India wants to make chips, but the US companies showing up aren't the ones doing the manufacturing.
The Summary
- Modi opened a major semiconductor conference in New Delhi, part of India's $30 billion push to build a domestic chip industry
- Applied Materials and Lam Research pledged billions in commitments, but these are equipment makers, not fab operators
- India is late to chipmaking and betting on subsidies to close a gap that Taiwan, