Sri Lanka's Dairy Sector in 2026: Domestic Production, Import Dependence and the Limits of Self-Sufficiency

Sri Lanka produced a provisional 524.044 million litres of cow and buffalo milk in 2025. That was the highest total in the Department of Census and Statistics series beginning in 2015, but it was only 0.45% above 2024. In the same year, customs-based records show imports of milk and milk products valued at Rs 107.529 billion.

Those two observations describe the sector more reliably than a long-range market forecast. Domestic output has recovered from the 2022–2023 decline, yet the latest annual increase was small and a large imported product flow remains. The practical question is therefore not whether Sri Lanka has a predetermined path to dairy self-sufficiency. It is which production, collection, processing and measurement constraints must change before dependence can fall without compromising affordability, quality or farmer livelihoods.

Data cut-off: 14 August 2026. The 2025 production, livestock and farm-size figures used here are provisional.

Scope and measurement

This report distinguishes four kinds of evidence:

1. Observed official data from the Department of Census and Statistics and the Department of Animal Production and Health.
2. Institutional targets and announced plans from the Ministry of Agriculture, Livestock, Land and Irrigation.
3. Peer-reviewed historical evidence on production constraints.
4. Author calculations and interpretation, identified as such and derived from the published aggregates.

The distinction prevents a common measurement error. Domestic production is reported in litres of raw cow and buffalo milk. Trade data are reported in kilograms across products such as milk powder, liquid milk, condensed milk, cheese, butter and whey. These products have different water and milk-solids content. Their weights cannot be subtracted from, or divided into, raw-milk litres to create a defensible self-sufficiency ratio without conversion factors and a complete supply-and-use account.

The Ministry stated in March 2026 that domestic production was around 40% of national need, leaving a 60% gap. That is relevant as the government’s operating baseline, but the public tables reviewed here do not expose the calculation well enough to reproduce it independently. This report therefore cites it as a policy statement, not as a verified result calculated from the trade tables.

Domestic production: recovery followed by near-stagnation

The official annual series shows a decline through 2023, a stronger rebound in 2024 and a much smaller rise in 2025.

Year	Cow milk	Buffalo milk	Total	Annual change
2021	425.370m L	87.936m L	513.305m L	—
2022	419.197m L	87.253m L	506.450m L	−1.34%
2023	411.099m L	93.055m L	504.153m L	−0.45%
2024	430.002m L	91.705m L	521.707m L	+3.48%
2025*	431.527m L	92.517m L	524.044m L	+0.45%

* Provisional.

Cow milk contributed 82.35% of provisional 2025 production, with buffalo milk contributing 17.65%. The series establishes output volume; it does not show how much was formally collected, processed, consumed on farm, sold informally or lost before processing.

The small 2025 increase also changes how policy ambition should be interpreted. Raising domestic supply materially above its current share requires sustained gains over several years, not one announced facility or one favourable season. A credible plan needs annual targets for saleable milk, quality, collection, productivity and farmer retention, with observed results published against each target.

Herd composition and the productivity signal

The national livestock table separates local and improved animals. In 2025 it recorded 248,670 local cows and 101,570 improved cows as milking. The same table reported average monthly cow-milk production of 13.788 million litres from local cows and 22.172 million litres from improved cows.

On those aggregates, improved cows represented about 29.0% of cows recorded as milking but accounted for 61.66% of monthly cow-milk output. This is a strong productivity signal, but it is not a controlled breed comparison. The table does not standardise for feed, climate, lactation stage, herd management, farm scale, animal health or recording quality. It supports closer investigation of productivity differences; it does not prove that genetic upgrading alone would reproduce the same result across farms.

Buffalo production has a different structure. The table recorded 97,130 local and 16,390 improved buffaloes as milking, with average monthly output of 5.772 million and 1.938 million litres respectively. Buffalo milk remains a material part of national output and should not disappear inside a cow-focused policy narrative, particularly where product markets and agroecological conditions favour it.

Small farms dominate the producer base

The 2025 farm-size table records the following distribution:

Recorded farm size	Cow-milk farmers	Share of cow records	Buffalo-milk farmers	Share of buffalo records
Fewer than 10 animals	104,815	85.55%	11,390	69.92%
10–50 animals	15,363	12.54%	4,158	25.53%
More than 50 animals	2,337	1.91%	742	4.56%

The cow and buffalo columns should not be added as a unique national farmer count because a household may appear in both. Even with that limitation, the distribution is clear: dairy policy is largely small-farm policy. Collection density, extension, feed access, veterinary coverage, working capital and price transmission will affect national output because they affect many dispersed producers, not merely a few industrial farms.

