Guest Column – Semiconductor Industry in Turkey: Key Lessons

Birender Singh

Birender Singh

The Turkish semiconductor sector has, regrettably, not developed as much as it could have over the past thirty years. There are a couple of reasons for this comparative underdevelopment vis-a-vis other sectors, such as construction, where Turkish companies have now the second highest volume of international portfolio of projects after Chinese companies.

One reason has been the relative economic instability the country has experienced since the 1970s to early 2000s. Very high levels of inflation with intermittent financial crises (in 1979, 1994, 1998 and 2001) have restricted the scope of risk-taking behaviour required for high-technology investments.

Another reason was the mindset in the private sector as a legacy from the time of the establishment of the Republic of Turkey in the 1920s where the leadership for industrialization was provided by the state. In fact, most of the private sector companies’ senior leadership was provided by so-called “State Economic Enterprises”. During the early 1970s, more than the half Turkish industrial production output came from these enterprises.

The prevailing “import substitution” industrial policies where industrial investments were primarily aimed at reducing and replacing the country’s largest import items, but de-emphasized exports to global markets, also contributed to the underdevelopment of the electronics sector, in general, and the semiconductor sector, in particular.

Nevertheless, the Turkish government foresaw the need for a domestic semiconductor fab in early 1970s, during the heydays of the state-planned industrialization drive. In the early 1980s, a technology transfer agreement was reached with US company EXAR where it would licence its bipolar transistor fabrication technology to state-owned company, TESTAS. An R&D lab for this company would be located in another location close to Istanbul, called “semiconductor technology research lab (YITAL)”, co-located at a government-sponsored research institute complex. TESTAS has never really reached full-scale production milestones due to the changing priorities of the government and inadequate levels of funding.

On the other hand, YITAL managed the transition from the bipolar fabrication technology transferred from EXAR to 3um MOS technology in 1988, using US$ 600,000 funding from NATO’s Science for Stability program. Another round of funding provided by NATO and some larger Turkish electronics manufacturers amounting in total to US$ 1 million in 1994 allowed the upgrade of YITAL’s CMOS technology to1.5um. Over the following years, further investments in this foundry reduced the feature size to 0.7um by 2005.

YITAL could not compete with the global semiconductor fabrication giants due to its small scale and the comparative lag in its technology. However, it found a niche for itself in fabrication of the military-grade cryptology chips for the Turkish armed forces. Such chips required a complete national solution from the circuit design to the fabrication due to obvious security implications in transferring designs to foundries in foreign countries. Thus, a separate research institute established in the same location as YITAL provided military-grade crypto algorithms and chip design, while YITAL fabricated the chip designs.

The success of this niche production strategy has led to further investments in YITAL by the central government. The difficulty in procurement of the RFICs and MMICs for military communications from Turkey’s nominal allies such as the United States, France and Germany has led policy makers to finance a leap in the technological capacity of YITAL to 0.25um SiGe BiCMOS technology.

Similarly, TESTAS’s facilites in Ankara were transferred to the Middle East Technical University in Ankara in 1998 to transform it to a microelectro-mechanical systems (MEMS) research fab. Over the recent years, more than US$ 60 million was spent to upgrade the facilities 0.35um feature size.

In summary, Turkey has decidedly failed in attracting any significant investments from global semiconductor IDM or foundry companies over the last three decades. At the same time, the attempts in establishing domestic foundry capacity were first hindered by the lack of investment capital and then the size of the domestic market. At the moment, its operations are restricted to niche markets that require totally indigenous solutions…