Summary

  • Farknet's economic test starts with the payer: households and small businesses paying monthly lira bills for local access, often on a no-commitment, no-quota pitch. The company has enough public evidence to be treated as an active regional ISP, but the record does not yet prove that its local reach produces pricing power rather than thin resale margin.
  • The clearest boundary is between services riding on Turk Telekom infrastructure and services Farknet describes as available only in pilot areas on its own GPON or wireless infrastructure. That boundary matters more than the brand name, because wholesale access can reset faster than retail ARPU while owned local infrastructure carries replacement, backhaul, support and utilisation risk.
  • Network-resource evidence is credible but narrow: AS205935 is active, three IPv4 /24s are visible under Farknet's own origin, RPKI evidence is valid for sampled prefixes, and the public BGP view shows two observed neighbours. That proves an operating footprint, not customer scale.
  • The conclusion is conditional. Farknet can be a durable local operator if it converts branch-level and pilot-network reach into retained subscribers with low support cost and disciplined price pass-through. If most economics remain tied to wholesale access and small-scale field work, capital renewal will catch up before local brand equity compounds.

The payer is the household, not the registry

The first useful way to understand Farknet is to ignore the autonomous-system number and start with the monthly bill. A regional access network survives because local households, apartment blocks, small shops and offices keep paying after the installation crew has left. The customer's incentive is simple: stable internet, acceptable speed, a reachable support desk, no unnecessary phone line, and a price that looks fair against national brands. Farknet's own public offer is built around exactly that retail psychology.

It advertises no commitment, no quota or limit, no phone requirement, free switching from another operator in some packages, and an address-availability check before purchase.

That tells us more about the business than a generic label such as "ISP" would. Farknet is not selling a luxury enterprise network in the public pages reviewed here. It is selling consumer and small-business access in a local market where a buyer can compare prices quickly and where many physical paths may ultimately depend on the same incumbent infrastructure. The payer therefore rewards convenience and trust as much as raw bandwidth. If the installer appears when promised, the modem works, the support phone answers, and the monthly bill remains explainable, the customer may stay.

If the service feels indistinguishable from a larger operator but with less certainty, churn becomes the hidden tax.

The package evidence also shows how thin the pricing corridor can be. Visible FTTH packages sit in a narrow band around TL 549 to TL 599 per month, DSL packages in a similar range, while wireless "no infrastructure" packages sit much higher, at TL 720, TL 825 and TL 870 per month for 20, 35 and 50 Mbps tiers. A pilot GPON 100 Mbps offer appears at TL 350 per month, far below the mainstream fiber and DSL prices. Those figures should not be read as audited realized ARPU; they are public tariff evidence. But they reveal the shape of the commercial problem.

Where Farknet relies on an incumbent access line, retail price has to cover wholesale port or access charges, support, billing, tax-inclusive pricing, CPE, bad debt, payment fees and overhead. Where Farknet uses its own wireless or pilot GPON infrastructure, it may keep more control, but then it must finance towers or sites, radios, fiber drops, optics, power, cabinets, spares, civil work and replacement cycles.

The payer is also not paying for "number resources." Customers rarely care whether the operator has a RIPE LIR record, whether routes are RPKI valid, or whether a prefix is originated by one ASN rather than another. Those records are useful evidence for analysts because they prove operational capacity and network boundaries. They do not prove that a subscriber is paying, that an apartment building has high take-up, or that a radio sector has enough customers to pay for maintenance. Farknet's economic case therefore has to be judged in two layers: public network capability, and monetized local access.

The public evidence clears the first layer. It does not clear the second with the same confidence. There is enough to say Farknet operates a real retail service, bills customers, publishes packages, maintains a contact and branch footprint, and runs an active AS. There is not enough to say how many customers are active, how concentrated they are by neighbourhood or building, how many are on owned infrastructure, how many are on Turk Telekom access, or how much cash remains after wholesale and field costs. That is why the central question is a capital-cycle question rather than a simple identity profile.

A local operator with two different control surfaces

Farknet's public footprint starts in Konya. The company contact page lists a head office in Meram, Konya, with a MERSIS number, and a Bozyazi/Mersin branch with a separate branch MERSIS identifier. The site also states that Farknet has been improving internet connections since 2014. The privacy notice and contract pages say the company operates under a BTK electronic-communications or internet-service-provider licence. RIPE records show the company as a RIPE LIR from 2017. Together, those facts point to a long-running local operator, not a temporary reseller page.

