Bangladesh ICT & Software Services Analysis

Bangladesh exports software talent, not software. Recorded ICT service exports of $724.6 million in FY2024-25 (EPB) reached 14.5 percent of the $5 billion target the ICT Division set under the previous government for 2025, a $4.28 billion shortfall, or 85.5 percent below the target. The miss is structural, not a forecasting error: the country has the second-largest online freelance workforce in the world but few firms able to win enterprise contracts, and its export mix is skewed toward low unit-value gig work. The fix is to convert individual gig income into institutional delivery capacity and to lift the per-worker value of what the sector sells.

The export figure is real, the gap is structural

Three numbers describe the same flow and reconcile. EPB recorded $724.6 million in ICT service exports for FY2024-25, up 7.7 percent year on year. World Bank balance-of-payments data report $715 million for calendar year 2024 (indicator BX.GSR.CCIS.CD, $715,028,938); the small difference is fiscal versus calendar year coverage. Industry stakeholders estimate the true figure is closer to $1.0 billion once export proceeds that clear through informal channels are added (Prothom Alo, 2025). Even that generous estimate reaches only one fifth of the $5 billion target.

The composition explains why the target was unreachable. Computer services (software, outsourcing, IT-enabled services) are 87 percent of the recorded total, telecom services 13 percent, and information services under 1 percent (EPB FY25 data via The Business Standard). Within computer services, the export mix skews toward low unit-value work: data entry, basic web development, graphic design, and content writing generate volume but compress revenue per worker. The shift toward enterprise software, cloud managed services, and AI-enabled analytics is the path to higher export intensity, but it requires delivery capacity the current firm base does not have. BASIS reports more than 1,100 member companies, most small and without the bench depth for large engagements.

Scale without depth in the freelance economy

Bangladesh is the second-largest source of online freelance labour after India, a position the Oxford Internet Institute established through its Online Labour Index (via BIGD and Dhaka Tribune). The country has about 650,000 registered freelancers and roughly 500,000 working regularly (USAID/OII), and their combined earnings are real income for hundreds of thousands of households.

The structure is fragile. Individual freelancers do not incorporate, accumulate intellectual property, or build the delivery track record enterprises require before awarding contracts. India crossed this divide through TCS, Infosys, and Wipro: fragmented individual work became institutional service delivery at scale. Bangladesh has not. The threat is now acute because AI code generation and content tools target precisely the low-skill segment where most Bangladeshi freelancers compete. If that income base contracts before an institutional base expands, the human capital pipeline weakens at both ends.

Connectivity is necessary but not sufficient

Headline reach looks adequate: 4G covers 97 percent of the population (BTRC, 2024). The operative gap is fixed broadband, which stood at 7.9 subscriptions per 100 people in 2023 (World Bank, IT.NET.BBND.P2), up from 4.9 in 2018. That is well above Pakistan but far below the densities that sustain bandwidth-intensive cloud development, remote enterprise collaboration, and video-based delivery. Mobile data cannot substitute for fixed broadband in professional environments that need sustained throughput and low latency.

The institutional layer compounds the problem. BHTPA has 39 Hi-Tech Parks planned or under construction, but activation has lagged: procurement delays, power reliability failures, and connectivity gaps make parks outside Dhaka unattractive. Talent agglomerates where connectivity is reliable, which concentrates IT employment in Dhaka and leaves the tier-2 and tier-3 talent pool stranded. The startup ecosystem reinforces the ceiling, held back by uncertainty over foreign equity, weak IP protection, and the absence of a functioning venture capital market.

Recommendations

1. BTRC and the ICT Division: set a binding fixed-broadband density target. Lift fixed broadband from 7.9 to 15 subscriptions per 100 people by 2028, enforced through last-mile build-out obligations on licensed ISPs with penalties for non-delivery in tier-2 cities. Success signal: the World Bank IT.NET.BBND.P2 series clears 12 per 100 by 2027. Without adequate broadband outside Dhaka, STEM investment and Hi-Tech Parks produce workers who emigrate.
2. BHTPA: activate parks with guaranteed services before opening. Sequence the 39 planned parks behind non-negotiable power and fiber service-level agreements in place from day one, and publish occupancy and uptime per park quarterly. Success signal: average tenant occupancy across activated parks exceeds 70 percent within two years. Fewer parks that work beat more parks that underperform.
3. UGC and public universities: tie CS curricula to industry sign-off. Require every public university computer science program to run an industry advisory board with binding curriculum approval. Success signal: a measurable rise within three years in graduates who pass a standardized technical screen used by BASIS member firms.
4. Finance Ministry and BIDA: stand up a blended-capital digital economy fund. Anchor a $150-200 million fund co-invested with multilateral development banks for seed through Series A capital, paired with regulatory sandboxes for fintech and healthtech. Success signal: at least 20 ICT firms reach Series A within four years, so domestic digital demand converts into firm formation, not just transaction volume.
5. BASIS and a2i: build a freelancer-to-firm conversion track. Channel the top tier of the 650,000-strong freelance base into registered service companies through tax incentives on incorporation, shared back-office and compliance support, and preferential access to the digital economy fund. Success signal: at least 5,000 freelancer-founded firms registered within three years, so individual reputation accumulates into institutional delivery capacity.

What would change this view

If EPB and industry reconcile the recorded and informal export figures upward and the true total approaches $1.0 billion on a sustained, formally banked basis, the gap to a revised target narrows materially. If fixed broadband density rises faster than the 2028 target implies, the infrastructure constraint eases and the value-mix shift becomes feasible sooner. Conversely, if AI tools erode low-skill freelancing income before institutional firms scale, total ICT earnings could fall even as headcount holds, and regional competitors such as Pakistan, whose IT exports reached a record $3.8 billion in FY25, would capture the demand Bangladesh is positioned to serve.

Data sources: EPB via The Business Standard, World Bank (BX.GSR.CCIS.CD, IT.NET.BBND.P2), Oxford Internet Institute via BIGD and Dhaka Tribune, Prothom Alo, BASIS, BHTPA, BTRC.