Opinion
There are no shortcuts to leadership in the field of semiconductors
Summary
- Chip fabrication plants are where silicon wafers transform into the semiconductors that power numerous devices. While much of the world focuses on research papers for innovation, advancements in the production process make them a reality. We have to learn by doing.
The semiconductor industry, the bedrock of modern electronics, operates on a scale both microscopic and colossal. Fabs, or semiconductor fabrication plants, are where silicon wafers undergo a complex metamorphosis, transforming into the chips that power our smartphones, laptops, cars and countless other devices.
Each new fab generation, denoted by nano- metre measurements (28nm or 2nm), signifies a leap in technological prowess, enabling smaller, faster and more energy-efficient chips.
The journey from one node to the next is arduous, demands billions of dollars in investment and years of painstaking research and development.
It’s estimated that moving from a mature 28nm node (where India is starting) to the cutting-edge 2nm kind could take upwards of a decade and over $100 billion in current times.
Also read: Tata Electronics plans to build two more chip fabs in Dholera
Few nations possess the technological and financial muscle to embark on this semiconductor odyssey. Yet, India, with its burgeoning tech sector and ambitious leadership, is poised to join this exclusive club.
Prime Minister Narendra Modi’s recent agreements in Singapore, alongside the landmark announcements by Adani and Tower Semiconductor, signal a promising start. These initial steps represent India’s intent to climb the semiconductor Everest, one innovation at a time.
The complexity of process innovations: The reason that fabs are so difficult for all but a handful of corporates is the accumulating nature of the innovations required. Moving from one cutting-edge fab generation to the next involves a massive ecosystem of innovations, research and refinements.
This journey is not just a leap in transistor density, but an intricate process requiring the collaboration of thousands of experts, engineers and corporations.
While much of the world focuses on flashy research papers announcing groundbreaking discoveries, the less celebrated process advancements make transitions possible.
Each company, and each fab, will have hundreds of undocumented unique features, and even tiny differences can lead to two fabs making the same products with differing yields, say, one of 70% and the other 10%.
Also read: TCS working with Tata Electronics to build first made-in-India chips
These process innovations are in the thousands, representing the lessons learnt during the day-to-day development within semiconductor fabs, involving everything from yield management and defect reduction to optimizing supply chains. Together, these innovations enable each new generation of semiconductors.
In other words, it is not just about buying expensive machines, like for extreme ultraviolet (EUV) lithography, although these also represent an accumulation of innovations.
Our engineers will learn about transistor architectures like FinFET or gate-all-around (GAA) and the integration of advanced materials like high-k metal gates in research papers. But, what makes everything work in the fab-building business goes beyond what could be learnt from books or bought with money.
Why you can’t skip ahead in the fab journey: For nations or companies looking to make semiconductors, there is no easy skipping ahead. Jumping straight to advanced nodes without the cumulative knowledge gained from mastering older generations is like trying to run a marathon without training.
Consider the advancements made in transistor architecture. At the 7nm node, the FinFET transistor was the standard, but now, at the 3nm and 2nm nodes, gate-all-around (GAA) architectures are being explored.
These changes require innovations in everything from transistor design to how these components are integrated into chips. The refinement of these technologies took years and thousands of process papers, experiments and real-world testing.
But short-cuts are possible in alliance with industry leaders: While starting with mature technologies is a good strategy, the most effective short-cut to cutting-edge nodes is partnering with industry leaders like TSMC of Taiwan or Samsung Electronics of South Korea.
This is precisely what the US, Germany, and Japan have recognized as they provide incentives in billions and tens of billions of dollars to these giants to set up fabs within their borders.
The evolving geopolitical landscape, particularly India’s warming ties with South Korea and Taiwan, opens new doors for strategic partnerships. Leveraging these relationships, India can leapfrog years of development, gaining access to cutting-edge technologies and expertise.
Also read: Is Adani's semiconductor play a long-term game?
The onus lies on industrialists and the government to seize this moment, aligning their visions and resources to accelerate India’s semiconductor ascent.
In the world where chips are the new oil: One destination is a self-reliant India capable of designing and manufacturing its own chips to fuel its digital economy and bolster its geo-strategic autonomy.
This could also be a way to harness the nation’s top-quality engineering talent, particularly if the technology sector overall witnesses a change in the balance between the hardware and software segments, as this author expects.
We must encourage higher ownership of hardware assets by Indians, and there are many avenues that policymakers and regular investors could explore to that end.
