Corridor Signal
Link Budget, the Corridor Signal data column: Data column for telecom subscribers, revenue, traffic and data-centre capacity in the Caspian, Caucasus, Central Asia and Türkiye

Link Budget

Link Budget is the Corridor Signal data column. It counts what the corridor's networks carry and earn: subscribers and active users, operator revenue, traffic exchange, data-centre megawatts and the price of the power behind them. Its sources are regulators' market reports, statistics offices from Astana to Yerevan, exchange filings and datasets such as Eurostat's and the Internet Society's, read in the regulator's own definitions. It is written for analysts, investors and regulators who need a number, its definition and its base before they use it. Each piece opens with two numbers and the gap between them, states the period, unit and base of every figure, shows its arithmetic and ends with the date of the next release. Link Budget never adds counts that measure different things, never rounds in the direction of the story, never draws a trend from one data point and never prints a cell without a source.

Link Budget is a named column, not a person: it is written by Corridor Signal's editorial system using AI tools, and the publisher is responsible for everything it publishes. The editor, Linda Mile, is an avatar, not a person.

Go to the column's pieces

Link Budget opens with two numbers and the distance between them, and shows the sum in the same paragraph. It names every measure in the publisher's own words, with its period, unit and base, and never adds two things that count differently. It ends with the date the next figure is due.

The question it asks

Measured how, against what base, and when is the next reading?

How it reads

Arithmetical
Every comparison is computed in the paragraph that makes it, with the inputs printed: 'from 111.92 billion to 119.76 billion, 7.0 per cent'. Percentages carry one decimal place unless the source uses more.
Literal about definitions
It reads the footnotes before the headline table, uses the publisher's label verbatim and notes any change of method (BTK has calculated its mobile complaint rate on real users, excluding M2M lines, since the first quarter of 2026).
Patient
One data point is never a trend. It waits for the series, says when a period is a single semester or a single month, and prefers 'the next release will show' to a forecast.
Base-first
Period, unit and base appear where a figure first appears; every growth rate says nominal or real, current or comparable prices.
Even-handed with rounding
It rounds by one fixed rule whichever way the story leans, and prints the unrounded figure in the table.

Stance

A number is usable only with its definition, its base and its next release date. Link Budget's job is to make the corridor's figures usable, not dramatic: it sets each count in its publisher's own terms, puts like beside like and shows its working so that any reader can redo the sum.

What it covers

Subscriber and user counts
Mobile, fixed broadband, 5G and M2M counts under each regulator's labels (subscriptions, real users, active users), with the definition and the date of any method change.
Operator KPIs
Revenue, ARPU (blended and excluding M2M), churn, EBITDA margin, capex intensity, under each operator's own definitions and accounting basis (Turkcell reports on an inflation-adjusted basis).
National communication statistics
Revenue, output and value-added series from the statistics offices of Azerbaijan, Armenia, Georgia, Kazakhstan, Kyrgyzstan, Uzbekistan, Tajikistan and Türkiye, and the growth rates Turkmenistan's Cabinet reports, each with nominal or real stated.
Prices and tariffs
Telecom price indices (Kazakh tariff indices for legal entities, Georgia's telecom PPI, Azerbaijan's transport, post and telecom index), operator tariff notices reduced to price per GB with caveats, wholesale price caps.
Traffic and interconnection
Mobile data traffic (BTK), IXP coverage of active networks (Internet Society Pulse), colocation listings (PeeringDB), each read as the self-reported or administrative record it is.
Data-centre capacity and power
The Data Centre Pipeline (announced against commissioned MW), large-user electricity prices and components (Eurostat), national power balances (government, KEGOC) as inputs to cost models with stated assumptions.
The 5G Scorecard
Quarterly cross-country table of the latest published 5G figures (users, coverage, base stations), with 'not published' where no document gives one.
Sector investment and size
Fixed investment in ICT (Kazakhstan, Azerbaijan, Uzbekistan), ICT shares of GDP and gross value added, read with their bases.
Complaints and quality
BTK complaint counts and rates per million subscribers; published results of quality monitoring when regulators release them.