But the operating boundary is not one surface. It is at least two.

The first surface is wholesale-dependent access. Farknet's own footnotes state that FTTH subscriptions are supplied over Turk Telekom infrastructure, and that ADSL/VDSL subscriptions are also supplied over Turk Telekom infrastructure. That is the access-resale or wholesale-access economics familiar across many broadband markets. Farknet can own the customer, set retail packaging, operate local support, manage billing, and perhaps differentiate by responsiveness. But the physical local loop and much of the regulated access cost base sit outside its direct control.

The customer may see Farknet; the line economics may still be anchored by Turk Telekom.

The second surface is local or semi-owned access. Farknet's package page and footnotes describe "FIBER GPON INTERNET" as Farknet infrastructure available only in pilot areas. The same pilot-area limitation appears for "no-infrastructure wireless internet." Those phrases matter because they imply an attempt to escape pure resale economics in selected locations. Owned or locally controlled GPON and fixed wireless can create a real edge if the operator chooses streets, apartment blocks or underserved districts where national operators are slow, congested or expensive.

It can also destroy cash if deployment is scattered, installation cost is high, or maintenance density is too low.

The difference between those two surfaces drives the whole judgment. Wholesale-dependent access can scale more quickly because the civil-work burden is reduced. It can also be commoditized quickly because rival retail ISPs may sell over the same underlying network. Owned local access can create differentiation and better gross margin if utilisation rises, but it requires upfront capital and operational discipline.

A small operator can afford one or two bad cabinets, a few low-take-up streets, or some expensive support calls; it cannot afford a repeated pattern where every new pocket of coverage needs fresh capex before the last pocket has paid back.

The package prices hint at that tension. The wireless plans are much more expensive than DSL and FTTH, and the installation fee for wireless/no-infrastructure service is much higher than for mainstream fiber or xDSL. That pricing is economically coherent. A radio-based local access service has to recover site hardware, installation labour, alignment, maintenance truck rolls, customer equipment and eventual replacement. If it offers service where wired infrastructure is unavailable, it can command a scarcity premium.

But that premium is valuable only when customers have few substitutes and remain long enough to repay the installation and site costs.

The pilot GPON price is harder to read. A 100 Mbps symmetrical-style GPON package at a low monthly price can be a strong building-acquisition tool, a legacy tariff, a limited promotional offer, or a signal that the operator can serve certain dense sites cheaply. Without homes-passed and take-up data, it should not be treated as proof of superior unit economics. It does, however, show the strategic direction: Farknet wants to be more than a pure reseller in at least some local pockets.

Registry evidence proves capability, not cash

Farknet's network-resource evidence is useful because it is concrete. RIPE records show AS205935, as-name "farknet," assigned to ORG-FTVB1-RIPE. The organisation entity identifies Farknet Telekomunikasyon ve Bilisim Hizmetleri Ltd. Sti. as a Turkish LIR. The address allocation 185.201.132.0 through 185.201.135.255 is recorded as TR-FARKNET-20170501, and four /24 assignments inside that /22 are labelled for FarkNet Telekom or Farknet. Route objects exist for three /24s with origin AS205935.

Current route visibility narrows the picture. RIPEstat and public BGP views show AS205935 visible with three IPv4 /24s and 768 IPv4 addresses. The originated prefixes are 185.201.132.0/24, 185.201.134.0/24 and 185.201.135.0/24. There is no visible IPv6 BGP for the ASN at the query time, despite IPv6 route objects appearing in routing-consistency data. RPKI validation is valid for sampled AS205935-origin prefixes. Hurricane Electric and BGP.Tools corroborate the small active footprint and two observed IPv4 neighbours.

That is a meaningful operational record. An active ASN with originated prefixes, route objects, valid RPKI signals and stable visibility is not just marketing copy. It means Farknet has public routing control for part of its address space. It can number customers or infrastructure, maintain registry entities, and participate in interdomain routing.

