What a country buys: every tender Latvia published in the first half of 2026

What does a country actually buy? Not in the abstract, but in line items, with prices, week after week. We decided to find out for one country across one year, by reading the paperwork.

So we pulled every public tender Latvian buyers put on TED, the EU's official tenders journal, between 1 January and 23 June 2026 — every genuine call for competition, with award notices and duplicate corrections set aside. That leaves 2,154 distinct tenders: six months of a small EU state asking the market to build, supply and serve.

Two things to keep in mind about the numbers below. TED is only the visible tip: it carries procurement above the EU value thresholds, so Latvia's full market is much larger — around EUR 5.45 billion in 2024, roughly 13% of GDP, across some 11,400 procedures a year. And only 43% of these tenders attach an estimated value, so every euro total here is a floor, not the full sum.

What a country buys

Split the 2,154 tenders by what is being bought and a clear picture emerges. By number of tenders, Latvia's public sector is overwhelmingly a buyer of buildings and the services around them: construction work is 18% of all tenders, architecture and engineering services another 12%, and medical equipment and pharmaceuticals 10%. After that the long tail begins — vehicles, environmental and cleaning services, repairs, laboratory equipment, furniture, IT, food.

By money, the ranking shifts but construction still towers over everything. These are the sectors that soak up the disclosed value:

The split by contract type is almost perfectly balanced: 41% of tenders are for supplies, 41% for services, and 18% for works. But that 18% of works carries a third of the money, because a building costs more than a box of supplies. This is the basic physics of public procurement in a small country: lots of small purchases of goods and services, and a smaller number of expensive construction projects that dominate the budget.

A few giants, and a very long tail

The median tender is EUR 240,000 — a modest sum, the price of renovating a kindergarten wing or supplying a year of lab reagents. Yet the ten largest tenders alone account for 46% of all the disclosed value, and the top 25 for 62%. A handful of mega-deals sit on top of a very long tail of ordinary purchases.

The giants are worth naming, because they are a portrait of what a state invests in when it spends big. The largest single tender of the half-year was a EUR 170.8 million training-services framework from the State Employment Agency (Nodarbinātības valsts aģentūra), buying the providers of its active labour-market training. Then EUR 121.5 million for 120 city buses for Rīgas satiksme, 60 of them electric. After that: a EUR 90 million road-bitumen framework and a EUR 50 million road-surfacing framework from the State Roads company; a EUR 86 million dynamic purchasing system for power cables from the grid operator Sadales tīkls; a EUR 69.6 million Ro-Ro berth at the Freeport of Riga; EUR 68 million of railway modernisation; EUR 62 million of diesel fuel; and EUR 50 million for the onshore works of an offshore wind project. Buses, roads, the grid, the port, the railway, the wind farm — the big cheques go to moving people and power.

This concentration is not just a TED artefact; it runs through the whole national system. Latvia's Finance Ministry, making the case for reform, noted that in 2024 just 80 contracting authorities accounted for 90% of all procurement value, and only 4,321 suppliers won anything at all — barely 2% of the country's roughly 180,000 active firms.

Who is doing the buying

By sheer number of tenders, the most active buyers are municipalities: Liepāja, Jelgava, Daugavpils, Bauska, Limbaži and a dozen others, each running dozens of mostly small contracts for schools, streets, water and heating. By money, though, the marquee buyers are state-owned companies and agencies — Rīgas satiksme, Latvijas dzelzceļš, Sadales tīkls, Latvian State Roads, the Freeport of Riga. The municipalities buy often; the state companies buy big.

Geographically, Rīga is the single largest place of performance at 27% of tenders, but the regions together hold the clear majority — Zemgale, Vidzeme, Kurzeme, Latgale and Pierīga between them account for far more than the capital. Roughly a quarter of tenders are national or unspecified in scope.

