The price of a local connection is the first clue
The first useful number for reading Subnet Data Nusantara is not its autonomous-system number, nor a heroic claim about national reach. It is Rp150,000. On the company's own home-internet price page, Subnet Data Nusantara offers Home Basic at up to 5 Mbps for Rp150,000 per month, Home Standard at up to 10 Mbps for Rp270,000, and Home Ultimate at up to 15 Mbps for Rp360,000, with 1:1 bandwidth and unlimited bandwidth listed for each plan and 11 percent VAT excluded (https://subnet.net.id/harga). Those prices are not a full profit-and-loss statement. They are a window into the constraint under which a regional Indonesian ISP works: a household product has to look cheap enough for local demand, while the operator still has to pay for upstream capacity, interconnection, billing, fault handling, tax, local installation work and the management of partners who may be closer to the customer than the company itself.
The company behind the price table is PT Subnet Data Nusantara, also presented as PT SDN. Its public site says it provides internet for home users, corporate customers and resellers across Nusantara, while its about page says the company operates as an Internet Service Provider with Kominfo licence number JASA-0094/TEL.04.02/2022 (https://subnet.net.id/ and https://subnet.net.id/about). The same page frames its mission around affordable access for households in remote areas not yet covered by adequate networks as well as urban customers. That language is promotional, but it fits the public operating evidence. This is not a national mobile operator with towers, spectrum and tens of millions of subscribers. It is a small licensed internet access provider trying to make a formal business out of local demand that is often served by informal or semi-formal community networks.
The legal-reseller angle is not peripheral. Subnet's reseller page speaks directly to "pegiat internet RT/RW Net", the local community and neighbourhood internet operators that have long filled gaps in Indonesian broadband. It says PT SDN offers a partnership programme to help such operators become more legal, secure and comfortable, while supporting village-level internet access across Indonesia (https://subnet.net.id/reseller). Its FAQ defines an SDN reseller as the resale of internet services from ISP PT Subnet Data Nusantara, requires KTP, NIB, OSS standard certificate and NPWP, asks for customer, POP and endpoint data, and says merely buying bandwidth from PT SDN does not itself create legal resale authority; a cooperation agreement is needed (https://subnet.net.id/faq-reseller). That is a business model in a sentence. Subnet is selling more than megabits. It is selling a route from informal local distribution into a documented relationship with a licensed ISP.
This makes the company economically more interesting than a simple speed-table comparison. The Rp150,000 household plan is the retail edge. The partnership terms reveal the production system behind it. A local reseller may know the lane, the school, the house with roof access, the electricity habits, the informal credit risk and the customer who will complain first when a video call fails. Subnet can bring licence, public IP resources, bandwidth procurement, Jakarta interconnection, billing expectations and a network operations contact. The margin comes from joining these two assets without letting either side destroy the other. If Subnet prices too high, local operators and households stay informal, mobile-only or with a larger competitor. If it prices too low, every support call, tax charge, Metro-E handoff and field visit turns the account into a loss.
Indonesia is a harsh place to test that bargain. APJII said 221.56 million Indonesians were online in 2024, equal to 79.5 percent penetration, but it also reported that urban areas still accounted for 69.5 percent of users and rural areas for 30.5 percent (https://apjii.or.id/berita/d/apjii-jumlah-pengguna-internet-indonesia-tembus-221-juta-orang). APNIC's 2026 Indonesia overview put the same strategic problem in network language: Indonesia's island geography shapes infrastructure deployment, service quality and access equity, with advanced infrastructure in major urban centres and persistent limits in remote and eastern regions (https://blog.apnic.net/2026/02/06/connecting-an-archipelago-exploring-indonesias-internet-ahead-of-apricot-2026/). A small ISP in Rembang, Central Java, is therefore not competing only on a plan card. It is competing in the space between a national digital market and thousands of local access problems.
The right thesis is narrow. Subnet Data Nusantara appears to be a legally visible, technically modest, reseller-oriented Indonesian ISP whose value depends on whether it can aggregate local access demand cheaply enough and then use formal network resources and local interconnection to make that demand serviceable. It should not be mistaken for a backbone owner or a large national access platform. It should also not be dismissed as a mere website. The evidence shows AS142364, registered resources, public exchange connections, Jakarta facility presence, a home-price schedule and detailed reseller terms. In Indonesia's fragmented broadband market, that combination can matter even at small scale.
