In recent years, Taiwan vaunted semiconductor industry has stepped into the limelight, as policy wonks discuss lessons derived from its technological prowess, artificial intelligence, and now even export control regulations. Lastly, no debate on de-risking and onshoring is complete without examining the role that Taiwanese electronics and semiconductor companies play in the ongoing recalibration of global supply chains.
At the same time, India has been keen to position itself as an alternative for companies that are pursuing a “China plus one” strategy. In light of this, India is increasingly seeking investment from Taiwanese companies to offset Chinese influence on its own industry. This article examines what India could do to enhance its engagement with Taiwan’s semiconductor industry—and why Taiwan should look favorably to India as a destination for investment. It outlines specific focus areas where India could elicit further investment from Taiwan, and offers recommendations for how to boost economic cooperation between the two nations.
The ideas presented in this article reflect discussions held during a virtual discussion hosted by Carnegie India in March 2024 on economic relations between Taiwan and India. The themes and suggestions presented in this article also take stock of both countries’ industrial policy frameworks and how to boost bilateral economic cooperation accordingly.
Lessons From Taiwan
Before delving into the challenges that Taiwanese enterprises face in India, it would be apt to look at the factors that led to these enterprises flourishing within Taiwan. This is relevant because better understanding of the local trends that enabled the rise of Taiwan’s tech industry could well provide indicators of what they may be looking for when considering investments in India. The trends observed here can be reduced to two key factors.
The first key factor was a shift in technology and business models beginning in the 1980s—which, in the case of the semiconductor industry, saw increasing heavy capital expenditure by semiconductor firms. Companies like Taiwan Superconductor Manufacturing Company (TSMC) capitalized on what they perceived as an opportunity to move toward a “pure-play” model in this business environment. That meant they would only produce bespoke chips for a wide variety of customers, be it fabless chip design firms or, more recently, even integrated device manufacturers like Intel.
The second key factor was the availability of a talented workforce, mostly engineers, many of whom had studied abroad and gained experience before returning to Taiwan. There are clear lessons here for India, as it attempts to grow its nascent indigenous semiconductor industry into a self-sustaining ecosystem. Toward this end, the Indian government introduced a semiconductor incentive scheme in December 2021, which has seen uptake from certain Taiwanese companies such as Powerchip Semiconductor Manufacturing Company (PSMC). While there has been no comparable shift in terms of novel business models, as was the case in Taiwan in the 1980s, India is attempting to focus on onshoring more “mature node” chipmaking, owing to a better domestic market for goods that may use such chips. Additionally, as part of its focus on upgrading workforce training, the Indian Ministry of Electronics and Information Technology (MeitY) has also introduced parallel schemes to train approximately 85,000 engineers in semiconductor design under its Chips-to-Startup (C2S) program.
India’s Areas of Focus
During the March 2024 Carnegie India roundtable discussion, participants pointed out that the creation of a sustainable semiconductor ecosystem for India may require looking at four key elements: (1) consistent government support; (2) strategic positioning of the domestic semiconductor industry when targeting global semiconductor firms; (3) strategic spending of taxpayer money; and (4) the need to set up more industrial/science parks for overseas entrepreneurs, especially those from Taiwan.
First, roundtable participants highlighted that steady and enduring government support for the semiconductor industry is required, regardless of political fault lines and diverging ideologies. Here, there should be no vacillating when it comes to making a push for incentivizing the semiconductor industry.
Second, the strategic positioning of the industry is also significant. Countries like India should consider their strategy very carefully, and focus on specific segments of the value chain. Here, a key takeaway from the roundtable was that the Indian government has done well to recognize the potential of mature node/legacy chip production, not just advanced nodes. The adjustment of the semiconductor subsidy framework in 2022 to provide equal subsidies to both mature nodes and advanced nodes showed that the Indian government was listening and iterating accordingly.
