Irish Public Procurement Insights - Q2 2026

Each quarter TenderFlare publishes an analysis of Irish public procurement activity, drawn from the eTenders / OGP open dataset. Here is what the second quarter of 2026 looked like.

One thing before the numbers: this edition corrects an error in the Q1 edition, where we listed the Office of Government Procurement as the second most active buyer in the country. It is not a buyer. The correction, and what the buyer table looks like once it is fixed, are under “Who is buying” below.

More competitions, each of them smaller

Irish public bodies put 2,710 new contract notices to market in Q2 2026, up 25.9% on Q2 2025 (2,152) and 27.6% on Q1 (2,124). That is the third-highest quarter since the dataset begins in 2013, and the second-strongest second quarter on record, behind only Q2 2022 (2,810). The first half of 2026 is running 21.7% ahead of the first half of 2025.

The estimated pipeline attached to those notices is roughly €21.3 billion. Separately, €15.88 billion is recorded against the 1,565 awards that carry both a value and an award date inside the quarter. That award figure is unusually large, and it is worth being precise about why: five awards account for more than half of it.

The median award was €216,000. That is up 2.3% on Q1 but down 14.0% on Q2 2025 (€251,184). More contracts, each of them typically smaller.

How to read the value figures

Two observations;

Framework values are multi-year ceilings, not spend. A €2.5bn framework is the maximum that can be drawn down over its life, usually shared across several suppliers. It is not money leaving the exchequer this quarter. This is why we lead with notice counts and medians and treats the billions as directional.

Q2 is a young quarter. Award data arrives for two to three quarters after a notice publishes. Only 24.5% of Q2’s notices carry award data so far, against roughly 55% for a mature quarter. Every award figure here is a floor that will rise. Comparing Q2’s award total to an older quarter’s is comparing a partial count to a settled one.

What changed since the Q1 edition

Four figures here will not line up with our Q1 edition if you have both open;

The median. Q1 published €215,000. On today’s data the same quarter reads €211,140, and the 2.3% quarter-on-quarter change above is measured against that restated figure. Two things moved it in opposite directions. A further quarter of award records has loaded since we published in August, which pulled it down. And this edition counts only awards that carry a value, excluding 39 recorded at zero, which pulls it back up: on the wider basis Q1 reads €200,811. The two effects partly cancel.

Single-bidder share. Q1 reported 14%, measured against all 1,525 awards including those recording no bid count at all. This edition reports 21.8% against the 1,223 that record one. The two are not comparable, and competition has not tightened by half between quarters. This basis is the sounder of the two.

The buyer table. Q1 ranked the HSE first on 100 notices with OGP second. That was the old basis, and it was also a different quarter. Client-resolved, the HSE shows 63 this quarter and OGP does not appear at all. The two tables are built differently and should not be read row against row.

Notice counts. Q1 published 2,109 notices for Q1 2026 and 1,811 for Q1 2025. Both now read higher, at 2,124 and 1,821. The OGP dataset is republished each quarter and historic counts revise slightly as records are added.

We are not restating the Q1 post’s figures to match. We will keep restating figures here as the data fills out rather than freezing the first number we saw.

Who is buying

A correction from the Q1 edition

Our Q1 edition listed the Office of Government Procurement (OGP) as the second most active buyer in the country.

OGP is a shared procurement service. When it runs a competition, the resulting contract belongs to whichever public body asked for it, not to OGP. Counting its notices as its own purchases inflates OGP and, more importantly, hides the bodies actually doing the buying.

From this edition on, every notice is resolved to the body that receives the goods or services. Applied to Q2, OGP ran 94 notices, of which 88 name a different client - 93.6%. Resolved, OGP leaves the top ten entirely.

Its largest client groupings were the Department of Education and Youth (8), the Department of Agriculture, Food and the Marine (7), the HSE (5) and the Road Safety Authority (4), with the Departments of Justice and of Health, the Irish Prison Service, University of Galway and the CSO on three each. A further 49 were spread across smaller bodies, and six carried no client at all.

Getting buyer attribution right is the entire point of this series, so we would rather correct it in the open than let it stand.

The corrected table

Two things stand out. Cork City Council recorded more notices than any other single named buyer, ahead of every Dublin authority. Read that as “the highest recorded count”, not a settled ranking: because name variants are not folded (ongoing data cleaning is work in progress with us), the HSE’s 63 is spread across seven or more spellings and its true total is higher than shown, quite possibly higher than Cork’s. What is not in doubt is that a mid-sized city council is running procurement at the scale of a national body.

The second is that local government dominates outright: six of the ten are county or city councils.

We are not publishing a buyers-by-value table this quarter. 131 records carry pipe-joined multi-client values (ongoing data quality control and assurance activities by us), which would surface as fake buyer entities like “Kildare County Council|Kilkenny County Council” if ranked by spend. Value analysis below stays at contract-type and sector level, where it needs no buyer entity. We will return to buyers-by-value once that is resolved.

What is being bought

By count, this is a services quarter. By value, it is emphatically a works quarter, and those two facts are not in tension - they describe different things.