How Latvia buys

Almost everything goes out as a fully open procedure (97%), with restricted, negotiated and competitive-dialogue procedures making up the rest. Bidders get a median of 29 days to respond, and a third of tenders give 20 days or fewer — a tight window for anything complex.

On the question of how winners are chosen, Latvia has a reputation for buying on price, and the data only half-confirms it. Per the EU's Single Market Scoreboard, 51% of Latvian procedures are awarded on lowest price alone — actually a touch below the EU average of 54%. So Latvia is not unusually price-obsessed on paper. What the headline misses is that even where "quality" criteria exist, they are often a formality worth a few points, frequently nothing more than who offers the longest warranty. The competition is effectively about price far more often than the 51% suggests.

There is a genuinely positive number, though. The same scoreboard shows Latvia with a single-bidder rate of just 22% (against an EU average of 28%) and SME participation at 91%, far above the EU's 71%. The financial barriers to entry are low — most tenders ask for around one year's contract value in turnover, well under the twice-the-value ceiling EU rules allow, and many ask for no turnover at all. Low walls, lots of small firms getting in. That is the upside of a price-driven, lightly-gated system.

The defence line in the data

One trend jumps out of the buyer list. Among Latvia's most active procuring bodies are two dedicated defence-procurement centres, which together with the Ministry of Defence ran 74 tenders in the half-year. That is the post-2022 security build-up showing up as line items — and it is only the visible part, since a defence budget that keeps climbing buys plenty that never reaches an open European tender at all. When a country's threat environment changes, its procurement system is one of the first places you can measure it.

A system rebuilt in the middle of our dataset

Here is the twist that makes 2026 a strange year to measure. Halfway through our window, the rules changed. Latvia's State Audit Office had concluded, in a 2024 audit, that the country's procurement regime was overgrown and slow:

A Finance-Ministry working group, chaired by the finance minister, turned that verdict into a structural reform. The Cabinet approved it on 26 August 2025, the Saeima passed the amendments, and they entered into force on 9 June 2026 — two weeks before the end of our dataset. The headline changes: the number of mandatory exclusion criteria is cut from 12 to 2, and from 1 January 2028 the value thresholds for running a full procedure rise sharply — to EUR 140,000 for goods and services, EUR 750,000 for social services, and EUR 1,000,000 for construction — with low-value buys below those lines dropping their formal procedures entirely. The government expects procedures to run about 25% faster and 2–4% cheaper.

For a data analyst, this is a reminder that procurement systems are living things. The 2,154 tenders we counted were published under one regime; next year's will be published under another, with a chunk of today's smaller tenders dropping off TED entirely as the thresholds rise.

The EU's fingerprints

Two EU forces shape what we found. The first is regulation as a buying instruction: those 60 electric buses are not a green gesture but compliance with the Clean Vehicles Directive, which requires Latvia to hit 50% clean buses in public contracts for 2026–2030, half of them zero-emission. The second is money: a large share of Latvia's public buildings are co-funded from the country's EUR 4.6 billion 2021–2027 EU Cohesion envelope. Watch the construction tenders and you are partly watching Brussels' budget being spent.

What it adds up to

Six months, 2,154 tenders, at least EUR 1.7 billion. The portrait is of a small state doing methodical, slightly conservative things: pouring most of its big money into construction, roads, the grid and public transport; keeping the door open to small local firms; buying largely on price; quietly rearming; and rebuilding its own rulebook as it goes. The money is concentrated at the top, the volume is spread thin at the bottom, and the most interesting facts only surface once you read past the headline figures into the documents themselves.

Working out what each document asks for, what it is worth, and where it contradicts itself is the entire problem of reading procurement at scale. We did it here by hand, with scripts and a lot of reading, to write one article about one country for half a year. It is exactly the work we built MitigateTenders to do continuously: read every tender and every bid, tie each finding to the requirement it answers and the passage that proves it, and tell you what is actually in the documents. A country's spending is hiding in plain sight, in tens of thousands of files. The trick is reading them.