Manufacturing productivity and AI adoption in Central & Eastern Europe — 2026 data report
Seven countries, one set of Eurostat datasets, every figure traceable to its code. Written for plant leaders who need numbers they can put in front of a board, and for anyone who needs to cite them.
Eight findings from the 2025 data
Value added per manufacturing worker, as a share of the EU-27 level
In 2025 a manufacturing worker generated between €23.2k of gross value added (Bulgaria) and €61.0k (Slovenia). The EU-27 figure is €89.3k and Germany's €109.8k. Poland stands at €39.5k and Czechia at €49.8k. These are nominal euros, not adjusted for price levels, so the gap in physical output is smaller than the ratio. Our calculation from Eurostat nama_10_a10 and nama_10_a10_e, 2025.
Rise in Poland's hourly manufacturing labour cost, 2023→2025
Bulgaria (+27.5%) and Romania (+25.0%) follow closely. The EU-27 rose 9.0% and Germany 6.7% over the same two years. Eurostat publishes the series in EUR, so part of Poland's increase reflects a stronger złoty rather than higher wages in local currency. Eurostat lc_lci_lev, 2023–2025.
Romanian manufacturers (10+ employees) using at least one AI technology, 2025
Bulgaria is at 5.3%, Poland at 7.7% and Hungary at 7.8%. All four are below half the EU-27 manufacturing rate of 17.3%. Romania's 2024 figure was 1.3%. Eurostat isoc_eb_ain2, 2024–2025.
Slovenia's manufacturing AI use, the only one of the seven above the EU-27 rate
Czechia (16.7%) and Slovakia (15.8%) are close to the EU-27 figure after moving from 9.6% and 8.1% in 2024. Part of that jump comes from the broader 2025 questionnaire, explained below. Slovenia barely moved (21.6% in 2024), which suggests its level was already high on the old basis. Eurostat isoc_eb_ain2, 2024–2025.
Fall in Romanian manufacturing employment, 2024→2025
Employment fell in six of the seven countries. Poland was the exception, at +0.9% (our calculation). Hungary lost 2.0%. Romania's manufacturing value added also fell 3.7% that year, so fewer workers did not mean more output. Eurostat nama_10_a10_e and nama_10_a10, 2024–2025.
Hungary's electricity price for a mid-size industrial user, 2025-S2
The band is 2,000–19,999 MWh a year, excluding VAT and recoverable taxes. Slovakia (€0.219), Poland (€0.214), Romania (€0.207) and Czechia (€0.198) also pay more than the EU-27 average of €0.192. Germany pays €0.237. Slovenia (€0.173) and Bulgaria (€0.163) pay less. Eurostat nrg_pc_205, 2025-S2.
Manufacturing share of Slovenia's gross value added, 2025
Czechia (20.8%) is also above Germany (19.4%). Slovakia (18.1%), Hungary (17.7%) and Poland (16.3%) sit above the EU-27 share of 15.8%; Romania (13.1%) and Bulgaria (12.7%) below it. In absolute terms Poland dominates with €132.9bn, about twice Czechia and three times Romania. Eurostat nama_10_a10, 2025.
Romania's manufacturing job vacancy rate, 2025
Poland (0.8%), Slovakia (0.7%) and Bulgaria (0.7%) are also below the EU-27 rate of 1.5%. Only Slovenia (2.1%), Czechia (1.9%) and Hungary (1.8%) are above it. Industry bodies in Czechia (AutoSAP, 2026) and Slovakia (ZAP, 2026) describe a shortage of skilled workers, which is consistent with low headline vacancy rates. Eurostat jvs_a_rate_r2, 2025.
Every indicator on this page for 7 countries, EU-27 and Germany, with Eurostat dataset codes and links. Excel.
Country comparison, 2025
One row per country, one dataset per column. Value added and employment are 2025; AI use is 2025; labour cost is 2025 with its two-year change; electricity is the second half of 2025.
