Without the ability to design and produce custom semiconductors, India will forever be dependent on others. Therefore, the acid test of the NPE 2019 is whether it enables India to gain competence in semiconductors.
The policy on electronics was finalised in February. As someone with a little experience with policy (I co-wrote India’s first Intellectual Property Rights Policy in 2016) and someone who has had to look at the industry from an innovator’s point of view (I ran an incubator for electronics companies till 2017) I am both excited and apprehensive about it.
Why An Industrial Policy In Electronics?
On the one hand, India does not have a great track record of industrial policy success. Often, the bureaucracy has done more to stifle rather than encourage green shoots. A good example of fumbling a promising development is in electric cars. Reva, a tiny electric town car, was viable and available in India way before the Teslas and other Western electric cars and their Chinese competitors. But Reva didn’t become a breakthrough, world-changing concept.
However, this is not unusual, because there are few success stories in policy matters anywhere except perhaps Japan’s MITI in the 1980s. I just read of a former Chinese finance minister saying that ‘Made in China 2025’ is a waste of tax-payers’ money.
On the other hand, it is good that attention is being paid to this industry, finally. I cannot think of any technology more crucial for the nation today than electronics. If I were to be dramatic, I could argue that falling behind in electronics could lead to an existential threat, because strategic competitors will simply be able to shut the country down remotely; they can win a war without firing a single shot by bringing India to its knees through cyber-attacks.
In addition, there are good commercial reasons as well. The policy notes (1.5, preamble) that India’s share in global hardware electronics production is a measly 3 per cent, at Rs 387,525 crore (approximately $59 billion) in 2017-2018. The Indian demand for electronics is expected to rise to Rs 2,600,000 crore (approximately $400 billion) by 2025, entailing a gigantic trade deficit which will be greater than the bill for oil.
In any case, can India afford to depend on the tender mercies of a potential competitor? For instance, we saw how Chinese telecom maker ZTE almost ceased operations when the US imposed an embargo on American components (since lifted). We can see how Huawei is likely to be hurt by an American semiconductor embargo. If India is dependent on perennial foe China for its chips, they could easily impose a non-tariff-barrier on us, as Japan found out when China simply stopped shipping rare-earths to it without warning. Americans are also not all that reliable.
Besides, there is the perennial fear of bugs and trapdoors inserted into imported electronics. There are persistent rumours that Chinese telecom equipment illegally ships data to China (one of the allegations against Huawei). The African Union headquarters in Addis Ababa, it appears, was built by Chinese firms, who were accused of “hacking computer systems every night for five years and downloading confidential data” (‘African Union accuses China of hacking headquarters’, Financial Times, 29 January 2018).
The Building Up Of Core Competence
I use the term ‘electronics’ in a broad sense, to include the related areas of semiconductors and telecommunications. In a classic 1990 Harvard Business Review article, “The Core Competence of the Corporation”, C K Prahalad and Gary Hamel wrote about their study of two firms: GTE from the US and NEC from Japan. In the 1980s, the two were comparable in size (though GTE was 3x in terms of revenues) and in industry focus — telecommunications and electronics.
But by 1990, things had changed: NEC was 50 per cent bigger in revenues, and it had become a world leader in telecommunications, semiconductors and mainframe computers. GTE had stagnated, had become a telecom operating company, exited semiconductors, and divested its switching business. It was destined to become a marginal player.
The difference was in identifying and nurturing core competences. NEC had figured out that the common thread between its domains was semiconductors, and they had invested in it. They realised that there would be convergence between computing and communications, with the core semiconductor competence being leveragable in both. By systematically investing in chip technology, both through joint ventures and its own research and development, NEC was able to become a major player.
In an odd parallel, we are in a similar situation with regards to the US, China and India in terms of the future of technology. It appears that artificial intelligence (AI) is one of the core competences that will determine future success. And it requires algorithms, data and electronics. The US is strong in all three areas, and China lags only in electronics.
India has a potential competence in all three, but it is critically deficient in the underlying semiconductor technology. Without the ability to design and produce custom semiconductors, India will forever be dependent on others. Therefore, the acid test of the NPE 2019 is whether India will gain competence in semiconductors. This is no trivial task, given the fact that building a single fabrication facility costs at least $10 billion, and requires guaranteed power, water and logistics, which India has lacked.
