Summary
- Bangladesh Telecommunication Regulatory Commission's 2011 submarine-cable guideline defines exactly three capacities — design, initial equipped and equipped — and requires a licensee to file a tariff chart showing proposed maximum and minimum charges, approved in writing, before it sells or leases a megabit.
- Bangladesh Submarine Cables PLC uses none of those three words. It publishes “ultimate achievable”, “design”, “present lit up” and “utilisation”, prints SEA-ME-WE-5's lit capacity as 2,200 Gbps, then 1,300 Gbps, then 2,200 Gbps again, and reports its own price movement as a percentage with no base rate behind it.
The licence names three quantities
The instrument is the Bangladesh Telecommunication Regulatory Commission's Regulatory and Licensing Guidelines for Invitation of Offers/Proposals for Issuing License to Build, Operate and Maintain Submarine Cable Systems and Services in Bangladesh, dated 31 March 2011 (BTRC submarine-cable guideline). It fixes three terms a licensee is expected to use:
- Design Capacity — the ultimate capacity of a cable segment.
- Initial Equipped Capacity — capacity at the Ready For Customer Service date.
- Equipped Capacity — capacity physically provided in transmission equipment at any given time.
Three quantities, three moments: what the glass could carry, what was lit on the day the service opened, and what is lit now. The same document makes the price an instrument of record rather than a negotiating position. Before selling or leasing capacity or providing facilities, the licensee must file a tariff chart schedule showing proposed maximum and minimum charges and the justification for them, and may not sell or lease capacity before the Commission approves the tariff in writing.
On revisions the sentence is unambiguous: “The licensee shall obtain the written approval of the Commission before making any changes to the approved tariffs and charges.” The reporting duty runs in the same direction — installation type and capacity, subscriber numbers, traffic data, leased transmission links, an annual report within three months of the financial year end, and audited statements containing plant capacity.
Two limits sit on that reading. The retrieved text is OCR-quality, and the Commission's own index lists amendments to the same guideline dated 15 April 2021, 10 January 2022 and 14 February 2022 which I did not read. So this is the vocabulary of 2011, cited as such, not a claim about which tariff regime operates in 2026. What a reader can still check is the direction of the instrument: the capacity term and the tariff are both filed with the regulator, and the tariff is approved in writing before any capacity leaves the building.
The company names four
Bangladesh Submarine Cables PLC — the name the annual report prints, where the directory entry carries the older Bangladesh Submarine Cable Company Limited — reports the year ended 30 June 2024 in a letter dated 5 November 2024 (BSCPLC annual report 2023-24). SEA-ME-WE-4 at Cox's Bazar: about 4,650 Gbps of ultimate achievable capacity after consortium upgrade #6, up from about 850 Gbps, the upgrade completed in December 2023 and adding 3,800 Gbps. SEA-ME-WE-5 at Kuakata: design capacity 2,570 Gbps, present lit-up capacity 2,200 Gbps, of which 2,100 Gbps toward Singapore and 100 Gbps toward France. Combined final achievable capacity about 7,220 Gbps. Adding the two cables reproduces the combined figure; that check is mine, and it is the only arithmetic the document needs.
Then come the four words: ultimate achievable, design, present lit up, utilisation. Only one of them, design, is a term the licence defines — and the report uses it for one cable and not for the other. SEA-ME-WE-4 is described by an ultimate achievable figure, SEA-ME-WE-5 by a design figure and a lit figure, and the country's consumed bandwidth by a fourth word, utilisation. Nothing in the retrieved text says which of the licence's three definitions each column was meant to satisfy.
The five-year table carries the same vocabulary (same report):
| Financial year | Ultimate achievable (Gbps) | Lit up (Gbps) | Utilised (Gbps) |
|---|---|---|---|
| 2019-20 | 2,800 | 1,550 | 1,102.72 |
| 2020-21 | 3,000 | 2,000 | 1,563.91 |
| 2021-22 | 3,370 | 3,050 | 2,068.88 |
| 2022-23 | 3,420 | 3,050 | 2,556 |
| 2023-24 | 7,220 | 6,850 | 2,696 |
And it prints what was actually sold as a stock on the closing day, beside the revenue each line earned in the year:
| Product | Sold at 30 June 2024 (Gbps) | Revenue, FY2023-24 (BDT million) |
|---|---|---|
| IPLC | 2,146 | 3,125.57 |
| IPLC export | 863 | 30.39 |
| IP transit | 178 | 608.65 |
| IP transit export | 20 | 101.68 |
The prior-year column reads 2,115.91, 863.00, 272.10 and 20.00, so the stock of IP transit sold inside the country fell by about a third while IPLC barely moved. The report attributes a 22.69% revenue fall mainly to an approximately 20% tariff reduction and a 68-day outage of SEA-ME-WE-5, claims about 45% of the national international-bandwidth market, and names ITC operators as competitors that sometimes offer lower tariffs. It names no rate.