But the same evidence also limits the claim. A /22 allocation is 1,024 IPv4 addresses. The current AS205935 public view shows 768 IPv4 addresses originated by Farknet's own ASN. One /24, 185.201.133.0/24, is assigned in RIPE under the same Farknet block, but RIPEstat sees it currently originated by GIBIRNet, AS208972. That could be a benign operational arrangement, a route policy choice, a customer or upstream handling detail, or evidence that part of the numbered footprint sits behind a partner. It should not be converted into a durable relationship claim without more evidence. It does, however, reinforce the supplier-dependence theme.

There is another caution. RIPEstat routing consistency lists 213.14.244.0/24 and several IPv6 /32 route objects as present in WHOIS or IRR but not visible in BGP at the query time. This is normal enough in network operations: not every registry entity is actively announced. For economic analysis, it matters because registry scale can look larger than live routed scale. A company may hold, register, prepare or preserve resources that are not yet monetized. Analysts should not count those entities as subscribers, coverage, revenue, or resilience.

The clean conclusion is this: Farknet has real number-resource evidence and an active routed footprint, but the network is small and partly dependent on observed upstreams. The registry layer proves capability. It does not prove utilisation. It does not prove that Farknet has filled streets with paying customers, that its GPON pilot is dense, or that its wireless network is cash-positive. The economic essay therefore has to test cash conversion after acknowledging operational legitimacy.

Upstream dependence is the hidden price of reach

Farknet's public BGP view shows two observed neighbours: Turk Telekom AS9121 and GIBIRNet AS208972. RIPE IRR policy also references AS34984, but RIPEstat did not see that peer in BGP at the query time. In a small regional ISP, two observed neighbours can be adequate for service, especially if one is the dominant national infrastructure player and the other is a local or regional upstream. It is not the same as strong route diversity or meaningful peering bargaining power.

The difference matters for costs and resilience. A local access provider's retail customer judges service by outages, latency, video performance, gaming experience, and support responsiveness. If upstream choices are limited, Farknet may have less leverage over transit price, repair priority, congestion, and routing quality. If one partner originates part of the address footprint, the operational boundary is even more mixed. The public evidence does not show an exchange-point strategy, private peering portfolio, or visible IPv6 deployment. That does not mean none exists privately, but the public BGP record does not demonstrate it.

Turk Telekom dependence appears in two layers. One is upstream or routing. The other is wholesale access. The company's own pages say FTTH and xDSL are supplied over Turk Telekom infrastructure. Farknet's announcements repeatedly tell customers that Turk Telekom port, transmission, THK and other fee changes will be reflected in retail prices. This is not a vague risk. It is a public pass-through mechanism.

Pass-through is not necessarily bad. In a high-inflation market, hiding wholesale cost changes can burn cash quickly. A small ISP that updates retail prices when access costs move may protect gross margin. The problem is customer tolerance. If customers believe Farknet is selling the same physical infrastructure as larger operators, a price increase has to be justified by service quality, local attention, shorter repair loops, or a still-attractive final price. If the customer sees only a rising bill, the supplier's tariff reset becomes Farknet's churn event.

There is a second upstream problem: scale. A large operator can absorb procurement cycles, spare inventory, vendor terms, network-engineering staff and traffic growth across a broad base. A small operator faces lumpier costs. A new backhaul commit, a cabinet refresh, a billing-system issue, an outage truck roll, or a radio replacement can consume the margin from many subscribers. If the supplier base is narrow, the operator has less room to arbitrage prices or technical quality.

For Farknet, supplier dependence is therefore not just a technical fact. It is the main economic hinge. Wholesale Turk Telekom access offers reach without building every last metre. GIBIRNet visibility may offer routing or operational support. Those relationships let a regional ISP exist. They also keep the business from fully owning the cost curve. The company wins only if local customer acquisition and retention are cheaper than the margin surrendered to suppliers.

Unit economics begin at installation

The public tariffs point to a business where installation recovery matters. Farknet lists installation and activation charges for fiber and xDSL, relocation charges for fiber and xDSL, a pilot GPON installation fee, and a much larger installation fee for wireless/no-infrastructure service. It also publishes modem-related notes: modem rental included in some packages, and a sizeable charge if a fiber HGW is not returned after termination.

Those details reveal practical economics. In consumer broadband, the first month is often the least profitable period. There is address checking, sales handling, paperwork, possible number/operator transfer, site visit, line activation, CPE provisioning, customer education, and early support. If the customer churns quickly, the operator loses even when the recurring tariff looks profitable. Installation fees and modem-return charges are not merely punitive terms; they are tools to protect the payback period.

