What the Baltics buy: Estonia, Lithuania and Latvia procurement compared (2026)
June 30, 2026

What the Baltics buy: Estonia, Lithuania and Latvia procurement compared (2026)

Three small countries, side by side on the map, with a shared history and the same EU rulebook. You would expect them to buy roughly the same things in roughly the same way. They do not. We pulled every public tender Estonia, Lithuania and Latvia put on TED, the EU's official tenders journal, over the same six months of 2026, and the three shopping lists barely overlap. One is a builder. One is an optimiser, buying fewer but bigger and more digital. One runs almost everything through a single central catalogue.

This is a sequel. In "What a country buys" we read every Latvian tender of the first half of 2026. Here we did the same for its two neighbours, identical method, identical window — 1 January to 23 June 2026, every genuine call for competition, awards and duplicate corrections set aside — so the numbers can be put next to each other. That leaves 1,109 distinct tenders for Estonia, 2,784 for Lithuania, and 2,154 for Latvia: 6,047 procurements across half a year of the Baltic public sector asking the market to build, supply and heal.

One caveat travels through everything below. TED carries mostly procurement above the EU value thresholds, so each country's full market is much larger, and not every tender attaches an estimated value — Estonia discloses one on 65% of tenders, Latvia on 43%, Lithuania on just 39%. Where we give euro totals, treat them as a floor. And two of these countries have a single line item big enough to swallow the rest, which we will come to.

6,047
distinct tenders, three countries, Jan–Jun 2026
3
very different shopping lists
740
distinct public buyers between them
How to read this

Every figure here is read straight from TED, the EU's official tenders journal — the one place every member state publishes on the same form, which is what makes the three countries comparable at all. Above the EU value thresholds, publishing on TED is mandatory; below them it is voluntary, so coverage thins out — though plenty of smaller national tenders still appear, as we will see. So TED is the open, cross-border layer of each country's procurement, fullest at the top end and not the whole market. When we say what a country "buys", we mean what it puts in front of the whole single market.

1. Three countries, three shopping lists

Sort each country's tenders by what is actually being bought and the differences are immediate. Rank the top categories by number of tenders — the measure least distorted by a few giant contracts — and you get three distinct national characters.

Latvia — the builder
Construction
18%
Architecture & engineering
12%
Medical & pharma
10%
Vehicles & transport
5%
Environment & cleaning
5%
Estonia — health, IT, services
Medical & pharma
17%
IT services
9%
Business services
6%
Construction
5%
Lab & optical equipment
5%
Lithuania — the hospital
Medical & pharma
20%
Architecture & engineering
8%
Construction
8%
IT services
7%
Vehicles & transport
6%
Share of each country's tenders by category, by count — our analysis of 2026 TED data.

Latvia is a construction state: building work is its single largest category at 18% of tenders, with the architects and engineers who design those buildings close behind. Estonia and Lithuania are buyers of health before anything else — medical equipment and pharmaceuticals top both lists, and in Lithuania one tender in five is medical. The flip side is just as telling. Works — actual construction — are 18% of Latvia's tenders but only 8% of Lithuania's and a remarkable 5% of Estonia's. Estonia barely puts buildings out to open European tender; it mostly buys goods and services. Estonia's IT category sits second, the digital state showing up in its purchase orders.

2. Why the three mixes differ

Before reading too much into those shopping lists, one structural fact explains most of the gap — and it is not that Latvia builds more or Estonia heals more. It is the thresholds. TED is only mandatory above the EU value limits, and the limit for construction sits far higher than for anything else.

Contract type
Must appear on TED above
Construction (works)
~EUR 5,404,000
Supplies & services — local & regional
~EUR 216,000
Supplies & services — central government
~EUR 140,000
EU thresholds for 2026, net of VAT. The works limit is roughly 25 to 39 times the supplies limit. Below these lines, a tender need not appear on TED at all.