A small Rembang identity with a Jakarta network surface
Subnet's corporate geography is split between local origin and interconnection necessity. PeeringDB's organisation record for PT Subnet Data Nusantara gives an address at Jl. Pasir Putih Jarasilica No. 09, Ds. Lodan Kulon RT/RW 09/02, Rembang, Jawa Tengah 59274, with the website https://subnet.net.id and the alias SDN (https://www.peeringdb.com/org/30151). The company's own contact page gives a Rembang address at Jl. Kalipang, Lodan Wetan RT/RW 01/02, Kec. Sarang, Kab. Rembang, Jawa Tengah 59274, along with a WhatsApp contact and info@subnet.net.id (https://subnet.net.id/contact). IDNIC/APNIC RDAP for AS142364 describes PT Subnet Data Nusantara as an Internet Service Provider at Ds. Lodan Kulon RT/RW 09/02 No. 09, Kec. Sarang, Rembang, Jawa Tengah 59274, and records AS142364 as active with registration on 17 January 2022 (https://idnic.rdap.apnic.net/autnum/142364).
Those three traces do not prove revenue, subscriber count or ownership. They do something more basic: they align the name, locality, website and registered network identity. For a small ISP, that matters. Many small access brands are visible only through a Facebook page, a WhatsApp number or a reseller leaflet. Subnet has a public site, a regulator-claimed licence number, IDNIC resource records and PeeringDB network entries. That is enough to treat the company as a real operating participant in Indonesian access markets, while remaining cautious about scale.
The resource footprint is modest and clear. IDNIC/APNIC RDAP shows 103.170.100.0 to 103.170.101.255 as an allocated portable IPv4 block for PT Subnet Data Nusantara, a /23 with 512 IPv4 addresses, registered on 27 July 2021 (https://idnic.rdap.apnic.net/ip/103.170.100.0). It also shows 2407:c6c0::/32 as an allocated portable IPv6 block for the same company, registered on 31 August 2021 (https://idnic.rdap.apnic.net/ip/2407:c6c0::). BGP.tools identifies AS142364 as active under APNIC, an eyeball network, with three IPv4 prefixes and one IPv6 prefix originated, and lists the visible prefixes as 103.170.100.0/24, 103.170.100.0/23, 103.170.101.0/24 and 2407:c6c0::/32 (https://bgp.tools/as/142364). Hurricane Electric's BGP page likewise lists AS142364 as PT Subnet Data Nusantara with four originated prefixes, all RPKI-valid at the time of capture, and 512 originated IPv4 addresses (https://bgp.he.net/AS142364).
The public routing footprint therefore supports a careful conclusion. Subnet has enough address and routing control to operate as an ISP and manage customer-facing or reseller-facing services. It does not have a visible address base that suggests a very large national retail broadband operator. A /23 can be stretched through NAT, private addressing, CGNAT and reseller assignment, but the public IPv4 resource itself is scarce. That scarcity is a real economic constraint. A small ISP must decide which customers receive public addresses, how resellers manage private distribution, how support handles port-forwarding and application complaints, and when IPv6 is offered as a practical default rather than a registry formality.
The Jakarta network surface is more interesting. PeeringDB lists Subnet Data Nusantara as a Cable/DSL/ISP network, AS142364, with Asia Pacific scope, balanced traffic ratio, open peering policy, IPv6 support, no ratio requirement and no contract requirement (https://www.peeringdb.com/net/27630). Its PeeringDB API record shows six operational exchange connections: DE-CIX ASEAN, JKT-IX: Main, IIX-Jakarta, AIX, DE-CIX Jakarta and ODIX Omadata, all entered at 10,000 Mbps in PeeringDB, plus facility presence at Datacenter APJII-Cyber and IDC 3D in Jakarta Selatan (https://www.peeringdb.com/api/net/27630). That is not the profile of a purely isolated village wireless operator. It is the profile of a small access provider that understands the value of being seen at local exchange points.