Farknet's no-commitment positioning makes that payback problem sharper. A taahhutsuz offer, meaning no long commitment, lowers customer friction. It may be especially attractive to renters, students, small shops, or households burned by long contracts. But every reduction in lock-in increases the importance of either upfront installation recovery or low churn. If the operator cannot collect enough at installation and cannot keep the subscriber long enough, the no-commitment offer becomes an acquisition subsidy.

The pricing bands suggest differentiated economics by product. FTTH and DSL monthly prices cluster tightly. That suggests strong market comparability and limited headroom. Wireless pricing is higher, likely because it solves a different problem: no adequate wired infrastructure. A customer without a good fixed-line alternative may accept lower nominal speed at a higher price if the service is available and stable. That scarcity premium can make wireless attractive for a local ISP, but wireless also carries capacity limits.

The operator must manage contention, line of sight, weather exposure, spectrum conditions, rooftop/site access, and CPE alignment. A few high-usage customers in a small sector can change service quality for others.

Pilot GPON has the opposite profile. Fiber-to-the-building or fiber-to-the-home can offer better long-term capacity and lower unit operating cost after buildout, but the upfront civil and electronics costs are less forgiving. The economics depend on density and take-up. A GPON port with many paying customers on a short route can be attractive. A scattered build with low adoption ties up capital and creates maintenance liabilities.

Farknet's public record does not disclose premises passed or active split ratios, so the right posture is cautious: own GPON could be the route to defensible economics, but only if deployed in dense pockets with disciplined customer conversion.

The same applies to customer support labour. A regional ISP's strongest advantage can be local trust: the technician knows the district, the support line speaks plainly, the repair visit is faster than a national call centre can manage. That advantage has a cost. Support labour must be paid every month, not just when a new subscriber signs. If the customer base is dense, support labour is leveraged. If customers are scattered between Konya, Bozyazi and other pockets, travel and scheduling can eat the premium.

Customer concentration cuts both ways

The public evidence gives locations but not concentration. Konya appears as the core operating base, with Meram and other Konya geographies visible in company and network records. Bozyazi/Mersin appears as a branch. Social and map traces also point to local presence, but they are too thin to support hard claims about subscriber density.

Customer concentration is usually a risk in finance. For a small ISP, it can be an advantage if it means apartment-by-apartment or neighbourhood-by-neighbourhood density. One building with high take-up can pay back installation faster than many isolated customers. A street with multiple nearby wireless subscribers can make a radio site profitable. A local shop cluster can reduce truck-roll time. A known local brand can lower sales cost because neighbours talk.

The danger is that concentration without contracts creates fragility. If a building manager changes, a local competitor discounts, a backhaul segment fails, or a better national fiber offer arrives, the operator can lose a meaningful pocket of revenue quickly. A small operator may have fewer statistical offsets than a national carrier. That is why retention after price increases is a key missing fact. Farknet's announcements show repeated tariff updates across 2024, 2025 and 2026. The unknown is whether customers accepted those updates because alternatives were worse, or whether price-sensitive users churned.

The Bozyazi branch is strategically interesting. A branch outside Konya implies either a deliberate local expansion, a pocket of customer demand, a local relationship, or an attempt to serve an underserved area. It could improve economics if the branch anchors a dense wireless or fiber pocket. It could worsen economics if it adds fixed overhead without scale. The public page alone cannot decide that. It does show that Farknet's ambition is not limited to one office page and one ASN.

Payment infrastructure also points to retail seriousness. The company lists online payment, bank channels, automatic-payment options and alternative payment providers. That matters because collections can be underrated in small broadband economics. The best network is not useful if bills are late, payment friction is high, or customers churn because payment is inconvenient. Multiple payment channels make the company look like a normal retail utility. But again, they do not prove the number of payers.

The right inference is measured. Farknet has a local retail surface and likely a collection process suited to consumer broadband. It has not publicly disclosed concentration, churn, or revenue by location. The economics could be excellent in a few dense pockets and mediocre elsewhere. Without concentration data, the safer judgment is that local density is the opportunity and scattered expansion is the risk.