| Country | Value added 2025 (EUR bn) | Share of total GVA | Employment 2025 (thousand) | Value added per worker (EUR k) | AI use, manufacturing, 2025 | Labour cost 2025 (EUR/hour) | Labour cost change 2023→25 | Electricity 2025-S2 (EUR/kWh) |
|---|---|---|---|---|---|---|---|---|
| Poland | 132.9 | 16.3% | 3,362.5 | 39.5 | 7.7% | 17.1 | +29.5% | 0.214 |
| Czechia | 66.0 | 20.8% | 1,325.7 | 49.8 | 16.7% | 20.2 | +9.8% | 0.198 |
| Romania | 44.9 | 13.1% | 1,399.8 | 32.1 | 3.2% | 12.0 | +25.0% | 0.207 |
| Hungary | 33.1 | 17.7% | 856.2 | 38.7 | 7.8% | 15.6 | +13.9% | 0.234 |
| Slovakia | 22.1 | 18.1% | 496.5 | 44.6 | 15.8% | 19.3 | +13.5% | 0.219 |
| Slovenia | 13.3 | 21.1% | 218.3 | 61.0 | 22.7% | 29.8 | +14.6% | 0.173 |
| Bulgaria | 12.8 | 12.7% | 553.2 | 23.2 | 5.3% | 10.2 | +27.5% | 0.163 |
| EU-27 | 2,668.8 | 15.8% | 29,901.6 | 89.3 | 17.3% | 35.0 | +9.0% | 0.192 |
| Germany | 797.2 | 19.4% | n/r | 109.8 | 24.4% | 49.5 | +6.7% | 0.237 |
Sources: Eurostat nama_10_a10 (B1G, CP_MEUR), nama_10_a10_e (EMP_DC), isoc_eb_ain2 (E_AI_TANY, 10+ employees), lc_lci_lev (D1_D4_MD5), nrg_pc_205 (MWH2000-19999, excl. VAT). Value added per worker is our calculation. n/r = German manufacturing employment not retrieved for this release; the per-worker figure for Germany comes from the same calculation in our research file.
| Country | Productivity per hour, whole economy, 2025 (PPS, EU-27 = 100) | Manufacturing job vacancy rate, 2025 | Basic digital intensity or higher, all sectors, 2024 | AI use, all sectors, 2025 |
|---|---|---|---|---|
| Poland | 68.4 | 0.8% | 70.0% | 8.4% |
| Czechia | 79.6 | 1.9% | 71.9% | 17.6% |
| Romania | 75.6 | 0.5% | 69.9% | 5.2% |
| Hungary | 70.4 | 1.8% | 58.5% | 10.4% |
| Slovakia | 81.1 | 0.7% | 64.2% | 18.0% |
| Slovenia | 86.2 | 2.1% | 68.5% | 21.6% |
| Bulgaria | 59.0 | 0.7% | 50.9% | 8.6% |
| EU-27 | 100 | 1.5% | 73.7% | 20.0% |
| Germany | 124.2 | 1.7% | 80.5% | 26.0% |
Sources: Eurostat tesem160 (NLPR_HW), jvs_a_rate_r2 (NACE C), isoc_e_dii (DII v4, 10+ employees), isoc_eb_ai (E_AI_TANY, 10+ employees).
Labour cost grew faster than output per worker in 2025
Put the one-year change in hourly labour cost next to the one-year change in value added per worker. In all seven countries cost moved faster. The widest gaps are in Bulgaria, Romania and Poland, where value added per worker was flat.
| Country | Hourly labour cost in manufacturing | Value added per manufacturing worker |
|---|---|---|
| Poland | +10.3% | +0.4% |
| Czechia | +8.0% | +2.6% |
| Romania | +11.1% | +0.6% |
| Hungary | +6.8% | +3.4% |
| Slovakia | +6.6% | +5.6% |
| Slovenia | +7.2% | +4.2% |
| Bulgaria | +13.3% | −0.2% |
| EU-27 | +3.6% | +2.8% |
Our calculation from Eurostat lc_lci_lev, nama_10_a10 and nama_10_a10_e. The two measures use different denominators (hours vs persons) and both include exchange-rate effects, so read the direction and the size of the gap, not a precise unit-cost figure.
Whole-economy productivity per hour, measured at purchasing-power parity, did improve: Romania from 72.6 to 75.6 and Poland from 66.2 to 68.4 (EU-27 = 100, Eurostat tesem160). The convergence is real. It is slower than the wage bill.
Read the 2025 AI figures with care
Eurostat's indicator E_AI_TANY counts enterprises with 10 or more employees that use at least one AI technology. In the 2025 survey Eurostat broadened the AI questions. The EU-27 manufacturing rate went from 10.6% to 17.3% in one year, and Czechia's and Slovakia's roughly doubled. The Czech Ministry of Industry and Trade headlined that Czech firms had doubled their use of AI. Some of that is adoption. Some of it is a wider net.
- Compare countries within the same year. The 2025 ranking is sound because every country answered the same questionnaire.
- Treat 2024→2025 changes as an upper bound on real growth, not as a trend line.