The evidence also argues against treating farm consolidation as the only route to growth. Larger units can support specialised management and processing contracts, but a programme that overlooks the existing smallholder base risks losing supply, local livelihoods and regional coverage. The test should be whether an intervention raises saleable, safe milk per animal and improves farmer returns after input costs.

Trade exposure in 2025

Department of Animal Production and Health tables, sourced to Sri Lanka Customs, record the following annual totals:

Dairy trade, 2025	Quantity	Value
Imports	85.360m kg	Rs 107.529bn
Exports	1.432m kg	Rs 1.560bn

The largest import line was “milk cream fat greater than 1.5%”, at 69.817 million kilograms and Rs 91.342 billion. Cheese and curd imports were 2.859 million kilograms valued at Rs 4.867 billion; whey and whey powder imports were 2.023 million kilograms valued at Rs 874.9 million.

These categories show substantial foreign-exchange exposure, but they do not reveal one homogeneous market. Milk powder, liquid milk, butter, cheese and whey answer different industrial and household needs. Replacing one imported line may require raw milk, specific milk-solids composition, processing technology, packaging, cold-chain capacity and a price acceptable to consumers. “Import substitution” should therefore be evaluated product by product rather than treated as a single volume target.

Recorded dairy exports were small relative to imports: imports were about 69 times exports by nominal value. That comparison describes the customs categories, not the profitability or competitiveness of any company. It also does not show re-exports or the domestic value added in imported ingredients used for processed products.

Why a scored five-forces model is not retained

The earlier version assigned numerical scores to competitive forces without a current, auditable dataset for processor shares, channel margins, entry costs, customer switching or supplier concentration. A number such as “7/10” suggests measurement that the underlying evidence cannot support.

A qualitative structure remains useful when it is tied to observed conditions. The following sections retain the legacy anchors for continuity while replacing unsupported scores with evidence boundaries.

Entry and processing capacity

Food manufacturing requires capital, reliable utilities, quality systems, regulatory compliance, distribution and access to raw material. Those conditions create genuine barriers, but the reviewed official data do not support a standard entry-cost estimate or payback period.

The clearest current capacity announcement concerns the state-owned Milco system. In January 2026, the Ministry reported that the long-delayed Badalgama factory had received a Rs 3 billion allocation for rehabilitation. Milco’s chair stated an intention to complete work in the first half of 2027 and reach capacity to process 200,000 litres per day. These are announced plans. They should enter the sector baseline only when commissioning, utilisation, milk procurement and output are reported.

A plant can also redistribute existing milk rather than increase national supply. Capacity policy should therefore connect factory investment to farmer contracts, collection routes, quality testing and incremental saleable production.

Farm supply and bargaining conditions

The producer base is fragmented, yet processors depend on a time-sensitive raw material. That creates a more complex relationship than a simple “weak farmer, strong buyer” score.

Farmers need dependable collection and payment; processors need consistent volume and quality. Where chilling, testing or alternative buyers are scarce, farmers may have limited practical choice. Where processors face seasonal shortages, productive farmers and organised societies can gain leverage. Public evaluation should examine farmgate price formulas, rejected milk, payment delays, input-credit arrangements and the distance to competing collection points.

The peer-reviewed 2020 dairy-sector review identified forage quantity and quality, reproduction services, milk quality and food safety among persistent constraints. Its findings provide historical context, not a substitute for current national monitoring. The continued policy focus on pasture, animal health and production support suggests that these constraints have not been fully resolved.

Processors, retailers and consumers

The public evidence reviewed for this update does not provide current, comparable retailer margins, payment terms or processor market shares. Those claims are therefore omitted.

Consumer power is also uneven. Households can switch among powder, liquid milk and other foods, but choice is constrained by income, availability, refrigeration and product format. A cheaper domestic product may widen access; a more expensive import-substitution policy may reduce it. Affordability should be measured alongside domestic production rather than assumed to follow from it.