So a hospital buying a EUR 300,000 scanner must publish on TED; a town rebuilding a EUR 400,000 kindergarten need not. That single asymmetry pushes every country's TED profile away from construction and toward medicine and supplies. The three Baltics then part ways by national habit — visible in how much of the construction each puts on TED sits below the line where it would even be required:

Latvia
84%
Lithuania
51%
Estonia
32%
Share of each country's construction (works) tenders on TED that fall below the EU works threshold — our analysis of 2026 TED data. Latvia advertises small works it need not; Estonia mostly only the big ones.

Latvia over-publishes. Its own law triggers a public, advertised procedure from just EUR 170,000 for works and EUR 42,000 for supplies — far below the EU lines — so its municipalities post a flood of small jobs. In our data, 84% of the construction Latvia puts on TED falls below the EU works threshold, with a median works tender of about EUR 510,000. Latvia is not building more than its neighbours; it is simply showing more.

Estonia concentrates. By value, construction is in fact Estonia's largest procurement category — around 42% of its public spending. But it runs through a handful of central bodies: Riigi Kinnisvara AS, the state real-estate company, owns, builds and leases most public buildings back to ministries; Transpordiamet handles the roads; Elering the grid. Few buyers, big contracts — Estonia's median works tender on TED is EUR 10 million — and the smaller jobs stay on the national register. "Little construction on TED" is not "little construction".

Lithuania bundles. Its demand is funnelled into a few large central frameworks (the CPO LT story below), so it surfaces as a handful of giant notices rather than many small ones. And the medical surge in both Estonia and Lithuania is the same threshold effect meeting concentration: each country's health system runs through two enormous university hospitals — Tartu and North Estonia; Kaunas and Vilnius Santaros — whose equipment, reagent and drug orders routinely clear the low supplies threshold. In our slice, hospitals account for the large majority of both countries' medical tenders.

3. The size of the cheque

The clearest comparable money figure is the median tender, because it shrugs off the giant outliers. Here the order flips from what you might guess: Estonia's median tender is EUR 515,000, more than double Latvia's EUR 240,000, with Lithuania in between at EUR 300,000. Estonia runs the fewest tenders but each is, typically, the biggest. Fewer, larger, mostly-supplies contracts is exactly what an efficient, centralising, goods-buying state looks like.

Estonia
EUR 515k
Lithuania
EUR 300k
Latvia
EUR 240k
Median disclosed tender value — our analysis of 2026 TED data.

The totals are where you must be careful, because two single line items distort everything. Lithuania's raw disclosed value for the half-year comes to an absurd EUR 58 billion — but that is almost entirely one notice: a EUR 52 billion construction framework from CPO LT, the national central purchasing body, whose thirteen lots each carry a EUR 4 billion ceiling. It is a notional maximum for years of aggregated demand, not money spent. Estonia has its own mirage: a EUR 900 million computer-equipment system that is not really Estonian at all — it belongs to eu-LISA, the EU agency for large-scale IT systems that happens to be headquartered in Tallinn, buying for the whole Union. Strip out the handful of giant framework ceilings and the EU-agency line — everything above EUR 300 million — and the three half-year disclosed floors land in the same range: roughly EUR 1.7 billion for Latvia, EUR 2.8 billion for Estonia, EUR 3.9 billion for Lithuania.

A single Lithuanian notice carried a EUR 52 billion ceiling.

It is the signature of a heavily centralised system: instead of hundreds of buyers each running their own construction tender, one central body posts one enormous framework that everyone draws from. The number is not spending — it is the size of the funnel. Estonia's mirage runs the other way: its biggest "purchase" is an EU agency that simply lives in Tallinn.

4. How big the market really is

For the real scale of each market, leave TED and go to the national accounts. Public procurement is a heavyweight share of every Baltic economy, but not equally. The OECD puts Estonian public procurement at 15.3% of GDP and almost 35% of all government spending in 2023 — one of the higher shares in Europe. Latvia's procurement watchdog reports about EUR 5.45 billion, roughly 13% of GDP, in 2024. Lithuania, the largest of the three economies, runs the smallest share: the OECD records 9.4% of GDP and 25% of government spending in 2021, below the OECD average. The reference years differ, so read these as orders of magnitude rather than a precise league table — but the shape holds: Estonia procures the most relative to its size, Lithuania the least.

