapes both the opportunity and the ceiling. The licensing framework creates legal local markets and enables small entities — including sole proprietorships, partnerships, and companies — to operate ISPs. It also prevents unconstrained vertical integration: entities holding NTTN, IIG, IGW, submarine cable, or ITC licenses are generally ineligible to apply for ISP licenses, with specified exceptions and transition rules. This keeps many retail ISPs structurally separate from major wholesale infrastructure.

The five-year license term creates renewal risk. The BTRC guideline stipulates that ISP licenses are initially valid for five years, and licensees must apply 180 days before expiry for renewal or the license is canceled after expiry, with penal actions if business continues without a valid license. The listed validity of Pabna Online runs to 2 March 2027, and that of Pabna Online Network to 11 April 2027. For a small provider, renewal is not just a compliance date; it is a funding and supplier-confidence event. Upstreams, payment partners, landlords, and business customers will care whether the license survives.

The tariff regime limits independent pricing. The BTRC guideline stipulates that a licensee shall not start providing a service before obtaining written approval for its tariff and must obtain approval before changing approved tariff fees. It also stipulates that the commission has the right to determine tariffs when necessary. In a market with rising speed expectations, this matters: an operator may not be able to pass wholesale cost increases on to customers quickly or fully, especially if competitive and regulatory pressures keep retail prices bunched.

A 2025 Daily Star opinion article warned that proposed BTRC fixed-telecom guidelines could impose a 5.5% annual revenue sharing plus a 1% Social Obligation Fund contribution on all broadband service providers, potentially hitting small and medium thin-margin operators. This is an opinion analysis, not a final rule enacted in the cited text, so it should be treated as a risk signal rather than a binding current cost item. Economically, however, it identifies the exact sensitivity: a low-ARPU local ISP can be pushed from viable to non-viable by a few extra percentage points of revenue sharing if it cannot increase prices.

Local authority costs and point-of-presence rules add further friction. The BTRC guideline requires licensees to pay fees approved by local authorities and stipulates that they cannot build or operate a PoP within one kilometer of an existing PoP without prior written approval. It also contains content-blocking obligations with the help of IIG or NIX providers and anticompetitive restrictions between license categories. These rules increase compliance burden and can slow dense network expansion.

Demand: why district broadband exists despite mobile scale

Bangladesh is a mobile-driven internet market, but fixed broadband has a distinct demand niche. AMTOB industry statistics from the BTRC show 188.60 million mobile subscribers at end-May 2026, 119.12 million mobile internet subscribers, and 14.95 million ISP plus PSTN internet subscribers. Mobile internet dominates in subscriber count. Fixed broadband, however, serves a different usage pattern: households, shared Wi-Fi, streaming, gaming, work-from-home, students, small shops, and multi-device families.

The national fixed broadband market still has room to grow but faces infrastructure cost barriers. A BTRC-linked broadband connectivity report indicates that fixed broadband, especially FTTH, requires significant infrastructure investment; it cites World Bank data showing only 6.9 fixed broadband subscribers per 100 people in 2022 against 105.3 mobile network subscribers, and notes an average fixed speed of about 48 Mbps in 2024. The same report states that ISP and PSTN users reached 13.74 million in October 2024, up from 12.49 million a year earlier, with 173,845 km of fiber deployment and total network bandwidth of 6,600 Gbps.

The report also indicates that Bangladesh had 2,715 ISPs and describes a market with many ISPs, overall poor quality of service, and reasonable fixed broadband prices relative to income. It calculates that a 10 Mbps service at Tk 800 per month represented 3.55% of average monthly income based on FY2022-23 per capita income. Pabna-area plans at Tk 500–1,500 for 20–80 Mbps sit within this affordability framework but also show the pressure: consumers are highly price-sensitive, and providers must deliver increasingly high headline speeds at prices that remain significant for middle- and lower-income households.

Fixed broadband churn dynamics differ from mobile. The Financial Express, citing the BTRC, reported that Bangladesh added 750,000 internet subscribers in March 2025 after eight months of decline, with 560,000 of the additions from broadband, and noted that the BTRC updates broadband data quarterly because of the large number of ISPs and the low typical monthly churn of fixed broadband connections. This low churn is the small ISP's saving grace. Once a household is wired and service is acceptable, inertia can protect the revenue stream.

