Semiconductor Atmanirbharta: What India Can Learn From Its Global Peers

India has re-ignited its semiconductor manufacturing aspirations with the announcement of a Policy Framework taking shape under Modi’s Government.

In this context, let us explore the key ingredients that are at play in shaping and sustaining a Fab ecosystem, drawing lessons from Semiconductor Policies and the role of governments around the globe.

The Semiconductor Fabrication/Manufacturing plant (Fab) forges many links of a complex supply chain. Fabs rely on a highly specialized and complex supply chain with interdependencies, whether operated by an integrated device manufacturer (IDM), such as Intel, or a pure-play foundry, such as TSMC.

These interdependencies in the ecosystem include: A Fabless company’s design realized on a specific technology node, Electronic Design Automation (EDA) Tools supporting the fab’s process nodes, manufacturing equipment, the chemicals, wafer vendors and interactions of various R&D institutes for process optimization and future node developments. Establishing such an ecosystem is challenging and takes significant time, focus and effort.

This article tries to collate how various governments around the world have tried to nurture semiconductor manufacturing as a “strategic asset” and build their economies around this industry.

The semiconductor supply and manufacturing chain includes a handful of key players namely: United States, Taiwan, South-Korea, Japan, Europe and relatively recent entrant, China. The focus in this part is on EU and future parts will detail others namely Japan, South-Korea, and Taiwan.

Why Begin With EU And Not US?

Even though the origins of the semiconductor industry have roots in the US and Europe of the late 1800s, jointly developing thermionic value (aka. vacuum tubes), the 1930s (due to political tension in Europe) saw a revived interest in solid state electronics to explore possible military applications for solid state electronics.

Post-war (until ‘70s), the US government expenditure in semiconductor research and manufacture (space-race) was to gain pole-position in arms race vis-à-vis the Soviet Union.

The US military provided an assured market, driving research, development and most importantly production technology.

US military was a significant factor in creating a semiconductor industry within the United States. The US military dictated the switch from thermionic value manufacturing to solid-state device fabrication as emphasized by the then (US) Director of Defense Research on Engineering in 1963 (Bridges, 1964) -

Unlike other countries, the US semiconductor industry has arisen and flourished in conditions which offered little external competition (P, 1990). The Western European countries had to operate in a landscape with an incumbent US, which is comparable to India’s current situation.

Hence European nations’ experiences have relevance to the Indian context. Italy, France, Germany and the United Kingdom were the early entrants trying to establish a sovereign semiconductor industry.

Italy

The Italian semiconductor industry emerged out of the Societa Generale Semiconduttori (SGS) founded in 1957 backed by Olivetti (Business equipment manufacturer) and Telettra (Telecommunication equipment manufacturer) to supply germanium diodes to their parent companies.

In 1961, Fairchild (US-company) negotiated a 1/3rd share in exchange for silicon planar technology. The re-named SGS-Fairchild operating from Agate, even though managed to capture a significant portion of the European market (12 per cent) in 7 years.

Fairchild’s intentions were only to use SGS as a selling conduit to the European market with no plans to invest in R&D or reluctant to expand production facilities culminating in buy-back of Fairchild’s stake by Olivetti in 1968.

In 1971, when Olivetti decided to withdraw IRI (Istituto per la Ricostruzione Industriale) — a state-owned telecommunication holding company — it averted Motorola’s (US company) takeover attempt by merging it with ATES (A Siemen’s subsidiary operating in Catania).

In 1977, the Italian government passed an Act titled ‘The Industrial Restructuring, Rationalization and Development Law’, providing $1.2 billion financial support to the electronic industry — the principal recipient was ‘the electronics component sector being SGS-ATES and Texas Instruments Italia (US company)’.

This policy led to some measure of success by 1981. SGS-ATES became the third largest European semiconductor manufacturer in size. In 1982, SGS-ATES signed an agreement with Toshiba (Japanese company) to transfer CMOS technology. This was the first major agreement of its kind between a Japanese and a European manufacturer.

The late ‘80s saw a decline in its fortunes due to increased competition from American companies. The reasons for this decline have parallels in all other European countries which will be detailed in a common subsequent section. Today, SGS-ATES survive as STMicroelectronics (A French-Italian company).