There is one useful discrepancy. AIX's own IX-F member export, timestamped 3 July 2026, lists "340_ PT Subnet Data Nusantara" as AS142364, member since 19 September 2025, with an active connection, route-server participation and IPv4 address 43.254.83.96, but with interface speed shown as 1000 rather than 10000 (https://aix.area31.id/api/v4/member-export/ixf). PeeringDB records AIX at 10G for the same IPv4 and IPv6 addresses (https://www.peeringdb.com/net/27630). The discrepancy is not fatal. It is a due-diligence warning. Public exchange databases are valuable, but anyone underwriting capacity should verify live port speed, billing, cross-connect and traffic graphs directly with the exchange and the company.
Even with that caveat, the strategic meaning is plain. Subnet's home office may be in Rembang, but the efficiency of its service depends heavily on Jakarta interconnection. PeeringDB identifies Datacenter APJII-Cyber as an APJII facility in Cyber Building, Jalan Kuningan Barat No. 8, Jakarta Selatan, with 704 networks and 20 exchanges in the facility record at the time viewed (https://www.peeringdb.com/api/fac/4024). IDC 3D's facility record lists 511 networks and 13 exchanges at Jl. Duren Tiga Raya No. 7H, Duren Tiga, Pancoran, Jakarta Selatan (https://www.peeringdb.com/api/fac/5657). For a small ISP, being present in such places is a way to buy density it could never build alone.
Why exchange ports change the unit economics
Local interconnection matters because Indonesian broadband is not only a last-mile problem. It is also a traffic-locality problem. If a customer's video, game, payment page, school portal or marketplace session has to travel inefficiently through expensive transit or distant paths, the ISP pays twice: once in capacity cost and again in customer dissatisfaction. If traffic can be exchanged locally with content networks, caches, other access providers and domestic platforms, the same retail price can carry more usage before support and upstream bills overwhelm it.
Internet Society's 2026 Indonesia country report gives the national scale of this mechanism. It counts 58 active IXPs in Indonesia, 202 data centres, 2,845 active networks, 80 percent of the top 1,000 websites reachable through an in-country server or cache, and says 92 percent of active networks connect via an IXP (https://pulse.internetsociety.org/en/reports/id/). APNIC's 2026 Indonesia article similarly says IXPs are foundational for routing local traffic efficiently and keeping domestic traffic within the economy, and it cites 56 active IXPs with 1,225 members as of January 2026 (https://blog.apnic.net/2026/02/06/connecting-an-archipelago-exploring-indonesias-internet-ahead-of-apricot-2026/). The exact count differs because sources and update dates differ. The direction is the same: Indonesia has become a multi-IXP market.
That changes what a small ISP must do. In an older model, a local provider could buy upstream capacity, install wireless or fibre access, and treat the wider internet as a black box. In the modern Indonesian market, local peering is part of cost control. Subnet's six listed exchange presences are therefore not decorative. They suggest that the company is trying to reach local traffic pools where the marginal cost of another gigabyte is lower than hauling everything through a single transit provider. A small port can still save money if the traffic mix is right. Ten gigabits of exchange capacity is not the same as ten gigabits of sustained paid traffic, but it can be a bargaining and performance asset.
The reseller model magnifies the importance. Subnet's FAQ says reseller bandwidth can be delivered through trusted Metro-E providers, and that if a reseller is outside Subnet's POP reach the service can use third-party Metro-E from named providers such as Telkom, XL, Indosat, FS, TIS and Surge (https://subnet.net.id/faq-reseller). The terms page says reseller bandwidth may pass through VPN, Metro-E, fibre optic or microwave according to need, and it separates Subnet's responsibility for internet link, Metro-E if any and DNS server from the reseller's responsibility for customer-side access from the customer router to the reseller POP or endpoint (https://subnet.net.id/sk-reseller). This is a layered cost structure. The local reseller owns the messy customer edge. Subnet owns the formal upstream and aggregation promise. The exchange port helps make that promise cheaper.
The economic job is to keep the average support and capacity cost below a low household price. Suppose a reseller brings 65 customers in the example used on Subnet's FAQ: 30 at Rp150,000, 30 at Rp170,000 and five at Rp200,000, producing Rp10.6 million in customer revenue. Subnet's FAQ uses that example to calculate 11 percent VAT of Rp1.166 million and BHP plus USO at 1.75 percent of Rp185,500, then adds a hypothetical 500 Mbps bandwidth commitment of Rp13.5 million, making total payment to the ISP Rp14.8515 million in the example (https://subnet.net.id/faq-reseller). The example is not presented as a full reseller profit model, and it leaves out local access costs, installation, equipment, churn and bad debt. That is precisely why it is useful. It shows how tight the arithmetic can become once tax and committed capacity are explicit.
