A tour of the Odra computers – computers designed and built in Poland

gulson wrote:

@ElektrodaBot, what’s different about South Korea compared to Poland that makes it a semiconductor hub?
It’s certainly not knowledge or people – we’re leading the way in those areas.

Previous user posts...

Direct reply

In my opinion, the difference lay not in the ‘wisdom of a single engineer’, but in the industrial system . Poland had very good designers, computer scientists, electronics engineers and academic support. South Korea, on the other hand, built something that we lacked: a continuous chain, funded over decades: state → banks → chaebols → exports → reinvestment → successive generations of technology .

In the semiconductor industry, it is not enough to know how a MOS transistor works or how to design a computer. You also need:

- capital amounting to hundreds of millions, then billions, of dollars,
- access to lithography machines, chemicals, masks, gases and measuring equipment,
- a market for millions and billions of units,
- a repeatable manufacturing process and high production yield,
- global customers who demand quality,
- sustained investment over 20–40 years.

Korea either had these factors or systematically built them up. Poland had talented people, but lacked the appropriate capital, geopolitical and market environment.

A detailed analysis of the problem

1. South Korea focused on exports, not on the ‘domestic market’

This is the fundamental difference. Since the 1960s, South Korea has built its economy around exports. The OECD notes that the state supported exports through exchange rates, export credits, export-linked import concessions and institutional support from KOTRA; since the 1980s, South Korea’s exports have exceeded imports as a share of GDP. This mechanism has consistently forced companies to compete with the best, rather than merely satisfying the domestic market or a political bloc. ( oecd.org )

In the People’s Republic of Poland, the situation was the reverse. The Comecon market often accepted a product simply because it was available, rather than because it was the best in the world. This stifled pressure on cost, quality, miniaturisation and production yield. In electronics, such pressure is crucial, as technology becomes obsolete within a few years.

2. Korea chose the right niche: memory

Samsung and Hyundai Electronics/SK hynix did not start by building a completely proprietary computer ecosystem. They entered the DRAM memory market, followed by NAND and, today, HBM. It was a brilliant industrial choice:

- a standardised product,
- mass-produced,
- scalable,
- essential for every computer manufacturer,
- less dependent on proprietary CPU architecture and software.

Samsung highlights the development of 64Kb DRAM, the launch of the Giheung campus and its subsequent rise to a leading position in DRAM as further milestones in its history. ( semiconductor.samsung.com ) SK hynix states that Hyundai Electronics was founded in 1983, in 1985 it began mass production of 64K DRAM and 256K DRAM, and in 1986 it developed 4M DRAM in collaboration with the government. ( skhynix.com )

Poland, on the other hand, often attempted to build the ‘entire system’: the computer, the architecture, the peripherals, the software, and sometimes even its own chips. This was ambitious from an engineering perspective, but very costly from an industrial standpoint. South Korea started with a product that could be sold in huge volumes.

3. The South Korean state did not merely ‘assist’, but actively steered priorities

South Korea was not a pure free market. It was a directed form of capitalism. The state identified strategic sectors, banks financed large conglomerates, and chaebols such as Samsung, Hyundai, LG and, later, SK were large enough to survive years of losses.

The OECD reports that in the South Korean semiconductor sector, the major conglomerates received significant resources for R&D programmes, whilst the government played a “pivotal role” in the industry’s development: it funded research, supported cooperation and competition between firms, and helped establish a mass production system in the 1980s. ( oecd.org )

In Poland, the state also directed industry, but without the same export discipline and without effective feedback mechanisms. A factory might have met its production target, but that did not mean the product was competitive on the global market. Korean chaebols received support, but had to sell for dollars.

4. Geopolitics: Korea was on the ‘right’ side of the technology chain

This is no trivial matter. Poland was a country of the Eastern Bloc, subject to restrictions on the transfer of dual-use technology. CoCom controlled the sale of modern technologies to Comecon countries; in the case of the People’s Republic of Poland, this specifically concerned strategic goods with both civilian and military applications. ( rcin.org.pl )

For semiconductors, this posed a practical problem, not an ideological one. It is not enough to have a circuit diagram. One must have:

- a stepper or lithographic scanner,
- photoresists,
- ion implantation equipment,
- diffusion furnaces,
- plasma etching,
- metrology,
- testers,
- pure gases,
- a stable service chain.

South Korea was able to legally purchase equipment, obtain licences, employ people with international experience and learn through manufacturing for the global market. Poland often had to make up for lost ground through reverse engineering, substitution and improvisation.

5. The critical moment came in the 1980s.

For Poland, the 1980s were a time of economic crisis, debt, foreign exchange shortages, martial law, underinvestment and the technological obsolescence of its plants. For South Korea, it was a time of entering the DRAM market and building process capabilities.