France

Unlike a single company-dominated Italian semiconductor backdrop, France in the ‘60s, managed to spring up two all-French companies namely Societe Industrielle de Liasions Electriques (SILEC) and the Compagne Generale des Semiconducteurs (COSEM) and three subsidiaries namely Soveur Electronique SA (minority stake by US company Corning Glass), La Radio Technique SA (controlled by Dutch company Philips) and Societe Europeanne des Semiconducteurs (SESCO) — a joint venture of Thomson Houston and the General Electric (US company).

The early indigenous integrated circuits were manufactured primarily for the French military use. An early French rapprochement with American companies like IBM and General-Electric (in collaboration with French company Bull) ensured that indigenous computer industry (primary driver of semiconductor manufacturing) could never take-off and eventually collapsed with GE buying Bull in 1964.

By 1966, facing the American competition, COSEM and SILEC made an agreement to co-operate on research and production. Also, COSEM and SESCO were funded by the French government to collaborate on semiconductor research.

In spite of French government support, indigenous French companies couldn’t break the American strangle hold of the domestic market, which amounted to more than 50 per cent in 1977.

French government intervened with a policy encouraging mergers, forming two entities Thomson-CF and SESCOSEM concentrating on bipolar integrated circuits.

A licensing agreement was made with General Instruments, National Semicon and Motorola to gather American know-how in this field.

In the same year, French government produced a $150 million plan for the development of integrated circuits entitled ‘Le plan circuits integres’. The decision following the plan paved the way for foundations of French microelectronics industry.

The government extended support in 1978 to include the construction of new research center (CEA-LETI) situated at Grenoble to specialize in development and production of integrated circuits.

Today, CEA-LETI is a key research facility providing R&D support to STMicroelectronics. The French government policy on electronics industry was to assist large manufacturing companies and encourage amalgamation. Most Semiconductor firms were in the vicinity of Paris, Grenoble and Provence, forming skill clusters — a key ingredient to sustain an industry.

Germany

The German semiconductor industry was the strongest in Western Europe with the exception that it had virtually no military market to sustain its activities unlike their other European counterparts.

Three vertical integrated companies with wide range of interests in electronics existed in the early 1960s namely Siemens, AEG Telefunken & Bosch.

The government support for semiconductor industry was through the Ministry of Research and Technology of the Federal Republic whose budget consistently increased throughout ‘70s. Even though the domestic German market was dominated by American companies, Siemens continued to invest in the American semiconductor industry to get access to technology and Telefunken obtained semiconductor technology license from US company RCA and set up continuous exchange of information via technical bulletins.

International investment in German semiconductor industry included US companies namely: Texas Instruments in Freisling, Motorola in Unterfahring and Intermetall in Frieburg.

Japanese companies namely: Hitachi in Landshut Bavaria, Toshiba at Braunschweig. Siemens in 1999, spun-off its semiconductor business into Infineon Technologies.

Increasing reliability of automotive technology necessitated Bosch to venture into Semiconductors because buying components that were developed for consumer electronics did not meet the tough standards that an automobile would have to withstand.

The Jetronic electronic gasoline injection system was the first step in 1967 that propelled Bosch in setting-up manufacturing plants in Reutlingen. Apart from common European problems summarized later, a unique German drawback was lack of constant upgrade driven by the attitude Dummer summarizes as follows (Dummer, 1964):

The German automotive industry has kept German semiconductor manufacturing afloat but It still relies heavily on the Asian Fabs for processing i.e. Engine Control Units which form the brains of an automobile.

The United Kingdom

Already in 1950s, a number of firms were manufacturing semiconductors in the United Kingdom, major ones being Plessy, GEC, Ferranti and Mullard (a subsidiary of Philips).

With the exception of Mullard, the overall picture in the ‘50s was that of cautious, inward-looking, resistant to change and slow to respond to new economic conditions.

Even though Harold Wilson under the National Plan gave immediate priority to the computer industry, essential building blocks (i.e. semiconductors) from which computers are made got scant attention.