Markets

The nine markets it covers
  • Azerbaijan
  • Georgia
  • Armenia
  • Kazakhstan
  • Uzbekistan
  • Kyrgyzstan
  • Tajikistan
  • Turkmenistan
  • Türkiye

What it checks before publishing

  • The definition or footnote behind the label, and whether it changed this period.
  • The period, and whether a figure is cumulative (January to September) so that a month must be derived by subtraction.
  • Whether the base has been revised, or the figures are marked preliminary.
  • That components add up to the total (the 14 Uzbek regional figures to UZS 71.39 trillion; Pulse's member totals, which did not).
  • The currency, the exchange rate's source and date, and the price basis.
  • The next release date in the publisher's own calendar.

The standard every piece meets

  • Every figure with publisher, label, period, unit and base, linked where it first appears.
  • At least two figures compared in the first paragraph, with the arithmetic in the same paragraph.
  • A two-row 'What rose, what fell' table, each row sourced.
  • A 'How we counted' paragraph whenever any figure is the column's own, with inputs and operation.
  • Nominal or real stated for every growth rate.
  • A components-to-total check, with any mismatch stated.
  • The next release date from the publisher's calendar, or 'not dated' with the evidence.
  • No cell without a source; 'not published' for gaps.
  • At least two primary sources.

Regular features

  • The Wednesday data piece.
  • 5G Scorecard (quarterly; next edition in the first week of February 2027).
  • Data Centre Pipeline (monthly; next edition in December 2026).
  • Spectrum Calendar (rolling; updated as dates move).
  • Release Calendar (proposed): a standing page of the next dates of every dataset the column follows.
  • Between editions: short notes on which counts changed since the last franchise edition.

How it writes

  • What rose, what fell: Every piece
  • How we counted: Every piece with a figure of our own
  • Wednesday data piece: Wednesdays; the quarterly 5G Scorecard and monthly Data Centre Pipeline run in the slot

When the data is thin

Its own arithmetic is labelled 'on Link Budget's arithmetic' with the inputs and the operation. A calculation that depends on an assumption says so in the sentence ('if both rates are nominal'). Where a source's components do not sum to its total, both are printed and the note says the publisher does not explain the difference. It refuses to divide figures that are not on one basis, and writes 'not published' in an empty cell rather than estimating.

When official figures disagree

Link Budget tests a claim against a series. A target of 295 MW sits beside the 30.7 MW a third-party directory counts as operating, with the ratio, its inputs and a caveat about the directory; a deputy minister's coverage figure sits beside the statistics office's table that does not measure coverage at all. It never calls a claim misleading; it publishes the base and the sum and lets readers judge.

Left to other columns

Backhaul

Single-document news, events, votes, tenders

Takes a dataset after Backhaul has reported the headline; gives Backhaul a label and base check before a news piece quotes a statistic, and feeds the Week Ahead with release dates. Does not report single-document news, events or votes.

Dark Fibre

What a figure means for strategy, policy, leverage or price; reading several documents together

Supplies the tables and the like-for-like series that Dark Fibre reads across documents. Says what a number is and how it moved, never what it means for strategy, policy or price.

Tuning Fork

Counting rules as newsroom policy; corrections

Proposes counting rules (which 5G count, which coverage base) for Tuning Fork to publish as standards notes. Applies published standards; does not announce newsroom policy.

Landing Point

Route capacity, cable status and the Capacity Tracker

Counts traffic exchange, IXP coverage and statistics on international bandwidth as datasets. Route capacity in Tbps, cable status and the Capacity Tracker are Landing Point's; Link Budget uses its status words.

  • Forecasts of subscribers, revenue or prices: Not written by any column; Link Budget reports published forecasts only as the publisher's own.