Third, it was also observed during the discussion that the Indian government should be careful how it spends the $10 billion pool of subsidies under the December 2021 semiconductor incentive scheme. The semiconductor industry is very capital intensive already, and a spending spree should be avoided. The announcement of a new Bharat Semiconductor Research Center (BSRC) may help in making strategic bets. The BSRC is intended to be a public-private-academic partnership that will seek to incubate next-generation semiconductor companies by focusing on specific segments including advanced packaging, compound semiconductors, and chip design. The MeitY minister of state even announced that the BSRC would “be the Indian equivalent of ITRI [Taiwan’s Industrial Technology Research Institute].” With a proposed appropriation of approximately $2.5 billion, this is a welcome and necessary institution. India’s relative absence from various stages of the semiconductor value chain means that it could face a market failure problem—left to its own devices, the market has not felt the need to invest in long-term, deep-tech indigenous ecosystem for semiconductors. Accordingly, limited government intervention may be welcome in terms of looking at the specific segments highlighted by the BSRC.
Fourth, it was explained during the discussion that while subsidies are important, the creation of industrial/science parks will also play a key role. Here, participants pointed out that industrial/science parks provide the necessary infrastructure and co-locate companies together, allowing for collaboration and the development of systems and products. This may be especially critical for the semiconductor industry, which is heavily siloed. Technical R&D projects often happen in isolation, without an understanding of what other parts of the value chain may need. In this light, participants noted that Taiwan’s strong design capability and relentless focus on manufacturing had created a positive feedback loop of further manufacturing, leading to innovation and growth.
These suggestions represent clear-eyed assessments by experts on what may be needed for India to emerge as a destination for Taiwanese semiconductor and electronics companies. However, some of the suggestions may need further consideration in the Indian context, especially with respect to bipartisan political buy-in to the semiconductor scheme. Academics as well as a former central bank governor have questioned the validity of India’s semiconductor scheme. Their reasoning ranges from the scheme not necessarily being the most “cost-effective” way to bolster the semiconductor ecosystem—since India does not enjoy a head start in fabrication—to why India’s focus should instead be on enhancing R&D spending. It appears, at least to the author of this article, that the semiconductor incentive scheme’s focus is not solely on onshoring foundries. Subsidies are also provided for outsourced semiconductor assembly and test (OSAT) and assembly, test, and packing (ATP) plants. In addition, the C2S scheme also focuses on training the Indian workforce.
Other criticisms, such as how the Indian government should focus on “sophisticated logic chips” rather than memory chips, are not entirely clear. For instance, a press release from India’s Tata Group about the Tata-PSMC plant says that they will also manufacture “high-performance computing logic” chips. But memory chips also present a clear market opportunity. According to the Semiconductor Industry Association, global market demand for memory chips is on par with logic chips—around $154 billion. Furthermore, the chip industry is upstream from the market for consumer electronics and other products that use these chips. Memory chip–powered devices enjoy robust demand in India—reflecting the market dynamics that show a greater demand for “higher capacity and faster storage” solutions. Why should the Indian government choose to ignore this trend and prioritize only logic chips, when there is still decent demand for memory chips?
Why India, Though?
Of course, there are larger questions surrounding India’s viability as a destination for Taiwanese semiconductor firms. These can be segmented into three categories. First, where does India figure in the larger scheme of the semiconductor ecosystem? For instance, what are the benefits for India and the world from Taiwanese semiconductor firms diversifying to India? How would this compare to the onshoring already taking place in the EU and the United States by other semiconductor firms? Second, how would the semiconductor scheme alleviate India’s unemployment issues? This is a massive challenge that India is currently facing. And third, can India provide semiconductor firms with the necessary resources—be it physical infrastructure or engineering talent—to sustain its semiconductor environment beyond the deals that have already been announced? Let’s address these questions one by one.
Taiwanese semiconductor firms, such as TSMC, have already made investments in both the United States and Germany. However, these investments are all in advanced fabrication plants. For instance, all of TSMC’s three investments (totaling approximately $65 billion) made since 2020 in the United States have focused on advanced semiconductor manufacturing fabrication. Similarly, TSMC’s co-investment in Germany has focused on “specialty technology fab.” Furthermore, TSMC’s majority-owned subsidiary, Japan Advanced Semiconductor Manufacturing (JASM), has announced plans to focus on creating 6 nanometer (nm) and 7 nm process technologies for automotive, industrial, and high performance computing–related applications, although it will continue to create mature node process technologies up to 40 nm as well. The common thread running through all these plants is that they are advanced fabrication facilities.