Competition and substitutes

Competition in Pabna is both formal and physical. The BTRC license list places Pabna Online, Pabna Online Network, and Pabna Fiber Net in Pabna Sadar. The ISPAB member directory and public ISP websites show additional operators in the Pabna area. BCF Online, for example, markets itself as a Pabna ISP with BDIX and FTP services and advertises aggressive plans: 30 Mbps for Tk 525, 45 Mbps for Tk 630, 75 Mbps for Tk 850, 90 Mbps for Tk 1,050, 120 Mbps for Tk 1,250, and 150 Mbps for Tk 1,550, with higher advertised speeds for YouTube, Facebook, BDIX, FTP, streaming, and gaming.

It lists offices or contact points in Bhangura, Chatmohar, and Faridpur in Pabna.

BCF's pricing scale is important because it shows the competitive frontier. If a Pabna-area provider can advertise 75 Mbps at Tk 850 or 150 Mbps at Tk 1,550, then a customer comparing plans will pressure any operator still selling 40 Mbps at Tk 1,000 or 80 Mbps at Tk 2,000, unless the more expensive operator has materially better reliability, support, coverage, or local trust. Competition is not just "more ISPs"; it is Mbps inflation at near-fixed household budget levels.

Mobile broadband and fixed wireless access are substitutes at the margin. Mobile data is easier to activate and does not require home fiber installation, but it can be more expensive per gigabyte for intensive use, less stable for multi-device home use, and weaker for gaming or work sessions. Fixed wireless access is also a strategic substitute. The BTRC-linked broadband report notes that fixed wireless access can be more flexible, faster to deploy, and less constrained by geography and buildings than FTTH.

If 5G or another FWA offers stable home broadband in Pabna at competitive prices, it could attack the local ISP's core advantage: the already-installed physical last-mile line.

Large nationwide ISPs are another substitute, but only where their last-mile economics work. National brands may have better procurement, IPv6 capability, monitoring, and corporate processes, but they still need local fiber routes, technicians, permissions, and customer support. Hyper-local ISPs survive because proximity is operationally valuable. A local technician who can fix a cable quickly can beat a national brand with better marketing but slower field response.

Local service trust as an economic asset

The most defensible asset of a district ISP is not the ASN. It is trust embedded in a service routine. PON's site emphasizes the local team, fast support, transparent billing, 24/7 support claims, seven-day contact availability, fast payment by customer ID, and SMS bill reminders. Its news section includes notices of fiber expansion, scheduled maintenance, and celebration of a 5,000th customer connection in June 2026.

These claims are self-published, and the customer count is not independently verified, but the format is commercially significant: the provider is trying to convert a local utility service into a documented customer relationship.

Trust lowers support cost and churn. When customers believe the operator answers the phone and dispatches a technician, they tolerate occasional outages. When they believe bills are transparent, they are less likely to churn over a small advertised speed difference. When local shop owners know the account manager, the provider can sell a static IP or a business plan. Trust also supports informal credit and collection practices, though public records do not show PABNA ONLINE's specific credit terms.

The economics of local trust is fragile. It takes years to build but can degrade quickly after repeated evening congestion, unresolved fiber cuts, payment disputes, or slow support. Because social proof circulates in neighborhoods, a highly visible outage can trigger cluster churn if a competitor is already present in the same building or street. The monitoring point is not just technical uptime; it is whether the operator preserves the belief that local accountability compensates for small-provider limitations.

Ownership, funding, and control

Public archives do not disclose a robust ownership structure for PABNA ONLINE. APNIC identifies PABNA ONLINE as the organization for AS150745 and provides role contacts, but registry contacts are not corporate control records. The AbuseIPDB APNIC-derived route output for 103.81.29.0/24 includes the description PABNA ONLINE and a route description entity naming Md Jasim Uddin in the public WHOIS material examined earlier; this is useful as an operational trace but not sufficient to establish ownership or beneficial control.

PABNA ONLINE's older website does not provide a board of directors, shareholder list, or funding history. It gives operational contact numbers, an email, an address, and a site developer credit. The PON site lists named managers, including a founder, CEO, and COO, but this is a website statement, not a corporate registry extract. The ISPAB public entry for Pabna Online Network lists the company, license type, license number, and trade license number, but states that no director information is on file.

No credible public evidence examined here proves external funding, acquisition by a larger ISP, a merger with another Pabna operator, litigation involving PABNA ONLINE, or a formal parent-subsidiary structure. This absence should not be read as proof that such events do not exist. Small ISPs often transact through informal asset sales, local partnerships, route leasing, customer book transfers, or management changes that leave weak public traces. The economic report therefore treats control as unresolved and focuses on observable operational dependencies.

