Sri Lanka Infrastructure Market Audit: Funding, Delivery and Operation

Funding has moved further than some projects have

Sri Lanka’s infrastructure opportunity is best assessed at the point where finance becomes usable service. The inspected record contains three different states: a financed transmission link, a transport terminal still under construction, and a port terminal handling vessels while its wider works programme continued. Treating these as equivalent additions to a national market would conceal the delivery question.

The fiscal evidence is also mixed. Capital expenditure and net lending increased in January–August 2026, but the recorded amount remained about a quarter of the annual estimate. That does not establish a full-year shortfall, and it does not measure physical infrastructure completed. The useful commercial question is which named works have financing, procurement authority, executable contracts and a route to operation.

Evidence checked through 20 October 2026. Limited coverage: this audit uses a purposive three-project panel and separately labelled national fiscal aggregates. Project observation dates run from October 2025 to May 2026; access in October does not make those physical statuses current to October. No national infrastructure revenue total, contractor ranking, project margin or representative user outcome is estimated.

The boundary is civil and utility infrastructure

The audit covers construction and major rehabilitation of Sri Lankan transport and utility infrastructure, with urban interchange, electricity transmission and port works as the observed segments. Preparation, finance and operation are examined as delivery context; ordinary operations, pure finance and design-only activity are not added to the construction market. Roads, water and irrigation also fit the broader boundary but are not represented in the admitted panel. Residential and nonresidential building construction, building-related specialised work and real-estate holding or sales belong to separate assessments. A transport terminal may contain buildings; here its integrated transport function determines the project classification. Its incidental building works are not counted again as an independent market total. ISIC Rev.4 classes 4210, 4220 and 4290 provide partial activity anchors for transport works, utility projects and other civil engineering. They do not allocate the fiscal aggregates or every mixed project package to a single statistical class. [12]

Three measures must remain separate. Construction GVA combines building and civil-engineering value added. Public investment covers a wider fiscal population than this project panel. Contractor consolidated revenue, receivables and debt may span several businesses. None allocates an infrastructure-only total without a compatible activity or project bridge. The same restriction applies to the all-construction PMI: a direction survey cannot allocate project expenditure or completed assets.

This boundary changes how demand is read. A planned asset describes an intended service. A signed loan identifies financing. An awarded contract identifies a procurement commitment. Certified work, payments, completed construction and actual service output answer different questions. The audit records each independently; an unknown stage is neither a zero nor evidence of failure.

Fiscal recovery has not removed the conversion problem

The Treasury’s Fiscal Strategy Statement 2027 gives the following public-investment comparison. Its 2025 outturn is provisional; the 2026 figure is an annual estimate against an early-year observation. [1]

The 2025 gap was LKR299 billion, or 22.7% of budget. The Treasury identifies procurement delays and the compressed implementation period associated with the Vote on Account among the explanations for capital underspending. These are aggregate explanations, not findings that every project suffered the same delay. [1]

The later January–August report uses a different measure: capital expenditure and net lending. It records LKR435.9 billion in 2026 against LKR331.2 billion a year earlier. Recalculation gives 31.6% growth and 25.4% of the LKR1,719 billion annual estimate, matching the rounded table. The narrative rounds growth to 32%; the table rounds the amounts to 436 and 331. [2]

The later figure cannot simply replace LKR161 billion in a public-investment series. Scope and period both changed. Neither execution percentage can be linearly extrapolated into a year-end outcome: the evidence here does not establish the seasonality of expenditure or the timing of large certificates. The figures support closer monitoring of conversion from allocation to spending; they do not prove that three-quarters of infrastructure projects are unbuilt.

Three projects show why the stages matter

The panel deliberately includes different asset types and delivery stages. Kandy and Kerawalapitiya were financed examples in the earlier article. East Container Terminal provides counterevidence to an undifferentiated account of stalled delivery. Selection was based on inspectable official evidence and analytical contrast; the panel is not representative of national investment or the probability of completion.