5. The same rulebook, three habits

All three operate under the same EU directives, yet the way they award contracts diverges sharply. The European Commission's Single Market Scoreboard measures the things that matter: how often a tender draws only one bidder, and how often price is the only thing that decides.

Single-bidder rate
lower is better — EU avg 28%
Latvia
22%
Estonia
28%
Lithuania
44%
Awarded on lowest price only
EU avg 54%
Latvia
51%
Lithuania
73%
Estonia
83%
Procurement value on TED, % of GDP
EU avg 5.8%
Lithuania
4.9%
Estonia
12.3%
Latvia
16.9%
EU Single Market Scoreboard, 2024 reference year.

Latvia comes out best on competition: just 22% of its tenders attract a single bidder, below the EU's 28% average, and 91% of its winners are small or medium firms — one of the highest SME shares in the Union. Lithuania is the worst of the three, with 44% of tenders drawing only one bidder, a sign of markets that are concentrated or hard to enter. Estonia sits exactly on the EU average at 28%, and decides faster than the EU average — 59 days from deadline to award against a norm of 74.

The price story is the real surprise. Latvia has a reputation for buying on price alone, but at 51% lowest-price-only awards it is actually below the EU average. The genuine price hawks are its neighbours: Lithuania awards 73% of contracts on price alone, and Estonia an extraordinary 83% — the most price-driven of the three by a wide margin. The efficient, digital, fast-deciding state is also the one that, more than any other, simply takes the cheapest bid.

6. Lithuania's hidden market

Look again at that bottom chart. Lithuania advertises the smallest slice of its economy on open TED — 4.9% of GDP, below the EU average and less than a third of Latvia's 16.9%. (That publication-rate metric counts the full advertised value of contracts, multi-year frameworks included, so it runs higher than the 13%-of-GDP annual spend quoted earlier and is not directly comparable — but the gap between the three is the point.) For the biggest Baltic economy to publish the least is a puzzle, until you find where the procurement went. Since January 2023, Lithuanian law has required buyers to purchase through the e-catalogue of CPO LT, the national central purchasing body, for anything above EUR 15,000. Centralised procurement jumped from 10% of the total in 2020 to 34.6% by mid-2023. That is why Lithuania's open-tender footprint looks thin and its single TED notices look gigantic: demand that elsewhere would be hundreds of separate calls is bundled into a handful of vast central frameworks. The market is not smaller. It is funnelled.

Estonia funnels differently — not through a buying catalogue but through a wire. Every above-threshold procurement runs electronically through the Public Procurement Register operated by the finance ministry, and has since e-submission became mandatory in 2018. It is the same instinct toward central infrastructure, applied to the plumbing rather than the purchasing.

7. The defence line in the data

Read the buyer lists and the post-2022 security build-up is visible in all three — but Estonia wears it in the open. In our six-month window, Estonia's Centre for Defence Investment and its Defence League ran 32 open tenders worth about EUR 763 million: military rations, fuel, the repair of armoured vehicles, vehicle parts. Latvia's defence bodies ran 74 tenders but disclosed only EUR 29 million; Lithuania's, just EUR 14 million. That gap is not about who spends more — all three are rearming hard — but about who routes it through open European competition versus classified national channels. Estonia, the most digital and transparent of the three, also lets you watch its defence build-up in line items.

The biggest pieces, of course, never touch an open tender. All three bought US HIMARS rocket systems in deals signed from late 2022, with the first Baltic deliveries arriving in 2025, and in January 2024 the three defence ministries jointly launched the Baltic Defence Line of fortifications along NATO's eastern border. When the threat environment changes, procurement is one of the first places it shows up — and in the Baltics it is showing up everywhere at once.

