55 Startups, venture capital and innovation

Chapter 55 of 60 in the Bangladesh 2036 research base. Contents of the series.

A founder base and a sandbox state, and the capital industry that does not yet exist

Chapter 13 closed its record with the one hole it could not open on the machine: startup finance, where it found the state venture vehicle in the annual report record and no series behind it, and handed the question to this chapter. This chapter opens that space with the primary records that do exist, and the thesis it defends about the FY27 to FY36 window, the horizon the chapter 15 scenarios price, is this: Bangladesh has built the demand side of an innovation economy and the sandbox paperwork of one, and has not built the capital industry or the measurement system that would make either compound. The measured base at drafting is a registered alternative investment sector worth 1,495 crore BDT across 12 funds a decade after the enabling rules, of which the venture capital class holds 480 crore BDT, roughly 39 million USD or about 0.13 percent of FY25 remittances by this chapter's arithmetic [BSEC 2024] [BB Econ 2025, table XVIII] [BIS 2026]; a central bank refinancing line of 500 crore BDT whose utilisation is unpublished [BB AR 2024]; a state venture company registered on 16 March 2020 that disburses grants averaging 9.2 lakh BDT per startup by this chapter's arithmetic [ICT Division 2020]; and a science ministry whose FY26 revised development allocation puts 11,374.45 of its 12,029.50 crore BDT, 94.6 percent, into the nuclear power programme chapter 10 tracks (chapter arithmetic on the programme table) [Planning Commission ADP 2026]. None of this is hidden; every element sits in a published government record no compilation connects. The decade decides three things: whether the fund register compounds into an industry, whether the state ever measures its own research input, and whether an exits venue opens. If they do not move, the 2,500 plus startups the government counts [ICT Division 2020] remain a consumer distribution scene financed from abroad on the rails chapter 13 documents, and the innovation system the 2036 endpoint the chapter 15 scenarios price assumes does not get built.

The record: consumer internet scale, a venture register the size of one regional fund, no measured research input

The innovation indices set the frame, and they have not moved in a decade of startup policy. The Global Innovation Index score fell from 26.7 in 2013, rank 130 of 142, to a trough of 19.7 in 2021, then recovered to 21.4 in 2022, 21.0 in 2023 and 21.4 in 2024, rank 106 of 133 [WIPO GII 2024]. The rank is similar to a decade ago and the score is lower, which is the arithmetic of a country that climbed mostly because the index expanded. The appropriable output behind the score is thin: resident patent applications were 74 in 2021 against 68 in 2019 and 40 in 2020, while nonresidents filed 373 in 2021 [WB WDI 2026], high technology exports were 0.54 percent of manufactured exports in 2018, the last year the series carries [WB WDI 2026], and receipts for the use of intellectual property were 1.3 million USD in 2024 [WB WDI 2026]. The one bright line is publication volume, chapter 33's finding, which this chapter inherits: scientific and technical journal articles rose from 1,426.96 in 2010 to a peak of 7,551.78 in 2021 and 7,056.17 in 2022 [WB WDI 2026]. The input side has no line at all: no series available to this chapter carries a national research and development expenditure figure, and any number for that share could not be confirmed; the Ministry of Science and Technology and the UNESCO institute the chapter 33 note names would resolve it.

The startup stock itself is a government claim, not a measured series. The ICT Division's annual report for FY20 put the count at more than 2,500 startups, said they had brought in more than 300 million USD of direct foreign investment and about 1.5 million jobs, and stated elsewhere in the same report that young entrepreneurs had brought 4,000 crore BDT of foreign investment in three years [ICT Division 2020]. The two investment claims cannot both describe the same stock at the same exchange rate, the count has never been re-measured in any series available to this chapter, and the compilation figures quoted in trade press since could not be confirmed; the ICT Division would resolve them. What is measured is the capital that the state's own institutions hold against the sector, and it comes from three regulators.