Given that India does not have a single semiconductor fab worth the name, why do I say India has a potential competence there? It’s because there is a design competence, not a manufacturing competence, although in the long run, one cannot be sustained without the other. Indian engineers can design chips that are then shipped out to fabs in Taiwan or South Korea, which make the actual physical chips. This is what major players ARM, Qualcomm, even Apple do.
In the May 2015 issue of Swarajya I wrote about “What ‘Make in India’ really means: it is ‘Design in India’ (and I could have added, ‘make anywhere’) because that is the skill set we have. A fair number of designs have been created in the Bangalore ecosystem originally set up by companies like Texas Instruments and Motorola. But there’s a fly in the ointment: India created a special intellectual property rights (IPR) category of semiconductor designs for creating and registering the intellectual property in them some 10 years ago, and if I am not wrong, only a very few (less than 10) designs have been filed.
The NPE 2019 talks about “develop[ing] core competencies in all the sub-sectors of electronics, including inter alia electronic components, sub-assemblies and semiconductors, telecommunication and broadcasting equipment, IT hardware, medical electronics, defense and strategic electronics, automotive electronics, industrial electronics, consumer electronics, etc., and fabless chip design.” [4.7, Objectives]
Right there I begin to worry. Where is the competence in semiconductor manufacturing, likely to be the crucial element? Besides, this does look like a mish-mash: it has both core areas such as a) discrete components, b) semiconductors, by which I presume the policy is talking about integrated circuits (ICs), and then c) a whole lot of application areas like automotive electronics.
And then fabless chip design is tacked on, sort of as an afterthought.
I think these need to be prioritised. Fabless chip design we can do today. Once we have that, we can focus on how to manufacture the devices. Which application domain or industrial sector these devices belong to is a secondary matter, and should not be confused with the technology and production issues.
Fortunately, there is an entire section, 5.22.1, devoted to the “Promotion of a Fabless Chip Design Industry”. Incidentally, there is a move globally towards open-source chip design, such as the RISC-V instruction set architecture from the University of California, Berkeley. IIT Madras announced that it had produced India’s first indigenously designed microprocessor, ‘Shakti’, in November 2018.
The Shakti chip was fabricated in India at Indian Space Research Organisation’s semiconductor laboratory in Chandigarh. This fab runs on antiquated 180 nanometer technology, while the current state of the art is close to five nanometers, but it is still important to control the entire supply chain of chips, and have a fully integrated process within the country. Such a trusted value chain, section 5.11, is necessary in light of persistent concerns about Chinese hacking of components and boards.
There is recognition of the need for ‘mega projects’, section 5.18, which suggests “special package[s] of incentives” for “semiconductor facilities (including trusted foundries), display fabrication, photonics and LED chip fabrication units, including according infrastructure status to these units”. That is a good beginning, but it doesn’t show sufficient urgency in building indigenous capability.
In fact, section 5.19 does say, reasonably so, that it is appropriate to “promote investment in mega facilities abroad, such as an existing semiconductor facility… where an ecosystem exists for a particular technology”. One cannot quarrel with that as a way to make rapid entry, but it would have been good to acknowledge that technology transfer is a goal, especially for trusted foundries.
An earlier presentation from 2018 from Ministry of Electronics and Information Technology (MeitY) also showed where India has succeeded in reducing imports: mobile phones, set top boxes, monitors, PCs, electric inverters, printers, video game consoles, audio amplifiers, and microwave ovens. This is the result of the earlier NPE 2012 and incentives provided.
However, there is a list of areas where imports have increased (substantially in some cases) between 2015-16 and 2016-17: telecom base stations (227 per cent), routers (78 per cent), telecom equipment parts (39 per cent), solar cells and modules (36 per cent), populated printed circuit boards (118 per cent) and bare PCBs (43 per cent). This is ominous, especially on the telecom side, with the specter of 5G domination by Huawei and other Chinese manufacturers. After AI, 5G is possibly the most important technology down the road, and we really cannot afford to keep outsourcing all that to the Chinese.