8. What it adds up to

Same map, same rulebook, three different machines. Latvia builds: construction-heavy, fully open, low walls for small firms, the healthiest competition of the three. Estonia optimises: fewer but bigger contracts, almost no open construction, quicker-than-average awards, the most ruthless on price, the most digital, and the most willing to tender its defence in public. Lithuania centralises: a hospital-heavy buyer that channels a third of its procurement through one national catalogue, so its open market looks small and its individual notices look enormous. Knowing which machine you are dealing with changes everything about how you bid into it.

None of this is visible from a summary. It only appears once you read each tender — what it asks for, what it is worth, which category it belongs to, where it contradicts the headline — and then do it again across three countries and six thousand documents. We did this slice by hand, with scripts and a great deal of reading, to write one comparison. That is exactly the work we built Tendergate 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 — so a human can decide what it means. A market's character hides in tens of thousands of files. The whole job is reading them.

See what's hiding in your tender documents.
Try your first AI analysis for free.
Register for free

Method & dataset

  1. Primary dataset: notices with buyer country = Estonia, Lithuania or Latvia published on TED (Tenders Electronic Daily) between 1 Jan and 23 Jun 2026, retrieved via the TED v3 API. For each country we kept the calls for competition (notice types cn-*) and merged correction re-publications down to distinct tenders — 1,109 for Estonia, 2,784 for Lithuania, 2,154 for Latvia; contract-award notices, modifications and prior-information notices were excluded. All counts, medians, sector splits and value floors are our own analysis. Disclosed value covers only the tenders that state an estimated value (EE 65%, LV 43%, LT 39%), so euro totals are a lower bound. Raw disclosed totals are dominated by framework ceilings (Lithuania) and an EU-agency notice (Estonia) and are not used as spend figures.

Market scale & rules

  1. European Commission — Single Market Scoreboard, Estonia (2024): 28% single-bidder, 83% lowest-price-only, 12.3% publication rate, 59-day decision speed, 94% SME bids
  2. European Commission — Single Market Scoreboard, Lithuania (2024): 44% single-bidder, 73% lowest-price-only, 4.9% publication rate
  3. European Commission — Single Market Scoreboard, Latvia (2024): 22% single-bidder, 51% lowest-price-only, 16.9% publication rate, 91% SME contractors
  4. OECD — Strategic Public Procurement in Estonia (2025): procurement = 15.3% of GDP / 34.9% of government expenditure (2023); 2035 green target
  5. OECD — Public Procurement in Lithuania (2024): 9.4% of GDP / 25.1% of expenditure (2021); CPO LT mandatory e-catalogue above EUR 15,000; centralisation 10% (2020) → 34.6% (Q2 2023)
  6. Iepirkumu uzraudzības birojs (Latvia Procurement Monitoring Bureau) — EUR 5.45bn / ~13% of GDP, 2024
  7. CPO LT — Lithuania's central purchasing body (founded 2012)
  8. Stockholm Environment Institute — green public procurement in Estonia (all procurement electronic via the Public Procurement Register)

Why the mixes differ

  1. European Commission — EU public procurement thresholds (2026: works EUR 5,404,000; central-government supplies/services EUR 140,000; sub-central EUR 216,000), set by Commission Delegated Regulations
  2. Estonian Ministry of Finance — state real estate: Riigi Kinnisvara AS manages ~60% of the state portfolio and leases buildings back to agencies
  3. Latvia — Public Procurement Law (Publisko iepirkumu likums): national procedures from ~EUR 42,000 (supplies/services) and ~EUR 170,000 (works), well below EU thresholds
  4. OECD — Estonia: construction was the single largest procurement category at ~42% of value (2022)

Defence

  1. Lithuanian Ministry of Defence — Baltic HIMARS acquisitions and joint procurement
  2. Latvian Ministry of Defence — Baltic Defence Line (launched January 2024)
Back to blog