Company Outsider is a weekly newsletter by Sundeep Khanna. Subscribe to Mint's newsletters to get them directly in your email inbox.
Erstwhile semiconductor giant Intel has shed 23% of its global workforce over the past two years. That’s nearly 30,000 people laid off despite receiving billions in government subsidies to boost domestic manufacturing. When CEO Lip-Bu Tan admitted that his company is "not in the top 10 semiconductor companies" by valuation, he was acknowledging what the market already knows: the company that once defined the industry now holds less than 1% of the discrete GPU market, which is the very architecture powering the AI revolution, compared to Nvidia's 92%.
The brutal lesson? Throwing money at semiconductors cannot solve what is fundamentally a technology and execution problem. For India, currently committing over ₹1.6 trillion to its semiconductor mission, Intel's downfall offers uncomfortable truths.
Intel's decline becomes particularly instructive when contrasted with Nvidia's rise. Nvidia didn't stumble into AI chip dominance; it built toward it for over a decade before the generative AI boom. While Intel maximized quarterly profits from its CPU monopoly, Nvidia took a massive risk by investing in specialized software and chips that seemed like a niche side project at the time, but eventually became the foundation for the AI revolution. Its founder, Jensen Huang, famously called it a “zero-billion-dollar opportunity”, since there was no telling when its commercial value would be realized. By the time ChatGPT created explosive demand for AI training chips in late 2022, Nvidia had spent 15 years developing exactly the technology the market needed. Meanwhile, Intel was busy protecting its legacy business.
This contrast exposes the flaw in India's semiconductor strategy. The Production Linked Incentive (PLI) scheme promises to cover up to 50% of fabrication costs, subsidizing construction as if buildings alone create competitiveness. However, the Tata-PSMC (Powerchip Semiconductor Manufacturing Corporation) facility targets 28 nm nodes, a workhorse technology which Taiwan Semiconductor Manufacturing Company (TSMC) mastered in 2011. While these chips are essential for local electronics, the reality is that by 2027, when India reaches volume production, the global frontier will have shifted to 1.6 nm. India isn't just starting late; the gap is widening even as it builds.
Intel's workforce reduction reveals why hiring alone cannot solve this. Despite thousands of experienced engineers, Intel consistently trails TSMC in yield rates and process reliability. Industry veteran Burn J. Lin identifies the core issue: in modern semiconductor manufacturing, knowing the technology is distinct from successfully producing it. TSMC's advantage lies in an "organizational muscle memory" that comes only from decades of iterating at scale.
India's challenge mirrors Intel's, but is significantly steeper. Intel at least possessed decades of institutional knowledge; India lacks the ecosystem of specialized suppliers, equipment manufacturers, and experienced technicians that Taiwan cultivated over forty years.
For that, it has a lack of foresight to blame. In the mid-1990s, after Vinod Dham left Intel having led the successful Pentium processor development, he made frequent trips to India. At that moment, the country should have made him an offer he couldn't refuse: to lead India's semiconductor program. Taiwan did exactly that with Morris Chang in 1987, creating the world-beating TSMC. India's failure to envision a future around a pioneer like Dham condemned it to the backwaters of the semiconductor universe. Even five years ago, as the criticality of chips became obvious, India should have made a bold, all-costs-covered offer to TSMC to establish a local plant. Instead, India reached out in late 2021, after TSMC had already committed to the US, Japan, and Germany. It was too little, too late.
Having lost that race, India must pick its next moves carefully. It already hosts 280 academic semiconductor institutes and 72 chip design startups. Design is where Indian talent has demonstrated global competitiveness. While manufacturing investments are necessary, given the country’s import intensity, they shouldn't come at the expense of looking ahead.
Ultimately, India is pursuing a settled market while the industry’s winners move toward "zero-billion-dollar" frontiers. By subsidizing the 28 nm nodes of 2011, the PLI scheme treats semiconductors like a 20th-century commodity rather than a high-conviction technology bet. If India continues to prioritize filling the gaps of the past over creating the markets of the future, it will remain a technological follower, discovering that no amount of capital can close a gap that widens by the second.