Services And Products In Electronics
Electronic system design and manufacturing (ESDM) is a big area, and the stated objective of the NPE 2019 is to “position India as a global hub” (section 2, Vision) with a view to promote a turnover of $400 billion by 2025, including 1 billion mobile handsets, of which 600 million will be exported (Section 4.1, Objectives).
There are several ways of taking this policy directive forward, both in the provision of services and in the creation of end products. An immediate low-hanging fruit is in developing a design competency and offering that as a service worldwide, in analogy with the software services sector. I dwelt on this in the April 2017 issue of Swarajya magazine in “It’s Time We Turned Our Focus Towards High Value Added Emerging Services”, where I suggested that “India’s core competence is design” with “open, standard hardware such as Arduino and Raspberry Pi”.
The first level of proof of concept design is feasible with open hardware, which can enable entrepreneurs to get to rapid prototyping. Once the prototype is accepted, higher-level design services can be offered using electrical CAD, and also mechanical CAD for the enclosures. Thus, Indian design houses can become outsourced design partners in all the emerging areas such as the Internet of Things, 5G, AI/machine learning, augmented reality, drones, robotics, wearables, and gaming.
There are already Indian firms doing this sort of design outsourcing. With the centres of excellence (section 5.7.2, Industry-led R&D and Innovation) and other measures, Indian skills at frugal design may become quite attractive, and electronic design houses may become big business (an analog is firms like frogdesign and IDEO that have made a success out of industrial design). This is where targeted engineering education (section 5.9) can be useful.
The next step up the value chain is in becoming original equipment manufacturers (OEMs). If you have the design skills, build up a supply chain of component manufacturers, generally in Taiwan, China, South Korea and Japan, and have access to an industrial-strength manufacturing line, then it is possible to be a contract manufacturer for large brands. This is precisely how Chinese firms in Shenzhen evolved into a tightly knit ecosystem (even without the design skills per se).
The final step is to become full-fledged product developers, and that is a quantum leap again. It takes a lot of sharp manoeuvring to understand markets, exploit weaknesses in big multinational companies’ strategies, and step in as an alternative. It also requires a good deal of marketing and branding. Chinese firm HTC made this transition in handsets: from contract manufacturer to brand. It was less successful than Lenovo, Asus, and Acer, and other Chinese firms that made the same leap in PCs. Before the Chinese, the Japanese, including Sony, Toshiba successfully created global brands.
That is the eventual prize: full-fledged electronic firms from India that do everything from market need analysis to product concept generation including R&D, prototyping, market validation, manufacturing, market entry, distribution and customer support. At the moment, Indian firms cover only some of the segments of this full value chain. The objective of the NPE 2019 is to allow integrated value chains to arise in India.
It is in the nature of policy documents to be necessarily vague and non-specific. After all, a policy is a guideline to the way ministries may act, but it is no roadmap. Within its limits, this policy does cover a lot of ground, but I am concerned it does not quite capture the urgency of chip fabrication in India.
What is also missing is more than a nod to the concerns of startups and small entrepreneurs already in the field: the availability of testing, prototyping and manufacturing equipment, the ability to get certifications, access to mechanical design for attractive enclosures and manufacturability, access to component supply chains, medium-term funding, and above all, marketing and partnering support.
The entrepreneurs whom I supported in an electronics incubator had access to only about half of these vital elements, and we tried to find industry partners who could license IPR, offer mechanical design, manufacture, distribute and brand products: indeed, for many electronics startups, they would be better off trying to create patentable intellectual property, and not attempt to do the customer-facing sales activity and channel building.
Throughout the policy, there is an emphasis on R&D, upskilling and industry-academia partnerships. These are good gaps to identify, especially the fact that both academic and industry research in India has been far below its potential: we are nowhere close to matching the quality and quantity of what has been achieved in countries with similar gross domestic product. That is the elephant in the room: the lack of leading-edge R&D. The policy can highlight it, but actual rules and actions have to be taken by the concerned ministries, and by past experience, there may continue to be a lacuna.
If India can ramp up its creativity and innovation, and if the NPE 2019 helps create a more appropriate ecosystem, then the critical electronics industry may well be on a better footing. We can only hope that this will happen.