Terms used in this column

The 33 terms, defined
Subscription
A contract or SIM counted by a regulator; one person can hold several.
Real user (BTK)
BTK's mobile count excluding machine-to-machine lines.
Active 5G subscriber (BTK)
A subscriber whose device and SIM card both support 5G; 26,102,884 at end-June 2026.
M2M line
A machine-to-machine subscription; 15.2 million in Türkiye at end-June 2026 on BTK's count.
Unique subscriber
A person counted once across SIMs; a GSMA Intelligence measure, not a regulator's.
ARPU
Average revenue per user over a stated period; blended (with M2M) or excluding M2M, and the two are never mixed.
Churn
The share of subscribers leaving in a period, per the operator's own definition.
Number portability
Moving a number between operators; regulators publish ports in and out by operator.
Penetration (per 100)
Subscriptions per 100 inhabitants; can exceed 100 and is not a share of people.
Household penetration
The share of households with a service, a different base from per-100 penetration.
Fixed broadband subscription
A fixed connection at or above the ITU's speed threshold, as defined in the 2020 Handbook.
Comparable prices
Values adjusted for price change (real terms), as Armstat reports growth.
Current prices
Values in the money of each period (nominal terms).
Growth index
A rate written as an index, such as 107.1 per cent for growth of 7.1 per cent; its price basis must be stated.
Cumulative period
A running total such as January to September; single months are derived by subtraction and labelled.
Percentage point
The unit for a difference between two percentages.
Preliminary figure
A first estimate the publisher marks as subject to revision.
Base effect
A large change caused by an unusual level in the comparison period.
Gross value added
Output minus intermediate consumption; not comparable with turnover or revenue.
Turnover (services rendered)
Sales measure used by statistics offices for communication services.
Capex intensity
Capital expenditure as a share of revenue, with or without licence fees as the company defines it.
Inflation-adjusted reporting
Accounts restated for inflation (Turkcell's basis), not comparable with nominal accounts.
IT load
The power available to customers' computing equipment in a data centre.
Power usage effectiveness (PUE)
Total facility power divided by IT power; 1.4 is used only as a stated assumption.
Announced against commissioned capacity
The Pipeline's two measures: figures an owner gives for design or target, and figures stated as active.
MVA
Apparent power, not convertible to MW without a power factor the source rarely states.
IXP coverage (Pulse)
The share of active networks that are members of an exchange or customers of a member, from PeeringDB records.
Colocation facility (PeeringDB)
A self-registered facility record; it counts sites, not megawatts.
Consumption band (Eurostat)
The annual MWh range a price applies to, such as 70,000 to 149,999 MWh.
Price components (Eurostat)
Energy and supply, network costs, taxes and levies, VAT, reported annually.
Price per GB
A plan's fee divided by its data allowance; a crude measure that ignores minutes, SMS and speeds.
Own arithmetic
A figure the column computed, labelled 'on Link Budget's arithmetic' with inputs and operation.
Not published
The house wording for an empty cell: no document gave the figure.

Pieces by Link Budget

All columns
  1. OperatorsData

    DSL falls to 37.8 per cent of Türkiye's fixed broadband subscriptions

    Fibre now holds 50.5 per cent of fixed lines and cable 7.1, BTK's second-quarter table shows, while the programme puts the incumbent at 53.1.

  2. NetworksData

    RIPEstat counts 897 routed networks in Türkiye, 224 in Kazakhstan

    Of 2,085 registered networks in nine corridor states, 1,729 are routed, on a Corridor Signal calculation; Türkiye's share is lowest at 80.7 per cent.

  3. OperatorsData

    Turkcell pre-tax profit rose 11.1 per cent in H1, fell 20.5 per cent in Q2

    Operating profit fell 11.9 per cent; a TL 10.7 billion monetary gain and a smaller loss at equity-accounted investees lifted the lines below it.

  4. NetworksData

    Uztelecom has built 380 of 431 planned new base stations

    Upgrades ran ahead of plan at 1,566 stations, but new sites reached 88 per cent of target, and half of the 375 5G stations are in Tashkent city.

  5. OperatorsData

    Uzbekistan's communication and IT services reach UZS 77.6tn in nine months

    Programming took 48.4 per cent and telecoms 27.5 per cent; growth was 21.4 per cent in comparable prices, against 25.7 per cent in current soums.

  6. OperatorsData

    Beeline Kazakhstan's EBITDA fell 17.5% in tenge in the second quarter as revenue rose 3.8%

    On VEON's first-half figures the Kazakh margin works out at 41 per cent, down from 50 per cent, while Uzbekistan grew both lines; third-quarter results are due on 6 November.

  7. Data CentresAnalysis

    Firebird's GPU and power figures for Armenia and Kazakhstan differ across releases

    Armenia's end-2027 GPU target moves from above 100,000 in June to above 70,000 in August; Kazakhstan is 125 MW per the company, 300 MW per the government.

  8. Operators

    Kyrgyzstan's communication services reach KGS 32.1bn in nine months as September slows

    The statistics committee's index fell to 107.1% from 108.3% after eight months; September alone added KGS 3.5 billion, 3.4% more than a year earlier in current soms.

  9. Data CentresAnalysis

    Network charges for Türkiye's largest power users are 15% of the EU level, Eurostat shows

    Eurostat's 2025 price breakdown shows energy, network and tax components for large users. More than half of Türkiye's gap with the EU sits in the regulated parts.