Pricing power depends on being more than cheaper resale

Farknet's public copy contains a revealing argument: why pay more for the same infrastructure? That is a sharp retail proposition, but it also defines the trap. If the product is "same underlying infrastructure, lower price," then Farknet has to protect margin through lower overhead, better local acquisition, faster support, or careful pass-through. It does not own a monopoly on the physical line in those wholesale cases.

In Turkey's fixed broadband market, the national context is moving toward more fiber, but xDSL remains large. Turk Telekom investor materials using regulator data show more than 21 million fixed broadband subscribers at the end of Q1 2026, with FTTH/FTTB above 10 million, xDSL still above 8 million, and cable above 1 million. Alternative operators have meaningful but much smaller fiber length than Turk Telekom. That context says two things at once. Demand for fixed broadband is real and growing into higher-speed tiers. The incumbent's infrastructure remains central.

For a local ISP, that creates a narrow but real opening. National operators may not optimise every district. They may have slower local sales, less flexible support, or weaker presence in small pockets. A regional operator can win customers by being available, answering calls, and solving edge cases. But national operators can also discount, bundle mobile, advertise heavily, and absorb short-term margin pressure. Pure price competition against a larger balance sheet is dangerous.

Farknet's no-commitment packages can attract customers who dislike lock-in. That is useful, but not a moat. A no-commitment promise can become a churn invitation if service is merely adequate. Pricing power comes when customers stay despite cheaper offers because the local operator reduces hassle. In broadband, hassle includes installation delays, recurring faults, opaque bills, slow repairs, and the fear of being ignored. If Farknet's local support genuinely lowers those costs for customers, it has pricing power beyond speed. If not, it is exposed.

The repeated wholesale-related price notices show Farknet is willing to move retail prices. That is financially rational. It also means the company cannot indefinitely win on cheapness. A customer who joined because Farknet was lower than alternatives may reconsider when pass-through increases narrow the gap. The more Farknet can shift the value proposition toward reliability and local accountability, the more resilient it becomes. The less it can do that, the more every wholesale reset becomes a competitive shock.

Pricing power is also product-specific. Wireless no-infrastructure service may have more pricing power where wired alternatives are absent. Pilot GPON may have pricing power where Farknet controls the in-building or local last-mile experience and can offer better upload or stability. Turk Telekom-based FTTH and xDSL have less structural pricing power, unless Farknet's local service quality is markedly better.

Capital renewal is the real maturity test

The economic danger for regional ISPs is not always the first build. It is the second cycle. The first cycle can be funded by founders, supplier credit, customer installation fees, or local enthusiasm. The second cycle arrives when radios age, power supplies fail, CPE is not returned, fiber drops are damaged, upstream commits need renewal, software systems require maintenance, and customers expect higher speeds without accepting proportional price increases.

Farknet's public evidence shows active service, but not capital age. The company says it has operated since 2014; RIPE records begin in 2017. That is long enough for multiple equipment cycles. Routers, switches, radios, ONTs, modems and batteries do not remain new. A small operator has to decide whether to sweat equipment, refresh it, or focus new money only on visible expansion. Sweating equipment preserves cash until faults rise. Refreshing equipment protects service but consumes cash before customers notice. Expanding while under-refreshing the base can create growth that later turns into reliability debt.

The incumbent sector context underscores the pressure. Turk Telekom's own 2026 guidance points to high capex intensity, including fiber spending. A large incumbent can spread that over millions of customers and a broad balance sheet. A regional operator cannot copy that scale. It has to choose precise local investments where customer density and retention justify the spend. In that sense, Farknet's smaller scale is not automatically a weakness. It can be an advantage if investment is disciplined and local. It is a weakness if capital is spread across too many lightly penetrated pockets.

Currency and inflation add another layer. Turkey's June 2026 CPI was above 32% year on year. Farknet's customers pay in lira, and retail prices can be adjusted, but imported active equipment, optics, radios, CPE, software, IPv4 scarcity costs and some upstream inputs are often linked directly or indirectly to foreign currency. Even when a supplier invoices in lira, replacement cost may reflect exchange-rate reality. A small ISP that underprices today may discover that the modem, radio or router it needs to replace tomorrow costs much more than the original payback model assumed.

This is why installation fees and modem-return charges matter. They reduce the capital leakage from short-tenure customers. But they cannot solve long-term renewal alone. Renewal requires recurring margin. A customer base that produces enough monthly cash after wholesale fees and support can finance upgrades. A base that only covers the immediate access cost leaves the operator dependent on new installation fees, delayed maintenance, or owner funding.