Kandy exposes an interface risk. The World Bank reports delays to the design-and-build contract, while urban integration activities were held for insufficient funds. The indicator record also lacked a robust operating-and-maintenance financial plan and institutional setup. The revised credit closing date is 30 May 2027; it is not an opening certificate. Financing the principal civil contract therefore did not establish that all complementary elements were funded and ready. [3]

A financial discrepancy remains visible. The same report’s narrative gives USD28.89 million disbursed; the two loan-table rows show USD24 million and USD5 million, totalling USD29 million. The USD0.11 million residual is not explained by the inspected report. Neither amount measures certified contractor work, cash received by suppliers or the value of completed service. [3]

The transmission project is a financing-stage observation. AIIB describes a 220kV underground cable system with design, supply, construction, testing and commissioning in its scope. Its announcement establishes a signed agreement, while the operational benefits are prospective. That creates a defined project to monitor, not evidence that the new connection was supplying electricity at the audit cut-off. The original line’s earlier commissioning must not be attributed to the second line. [4]

The port evidence shows delivered activity, with a narrower completion claim. SLPA reports commissioning the third berth and an actual vessel call. It also reports more than 500,000 TEU handled during 2025, an operator statement rather than an independently audited utilisation measure. The January article still describes further jetty completion and systems integration as planned. A commissioned berth and a completed entire terminal are therefore different claims. [5–6]

Scope discrepancies also matter: the November 2025 article describes 900 metres completed plus a 300-metre extension, whereas the January account describes an approximately 1,300-metre quay. Those descriptions are retained without deriving a construction-completion percentage from unlike lengths. [5,7]

Buyers, suppliers and users occupy different parts of the chain

The project owner specifies and procures works; a financier may supply credit; the contractor delivers a defined package; the operator maintains the asset and serves users. These roles do not collapse into one customer. Kandy’s relevant service users are passengers and bus operators. The transmission asset supports network service to connected loads. At the port, shipping-line terminal demand differs from the public owner’s demand for construction. The observed project scope supports this role map, but it does not quantify willingness to pay, user satisfaction or social benefit. [3–6]

This distinction gives a practical explanation of where value can be lost. An owner can acquire a physical asset without completing connections, operating arrangements or maintenance funding. A supplier can have a contract without having received cash. A lender can report disbursement before a user receives the intended service. Conversely, phased operations can deliver a real service before every construction package is complete. The panel supports these differences; it does not show which procurement model has the highest return.

Competition is consequently contract-specific. CIDA’s registration scheme screens financial, technical and experience capability across different construction fields, and explicitly does not replace major-contract prequalification. Its financial-limit table is labelled effective from 1 April 2025. A registration grade alone cannot demonstrate that a firm can execute an underground cable, transport interchange and maritime package equally well. [8]

No reconciled bidder, award or supplier-share dataset was established for a common project population. The audit therefore does not rank contractors, calculate concentration or attach numerical Five Forces scores. Maintenance, rehabilitation and new construction may compete for an owner’s budget, but their relative lifecycle value requires project appraisal. A PPP changes financing and risk allocation; it is not a physical substitute for the asset or proof of lower public cost.

Authority to spend and readiness to deliver remain separate

The original Public Financial Management Act No.44 of 2024 distinguishes authorisation, approval, commitment, certification and payment. Its public-investment provisions establish appraisal, programme and budget gates, monitoring, reporting and a project repository; PPP provisions address fiscal affordability and contingent liabilities. Those are rules in the inspected instrument, not evidence that each panel project complied or that every implementation system is complete. [9]

The National Procurement Commission’s current manual index identifies the 2024 manual issued on 1 January 2025 and amended on 18 July 2025. It also displays a revised-manual consultation with a 30 September 2026 comment deadline. That deadline does not establish that the consultation version became operative. The earlier generic guidelines URL now returns an error, while the manual index has a new location. Tender decisions need the applicable current instrument and financing conditions, not an assumption drawn from an old link. [10]

The relevant stakeholder map includes the Treasury and responsible public owner for spending authority, the financier for financing conditions, the procurement authority and qualified suppliers for award, and the operating entity for service delivery. Users, workers and affected communities bear consequences that are not measured by financial close. This audit has not completed project-by-project land, environmental, labour or safeguard compliance assessments and draws no compliance finding from the panel.