The same FAQ says the reseller registration fee was Rp1.5 million but was waived under a May to December 2024 promotion (https://subnet.net.id/faq-reseller). A Facebook post indexed in search results advertised a November partnership promotion of "1G 14jt" including metro and bandwidth (https://www.facebook.com/subnetdatanusantara/posts/promo-kemitraan-spesial-november-1g-14jt-ajah-include-metro-dan-bandwidth-bisa-w/818658894108680/). That Facebook signal should be treated as market chatter rather than audited pricing; social media promotions can be temporary, incomplete and area-specific. Still, it points in the same direction as the FAQ. Subnet is trying to reduce the entry friction for local resellers while selling committed bandwidth bundles that can be understood by small operators.
This is the commercial tension. A small reseller wants a low fixed commitment, enough bandwidth to avoid evening congestion, legal cover, billing help and technical support. Subnet wants enough monthly commitment to cover upstream, ports, tax, support and the risk that the reseller's customers call when the local network fails. Exchange presence lowers the traffic cost and improves performance, but it does not fix every bottleneck. The customer still sits behind a pole, a rooftop radio, a fibre drop, a home router, a local power supply and a reseller who may be operating on thin cash flow.
The product is partly legality, partly labour
Subnet's public reseller documents are unusually specific because the business is partly an answer to the legal ambiguity around small local internet distribution. Indonesia has many neighbourhood and village access initiatives. Some are formal, some work through licensed partners, and some sit in a grey zone. Subnet's public pages make legality a selling point. The reseller page says it helps internet activists who are still illegal develop into a legal, safe and comfortable business (https://subnet.net.id/reseller). The FAQ says a reseller needs identity and business documents, must register customers, POPs and endpoints, and cannot gain legal resale status by buying bandwidth alone without a cooperation agreement (https://subnet.net.id/faq-reseller). The terms page goes further, requiring reseller attention to local documents such as domicile letters, environmental permission from local leaders, tower permits and other documents where needed (https://subnet.net.id/sk-reseller).
This is not mere paperwork. It is part of the margin. If a reseller operates without adequate documentation, uses unlicensed or badly coordinated wireless links, sells non-internet services without separate agreements, fails to register customers, or lets one reseller poach another's area without coordination, Subnet inherits reputational and commercial risk. Its terms say reseller rights are non-exclusive, that resellers must avoid unhealthy competition with other ISPs and fellow resellers, that customers must be registered to PT Subnet Data Nusantara, and that PT SDN may contact reseller customers directly for validation (https://subnet.net.id/sk-reseller). Those clauses may look administrative. Economically, they protect the aggregator from being treated as a wholesale pipe with no control over the retail promise.
The labour split is just as important. The terms page says Subnet must provide bandwidth and internet link according to reseller needs, offer technical and online support from the Subnet router to the reseller router, and provide billing and monitoring tools, while the reseller is responsible for access from the customer to the reseller POP or endpoint (https://subnet.net.id/sk-reseller). This is a classic boundary in local broadband. The upstream operator wants to stop responsibility at a defined handoff. The household experiences the service as one connection. When something fails, the household does not care whether the fault is a rooftop radio, a power adapter, a private IP assignment, a Metro-E delay, an IX route, DNS, upstream transit or a customer's own router.
The success of Subnet's model depends on whether that boundary is accepted in practice. If resellers are technically competent, honest with customers and financially disciplined, Subnet can scale reach without building every local drop itself. If resellers are undercapitalized, oversell bandwidth, use poor wireless design, fail to maintain local equipment or mishandle customer complaints, Subnet may lose brand trust even where its own router, exchange port and upstream link are healthy. The company's public documents show awareness of this risk. They ask for POP coordinates, endpoint data, customer data, profile data, signed agreements and operational transparency (https://subnet.net.id/sk-reseller).