The securities register is the core record. The commission formulated its alternative investment rules in 2015 and records the goal as launching alternative investment for supplying capital to start-up and new companies, implemented June 2015 [BSEC 2021]. The first private equity fund registered on 22 November 2015; the venture capital class waited until 14 August 2019, when BDVL Venture Fund One registered, followed by a second VC fund in December 2019, a third in January 2020 and a fourth in August 2021 [BSEC 2023]. By the FY23 annual report the register held 7 funds with 880 crore BDT of corpus, 650 crore BDT of it private equity and 230 crore BDT venture capital [BSEC 2023]. By the FY24 report the register held 12 funds and 1,495 crore BDT, of which 6 private equity funds hold 1,015 crore BDT and 6 venture capital funds hold 480 crore BDT, sums this chapter takes from the commission's own fund table [BSEC 2024]. The largest VC commitments are UCB Venture Fund at 150 crore BDT, UFSEPL Venture Capital Fund at 125 crore BDT and the Angel fund at 100 crore BDT; the smallest, Maslin VC Fund, is 10 crore BDT [BSEC 2024]. Convert the venture class and the scale reads directly: 480 crore BDT is roughly 39 million USD at the end May 2026 rate of 122.75 taka per USD [BIS 2026], about 0.13 percent of FY25 remittances of 30,328.80 million USD, chapter arithmetic [BB Econ 2025, table XVIII]. The whole register, both classes, is less than 0.1 percent of one year of mobile money transaction value, 1,737,411 crore BDT in 2024 [IMF FAS 2024].

The bank channel and the state channel complete the measured stack. Bangladesh Bank reports that it created a 5 billion taka, 500 crore BDT, refinancing revolving fund from its own source for a five year tenure, and required commercial banks to set up their own startup funds at 1 percent of net profit counted from 2020 [BB AR 2024]. No published utilisation figure for either exists in the sources available to this chapter, and any figure for funds disbursed under the line could not be confirmed; the central bank's refinance statistics would resolve it. The state's own vehicle, Startup Bangladesh Limited, completed company registration on 16 March 2020 [ICT Division 2020]; the same report records the prime minister's 100 crore BDT startup allocation, 40 startups already supported under it, and a plan to bring 1,000 new startups under government support by 2021 [ICT Division 2020]. The grant machinery beneath it, the Innovation Design and Entrepreneurship Academy, held 12 selection meetings, selected 135 startups and recommended 12.42 crore BDT of grants, paying a first instalment of 5.27 crore BDT to 100 startups and a second of 1.5550 crore BDT to 53 startups, an average recommended grant of 9.2 lakh BDT by this chapter's arithmetic [ICT Division 2020]. The academy trained 1,159 people, and the a2i innovation fund line in the same report records more than 5,350 online proposals, 185 awardees, 3.1 million USD awarded and 247 ideas funded [ICT Division 2020]. The hi-tech park programme chapter 13 cites, 39 parks under construction with 5 at commercial operation and 120 companies at about 50 million USD of investment as of FY20 [ICT Division 2020], continues in the FY26 revised programme: the district IT and hi-tech park project for 12 districts carries 1,846.09 crore BDT of committed cost, a 179.00 crore BDT FY26 allocation and 719.45 crore BDT of cumulative expenditure, and the IT training and incubation centre lines together carry FY26 allocations of 156.07, 60.00 and 45.50 crore BDT [Planning Commission ADP 2026].

What the state prices overall is the deepest finding in the record. The FY26 revised programme allocates the ICT Division 2,286.25 crore BDT across 18 projects, of which the startup facing lines, the digital entrepreneurs and innovation ecosystem development project at 122.69 crore BDT and the innovation and entrepreneurship development academy at 30.05 crore BDT, are small fractions [Planning Commission ADP 2026] [chapter 13's division total, same table]. The Ministry of Science and Technology holds 14 projects and 12,029.50 crore BDT, of which the Rooppur nuclear lines take 11,374.45 crore BDT and everything else, the oceanographic institute, the genome sequencing facilities, the tissue bank, the light engineering and e-waste facility and the Institute of Nanotechnology, shares 655.05 crore BDT, chapter arithmetic on the programme table [Planning Commission ADP 2026]. The Institute of Nanotechnology's FY26 allocation is 0.35 crore BDT against 380.78 crore BDT of committed cost [Planning Commission ADP 2026]. The university side of the innovation system expanded in campus count, 49 public universities in 2022 against 31 in 2010 [UGC 2023], while no measurable technology transfer or industry link record exists anywhere in the sources available to this chapter, and could not be confirmed; the commission's yearbook would resolve it, and chapter 33 carries the research financing record that sits behind the article count.

Mechanism: the state built access and sandboxes while the capital never got institutionally priced

Three mechanisms produce the shape of this record, and each is a policy choice the decade inherits rather than a market outcome.