Financial pressure is context, not a project-debt estimate

The supplied CBSL Financial Stability Review 2026 reports 1.3 times interest coverage for its listed Construction & Engineering category in the surrounding H1 2026 discussion. That source category is retained without treating it as an infrastructure-only population. Sector membership counts, weights, detailed aggregation and a company-to-project mapping are not disclosed; the supplied edition’s publisher byte match and publication day remain unresolved. [11]

The ratio cannot be assigned to infrastructure projects, unlisted contractors or every firm. Nor does it establish default probability, project bankability or the amount of debt at risk. Contractor group receivables, operating cash flow and borrowing belong to the separate company-finance evidence in the building-construction assessment, with its mixed-business limitations. Reusing them as project finance would count a different economic object under the same name.

For infrastructure, the admitted evidence supports a narrower concern: finance for one stage may coexist with unresolved funding for another. Kandy makes that concern observable. It does not establish that the FSR ratio caused the funding gap, or that all public infrastructure faces the same constraint.

The outlook depends on funded interfaces and demonstrated service

The evidence supports a conditional outlook. If funding, procurement and systems integration convert into accepted assets with an identified operator, the opportunity moves from construction toward service and maintenance. If cost increases absorb complementary budgets, civil progress can coexist with delayed use. New project-status releases, actual acceptance records and service measures can distinguish these paths. This is an interpretation of the panel, not a revenue forecast or a probability model for national completion.

Three requirements follow for participants. First, screen opportunities using named financing and procurement records, because a pipeline entry alone cannot establish addressable work. Second, price and assign interfaces, testing and operating responsibilities: an apparently peripheral package can determine whether the asset is usable. Third, reconcile certified work and payment evidence rather than interpreting loan disbursement as contractor cash conversion. These requirements are supported by the legal spending stages, the project delivery evidence and the different customer roles. They do not imply a universal margin, cash-buffer target or preferred contractor.

The prior article’s unsupported dollar market forecast, segment shares, national government-demand percentage and company-specific financial targets remain excluded. No recommendation here asserts Access Engineering’s capabilities, investment priorities or market share. A company strategy would need its own eligible assessment and internal evidence where required.

Methods, limitations and evidence updates

This is a methodology 1.1.0 external audit with limited empirical coverage. Every E0–E11 module was assessed; missing national size, competition, customer outcomes and technology adoption evidence constrains the conclusions. Public fiscal data, lender implementation records, regulator pages and operator announcements have different populations and reporting incentives. Project observations are not an October census, and an operator announcement is not independent validation of performance.

Calculations use the displayed Treasury narrative or table inputs: 1,016/1,315; 161/1,730; (435.9/331.2−1); and 435.9/1,719. The Kandy residual is (24+5)−28.89. Percentages are rounded only after calculation. No currencies or unlike fiscal series are combined. The retained research package records source locators, date uncertainty, raw observations, the full stage panel, failed access and excluded conclusions.

The next material evidence is the next Kandy implementation report, transmission procurement and progress records, ECT whole-project acceptance and audited service data, and the next like-for-like Treasury outturn. Revised procurement instruments and clarified FSR classifications can change the relevant bounded findings. These are source obligations; they do not justify treating unavailable evidence as already observed.

Sources

1. Treasury, Fiscal Strategy Statement 2027, Tables 2–3, printed pp.8–10. Main title page: June 2026; exact publication day unconfirmed.

2. Treasury, Fiscal Review Report January–August 2026, p.8. Publication day unstated.

3. World Bank, Kandy project ISR07603, 10 May 2026, pp.2,7,9,11–12.

4. AIIB, transmission loan signing, 8 October 2025.

5. SLPA, ECT third berth commissioning, 23 January 2026.

6. SLPA, Ever Eon vessel call, published 23 January; event 22 January 2026.

7. SLPA, second-berth pre-test operations, November 2025.

8. CIDA, registration and grading scheme, accessed 20 October 2026; financial limits labelled effective 1 April 2025.

9. Public Financial Management Act No.44 of 2024, original instrument, ss.28–29 and 40–45.

10. National Procurement Commission, manuals on goods and services, accessed 20 October 2026.

11. CBSL Financial Stability Review, supplied 2026 edition, Figure 4.25, printed pp.81–82; sector membership and aggregation are undisclosed; publisher byte match and publication day remain unresolved.

12. United Nations Statistics Division, ISIC Rev.4 4210, 4220 and 4290, inspected 20 October 2026.