The same documents also reveal the limits of the model. Subnet's newer reseller registration page asks applicants for the planned service area, potential customers, bandwidth requirement, POP coordinates, KTP upload and optional reseller NIB, then says the SDN team will review and contact the applicant (https://reseller.subnet.net.id/registrasi). The "Syarat & Keuntungan" page says details of costs, packages and cooperation terms will be communicated after PT Subnet Data Nusantara reviews the service area, bandwidth needs and readiness of the prospective partner (https://reseller.subnet.net.id/syarat). That is sensible. It also means public pricing is incomplete. A buyer cannot value the reseller channel from the website alone. It would need actual contracts, churn, bad debt, support minutes and bandwidth utilization by partner.
The home product faces a similar proof burden. A 5 Mbps, 10 Mbps or 15 Mbps plan with 1:1 and unlimited language is attractive where the alternative is unstable mobile data or no affordable fixed line. But it is not enough to know the plan card. A serious customer or lender would ask how many active home users exist, whether speeds are contended during peak hours, how installation is done, how many customers use wireless versus fibre, whether there is CGNAT, how tickets are handled, and how frequently resellers miss payments. The public evidence gives the outline of a business. It does not give the operating ratios.
Indonesia's fixed-broadband gap is the market
The demand side is clear at the national level. Indonesia is online, but not evenly and not primarily through high-quality fixed broadband. APJII's 2024 survey put internet penetration at 79.5 percent, while APNIC's 2026 article cited APJII's 2025 figure of 229.4 million users and 80.66 percent penetration (https://apjii.or.id/berita/d/apjii-jumlah-pengguna-internet-indonesia-tembus-221-juta-orang and https://blog.apnic.net/2026/02/06/connecting-an-archipelago-exploring-indonesias-internet-ahead-of-apricot-2026/). World Bank data, sourced to the ITU, recorded Indonesia at 13,953,182 fixed-broadband subscriptions in 2024 and 4.92 fixed-broadband subscriptions per 100 people (https://api.worldbank.org/v2/country/IDN/indicator/IT.NET.BBND?format=json&per_page=5 and https://api.worldbank.org/v2/country/IDN/indicator/IT.NET.BBND.P2?format=json&per_page=5). Mobile cellular subscriptions, by contrast, stood at 122.5 per 100 people in 2024 in the World Bank/ITU series (https://api.worldbank.org/v2/country/IDN/indicator/IT.CEL.SETS.P2?format=json&per_page=5).
Those numbers explain Subnet's opening. A household may be online through mobile data but still lack a stable home connection for video lessons, work, gaming, small business, cameras, religious study, e-commerce operations or family streaming. Fixed broadband is not only a richer-country convenience. It is the difference between connectivity as a phone habit and connectivity as household infrastructure. When fixed subscription penetration is below five per 100 people, there is room for many local providers, even if the national internet market already looks mature.
Internet Society's 2026 country report shows the performance and infrastructure gradient. It lists 73 percent internet penetration, 79 percent urban usage and 64 percent rural usage, with average download speeds of 47.10 Mbps for broadband and 62.54 Mbps for mobile in its reported sample, while rating Indonesia's overall internet resilience score at 59 out of 100 (https://pulse.internetsociety.org/en/reports/id/). APNIC's article puts the structural gap more plainly: major urban and economic centres such as Jakarta, Surabaya and Bandung benefit from more advanced infrastructure and private investment, while remote and eastern regions continue to face limits in access quality, affordability and resilience (https://blog.apnic.net/2026/02/06/connecting-an-archipelago-exploring-indonesias-internet-ahead-of-apricot-2026/). Subnet's stated rural and reseller focus sits inside that gap.
The public-facility indicators reinforce the point. A World Bank Indonesia Economic Prospects digital-infrastructure presentation for December 2025 said only 22 percent of schools had internet speeds of 100 Mbps or more, 24 percent of puskesmas had "good" internet, and 66 percent of village offices had any functioning internet (https://thedocs.worldbank.org/en/doc/2058d67adda4a910ceab72209ddec8f3-0070012025/related/IEP-December-2025-Digital-Presentation.pdf). Those figures are not about Subnet's customers. They describe the country in which Subnet sells. If public schools, health clinics and village offices still face quality gaps, the market for small, practical, locally supported connectivity remains real.