- Use the manufacturing series (isoc_eb_ain2, NACE C), not the all-sector series (isoc_eb_ai), when talking about factories. In Hungary the two differ by 2.6 points (7.8% vs 10.4%).
- The indicator counts firms, not use cases. A plant that uses a text-generation tool in its office counts the same as one running a model on machine data.
What the numbers mean for a plant
The cost advantage is shrinking where it was largest. Labour cost per hour in Polish manufacturing reached €17.1 in 2025, above Hungary's €15.6. Romania (€12.0) and Bulgaria (€10.2) are still the cheapest, but they also rose fastest. A plant that won contracts on labour cost in 2022 is quoting against a different base now.
Output per worker has not kept pace. In 2025 hourly labour cost rose 10–13% in Poland, Romania and Bulgaria, while value added per worker moved between −0.2% and +0.6%. Price increases to customers can close part of that gap once. After that the only lever left inside the plant is more good output from the same hours.
AI adoption is low outside Czechia, Slovakia and Slovenia. In Romania, Bulgaria, Poland and Hungary fewer than one manufacturer in twelve uses any AI technology. That is not by itself a failure. Most AI methods in production need a clean history of machine states, stop reasons and part counts, and many plants do not yet record those automatically.
Data capture comes first. Before any AI project, a plant needs to know when each constraint machine stopped, for how long and why. That can start with a paper log or a spreadsheet on one line. It can move to a local integrator, an MES module or a dedicated monitoring tool once the stop categories are stable. TeepTrak, the publisher of this site, sells such a tool; the order of steps is the same whichever you choose.
Energy is a second cost line worth measuring. Mid-size plants in Hungary, Slovakia, Poland, Romania and Czechia paid more per kWh than the EU-27 average in the second half of 2025. Idle-machine consumption is measurable with the same data you collect for stops.
Public money exists for the first steps. Several CEE digitalisation schemes treat software, sensors and implementation as eligible costs; status and conditions change often, and we track them with dates on factoryoptimizationai.com/cee/.
Methodology
All indicators come from Eurostat and were retrieved in September 2026 from the dissemination API. Dataset codes are given in every table so each figure can be checked in the Eurostat Data Browser.
| Indicator | Dataset and filter | Year |
|---|---|---|
| Gross value added, manufacturing and total | nama_10_a10, B1G, CP_MEUR, NACE C and TOTAL | 2024–2025 |
| Employment, manufacturing | nama_10_a10_e, EMP_DC, THS_PER, NACE C | 2024–2025 |
| Value added by manufacturing branch | nama_10_a64, B1G, CP_MEUR | 2023 |
| Number of enterprises | sbs_ovw_act, ENT_NR, NACE C | 2023 |
| AI use, all sectors / manufacturing | isoc_eb_ai / isoc_eb_ain2, E_AI_TANY, 10+ employees | 2024–2025 |
| Digital intensity | isoc_e_dii, DII v4, at least basic level, 10+ employees | 2024 |
| Hourly labour cost | lc_lci_lev, D1_D4_MD5, EUR, NACE C | 2023–2025 |
| Job vacancy rate | jvs_a_rate_r2, NACE C, annual average | 2024–2025 |
| Productivity per hour worked | tesem160, NLPR_HW, PPS, EU-27 = 100 | 2024–2025 |
| Industrial electricity price | nrg_pc_205, MWH2000-19999, excl. VAT and recoverable taxes | 2025-S2 |
Eurostat Data Browser: ec.europa.eu/eurostat/databrowser/view/[dataset code]/default/table
Value added per worker is manufacturing gross value added at current prices (nama_10_a10) divided by persons employed in manufacturing, domestic concept (nama_10_a10_e), for the same year. It is nominal and in EUR. It is not adjusted for price levels, so it overstates the physical productivity gap between low-price and high-price countries. For a price-adjusted view use tesem160, which is per hour and in purchasing-power standards but covers the whole economy.
Share of value added is manufacturing gross value added divided by total gross value added. It is close to, but not the same as, a share of GDP, which is typically 1–2 points lower because GDP includes net product taxes.
- Enterprise counts. Czechia (189,479) and Slovakia (89,803) register many sole traders as manufacturing enterprises. Their counts are not comparable with Poland's or Hungary's as a measure of factories. A breakdown by 10+ or 50+ employees was not retrieved.
- Exchange rates. Labour cost and value added are published in EUR. Currency movements change the EUR figures for Poland, Czechia, Romania and Hungary without any change in local wages or output. We did not convert to local currency.