Outages, abuse, and security signals

No major public outage, legal dispute, procurement conflict, license cancellation, or security incident specific to PABNA ONLINE was found in the cited public record. The evidence that exists is indirect. The PON site posts a scheduled maintenance notice for 18 June 2026 and a fiber expansion notice for Shalgaria on 10 June 2026. IPinfo labels at least one IP assigned to AS150745 as VPN and characterizes the network as a consumer ISP with day-night activity patterns. AbuseIPDB identifies the IP as fixed-line ISP usage and provides WHOIS details, but the lines examined do not establish a material abuse history.

The VPN label must be treated with caution. A single VPN classification can result from a customer using a VPN endpoint, a classification error, a reseller, a proxy service, or a small number of addresses used for remote access. It does not prove the ISP operates a VPN business or has systemic abuse problems. For a small ISP, the most relevant security risk is operational: poor router hygiene, shared NAT attribution issues, abuse ticket management, exposed billing portals, customer premises equipment compromise, and lack of 24x7 network operations depth.

The absence of visible hosted domains is also informative. IPinfo reports no hosted domains found for AS150745. This supports the interpretation that PABNA ONLINE is primarily an access/eyeball network rather than a hosting provider or data center operator. Hosting would create a different abuse profile and revenue model; the public data point instead toward connectivity for households or small businesses.

What the evidence suggests but does not prove

The evidence suggests that PABNA ONLINE's strategic asset is local last-mile density. The Hazi Akbar Ali Super Market address, the Pabna Sadar license, the small ASN, the /24, the retail pricing, and the FTP/IPTV/BDIX product design all point to a neighborhood or district access operator. It is not a national backbone and does not appear to be a heavy managed services provider for enterprises. Its most valuable capabilities are likely customer acquisition, installation, support dispatch, local content performance, and collection discipline.

The evidence suggests that upstream dependency is acute. The observed route-visible upstream is Windstream only. Bangladesh regulation requires ISP licensees to connect to licensed IIGs for leased internet bandwidth and to the NIX for domestic inter-operator traffic. Therefore, PABNA ONLINE's wholesale cost and quality of service depend heavily on contracts and technical paths outside its direct control.

The evidence suggests that speed-price competition compresses margins. The older PABNA ONLINE scale, the newer PON scale, and BCF Online's aggressive scale in the Pabna area all show the same market direction: more advertised Mbps for roughly the same household monthly budget. Because customers' willingness to pay is bounded by income and competing mobile/fixed offers, operators must rely on oversubscription, cheaper domestic traffic, caches, and procurement improvements to keep gross margins from collapsing.

The evidence does not prove customer numbers for AS150745. PON claims over 5,000 connected users, but the legal and operational link between PON and AS150745 remains unresolved. A /24 could support thousands of NATted customers, so the claim is technically possible, but it is not independently validated. Conversely, the public ASN footprint could represent only a subset of a larger retail network if customers are routed via upstream address space or another ASN.

The evidence does not prove fiber ownership. The marketing language says fiber internet, and BTRC rules require local access and leased transmission arrangements, but no public map, pole permit, NTTN contract, OLT inventory, splice plan, or CAPEX filing was found. In economic terms, the safest claim is that the operator controls or operates last-mile service relationships and likely local distribution facilities, while long-distance and upstream transmission depend on licensed wholesale infrastructure.

The economics of fragility

A small ISP such as PABNA ONLINE lives inside a narrow gap. Household monthly plans of Tk 500–1,500 are affordable enough to generate demand, but they do not leave much room for error. A few cost shocks can erase surplus: upstream price increase, local fiber damage, higher pole or authority fees, payment gateway charges, more support staff, router replacement, backup power failure, or regulatory revenue sharing. The business is recurring revenue, but the cost base is not perfectly recurring; it arrives in bursts when lines break, customers churn, or regulators renew fees.

The business's gross margin depends on how oversubscription works. Not all households use peak bandwidth at the same time. Domestic and cached traffic reduces consumption of expensive international transit. Low-income households may buy lower tiers but use intensively at night. Gaming and streaming can create evening congestion. If oversubscription is too high, customers complain and churn; if too low, the operator buys more bandwidth than retail ARPU can support. The art of the business is to engineer a tolerable congestion frontier.

The business's pricing power is asymmetric. It can lower installation price, offer referral credits, or add temporary speed bonuses. It is much harder to increase the monthly bill. Customers compare advertised Mbps and monthly taka, and regulator-approved tariffs may restrict unilateral changes. Local competitors can overbuild profitable streets. Mobile data can substitute for light users. The operator can earn loyalty through service, but loyalty has a ceiling when households face cheaper offers.