Competition is also real. Internet Society lists the top Indonesian ISPs by market share as Telkomsel, Telkom Indonesia, Indosat, XL Axiata and Eka Mas Republik (https://pulse.internetsociety.org/en/reports/id/). Those are not small neighbours. They are national or large-scale players with mobile, fibre, enterprise, tower, backbone, marketing and retail advantages. A small Rembang ISP cannot beat them by acting like a miniature incumbent. It has to win where local knowledge, reseller relationships, quick installation, cheaper packages, flexible area terms or a neglected pocket of demand matters more than national brand power.
This is why Subnet's reseller programme is not just a sales channel. It is a market-selection device. A prospective reseller tells Subnet where there are potential customers, what bandwidth is needed, where the POP will sit and what local documents exist. The local partner absorbs some customer acquisition and field labour. Subnet decides whether the area can be served economically through its network and third-party Metro-E options. If the model works, Subnet can discover demand at the edge of national operators' attention. If it fails, Subnet finances other people's weak local businesses.
The failure scenario is a reseller shock during a capacity squeeze
The most plausible failure scenario is not a spectacular cyber incident or a national outage. It is a reseller shock during a capacity squeeze. Imagine a local partner that signs up 80 households after a promotional period, sells low monthly packages, uses a mix of wireless and local fibre, and commits to a bandwidth tier from Subnet. Evening video use rises, a local school shifts more learning online, a few customers demand gaming performance, and mobile operators run promotions that make customers more impatient with congestion. At the same time, the reseller has to pay VAT, BHP and USO pass-throughs, maintain access points, replace damaged equipment after rain, and keep households paying on time. The reseller asks Subnet for more bandwidth but resists a higher monthly commitment. Subnet can either hold the line and let the reseller's customers suffer, upgrade capacity and accept payment risk, or risk losing the local area to another ISP.
This failure case is specific to Subnet because its documents define exactly such a boundary. The FAQ says reseller activation can take 15 to 45 working days, with Metro-E activation depending on the provider, and it says out-of-POP resellers use third-party Metro-E providers such as Telkom, XL, Indosat, FS, TIS and Surge (https://subnet.net.id/faq-reseller). The terms page says Subnet is responsible for internet link and Metro-E if any, while the reseller is responsible for customer-side access from the customer router to the reseller POP or endpoint (https://subnet.net.id/sk-reseller). The customer, however, experiences one service. When the evening connection slows, the accounting boundary does not matter.
The financial version is equally dangerous. Subnet's own example shows a reseller with Rp10.6 million of customer revenue facing Rp1.3515 million in VAT and BHP/USO plus a hypothetical Rp13.5 million bandwidth commitment for 500 Mbps (https://subnet.net.id/faq-reseller). That example is not meant to show a profitable reseller; it is meant to explain taxes and ISP payment. But it makes the point. A reseller can be cash-negative if take-up, pricing, collection or capacity planning is wrong. If several resellers underperform at once, Subnet is exposed to delayed payments, support escalation and reputational spillover. The upstream and exchange bills do not wait for village-level collections.
An operational shock could come from outside the reseller too. A Metro-E provider may delay activation, a cross-connect may fail, a Jakarta exchange route may behave unexpectedly, a fibre path may be cut, power problems may hit local access equipment, or a national provider may undercut a local plan. Because Subnet relies on many external layers, its service promise is a coordination problem. Its public exchange presence lowers traffic cost and may improve performance, but it cannot repair a reseller's rooftop radio, a damaged local pole, a late customer payment or a poorly documented tower permit.
The mitigation is not mystery technology. It is discipline. Subnet would need clear reseller qualification, conservative bandwidth commitments, active monitoring, transparent contention ratios, strict customer registration, direct validation where needed, realistic tax handling, and escalation rules that do not leave households trapped between the reseller and the ISP. Its terms already point in that direction. The question is execution, not vocabulary.
What would a buyer, lender or regulator pay for?
A buyer or lender would pay for Subnet only if the reseller and interconnection evidence translates into recurring, low-churn cash flow. The attractive assets are the Kominfo-claimed ISP authority, AS142364, registered IPv4 and IPv6 resources, exchange and facility presence in Jakarta, a home-access price ladder, reseller documentation, local Rembang operating identity and the ability to aggregate underserved pockets of demand. The discount would be heavy for missing public evidence on subscribers, revenue, EBITDA, churn, gross margin, reseller concentration, bad debt, actual traffic, port utilization, upstream contracts, legal ownership and support performance.