One cable, three lit capacities
SEA-ME-WE-5's lit capacity is printed three ways, and only the dates can separate them. The FY2021-22 report, in a letter dated 8 November 2022, already reported 2,200 Gbps lit on that system, split 2,100 toward Singapore and 100 toward France, alongside SEA-ME-WE-4 at about 800 Gbps and combined design capacity of about 3,000 Gbps (BSCPLC annual report 2021-22). An undated paper on the second and second-and-a-half light-up programmes puts SEA-ME-WE-5's ultimate achievable capacity at about 2,300 Gbps and lit capacity at 1,300 Gbps — 1,200 toward Singapore, 100 toward France (capacity and light-up paper). The FY2023-24 report puts it back at 2,200 Gbps lit against a design capacity of 2,570 Gbps (BSCPLC annual report 2023-24).
These are sequential light-up programmes, not a single series, and the three documents do not reconcile their own dates: a 2022 letter reports the higher figure before an undated paper reports the lower one. I report all three as printed and adopt none of them. The same applies to SEA-ME-WE-4, whose history the earlier report gives as 7.5 Gbps at service start in 2006, 44.60 Gbps after upgrades 1 and 2, 300 Gbps after upgrade #3 in 2011, and about 800 Gbps by 2022 — a ladder the later report's 4,650 Gbps extends by a factor of nearly six in two years, which is what an upgrade #6 adding 3,800 Gbps would do.
One number in the FY2023-24 report is worth guarding against a common misreading. The company states that SEA-ME-WE-5's initial system design capacity was 24 Tbps, and that BSCCL could initially achieve 1,500 Gbps of it. Twenty-four terabits is the consortium's whole system, not Bangladesh's share, and no article should lend the bigger number to the smaller landing.
Where the price went
The report's price story is one sentence: about 20% off, therefore 22.69% less revenue. A percentage is not a price. Twenty per cent of what, at which volume, on which of the four product lines? The document does not say, and no per-megabit figure appears anywhere in the FY2023-24 report I read.
Press reporting fills part of the gap, dated. In September 2017 the company was reported to charge a minimum of Tk 560 per Mbps while bandwidth imported from India cost about Tk 350, with a 20% wholesale cut planned after commercial launch (The Daily Star, September 2017). In February 2018 a 15% wholesale cut was reported to take the Cox's Bazar wholesale price to Tk 310 per Mbps (The Daily Star, February 2018). Those are two dated points on a wholesale ladder, one of them eight years old.
The only full price ladder I located is on a third-party mirror that reproduces BSCCL-branded tariff pages, and it arrives with a vintage problem of its own. The IPLC schedule is per circuit per month in taka: E1 at 15,000, DS3 at 75,000, STM-1 at 150,000, STM-4 at 450,000, STM-16 from 800,000 down to 550,000 as committed volume rises, 10G from 2,700,000 down to 2,000,000 above 350G in total, 100G from 23,000,000 down to 18,000,000 at 400G or more (IPLC schedule, mirrored). The customer columns are international internet gateway operators, academic and research institutions, BPO and ITES firms, call centres and software exporters — with no column for an international gateway operator and none for a generic corporate buyer. The page's own footer says last updated 26 January 2022 while its narrative dates the rates to 1 February 2018, and neither stamp makes the schedule an approved one.
The IP transit mirror is undated and just as specific: per Mbps per month, for international internet gateway customers 415 taka in Dhaka, 395 in Chattogram, 362 at Cox's Bazar and 367 at Kuakata between 100 and 4,999 Mbps, falling to 300, 280, 247 and 252 above 100 Gbps, with ISP rates higher at every tier, plus 10,000 taka of registration and a 20,000 or 100,000 taka port charge for 1G or 10G (IP transit schedule, mirrored).
The arithmetic a reader can do with the company's own numbers is a blend, and I label it as mine. IPLC revenue of 3,125.57 million taka against 2,146 Gbps sold is about 121 taka per Mbps per month. That divides a financial-year flow by a year-end stock, averages E1 circuits with 100G circuits, and mixes every customer class in the export column's shadow. It is not a tariff. It is what the published numbers will support.