The small provider's bargaining leverage increases with density, not geography. A dense cluster of wired customers on a few streets is more valuable than scattered coverage over a wide area. Density reduces field time per subscriber, lowers cable maintenance cost per account, improves referral efficiency, and concentrates traffic demand for upstream negotiation. The named service areas and expansion notices on the PON site are economically significant because they imply route densification rather than abstract coverage.

The address resource layer has signal value. An ASN, a RPKI-valid /24, and an APNIC LIR organization make PABNA ONLINE more visible and potentially more credible than a purely informal reseller. It can appear in BGP collectors, maintain route entities, receive abuse mail, and negotiate as a network. But the same visibility exposes the dependency: anyone can see the single /24, the absence of IPv6, and a single upstream. Public routing data both legitimates and limits the enterprise.

Alternative scenarios

In the conservative scenario, PABNA ONLINE is a small independent Pabna Sadar ISP with AS150745, a /24, a legacy website, and Windstream upstream. Pabna Online Network is a distinct operator with a similar name and a more modern website. Under this scenario, PABNA ONLINE's economics are modest and vulnerable. Its hard assets are the local license, customer relationships, limited address space, and existing facilities. It faces direct local competition, weak upstream leverage, and limited expansion scope.

In the affiliation scenario, PABNA ONLINE and PON are linked through ownership, management, asset transfer, brand evolution, or operational cooperation. Under this scenario, the combined economic unit may have greater customer density, better billing systems, and stronger procurement leverage than AS150745 alone implies. PON's claim of over 5,000 connected users would become material if verified. The 2027 license renewals for both names would be key control events, not separate compliance details.

In the network resource scenario, AS150745 functions as a visible routing and address resource layer while retail traffic may also use upstream-assigned addresses, another ASN, or separate local entities. Under this scenario, public BGP would understate operational scale. It would also mean that customer economics cannot be inferred from the /24 alone. Decisive evidence would be router-level path test from customer premises, billing records, upstream invoices, NTTN contracts, or customer CPE configuration.

In the consolidation scenario, a larger IIG, nationwide ISP, or regional operator eventually consolidates small Pabna access networks by acquiring customer books or operational control. Bangladesh's fragmented ISP market creates a natural consolidation logic: upstream buyers want local eyeballs; small ISPs need better procurement and compliance capability. But regulatory separation between license layers complicates full vertical integration. Consolidation might therefore occur through management contracts, wholesale dependency, brand migration, or asset purchases rather than a simple legal merger.