- AI survey break. The 2025 AI questions are broader than 2024; see the section above.
- Coverage gaps. Robot density for Czechia, Slovakia, Poland, Hungary and Romania, regional manufacturing shares and German manufacturing employment were not retrieved. We left them out rather than estimate.
- Publisher. Disclosure: manufacturingml.com is published by TEEPTRAK SAS, which sells production-monitoring and OEE software. The data are Eurostat's; our own calculations are labelled as such.
Country profiles
- CEE 2026 data report: all seven countries compared
- Poland country profile
- Czechia country profile
- Romania country profile
- Hungary country profile
- Slovakia country profile
- Bulgaria country profile
- Slovenia country profile
- CEE Manufacturing Data Pack 2026 (xlsx)
Every indicator on this page for 7 countries, EU-27 and Germany, with Eurostat dataset codes and links. Excel.
By country
- Poland: manufacturing productivity, cost and AI adoption in 2026 data
- Czechia: manufacturing productivity, cost and AI adoption in 2026 data
- Romania: manufacturing productivity, cost and AI adoption in 2026 data
- Hungary: manufacturing productivity, cost and AI adoption in 2026 data
- Slovakia: manufacturing productivity, cost and AI adoption in 2026 data
- Bulgaria: manufacturing productivity, cost and AI adoption in 2026 data
- Slovenia: manufacturing productivity, cost and AI adoption in 2026 data
Questions
- Which CEE country has the most productive manufacturing sector?
- Slovenia, on both measures used here. Its value added per manufacturing worker was €61.0k in 2025, 68% of the EU-27 level (our calculation from Eurostat nama_10_a10 and nama_10_a10_e), and its whole-economy productivity per hour was 86.2 against an EU-27 index of 100 (Eurostat tesem160, 2025). Bulgaria is lowest at €23.2k and 59.0.
- How fast are manufacturing labour costs rising in Central and Eastern Europe?
- Between 2023 and 2025 hourly labour cost in manufacturing rose 29.5% in Poland, 27.5% in Bulgaria and 25.0% in Romania, against 9.0% in the EU-27 (Eurostat lc_lci_lev). Czechia (+9.8%) was the only one of the seven close to the EU rate. Figures are in EUR and include exchange-rate effects.
- How many manufacturers in CEE use AI?
- In 2025, 22.7% of manufacturers with 10+ employees in Slovenia used at least one AI technology, 16.7% in Czechia, 15.8% in Slovakia, 7.8% in Hungary, 7.7% in Poland, 5.3% in Bulgaria and 3.2% in Romania. The EU-27 rate was 17.3% (Eurostat isoc_eb_ain2). The 2025 survey asked broader questions, so year-on-year increases overstate real growth.
- Why is value added per worker so much lower in CEE than in Germany?
- Partly because output per hour is lower, and partly because prices are. The figure is nominal EUR, so it does not adjust for lower price levels in CEE. On a purchasing-power basis the gap is narrower: whole-economy productivity per hour is 59.0–86.2 in the seven countries against 124.2 in Germany (EU-27 = 100, Eurostat tesem160, 2025).
- Can I download the data behind this report?
- Yes. The CEE Manufacturing Data Pack 2026 is an xlsx file with one sheet per indicator, all seven countries plus EU-27 and Germany where available, dataset codes, source URLs, the value-added-per-worker calculation and a notes sheet with caveats.
Sources
- Eurostat nama_10_a10 (2024–2025)
- Eurostat nama_10_a10_e (2024–2025)
- Eurostat nama_10_a64 (2023)
- Eurostat sbs_ovw_act (2023)
- Eurostat isoc_eb_ai (2024–2025)
- Eurostat isoc_eb_ain2 (2024–2025)
- Eurostat isoc_e_dii (2024)
- Eurostat lc_lci_lev (2023–2025)
- Eurostat jvs_a_rate_r2 (2024–2025)
- Eurostat tesem160 (2024–2025)
- Eurostat nrg_pc_205 (2025-S2)
- Czech Ministry of Industry and Trade (MPO): Czech firms doubled their use of AI (2025/26)
- AutoSAP via Kurzy.cz: Czech automotive 2025 (2026)
- The Slovak Spectator: ZAP warning on Slovak automotive (13 Jan 2026)
Published by TEEPTRAK SAS, which makes production-monitoring and OEE software, with an office in Bucharest (TEEPTRAK SRL). Figures are sourced on each page. Funding rules change: check the official call documents before you budget.