It is worth noting that India is not yet truly competing in advanced fabrication facilities. The Indian government’s professed strategy is to encourage mature nodes of more than 40 nm, where it has identified more potential for growth. Here, overseas semiconductor firms have found potential in, and also a way to stay connected to, the Indian market. For instance, while not much information is available in the public domain regarding the PSMC-Tata joint venture that invested in setting up India’s first commercial fabrication plant, it seems likely that PSMC may have opted solely for technology transfer to the joint venture rather than investing vast sums of capital, given their past pronouncements on this issue. In the meantime, the Tata-PSMC venture has reportedly secured orders from Tesla for its worldwide operations. Again, this venture will look at mature node technologies in the 28 nm, 40 nm, 55 nm, 90 nm, and 110 nm range. This hints at the possibility that India has chosen to compete (for now) at a different stage of the supply chain for semiconductors, where innovative ways have been found to address the massive sums still required to set up mature node fabrication plants.
The introduction of production-linked incentive (PLI) schemes—which provide financial incentives based on the incremental sales achieved over the overall base of the previous year at a rate of 4 to 6 percent—has also seen a healthy uptake in certain sectors. Many assembly operations, especially for electronic devices, having moved to India. This has led to a growing assortment of contract manufacturers moving to India, many of them Chinese. Since a lot of these assembled devices also use chips, imports of semiconductors have surged in India by 92 percent over the last three years. The majority of these are chips of Chinese origin, raising concerns in India over cybersecurity.
Accordingly, India may wish to procure chips locally given its rising import bill, as well as the cybersecurity implications of importing Chinese-origin chips. For example, the Indian government has recently issued a series of notifications aimed at cyber-auditing hardware related to Internet of Things devices and CCTVs—areas dominated by Chinese firms such as Hikvision. Specifically, an amendment to India’s Electronics and Information Technology Goods Order of 2021 required that there be on-chip debugging services to provide for appropriate protection mechanisms. Since India is increasingly exporting a large number of mobile phones (approximately a fourfold increase over the last four years), any cost savings by contract manufacturing firms when it comes to their assembly operations could be passed on to the original equipment manufacturers.
India’s semiconductor policies have also led to a shift in momentum on job creation. In February 2024, Tata Electronics announced that its OSAT facility in the Indian state of Assam would lead to the creation of 27,000 jobs. There are other OSAT plants that have also been announced. It should be noted that India’s IT sector, while productive, has not created many jobs. A new labor code was created at the national level to help encourage new factory openings, but the code is still perceived as being restrictive. Therefore, trying to kickstart manufacturing jobs should not be ruled out. Finally, the onus for job creation should ideally not be on the semiconductor sector alone. Other sectors would have to be transformed as well. Agriculture will be a key focus, although the recent Economic Survey of India expects that agriculture’s share of the Indian workforce will decline from the current 45.8 percent to 25 percent by 2047, with the workforce leaving agriculture gradually being added to the rise in the workforce.
Regarding the question of whether India has the physical infrastructure and engineering talent to build a sustainable ecosystem in semiconductors, let’s consider the case of Malaysia—which is arguably better positioned than India. A lot of things may appear to be working for Malaysia when it comes to semiconductors. It recently announced a $5.3 billion semiconductor scheme. An existing semiconductor ecosystem has been in existence since 1972, when Intel set up a semiconductor production facility in Penang. It is also one of the leading countries when it comes to OSAT capabilities. Finally, Malaysia’s inclusion in the Regional Comprehensive Economic Partnership (RCEP) is an incredible boost to its viability as a destination for countries keen on a “China plus one” strategy, as both countries are part of the RCEP.
However, India still manages to compare favorably with Malaysia despite the latter’s perceived advantages. First, a lot of the companies that have relocated to Malaysia in the wake of the U.S.-China trade war are Chinese firms that do chip packaging and are contract manufacturers for other global semiconductor manufacturing equipment firms. There are inherent risks involved with using such firms, ranging from export control sanctions to cybersecurity issues, as security backdoors are largely created during the chip packaging process. Furthermore, as the U.S. CHIPS Program Office (CPO) considered guardrails regarding the use of the funds disbursed under the 2022 CHIPS and Science Act, “advanced packaging” also came under review. Indeed, the September 2023 CPO guardrails clarified that “semiconductors utilizing advanced 3D integration packaging” will not be permitted by “foreign entities of concern,” which would include persons as well. This growing aversion to Chinese-origin products could play in India’s favor.