Evidence registry

APNIC Whois database, AS150745,https://wq.apnic.net/apnic-bin/whois.pl?object_type=aut-num&searchtext=AS150745— principal registry record identifying AS150745 as PABNAONLINE-AS-AP, description PABNA ONLINE, organization ORG-PO5-AP, Bangladesh, contact address Hazi Akbar Ali Super Market, and abuse mailbox[email protected].
RDAP starting record,https://rdap.org/autnum/150745— user-provided RDAP endpoint for the same AS150745 entity; the report uses the APNIC Whois output above as the accessible registry mirror text for the same infrastructure identity.
BGP.tools, AS150745,https://bgp.tools/as/150745— routing intelligence profile showing PABNA ONLINE, active APNIC allocation, eyeball classification, one IPv4 prefix, zero IPv6 prefixes, 103.81.29.0/24, valid RPKI indication, and Windstream Communication Limited as upstream.
Hurricane Electric BGP Toolkit, AS150745,https://bgp.he.net/AS150745— independent BGP view showing Bangladesh origin, one IPv4 prefix, zero IPv6, one observed peer, 256 IPv4 addresses, and valid RPKI-originated status.
IPinfo, AS150745,https://ipinfo.io/AS150745— classifies the network as a stub AS, single-homed, lists 103.81.29.0/24, notes consumer ISP activity, no downstream, and Windstream as sole peer/upstream in its view.
AbuseIPDB WHOIS, 103.81.29.25,https://www.abuseipdb.com/whois/103.81.29.25— APNIC-derived IP record connecting 103.81.29.0/24 to PABNA ONLINE, fixed-line ISP usage, domain pabnaonline.net, and Pabna, Rajshahi division.
PABNA ONLINE website,https://pabnaonline.net/— operational channel evidence for Pabna Online retail ISP services, FTP/IPTV links, plan scale, support contacts, and Hazi Akbar Ali Super Market address.
Pabna Online Network website,https://pabnaonline.net.bd/— adjacent live retail evidence by name for PON, fiber service claims in Pabna Sadar, connected user claim, support language, management labels, plan scale, service areas, online billing, and maintenance notices.
Pabna Online Network portal,https://portal.pabnaonline.net.bd/— customer portal trace showing a Pabna Online customer portal endpoint.
Pabna Online Network e-billing,https://ebilling.pabnaonline.net.bd/— billing system trace identifying Pabna Online Network and a login panel powered by Maestro.
ISPAB public member directory, Pabna Online Network,https://ispab.org/member/pabna-online-network— industry association record listing Pabna Online Network, Upazila/Thana license type, membership C-622, BTRC license number 14.32.0000.702.47.645.22.183, trade license 01515, and no directors on file.
BTRC Upazila/Thana ISP license list,https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/29e9f4bf494145f5bfee76bd1a384ddc.pdf— official license evidence distinguishing Pabna Online and Pabna Online Network as distinct license holders listed in Pabna Sadar with distinct addresses, license numbers, and 2027 validity dates.
BTRC Regulatory and Licensing Guideline for Internet Service Providers in Bangladesh,https://lims.btrc.gov.bd/uploads/service_guideline/Regulatory%20and%20Licensing%20Guideline%20for%20Internet%20Service%20Provider%20%28ISP%29%20in%20Bangladesh.pdf— principal regulatory framework for licensing, last-mile limits, NTTN leasing, license categories, eligibility, renewal, tariff approval, IIG/NIX connectivity, local authority fees, and IPTV conditions.
Bangladesh broadband connectivity report linked to BTRC,https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-btrc/2024/12/2553c9a48743467faaa8b420c2e6ecb5.pdf— market context for fixed broadband infrastructure cost, low fixed penetration, ISP/PSTN user counts, fiber deployment, ISP count, speeds, and affordability.
AMTOB industry statistics,https://www.amtob.org.bd/home/industrystatics— May 2026 national subscriber data from BTRC showing scale of mobile subscribers, mobile internet subscribers, and ISP plus PSTN internet subscribers.
Financial Express, 'Internet subscriber numbers rise after 8 months',https://thefinancialexpress.com.bd/home/number-of-internet-subscribers-rises-after-8-months— press context for subscriber rebound, broadband additions, and BTRC quarterly broadband data practice due to large ISP count and low monthly churn.
Daily Star opinion, 'Fixed broadband at risk',https://www.thedailystar.net/opinion/views/news/fixed-broadband-risk-btrcs-proposed-tax-measures-could-hurt-users-and-isps— policy risk commentary regarding proposed revenue sharing and Social Obligation Fund fees affecting small and medium broadband providers.
Daily Star, '29 IIG operators still owe Tk 205cr to BTRC',https://www.thedailystar.net/business/economy/news/29-iig-operators-still-owe-tk-205cr-btrc-3825121— evidence of financial pressure on the IIG market involving unpaid dues, including references to Windstream, Peerex, and Coronet.
BGP.tools, AS139009 Windstream Communication Limited,https://bgp.tools/as/139009— upstream scale evidence showing Windstream's peer, upstream, and downstream counts and PABNA ONLINE in the downstream table.
Windstream Communication Limited website,https://www.windstreamcommunication.net/— self-description of Windstream as a Bangladesh private limited company and IIG operator with contact information in Dhaka and Gazipur.
PeeringDB, AS139009 Windstream Communication Limited,https://www.peeringdb.com/net/26978— interconnection record listing Windstream as NSP with 1–5 Tbps traffic level, AS-WCL-BD, Asia-Pacific scope, CDN/IX notes, and contact information.
Internet Society Pulse, 'Don't put all your internet infrastructure in one basket',https://pulse.internetsociety.org/en/blog/2023/10/dont-put-all-your-internet-infrastructure-in-one-basket/— resilience context for Bangladesh IIG structure, transit diversity issues, and cost-resilience trade-off for small transit providers.
BGP.tools, AS137491 Peerex Networks Ltd,https://bgp.tools/as/137491— ecosystem-adjacent IIG evidence for Peerex, including upstreams, IPv4/IPv6 resources, and network role in Bangladesh.
BGP.tools, AS149765 Coronet Corporation Limited,https://bgp.tools/as/149765— ecosystem-adjacent network evidence for Coronet, including upstreams and AS-SET data.
PeeringDB, AS149765 Coronet Corporation Limited,https://www.peeringdb.com/net/32178— Coronet interconnection profile, AS149765:AS-CORONETIIG-BD, 1–5 Tbps traffic level, self-description as IIG and nationwide retail ISP, CDN/IX notes, and peering locations.