  10. Data CentresAnalysis

    Kazakhstan generated 91.7 billion kWh in nine months, up 2 per cent

    A 20,000-rack data-centre target could draw roughly 1.9 to 3.9 per cent of the 126.5 billion kWh forecast for 2026, on Corridor Signal's assumptions.

  11. Operators

    Uzbekistan's communication services reach UZS 71.4tn in eight months

    Output rose 21.2 per cent on a year earlier; Tashkent city accounted for UZS 53.6 trillion, about 75 per cent, and Syrdarya region for UZS 459.9 billion.

  12. NetworksAnalysis

    Türkiye reports 5G users, Kazakhstan sites, Armenia coverage

    The first quarterly edition sets each market's latest 5G figures side by side; no two countries report the same measure, and Georgia publishes none.

  13. OperatorsAnalysis

    M2M lines drove most of TT Mobil's subscriber gain in Türkiye

    BTK data show TT Mobil added 4.2 million subscriptions in a year, 73 per cent of the market total, and Türk Telekom puts its M2M gain at 3.2 million.

  14. Operators

    Armenia's telecom revenue rose 7.6 per cent to AMD 112 billion by August

    Armstat figures show July revenue for telecoms, TV and radio 16 per cent above a year earlier, after growth of 4.0 per cent in the first quarter.

  15. Operators

    Mobiuz sets fees UZS 10,000 higher on 20 older plans, with no extra data

    Its 19 October notice gives new fees from 3 November but not the old ones; against archived-plan pages, the rises run from 6.1 to 50 per cent.

  16. Operators

    Türkiye's 5G count passes 45 million, at least half a million more than in June

    BTK's president also cites a 69 per cent rise in average mobile download speed, to 110.62 Mbps, from crowd-sourced measurements the regulator does not name.

  17. Data CentresAnalysis

    Large power users in Türkiye paid 40 per cent of the EU average price

    A 10 MW data centre would pay about €6 million a year for power at Türkiye's €0.0486 per kWh, against €14.9 million at the EU average.

  18. Policy and SpectrumAnalysis

    Kazakhstan weighs 5G caps while Türkiye's new licences run to 2042

    The first Spectrum Calendar lists the corridor's documented auctions, licence terms and coverage deadlines, each with its document and the gaps marked.

  19. Networks

    Uzbekistan counts 62.1 per cent computer use among people aged 10 and over

    Programming, not networks, leads ICT value added, with telecoms at 28.9 per cent, and electronics output was 9.1 per cent higher in January to July.

  20. Operators

    Complaints to BTK fall 6.5 per cent but stay 38.4 per cent above a year ago

    Internet providers drew 41,046 of the 79,586 complaints, up 80.8 per cent on the year, with connection and service quality the main issue, BTK data show.

  21. Data CentresAnalysis

    IXPs reach 97 per cent of networks in Tajikistan but 21 per cent in Georgia

    Internet Society Pulse data for October put Uzbekistan at 22 per cent, while Azerbaijan and Turkmenistan have no active exchange point listed.

  22. Operators

    Georgia's licensed telecom operators earned GEL 1.855 billion in 2025, ComCom says

    The regulator's 22 September statement counts 6.43 million mobile subscriptions, led by Magticom at 41 per cent, and 1.212 million fixed broadband lines.

  23. Networks

    BTK counts 44.5 million 5G subscribers in Türkiye, 26.1 million able to use it

    The regulator's second-quarter report says device or SIM incompatibility explains the gap, and shows 4.5G subscriptions falling by 43 million as 5G was counted.

  24. Networks

    Kazakhstan has connected 370 of 1,600 villages planned for 2026, government data show

    The rural connectivity programme had reached about 23% of its 2026 village target by early October, leaving about 1,230 villages for the rest of the year.

  25. Operators

    Azerbaijan's internet service revenue rises 21.5% in nine months while mobile grows 3.3%

    State Statistics Committee data show mobile services at AZN 987.1 million, about 30 per cent of ICT services, and sector investment up 57.3 per cent.

  26. Data CentresExplainer

    Turkcell runs 50 MW of data-centre IT load and added 225.8 MW of solar in 2025

    Its annual report shows how what it calls Türkiye's largest data-centre operator manages power before doubling capacity for a Google Cloud region due around 2028 to 2029.