The firm question, then, is not whether Farknet can sign customers. It is whether its local pockets reach the utilization threshold where every added customer becomes cheaper to serve, rather than adding another support obligation. That threshold is different for wholesale FTTH, DSL, pilot GPON and wireless. The company has to manage all of them without letting one product's margin hide another product's capital drain.

Regulation and compliance are operating costs, not just badges

Farknet's public pages refer to BTK licensing, the 5809 Electronic Communications Law, KVKK obligations, safe-internet profiles, subscriber data, billing data, network and traffic data, and transfers to competent public bodies and service providers. Those statements are not just legal wallpaper. They identify a regulated operating surface that costs money.

Every ISP must manage identity, subscription contracts, billing, lawful data retention duties, customer complaints, safe-internet profile changes, public authority requests, abuse handling, and privacy processes. Larger operators spread those costs across a wide base. Smaller operators experience them as fixed or semi-fixed overhead. A privacy notice and safe-internet page do not prove compliance quality, but they show the categories of obligation Farknet acknowledges publicly.

Regulation can also protect and pressure Farknet at the same time. Wholesale access regulation creates the possibility for alternative providers to sell over Turk Telekom infrastructure. Without that framework, a local retail ISP might have fewer paths to customers. But the same regime also means many providers can compete over similar wholesale inputs. Regulation opens the door; it does not guarantee margin.

BTK and Turk Telekom wholesale decisions are therefore central to the business, even when they are not Farknet-specific. Farknet's own announcements make the connection explicit by linking price changes to Turk Telekom port and transmission costs. For a retail customer, that is an explanation. For an investor or strategic analyst, it is a margin map. The company is telling the market that a meaningful part of its cost base can be reset by a supplier and regulator-approved tariff structure outside its own control.

Geopolitics enters mainly through capital and suppliers rather than direct sanctions evidence. Turkey's macro environment, currency movement, imported network hardware, financing cost, and telecom-infrastructure policy all shape renewal economics. There is no public evidence here that Farknet faces a special sanctions problem or political-risk event. The broader risk is ordinary but severe: high nominal cost growth and a customer base that may resist price increases faster than wholesale and equipment inputs rise.

Unofficial signals are useful only when kept in their place

Public social and map traces around Farknet and Bozyazi show a local-facing brand, modest online presence, and consumer-service positioning. They are helpful because small ISPs often leave their strongest evidence in local channels rather than polished investor documents. A Facebook page, Instagram snippet, map listing, or transit/map result can show that customers or locals have encountered the brand. It can also mislead.

The discipline is to treat unofficial signals as directional, not determinative. A social post about fiber activation does not prove paying take-up. A page with a small follower count does not prove business weakness, because local ISPs may sell through phone calls, branches, neighbourhood referrals and apartment managers. A map listing with limited reviews does not prove low customer count. Conversely, polished website copy does not prove service quality.

For Farknet, unofficial signals reinforce the conclusion that the company is local and retail-facing. They do not resolve the economic unknowns. The hard facts remain in tariffs, branches, payment channels, RIPE records, BGP visibility, and wholesale pass-through notices. The soft signals can suggest where to look next: Bozyazi branch activity, wireless no-infrastructure coverage, local apartment/building penetration, and support reputation after price increases.

This distinction matters because a regional ISP can look either smaller or larger than it is, depending on which public trace one overweights. A network analyst may overvalue ASN evidence. A consumer researcher may overvalue social evidence. A financial analyst may dismiss both because no audited accounts are visible. The right approach is to combine them, then preserve uncertainty. Farknet is visibly operating; the scale and cash quality of that operation remain unproven.

What would make Farknet defensible

The strongest positive case would be dense local own-infrastructure pockets. If Farknet's pilot GPON areas have high take-up, low churn and low support cost, the economics could be materially better than the public record can currently prove. A dense GPON pocket lets the operator amortise fiber, splitters, OLT ports, installation labour and backhaul over many users. It also creates a local service reputation that wholesale-only rivals cannot easily replicate.