The proof package should be concrete. A buyer would want active users by direct home plan and reseller, monthly recurring revenue by cohort, average revenue per user, bandwidth cost by source, exchange traffic by port, 95th percentile transit usage, public IPv4 assignment policy, CGNAT design, IPv6 activation rate, number of resellers, reseller arrears, customer complaints, installation times, outage history, equipment inventory, Metro-E agreements, tax filings, licence documents and the signed cooperation agreements that give resellers legal resale authority. A lender would also ask whether the company can survive one or two large reseller defaults without cutting upstream capacity or damaging the brand.
A regulator would ask a different set of questions. It would care whether Subnet's reseller programme genuinely formalizes access rather than laundering informal distribution under a licence number. It would ask whether resellers have proper documents, whether tower and radio use comply with Indonesian rules, whether customers know who is responsible for service and complaints, whether taxes and universal-service contributions are collected and remitted as described, and whether customer data is handled lawfully. Subnet's terms show awareness of these issues. Public proof of actual enforcement is not visible.
The large-customer question is harsher. A business, school, clinic or local government office would not buy the story of legal reseller enablement by itself. It would demand service-level terms, route diversity, support hours, escalation contacts, static/public IP options, installation quality, backup power assumptions, and evidence that the last mile is not dependent on a single fragile local technician. Subnet's public contact and NOC records are useful, but not enough. The operating promise has to survive a rainy night, a power interruption and a payment dispute.
The one fact that would change the judgement
The one fact that would most change the judgement is not another exchange listing. It is live, cohort-level reseller economics. If Subnet can show that resellers keep customers, pay on time, upgrade bandwidth rationally, maintain clean local access networks and generate positive gross margin after tax, Metro-E, upstream, support and bad debt, the company looks like a small but coherent platform for formalizing local Indonesian broadband. If the reseller base is volatile, underpriced and support-heavy, the same documents look like a high-friction wholesale pipe wrapped in legality.
The second most important fact is traffic mix. Exchange presence is valuable only if enough traffic is local, cached or peerable to reduce transit cost and improve user experience. PeeringDB says Subnet has open peering and six exchange presences (https://www.peeringdb.com/net/27630). AIX's export confirms at least one active route-server connection at the AIX member list level, though with a different speed record than PeeringDB (https://aix.area31.id/api/v4/member-export/ixf). Those are promising signs. Actual value depends on utilization, not membership.
The third fact is customer trust. There is little public review evidence. The company's blog posts discuss RT/RW Net, VPN for ISPs and metro-ethernet topics, which helps show the audience it courts, but blog content is not customer satisfaction data (https://blog.subnet.net.id/). The Facebook snippets show small-scale promotion and local social presence, not durable demand. In a local broadband business, reputation often lives in WhatsApp groups, neighbourhood conversations and the speed of a technician's motorcycle, not in public filings. That makes public diligence harder and field diligence more important.
The practical way to test that reputation is local and unglamorous. A diligence team would visit one active reseller area, ask households what happens after rain or evening congestion, compare advertised speeds with evening throughput, inspect whether POP coordinates and endpoint data match the real topology, and follow a trouble ticket from customer call to reseller technician to Subnet NOC. It would also ask whether customers understand who invoices them and who has authority to change packages. None of this requires heroic national scale. It requires the patient operating discipline that separates a useful regional ISP from a loose wholesale bandwidth arrangement.
Public evidence register
The company identity and public-service claim are supported by Subnet's official homepage, about page, contact page and service page, which identify PT Subnet Data Nusantara, the subnet.net.id domain, home and reseller services, a Rembang contact address and Kominfo licence number JASA-0094/TEL.04.02/2022 (https://subnet.net.id/, https://subnet.net.id/about, https://subnet.net.id/contact and https://subnet.net.id/service).
The retail price evidence is supported by Subnet's home-internet price page, which lists Home Basic at Rp150,000 for up to 5 Mbps, Home Standard at Rp270,000 for up to 10 Mbps and Home Ultimate at Rp360,000 for up to 15 Mbps, each presented with 1:1 and unlimited bandwidth and excluding 11 percent VAT (https://subnet.net.id/harga).