The first is consumer internet substitution. The ventures that reached household names, the investment guide names Pathao among the fastest growing technology startups alongside Uber and Shohoz [BIDA 2023], monetise domestic consumption over the rails chapter 13 built: 238.68 million registered mobile money accounts, a payment pipe moving 34.7 percent of GDP in transaction value, and 53.42 percent of the population online in 2024 [IMF FAS 2024] [WB WDI 2026]. Foreign funds can price that demand; the domestic register cannot fund it. The 1,015 crore BDT private equity class is late stage money for established businesses, and the 480 crore BDT venture class is too small to carry even one regional scale seed portfolio. The result is visible in the output statistics: a distribution economy with 0.54 percent high technology export share and 74 resident patent filings a year [WB WDI 2026]. The 715.03 million USD of ICT service exports in 2024 [WB WDI 2026], chapter 13's plateau line, is the same story one layer up: services income without venture capital behind it.

The second is institutional sequencing. The instruments arrived in an order that could not price a pre revenue company. The alternative investment rules of 2015 created a vehicle that no venture fund used for four years [BSEC 2023]. The bank channel of 2020 routed startup finance through collateral lenders, the same institutions whose SME credit record chapter 07 traces down to 5.98 percent of GDP [IMF FAS 2024] and whose information problems chapter 22 documents; a refinancing line administered by banks that underwrite on assets cannot reach a company whose only asset is code. The state vehicle of 2020 disbursed grant scale cheques, 9.2 lakh BDT on average, subsidy arithmetic rather than portfolio arithmetic [ICT Division 2020]. Each instrument is real, none is measured, and the gap between them, equity capital priced by an investor who reads the technology, is the hole the sector sits in. The commission's own agenda items, a workshop on investments in startups by registered alternative investment funds in October 2023 [BSEC 2024] and the 2020 venture capital and start-up investment forum addressed by the Startup Bangladesh managing director [BSEC 2021], show the regulator and the state vehicle talking to an industry whose measured corpus stayed under 1,500 crore BDT.

The third is a state whose own innovation demand is capital goods, not research. A country's innovation system begins with what the state buys. The FY26 programme says the science ministry buys a nuclear power plant, 94.6 percent of its development allocation, and allocates 0.35 crore BDT to nanotechnology [Planning Commission ADP 2026]. Chapter 33's finding that no research expenditure series exists is the same fact seen from the statistics side: an input that is not measured cannot be managed, and the article count rose on project money and journal incentives without an appropriable base behind it. The sandbox layer repeats the pattern at the regulators. The revenue authority's startup provisions, summarised in the investment guide, offered a growth year regime to companies incorporated between 1 July 2017 and 30 June 2022 with turnover not exceeding 100 crore BDT, overriding the disallowance sections of the income tax ordinance and carrying losses forward 9 successive years [BIDA 2023]; the insurance regulator issued Regulatory Sandbox Guidelines in 2023 for insurtech product, distribution, claims and underwriting experiments [IDRA 2024]; the central bank runs a Regulatory Fintech Facilitation Office that assesses business models through simulations in a controlled environment, alongside 10 payment system operators and 8 payment service providers licensed and operational [BB AR 2024]; and the SME Policy 2019 action matrix carried a line to appoint a venture capital fund manager under the commission's 2015 rules for startup SME financing [SME Policy 2019]. Every sandbox is an official admission that the default regime does not fit innovation, and none of them creates capital; the tax sandbox closed to new entrants when its incorporation window ended in June 2022, and the SME Policy action line's execution status could not be confirmed; the ministry would resolve it.

The decade ahead: five decisions decide whether the register becomes an industry

No international projection series exists for this sector; the IMF, World Bank and ADB outlooks the macro chapters cite carry no startup line. The only named forecasts are the government's own target record, and it is the discipline of the decade: the 5 billion USD IT export target for 2021 that chapter 13 prices at a sevenfold miss [ICT Division 2020], and the 1,000 startups under government support by 2021 [ICT Division 2020], against the measured register above. The innovation system of 2036, the endpoint the chapter 15 scenarios price, is therefore not a forecast but five decisions, each with an author.

The corpus decision belongs to the securities regulator with the pension and insurance authorities of chapter 38. A decade of fundraising produced 480 crore BDT of venture corpus because the only permitted investors are institutions and the wealthy, and the long term domestic pools, the pension scheme chapter 38 examines and the insurers IDRA regulates, sit outside alternative investments. An allocation rule that lets even 1 percent of those portfolios into registered funds would reprice the register; the reform scenario the chapter 15 table assumes treats that conversion as done, and the stall scenario treats the register as capped. The indicator is a single line in the commission's annual report.

The measurement decision belongs to Bangladesh Bank. Publish the refinancing line's utilisation and the bank startup funds' balances, or retire the line. The same machinery appears in chapter 07's payment to credit conversion and in chapter 37's instrument agenda, and it applies verbatim at the venture layer: a channel that cannot show disbursement is a press release, and a published series is the cheapest fintech and startup policy the bank can buy.