The second positive case would be wireless scarcity. The wireless/no-infrastructure packages are expensive relative to mainstream FTTH and DSL, but that may be rational where customers lack wired alternatives. If Farknet can serve specific underserved localities with controlled contention and reliable support, wireless could produce attractive ARPU. The installation fee suggests the company is trying to recover upfront cost. The risk is capacity and maintenance; the upside is that unavailable infrastructure can create willingness to pay.

The third positive case would be local support as retention capital. A small ISP does not need to win every customer. It needs to win enough customers who value quick local help. If Farknet's support reduces outage duration, installation friction and billing confusion, then it can retain customers through price resets. That would turn a local branch from overhead into a moat.

The fourth positive case would be supplier discipline. If Farknet has stable upstream pricing, good wholesale terms, and enough routing redundancy for its size, it can keep service quality predictable while avoiding overinvestment. The public BGP record is small but not alarming. It becomes a concern only if traffic growth, outage exposure or pricing pressure outgrow the two-neighbour model.

The fifth positive case would be conservative capital allocation. A regional ISP wins by saying no to bad coverage. It should not build every street, serve every remote customer, or chase every low-ARPU installation. It should invest where density, local reputation and payback are visible. Farknet's public record does not show whether management behaves that way. The survival of the company since 2014 suggests some operating discipline, but survival alone is not proof of attractive economics.

What would reverse the view

Several facts would change this assessment quickly.

The most positive reversal would be verified subscriber data showing rising active customers, low churn after the 2025 and 2026 tariff updates, and a growing share of customers on Farknet-controlled GPON or wireless infrastructure. A payback schedule showing installations recovered within 18 to 24 months would materially strengthen the case.

A second positive reversal would be evidence of long-term backhaul and upstream contracts with price stability, direct peering growth, or exchange presence that lowers traffic cost and improves resilience. Visible IPv6 deployment would not transform the economics alone, but it would signal network maturity and future readiness.

A third positive reversal would be audited revenue, EBITDA, capex and debt data showing that recurring margin funds replacement capex without relying on constant upfront installation charges or owner injections. That is the cleanest test for a regional ISP entering a renewal cycle.

The negative reversals are equally clear. If most customers remain on wholesale Turk Telekom access and gross margin is thin after port, transmission, modem, support and tax-inclusive pricing, then Farknet's local brand has limited economic value. If tariff increases caused meaningful churn, the pass-through model is fragile. If wireless customers require frequent truck rolls or if GPON pilot areas remain low-density, owned infrastructure would become a capital sink rather than a moat.

Operational evidence could also reverse the view. Loss of active ASN visibility, loss of RIPE/LIR standing, invalid RPKI or route instability, unpaid wholesale disputes, or public evidence of chronic outages would weaken the active-operator case. None of that is established in the current public record. The current evidence supports an active but small operator with conditional economics.

Conclusion: local reach must become retained cash

Farknet's public evidence is stronger than a shell profile and weaker than a proven compounding infrastructure case. The company has a real local identity, published tariffs, customer-payment channels, BTK-facing regulatory claims, a Konya office, a Bozyazi branch, RIPE LIR status, an active ASN, visible IPv4 routing, and public price notices tied to wholesale cost changes. That is enough to treat Farknet as an operating regional ISP.

It is not enough to assume durable pricing power. Much of the visible retail offer depends on Turk Telekom infrastructure. The owned or locally controlled GPON and wireless surfaces are the more interesting part of the story, but they are described as pilot or limited-area services and lack public take-up data. The number-resource layer confirms operational capability; it does not count customers. The payment page confirms collection channels; it does not prove cash generation. The tariff page shows retail ambition; it does not show margin.

The investment and strategic judgment therefore turns on cash conversion. If Farknet uses local knowledge to choose dense, underserved pockets, recovers installation costs, keeps churn low after price resets, and limits capex to areas where support labour and backhaul can be leveraged, it can turn local fibre and wireless reach into a defensible regional business. If it remains mainly a lower-priced seller over the same incumbent infrastructure, with small owned-network islands and little route diversity, it will be squeezed between wholesale resets, imported equipment costs, customer price sensitivity and the next renewal cycle.

The firm conclusion is that Farknet has earned operational credibility, not economic inevitability. Its next test is not branding, licensing, or BGP visibility. It is whether each lira of local network reach becomes retained monthly cash before the equipment, backhaul and civil-work bills come due again.

Sources