What a buyer can read, and what they cannot
On 19 November 2026 the operator's tariff page returned a connection error at 11:36:25 UTC and no content (bsccl.com.bd/tariffrate). The Cox's Bazar landing-station page was readable: it describes the company as the root provider of submarine cable bandwidth, offers IPLC through co-location centres at Chittagong, Mohakhali Dhaka and Moghbazar Dhaka, and names international internet gateway operators, international gateway operators and BTRC-approved corporate users as its customers (IPLC page, Cox's Bazar). The retrieved text carries no numeric rate, and that text may be truncated.
So the perimeter looks like this. A reader can see what the gateway could carry at the end of June 2024 — 7,220 Gbps of ultimate achievable capacity, 6,850 Gbps lit, 2,696 Gbps used inside the country — and what it had contracted to sell: 3,207 Gbps across four product lines, my sum, not the company's. A reader cannot see the per-megabit price on any of those lines, because the operator's own documents reduce it to a percentage and the one schedule-shaped artefact in circulation is a third-party mirror with a 2022 stamp.
In September 2026 the operator was reported putting combined SEA-ME-WE-4 and SEA-ME-WE-5 capacity at about 7,200 Gbps with about 4,000 Gbps in use and 3,200 Gbps available, and SMW-6 in 2027 lifting its capacity to 34,000 Gbps against a regulator projection of 30,000 Gbps of national demand by 2030 (bdnews24, 3 September 2026). Those are news reports of a statement, and the statement's own definitions were not published with it. On 12 November 2026 BTW Media reported the FY2024-25 annual report as showing 3,123.4 Gbps of IPLC sold at 30 June 2025 and 323.84 Gbps of local IP transit, per-cable capacity unchanged, and no period-matched taka-per-megabit figure anywhere in the record it examined (BTW Media, 12 November 2026); I did not open that report first-hand.
The licence makes the missing artefact a specific thing rather than a mood: a tariff chart schedule, filed with the Commission, carrying proposed maximum and minimum charges, justified, and approved in writing before capacity is sold. If that schedule exists — and the licence says it must — it is the document a buyer, a competitor or a regulator would need to read a price. What the public record offers instead is a capacity vocabulary the regulator defined, replaced by four words the regulator did not, and a price expressed only as a change in itself.
Boundaries
- Everything here is a reading of documents retrieved on 19 November 2026 between 11:29 and 11:37 UTC, and the annual reports were read as extracted PDF text rather than as rendered pages.
- The 2011 guideline is OCR-quality text, and the Commission's index lists amendments dated 15 April 2021, 10 January 2022 and 14 February 2022 which I did not read; nothing here states which tariff-approval regime is operative today.
- The FY2023-24 figures are for the year ended 30 June 2024, on a letter dated 5 November 2024; capacity sold is a stock at 30 June, revenue is a flow for the year.
- The capacity paper that prints the lower SEA-ME-WE-5 lit figure is undated in the text retrieved; the storage path is not a publication date.
- Three printings of one cable's lit capacity are reported as printed. I inferred no chronology between them and adopted none.
- bsccl.com.bd/tariffrate returned a connection error at 11:36:25 UTC on 19 November 2026 — an access observation about one address at one instant, not a finding that no schedule is published anywhere.
- The Cox's Bazar landing page was readable and its retrieved text carried no numeric rate; that text may be truncated.
- The per-circuit and per-megabit ladders come from a third-party mirror, carry a January 2022 stamp against a February 2018 effective date in their own narrative, and were not confirmed as BTRC-approved or in force.
- Every division here is my arithmetic and labelled as such; the publisher prints no per-megabit IPLC rate.
- The September 2026 figures are a news report of the operator's statement, not a company document, and the statement's definitions are unpublished.
- FY2024-25 figures are as reported by BTW Media on 12 November 2026; I did not open that report first-hand.
- Nothing here asserts that a figure is wrong, that capacity is unsold, that a price is excessive or discriminatory, that a buyer was favoured, or anything about the conduct of any company, director, officer or official that a read document does not itself state.
Member Briefing
Read the full trend briefing
Members read the full briefing with the source notes behind every finding.
Strategic Circle
A daily briefing in your language, built from changes we have verified.
Join for USD 20/monthCancel anytime.
Members also get the full research library, a daily briefing on the networks they follow and the private forum.