Surveillance points

License renewal ahead of 2027 expirations. The listed license validity of Pabna Online runs to 2 March 2027, and that of Pabna Online Network to 11 April 2027. Renewal, delay, suspension, transfer, or non-renewal would directly alter legal operating risk and could clarify whether the two identities are distinct or linked.
Any new BGP upstream for AS150745. A second upstream in addition to Windstream would materially improve resilience and bargaining power. Continued single-homing would confirm a low-cost, low-redundancy model.
IPv6 origination. A visible IPv6 prefix from AS150745 would indicate network modernization, address scaling readiness, and potentially better technical support capability. Continued absence of IPv6 origination keeps the operator dependent on scarce IPv4 and a NAT-heavy access economy.
Additional IPv4 resources. A second /24, upstream-assigned static pools, or route-visible expansion would change the static IP and small business revenue ceiling.
Verified relationship between PABNA ONLINE and Pabna Online Network. Corporate filings, license transfers, shared invoices, common bank/payment accounts, shared NOC contacts, or customer migration notices would determine whether PON's customer count and billing system evidence belongs to the AS150745 economic unit.
Customer count validation. Independent confirmation of PON's 'over 5,000 connected users' claim would materially raise inferred scale, traffic purchasing power, and customer book value. Refutation would return the analysis to a smaller legacy ISP interpretation.
Plan scale changes. A further move toward 100–150 Mbps plans at Tk 1,000–1,500 would signal intensified price compression and need for better caching or cheaper upstream bandwidth. Price increases would indicate either improving local pricing power or cost pressure feeding through.
Overbuilding by competitors in Pabna Sadar. Expansion of BCF Online, Pabna Fiber Net, national ISPs, or mobile/FWA substitutes into the same streets would increase churn and reduce installation payback periods.
Windstream commercial or regulatory stress. BTRC dues, bandwidth caps, license disputes, upstream outages, or peering degradation at Windstream would directly feed into PABNA ONLINE's quality of service as AS150745 is currently observed as Windstream-dependent.
Migration to Coronet, Peerex, or another IIG path. If AS150745 begins routing via Coronet, Peerex, BSCCL, Fiber@Home, Summit, or another provider, the bargaining and resilience thesis changes. Live AS-SET hints make this a plausible monitoring item but not a current observed fact.
Self-service and billing platform maturity. More complete customer portals, automated payments, fault tickets, and usage visibility would reduce collection leakage and support cost. Billing system weaknesses or security incidents would damage trust and churn.
Local content and IPTV enforcement. Any regulatory tightening around IPTV, FTP libraries, content blocking, or tariff approval would affect the non-bandwidth value bundle that local ISPs use to defend ARPU.
NTTN and local infrastructure fees. Higher leased transmission costs, pole fees, municipal charges, or building access restrictions would directly pressure gross margin and slow neighborhood expansion.
Evidence of consolidation or asset sale. Acquisition by a regional ISP, a management agreement with an IIG, a customer book transfer, or a co-branding restructuring would likely improve procurement leverage but could weaken the local trust advantage if support centralizes.
Public complaint pattern. A cluster of complaints about evening congestion, billing disputes, slow repairs, or repeated outages would be more economically significant than a single technical outage, as neighborhood churn can spread quickly when competitors are physically present.
RPKI or route entity degradation. Loss of valid RPKI status, route leaks, stale IRR entries, or abuse contact failures would weaken the small but important credibility that comes from being a visible, well-registered access network.
Mobile and FWA pricing in Pabna. If mobile operators or fixed wireless access providers offer stable home broadband substitutes at comparable monthly costs, the switching cost from last-mile fiber drops and local ISP retention weakens.
Business plan uptake. Growth in customers with static IP and dedicated support would improve ARPU and reduce pure residential pricing exposure. Absence of such uptake leaves the business tied to low-margin residential broadband.
Power and resilience investment. Battery, generator, redundant aggregation, monitoring, and fiber ring upgrades would improve service trust but require CAPEX. Under-investment would preserve short-term cash but increase outage and churn risk.
Regulatory revenue sharing implementation. Any final rule imposing broad revenue sharing or Social Obligation Fund fees on small ISPs would be a direct margin shock unless offset by tariff relief, upstream price reductions, or consolidation.