This statement by then British Prime Minister sums up the government attitudes aptly (Morris, 1994). –

And another factor has been the constant change of bureaucracy with no background in industry or research establishments, as published in the House of Lords Select Committee Report:

Both these statements would find eerie resonance to many India policy watchers.

The common theme of government defense contracts did support the industry but there is an interesting British caveat.

The contracts were awarded on basis of estimated costs plus “incentives” (i.e. cost-plus contracts). The procedure was to place a Production contract for a limited number of components at an agreed price, which being turned into a longer-term Production contract at a later stage.

In the event of a long production run, the system worked strongly in favour of the company with the contract, which is subject to rapidly falling unit costs and rapid technical obsolescence.

This ensured higher return on capital and no incentive to invest, innovate or raise efficiency. Unlike its American counterpart, the British defense industry was too small to finance contracts on a scale as to enable large-scale production runs of devices to be initiated, and to subsidize these devices by providing a market during the early production period when unit costs are high.

Access to American defense industry contracts have, of course, been excluded on grounds of national security. A major source of finance has thus been denied to the British semiconductor industry and this has perhaps been the single largest factor in ensuring the rapid development of the corresponding American industry.

During 1978, Callaghan’s government was instrumental in the setting up a private semiconductor manufacturing company (INMOS). This was a decision at a time when British component manufacturing firms in private sector had withdrawn from highly competitive activity, in the face of overwhelming competition from American semiconductor industry.

Although the idea was conceived during the Labour Government, the incoming Conservative administration continued to support the scheme, somewhat reluctantly for a brief period.

The conservatives led by Thatcher abandoned the previous industrial policy strategy in 1979 and began a long period of industrial disengagement of the state.

The policy of state investment through shared ownership which had been vital in the creation of INMOS and continued existence of Ferranti were brought to an end. INMOS was sold to Thorn and then Thorn passed it on to STMicroelectronics.

The frequent changes in government (McMillan, Heath, Wilson, Callaghan, Thatcher) in the formative years and consequently policy changes couldn’t institute long-term industrial planning and finance, which are vital for the industry to take roots.

Also, the role of the commercial banks in Britain, geared to a 'seed time to harvest' mentality, has ensured that long term loans to industry from this source have not been forthcoming, particularly in the case of the semiconductor industry, as it is subject to long periods of initial financial loss.

Finally, an important factor contributing to European lack of performance in the semiconductor field, and perhaps applying to Britain in particular, has been lack of a successful management and marketing strategy of production engineering.

This has been recognized by overseas competitors and has been the subject of many discussion and analysis. The last of the fabrication facilities at Newport has been acquired by the Chinese-owned Nexperia in July 2021.

Summary

The policy makers face three primary challenges in devising an effective industry policy (Kleinhans & Baisakova, 2020): First, how to ensure access to foreign technology providers through trade and foreign policy? Second, how to establish domestic companies through strategic industrial policy that can be leveraged to hold key positions in the value chain? Third, how to increase the value-chain robustness and make it resilient to supply chain disruption?

European semiconductor industry suffered significant disadvantages in its nascent stage (P, 1990) to overcome these challenges namely:

1) A general lack of cohesion, each government pursuing a separate strategy with no attempt towards coordinated plan (Recently, there has been a renewed attempt to correct this, in 2020 with a joint declaration titled “A European Initiative on Processors and Semiconductor Technologies” and this year with “2030 Digital Compass” to increase semiconductor EU market share to 20 per cent).

EU is also trying woo Intel and/or TSMC to set up shop on the continent.

2) The financial institutions (in varying degrees) were inadequate in providing sufficient financial support to enable a healthy industry to develop.

3) The early foreign penetration of home market and no response to it from respective governments. 4) R&D facilities often underfunded and relatively paltry investment in area of production development. Although, a course correction on EU-wide funding of R&D institutes has been made albeit the early mover advantage of the ‘70s to build a sovereign industry frittered away.

Since most of the Indian policy is rehashed on a western template, these lessons from Europe are policy pitfalls that India needs to avoid.

The next part will detail the journey of the Asian giants (Japan, South Korea & Taiwan) in the semiconductor industry, how they navigated these pitfalls in their foundational years, and continue to succeed in the semiconductor industry through prudent policy.