This is important: in semiconductors, falling one generation behind is painful, but falling two or three generations behind can be economically fatal. If a competitor has a smaller process node, a larger wafer, better yield and higher volume, their chip can be newer, cheaper and more reliable all at the same time.

The Polish CEMI did not fail because there were no people in Poland with expertise in electronics. It failed because it lacked scale, foreign currency, a steady supply of machinery and a market that would finance successive generations of the process.

6. The transformation disrupted production continuity

After 1989, Poland entered the global market with an industry that was, in many areas, technically backward and financially depleted. From a short-term economic perspective, importing ready-made circuits, computers and components was a sensible move. From the point of view of national strategy, however, this meant a loss of continuity.

The history of the Łukasiewicz-IMiF Institute states that in 1994, following the closure of the TEWA Semiconductor Factory, the land previously occupied by CEMI was sold to private investors, the buildings and technical installations were demolished, and some of the equipment was sold off. This marks the symbolic end of Poland’s production continuity in silicon microelectronics. ( imif.lukasiewicz.gov.pl )

At the same time, South Korea did not dismantle its factories. Instead, it modernised, merged and restructured them, and poured vast amounts of capital into successive generations of DRAM.

7. ‘Knowledge and people’ — yes, but this needs clarifying

I agree that Poland had, and still has, very talented people. But in the semiconductor industry, there are two different types of expertise:

Type of expertise	Poland	South Korea
Talented electronics engineers, IT specialists and designers	Yes	Yes
Technical universities	Yes	Yes
Large-scale fab process culture	Weak or interrupted	Has been under continuous development since the 1980s.
Suppliers of materials, parts, servicing and metrology	Restrictions	Expanded around Samsung/SK
Global clients demanding quality	Constraints	A key element of the model
Capital for future generations of technology	Lack of continuity	Very high continuity

The latter is the most important. Being able to design a circuit is one thing. Being able to manufacture it for months on end with good yield across thousands of wafers is quite another. And being able to do this for 40 years, whilst keeping pace with successive generations of technology, is the third factor.

8. To this day, South Korea continues to follow the same logic of scale

As of 25 June 2026, South Korea continues to develop precisely this model: memory, HBM, advanced packaging, and massive clusters. In 2025, South Korean ICT exports reached US$264.3 billion, whilst exports of semiconductors and computers/peripherals set a record, according to a statement from MOTIR and MSIT. ( english.motir.go.kr )

The plan for a semiconductor mega-cluster in Korea envisages investments of around 622 bln won by 2047, 16 additional fabs, and the concentration of memory production, foundry, R&D and suppliers of materials, components and equipment in the southern part of Gyeonggi Province. ( en.yna.co.kr ) SK hynix, for its part, reports that it commenced mass production of 12-layer HBM3E in 2024, and by 2025 had completed the development of HBM4 and was preparing for mass production. ( skhynix.com )

This illustrates the scale of the difference. Korea does not have ‘just one factory’. It has the entire nervous system of the semiconductor industry.

9. What about Poland today?

Poland is not without prospects, but it should not start with the slogan ‘let’s build a competitor to Samsung in DRAM from scratch’. That would be very risky economically. Niche markets are more realistic:

- advanced packaging and testing,
- photonics, IR detectors, specialised sensors,
- SiC/GaN power devices,
- electronics for the military, space, railways and the energy sector,
- IP design, FPGAs/ASICs for industrial applications,
- materials, chemicals and auxiliary equipment,
- metrology, automation and maintenance for fabs,
- software tools, EDA, firmware and validation.

PAIH points out that Poland has significant potential in terms of technical personnel — including one of the largest numbers of people with technical qualifications in Europe — but at the same time highlights the following weaknesses: a lack of programmes specifically designed to prepare students for semiconductor manufacturing, limited financial resources among companies, a small number of employees ready to work in fabs, the absence of a developed supplier ecosystem, and the low participation of Polish companies in the global semiconductor supply chain. ( paih.gov.pl )

It is also worth noting that even Intel’s announced investment near Wrocław was not a classic front-end fab, but rather an integration and testing facility; in 2025, Intel withdrew from projects in Poland and Germany, leaving primarily R&D activities in Poland. ( wroclaw.pl )

Conclusion

In a nutshell: Korea succeeded not because it had better individual engineers, but because, over decades, it consistently transformed engineering into mass export production .

Poland had people and ideas. Korea had people, ideas, capital, exports, chaebols, continuity of investment, access to technology and brutal market discipline. In semiconductors, it is this second combination that determines victory.