Second, Malaysia has also shown a reluctance to remove sanctioned Chinese firms like Huawei from their local 5G networks, which could impact Malaysia’s reputation as a secure place to package, design, and manufacture chips.
Third, despite being an OSAT destination for decades, Malaysia has not migrated to the higher parts of the semiconductor value chain, as it lacks the skilled workforce necessary to service a chip design ecosystem or advanced packaging. India compares favorably here, too. It is home to approximately 25 percent of the global chip design workforce, which is now increasingly being trained in “digital twinning”—essentially the virtual representation of a semiconductor ecosystem. Firms like LAM Research have announced the creation of a “virtual nano-fabrication environment” in India—something that may become increasingly significant as the complexity of making integrated circuits becomes more pronounced. Practices like digital twinning can also run a massive number of parallel experiments and collect relevant data points to refine the chipmaking process. Furthermore, as most industries move toward the integration of hardware and software, India’s prowess in the software sector could play well. For instance, India is witnessing the growth of an impressive drone industry, which works not just on the basis of sound hardware specifications but also with AI algorithms that can sync with the hardware.
Fourth, Malaysia’s inclusion in RCEP may not work out as proposed, since it could be used as a conduit to route Chinese goods to other countries that have a free trade agreement with Malaysia. The leaky rules of origin in most free trade agreement don’t do Malaysia any favors.
Despite India’s late arrival to different stages of the semiconductor supply chain, the strides made over the last few years have been impressive—if modestly so—when compared to incumbents like Malaysia.
A Few Downstream Prescriptions
Based on the expert analysis and suggestions emerging from the roundtable discussion, a few direct policy prescriptions may be considered when looking at the downstream implications.
Investment Tax Credits in India
The semiconductor incentive scheme aims to provide upfront subsidies to proposals that are approved by the Indian government. However, the framework of financial incentives offered under the December 2021 semiconductor incentive scheme is largely missing investment tax credits (ITCs). These are essentially portions or percentages that can be offset by a firm from its tax liabilities. For instance, the U.S. CHIPS and Science Act offers an ITC equivalent to 25 percent of the qualified investment for the taxable year with respect to any advanced manufacturing facility of an eligible taxpayer. Companies are believed to usually prefer ITCs over grants or subsidies, because government grants require the government to direct the research area where the grant can be used. On the other hand, ITCs can be claimed by companies immediately and “interfere less with market mechanisms,” since the company itself decides the direction of R&D. In Taiwan, for example, it is very common to see policymakers and Taiwanese companies rely largely on ITCs. While establishing a direct linkage between innovation and such incentives is difficult, evidence by academics shows that tax credits “positively influence R&D spending” of firms. For ITCs to be truly effective, however, depends on whether the firm was engaged in this activity already.
Utilizing Taiwan’s Industrial Innovation Statute
In January 2023, Taiwan enacted its version of the U.S. CHIPS and Science Act: the Statute for Industrial Innovation. The statute seeks to attract competitive industry players in its semiconductor ecosystem. Interestingly, Article 21 of this law provides that central governing bodies are capable of extending support and guidance for industries aiming to undertake overseas investments or engage in global technological collaborations. The law also provides that any Taiwanese companies looking to make overseas investments are required to seek approval from the relevant authority, depending on the scale of the investment. Perhaps, given the role of the Taiwanese government in shepherding investments to other countries, India’s MeitY may consider frequent government-to-government discussions with Taiwan on what more India could do to secure investments from the Taiwanese semiconductor industry.
Conclusion
The Indian government looks set to continue building up the country’s semiconductor ecosystem. The semiconductor incentive scheme’s upfront provision of financial support and the growing downstream market may well act as accelerators for the government’s goals. One way to capitalize on the growing interest around India’s semiconductor scheme could be to think about leveraging new trends in semiconductor technology, such as advanced packaging or making chips for power devices. Currently, there is a lack of dominant players in these markets, providing ample opportunity for new entrants. Taiwan’s considerable expertise is especially valuable and ought to be utilized as Taiwan and India think about the future of their economic partnership.