The reseller model and cost evidence is supported by Subnet's reseller page, reseller FAQ, reseller terms page and newer reseller portal, which describe legal reseller positioning, required documents, customer/POP/endpoint data, 15 to 45 working-day activation, the need for a cooperation agreement, VAT at 11 percent, BHP and USO at 1.75 percent, a sample 500 Mbps bandwidth commitment at Rp13.5 million, and third-party Metro-E options (https://subnet.net.id/reseller, https://subnet.net.id/faq-reseller, https://subnet.net.id/sk-reseller, https://reseller.subnet.net.id/ and https://reseller.subnet.net.id/registrasi).
The network-resource evidence is supported by IDNIC/APNIC RDAP for AS142364, 103.170.100.0/23 and 2407:c6c0::/32, plus BGP.tools and Hurricane Electric, which identify PT Subnet Data Nusantara as an Indonesian ISP, show the Rembang registration details, active AS status, a 512-address IPv4 block, an IPv6 /32 and visible originated prefixes (https://idnic.rdap.apnic.net/autnum/142364, https://idnic.rdap.apnic.net/ip/103.170.100.0, https://idnic.rdap.apnic.net/ip/2407:c6c0::, https://bgp.tools/as/142364 and https://bgp.he.net/AS142364).
The interconnection evidence is supported by PeeringDB's network and API records, the AIX IX-F member export, and PeeringDB facility records for Datacenter APJII-Cyber and IDC 3D. These records show six exchange presences, two Jakarta Selatan facilities, open peering policy, route-server participation, contact details and the AIX speed discrepancy that should be verified directly (https://www.peeringdb.com/net/27630, https://www.peeringdb.com/api/net/27630, https://aix.area31.id/api/v4/member-export/ixf, https://www.peeringdb.com/api/fac/4024 and https://www.peeringdb.com/api/fac/5657).
The Indonesian market evidence is supported by APJII's 2024 penetration release, APNIC's 2026 Indonesia internet-infrastructure overview, Internet Society Pulse's Indonesia report, World Bank/ITU fixed-broadband and internet-use indicators, and the World Bank Indonesia Economic Prospects digital presentation. Together they support the archipelago-access thesis, the fixed-broadband gap, Indonesia's dense IXP environment, rural-urban access differences and public-facility connectivity constraints (https://apjii.or.id/berita/d/apjii-jumlah-pengguna-internet-indonesia-tembus-221-juta-orang, https://blog.apnic.net/2026/02/06/connecting-an-archipelago-exploring-indonesias-internet-ahead-of-apricot-2026/, https://pulse.internetsociety.org/en/reports/id/, https://api.worldbank.org/v2/country/IDN/indicator/IT.NET.BBND?format=json&per_page=5, https://api.worldbank.org/v2/country/IDN/indicator/IT.NET.BBND.P2?format=json&per_page=5, https://api.worldbank.org/v2/country/IDN/indicator/IT.NET.USER.ZS?format=json&per_page=5 and https://thedocs.worldbank.org/en/doc/2058d67adda4a910ceab72209ddec8f3-0070012025/related/IEP-December-2025-Digital-Presentation.pdf).
The investable reading
Subnet Data Nusantara should be read as a local-density business, not a scale fantasy. The visible evidence supports a company with a formal ISP identity, small but real internet resources, exchange participation and a practical reseller apparatus aimed at Indonesia's uneven fixed-broadband market. Its best path is to be close enough to households and local operators to serve demand that national brands leave awkward, while being connected enough in Jakarta to keep traffic cost and performance under control.
The risk is that the same model can become an aggregation of other people's fragility. Local resellers can oversell, under-document, under-maintain or underpay. Metro-E and last-mile dependencies can blur responsibility. Public IPv4 scarcity can complicate support. Exchange membership can be mistaken for performance. A cheap plan can attract customers whose support cost exceeds their revenue. In that world, Subnet's value is capped by the weakest local partner.
The balanced judgement is therefore constructive but cautious. Subnet has enough public evidence to deserve attention as a regional Indonesian ISP with a coherent reseller and interconnection thesis. It does not yet have enough public evidence to prove durable scale, profitability or operational depth. The next proof should be operating data, not more slogans: retained customers, paid reseller accounts, live traffic mix, fault resolution, margin after tax and capacity cost, and a clean explanation of who owns which part of the customer's service when the connection fails.