Separately, and on a different basis: of the 2,710 notices published in the quarter, 1,304 were Services, 569 Works, 413 Supplies, 141 Work Related Services, and 283 carried no type. Notices and awards are different cohorts. They should not be read as one series.

Competition and access

Most competition here sits below the EU threshold, where the administrative burden is lower.

Who was actually allowed to bid

The threshold split says how big the competitions were. A more useful cut is by procedure, because it sorts the quarter by who was allowed to enter.

2,482 competitions, 91.6% of the quarter, were open to any qualifying firm. A further 193, or 7.1%, were not open at all: 187 mini-competitions run under an existing framework or dynamic purchasing system, where only firms already admitted could bid, and 6 negotiated without prior publication, which involve no competition.

That leaves 35. These were qualification rounds, 25 for dynamic purchasing systems and 10 for other qualification systems, and they are the doors into that closed 7%. Thirty-five in three months, across every sector in the country.

A firm that misses one is not missing a contract. It is missing every call-off that framework runs for its duration (typically ~3yrs). A framework closes after its qualification round and does not reopen, whereas a dynamic purchasing system admits new applicants throughout its life, so a DPS missed in Q2 can still be joined.

Awarded contracts drew a mean of 4.49 bids and a median of 3. But a substantial minority drew no competition at all.

Bid counts publish with the award, not with the notice, so they tell you about a buyer’s history rather than about the competition in front of you. That is the more useful direction anyway. A buyer whose competitions repeatedly draw one or two bidders is describing its own market to you. Every awarded contract in TenderFlare shows the bid count it drew.

The biggest awards of the quarter

Five awards account for 56.1% of the quarter’s award value, and eight exceed half a billion euro. All are framework ceilings.

One buyer runs through that table. Four of the eight largest awards are Department of Education and Youth construction frameworks, worth €5.95bn of ceilings between them, or 37.5% of the quarter’s entire award value from a single department. If you build, fit out, wire or heat public buildings, that is where the decade’s work is being positioned right now.

The right-hand column is the other part worth dwelling on. Only three of the eight largest awards name any supplier at all. For the other five, worth €8.45bn between them, the eTenders record names none. Those names often exist elsewhere, in the TED notice or in the buyer’s own published framework list, which is precisely the reconciliation work a bid team should not be doing by hand. If you are trying to work out who is winning the biggest positions in this market, the largest contracts are frequently the least informative place to look.

Sectors, and why the category label are sometimes not reliable

Start with the classification problem;

Environmental work barely exists under the environmental CPV code. Division 90 carried just 80 notices and €30m in Q2. That is not the size of the environmental market in Ireland. It is an artefact of how the work is filed: water, drainage, remediation and environmental assessment are mostly procured inside construction and engineering notices instead. Search by category and you will not find that work. Follow the buyers and you will.

This is why TenderFlare profiles buyers mainly rather than categories. A category filter inherits whatever classification the buyer happened to choose. A buyer’s own history does not.

With that caveat attached, the three largest sectors by notice count:

Three things we are watching in Q3

Whether the award total holds. Q2’s €15.88bn rests on 24.5% award coverage. As that matures toward the ~55% a settled quarter reaches, the total will rise. The open question is whether the concentration holds with it, or whether the top five simply become a smaller share of a much larger base.

Whether the Education ceilings convert. Four framework ceilings worth €5.95bn are now in place with no suppliers named against them. Call-offs under those frameworks are where the actual work appears, and whether they surface as named awards in the data is the single biggest determinant of how useful the next few editions are.

Whether notice growth is real. 2,710 notices is a 25.9% year-on-year rise, and we have not yet separated genuine market growth from fuller publication of low-value notices that previously went unrecorded. A rising count with a falling median is consistent with either. Q3 gives us a third data point.

Methodology

Figures are drawn from the eTenders / OGP open dataset published via data.gov.ie and loaded into TenderFlare. Q2 2026 covers notices published 1 April to 30 June 2026. Notice counts and estimated values use publication date; award counts and values use award date and are approximately 24.5% complete for this quarter. Framework values are multi-year ceiling values, not committed single-contract spend.

Buyers are resolved to the client body named on the notice where one is recorded, rather than to the body that ran the competition. Buyer counts are floors, because name variants are not folded together.

Award totals and medians describe the 1,565 awards carrying a value; award counts describe all 1,604 dated in the quarter. Single-bidder share is measured against the 1,223 awards recording a bid count. Figures cover the eTenders source only, so a notice that also appears as a separate TED record is not counted here.

Two records in this quarter’s source data, both titled as tests and carrying €410m between them, were identified as non-genuine and excluded before these figures were produced.

Data as at 31 August 2026.

Pick a buyer from the table above and look it up. Every profile is free to browse. Cork City Council, to take the one at the top: 1,062 competitions on record, an average of 4.6 bidders, 141 days from notice to award, and a 6% cancellation rate. That is a buyer’s temperament, and it is the difference between a considered bid/no-bid call and a guess. Search the whole dataset the same way. TenderFlare Pro adds the part that takes longest to work out by hand, the ranked supplier list and spend breakdown behind each authority.

The Q3 2026 edition publishes later this year. Email us for the PDF of this edition and we will send the next one when it lands.