The vehicle decision belongs to the ICT Division. Startup Bangladesh Limited should publish an audited portfolio, move from grants to equity with follow on rights, and restate the 2,500 startup count on a definitional basis. The 100 crore BDT fund split across the 100 planned startups is 1 crore BDT each, chapter arithmetic, pre seed money by any regional standard; an equity vehicle that recycles exits funds more startups than a grant shop that must return to the budget each year.

The research measurement decision belongs to the Ministry of Science and Technology with the statistics agency chapter 53 examines. A first official research and development expenditure figure, on the UNESCO methodology the chapter 33 note names, is the single cheapest institution the decade can build, and the non nuclear share of the ministry's own programme, 5.4 percent at FY26 (chapter arithmetic), is the base it would measure. Chapter 33 carries the university financing half of this decision; the ministry half is the laboratory and institute estate the FY26 table prices at 655.05 crore BDT of allocation [Planning Commission ADP 2026].

The exits decision belongs to the exchanges with the commission. The SME platform has existed since March 2016 [BSEC 2021], and the register available to this chapter records no startup listing flow against it; that could not be confirmed, the exchange and commission records the resolving sources; chapter 07 carries the market depth problem. Without exit the corpus cannot recycle, and without recycling the fundraising record stays at the 480 crore BDT scale the register shows. The link to chapter 02 is indirect but real: graduation removes no startup privilege, while the GSP+ compliance agenda and the Personal Data Protection Act 2026 that chapter 54 documents are the trust infrastructure cross border funds underwrite before the first dollar arrives. The diaspora is the largest venture grade pool the country has, 30,328.80 million USD of remittances in FY25 [BB Econ 2025, table XVIII]; chapter 60 carries the diaspora finance agenda, and 1 percent of one year's inflow routed through registered funds is about 3,723 crore BDT at the May 2026 rate, about 2.5 times the entire register, chapter arithmetic [BIS 2026].

Three risks print in registers and the upside runs on rails the state already built

Risks. First, register stagnation: if no new fund above the 150 crore BDT of the largest current venture fund registers and total corpus holds below 2,000 crore BDT through the 2020s, the venture class stays a project, and the revealing indicator is the alternative investment fund table in each commission annual report, 12 funds and 1,495 crore BDT at FY24 [BSEC 2024]. Second, state substitution: if the grant machinery keeps writing 9.2 lakh BDT cheques, no audited portfolio appears, and the last published startup count remains the FY20 claim of 2,500 plus [ICT Division 2020], the state vehicle crowds out the measurement and standards a market needs; the revealing indicators are the next ICT Division annual report and the first audited accounts of Startup Bangladesh Limited, neither of which exists in the sources available to this chapter. Third, talent exit: chapter 33's settlement stock, 619,452 Bangladesh born people in the four settlement economies by 2020, compounds while chapter 54's AI compression removes the entry level services work the sector's on ramp runs on, and the revealing indicators are journal articles against the 7,551.78 peak of 2021 and resident patent applications against the 74 of 2021 [WB WDI 2026].

Upside. First, the payment rail as distribution: a venture that monetises per transaction reaches 238.68 million registered accounts without building its own pipe [IMF FAS 2024], and the revealing indicators are the licensed operator counts, 10 payment system operators and 8 payment service providers [BB AR 2024], and chapter 13's mobile money active share threshold. Second, diaspora and private capital conversion: the register is small enough that one decision moves it, and 1 percent of the FY25 remittance flow routed through registered funds is about 2.5 times the whole alternative investment sector, chapter arithmetic [BIS 2026] [BB Econ 2025, table XVIII]; the revealing indicator is the first diaspora branded fund in the commission's table [BSEC 2024]. Third, the compliance dividend: the data statute of chapter 54 and the conventions of the chapter 02 GSP+ agenda give the jurisdiction the governance certifications foreign funds underwrite, and the revealing indicator is the Global Innovation Index score, 21.4 in 2024 [WIPO GII 2024], moving above the 26.7 of 2013 with a rank inside the top 100, the trajectory the reform scenario targets.