For processors, the relevant buyer evidence would include realised prices by product, sales volume, returns, wastage and channel costs. Without those data, precise claims about bargaining strength remain interpretation.

Imports and product substitution

Imported dairy products are not one substitute for local raw milk. Some compete directly with locally processed products; others supply ingredients or formats for which local capacity is limited. Product-level analysis is therefore necessary.

A substitution programme should specify the import line, required raw-milk equivalent, processing yield, quality standard, capital requirement and expected consumer price. It should also test whether domestic production would displace imports or simply meet additional demand. The current customs table is a useful starting point, but it is not a conversion model.

Competition and the evidence gap

Sri Lanka has state, cooperative and private dairy participants, but a defensible concentration analysis requires company-level sales or procurement data on a consistent basis. The reviewed public records do not provide that series. Company descriptions of farmer networks or plant capacity can explain their operating model, but they should not be converted into market shares without an independent denominator.

This evidence gap matters for policy. Competition affects farmgate prices, collection access, product innovation and consumer prices. A regular public dataset on formal milk procurement and processed output by product would permit better analysis without requiring disclosure of commercially sensitive transaction details.

Regulation: distinguish food law from voluntary certification

Dairy products operate within Sri Lanka’s food-law framework. The Ministry of Health’s current register includes the Food (Labelling and Advertising) Regulations 2022 and subsequent amendments, alongside other product and public-health regulations. Producers and importers must identify the rules applicable to the specific product and activity.

The SLS Mark is a separate product-certification mechanism. The Sri Lanka Standards Institution describes its certification schemes as essentially voluntary, while noting that relevant authorities have made the mark compulsory for some products. It is therefore inaccurate to state that every dairy product automatically requires SLS certification. The correct question is which standard, regulation, import-control condition or procurement requirement applies to each product.

Regulatory compliance should be evaluated through food safety, composition, labelling, traceability and enforcement outcomes—not by treating possession of one mark as a complete quality system.

Policy ambition and implementation risk

The government’s stated objective is to narrow a large domestic supply gap through public and private action. The 2025 macroeconomic recovery provides a more stable setting for investment: the Central Bank estimated real GDP growth of 5.0%. Yet sector delivery remains exposed to feed costs, weather, animal health, farmer exits, utility reliability, public procurement and consumer affordability.

The Badalgama plan illustrates both opportunity and execution risk. The facility could add substantial processing capacity if it is completed and supplied. It also originated years earlier and was still described as abandoned in January 2026. Future reporting should therefore separate:

• money allocated from money spent;
• construction progress from commissioning;
• rated capacity from actual utilisation;
• milk procured from milk processed;
• and processed volume from incremental national production.

The same discipline should apply to self-sufficiency targets. A percentage target should identify the demand denominator, milk-equivalent conversion method, product scope and reporting frequency. Without those elements, different institutions can appear to discuss the same target while measuring different things.

What a stronger dairy evidence system should publish

A practical monitoring framework would add five linked series to the existing production totals:

1. Unique active producers and animals, with farm size, district and production system.
2. Formal and informal milk disposition, including on-farm consumption, collection, rejection and loss.
3. Quality and animal-health indicators, such as bacterial counts, somatic-cell measures where applicable, adulteration findings, disease incidence and withdrawal periods.
4. Economics, including farmgate prices, feed and veterinary costs, processor procurement prices and consumer prices by product.
5. Milk-equivalent supply and use, using transparent conversion factors for production, imports, exports and inventories.

These measures would allow policy to distinguish growth in gross production from growth in safe, marketable milk and from improved farmer welfare.

Assessment

Sri Lanka’s dairy sector entered 2026 with record provisional domestic output, but the latest annual growth rate was less than half of one percent. The producer base remained dominated by small farms, improved cows contributed a disproportionately large share of recorded cow-milk output, and customs data showed a dairy import bill above Rs 107 billion.

The evidence supports sustained work on productivity, feed, animal health, collection, quality and processing. It does not support the earlier US-dollar market forecast, numerical competition scorecard, company-specific market-entry recommendation or investment-return model.

Self-sufficiency is best treated as a measurable policy programme rather than an assumed destination. Progress should be judged through a transparent milk-equivalent account, higher saleable output per animal, resilient farmer returns, reliable quality and affordable products. Until those measures are published together, claims of transformation should remain conditional on observed results.