What to watch: five indicators whose thresholds mark the regime

1. Registered venture capital corpus. Current value 480 crore BDT across 6 funds in the FY24 register [BSEC 2024]. Threshold: corpus above 2,000 crore BDT, the line the reform scenario assumes, marks the arrival of an industry; a register still below 500 crore BDT in the next two annual reports confirms the project scale regime.
2. Startup refinance utilisation. Current value: a 500 crore BDT revolving fund created by Bangladesh Bank [BB AR 2024], utilisation unpublished and not confirmed; the bank's refinance statistics would resolve it. Threshold: a first published utilisation above 50 percent signals the bank channel functions; continued non publication through two more annual reports confirms the line is inert.
3. Resident patent applications. Current value 74 in 2021 [WB WDI 2026]. Threshold: prints above 200 a year signal research reaching appropriation, the threshold chapter 33 sets; a fall below the 40 of 2020 confirms a system that publishes but never files.
4. Global Innovation Index score. Current value 21.4, rank 106 of 133, in 2024 [WIPO GII 2024]. Threshold: a score above the 2013 reading of 26.7 with a rank inside the top 100 validates an innovation system; a score below the 19.7 trough of 2021 signals reversion under the stall scenario.
5. Measured state innovation input. Current value: 655.05 crore BDT of non nuclear research allocation in the science ministry's FY26 revised programme against 11,374.45 crore BDT for the nuclear programme [Planning Commission ADP 2026], and no national research expenditure measurement, not confirmed; the Ministry of Science and Technology and UNESCO UIS would resolve it. Threshold: a first official research expenditure figure plus a non nuclear allocation share above 25 percent of the ministry's programme marks the research regime chapter 33 defines; a share still below 5 percent at FY30, the window the scenarios assume, confirms the capital goods regime.

Sources used

[BB AR 2024] Bangladesh Bank annual report 2023-24, startup refinancing revolving fund of 5 billion taka and the 1 percent of net profit bank startup fund policy, Regulatory Fintech Facilitation Office, 10 PSO and 8 PSP licences, via ocr_text/bb/annual_report. [BSEC 2021] BSEC annual report 2020-21, alternative investment rules 2015 implementation record, SME platform launch of March 2016, venture capital and start-up investment forum, via ocr_text/bsec_deep. [BSEC 2023] BSEC annual report 2022-23, alternative investment fund register with corpus and registration dates, private equity 650 crore BDT and venture capital 230 crore BDT, via ocr_text/regulators_soe_bd. [BSEC 2024] BSEC annual report 2023-24, Alternative Investment Department fund register of 12 funds, corpus 1,495 crore BDT in taka crore, fund level detail, startup investment agenda, via ocr_text/bsec. [BIDA 2023] BIDA investment guide 2023, startup tax sandbox conditions and benefits, technology startup profiles, via ocr_text/bida_deep. [IDRA 2024] IDRA insurtech brochure 2024, Regulatory Sandbox Guidelines 2023 scope, via ocr_text/idra_deep. [ICT Division 2020] ICT Division annual report 2019-20, Startup Bangladesh Limited registration of 16 March 2020, startup count and foreign investment claims, prime minister's 100 crore BDT fund, iDEA selection and grant record, training counts, a2i innovation fund record, hi-tech park counts, via ocr_text/govtwin_min_ict. [SME Policy 2019] Ministry of Industries, SME Policy 2019 action matrix line on the venture capital fund manager under the BSEC (Alternative Investment) Rules 2015, via ocr_text/policies. [Planning Commission ADP 2026] Revised Annual Development Programme FY2025-26, ICT Division and Ministry of Science and Technology project costs, allocations and cumulative expenditure, lakh taka converted to crore, via the adp_projects table. [UGC 2023] University Grants Commission public university panel, counts 1985 to 2022, via lake/education. [WIPO GII 2024] Global Innovation Index, Bangladesh score and rank series 2013 to 2024, via the indicators/wipo_gii_bd parquet. [WB WDI 2026] World Bank World Development Indicators via bdpolicy parquets wb_full_bd and wb_sci_tech_bd: patents IP.PAT.RESD and IP.PAT.NRES, journal articles IP.JRN.ARTC.SC, high technology exports TX.VAL.TECH.MF.ZS, intellectual property receipts BX.GSR.ROYL.CD, ICT service exports BX.GSR.CCIS.CD, internet use IT.NET.USER.ZS. [IMF FAS 2024] Financial Access Survey, mobile money registered accounts and transaction value, SME loans share of GDP. [BB Econ 2025] Bangladesh Bank Monthly Economic Trends October 2025, table XVIII, FY25 remittances 30,328.80 million USD. [BIS 2026] BIS USD/BDT end of period exchange rate, 122.75 taka per USD, May 2026.

---

Verified line by line against primary sources: 46 claims checked, 0 corrected.

Previous: 54 AI adoption, data protection and cybersecurity

Next: 56 Tourism and aviation

All sixty chapters and the ten narrative chapters.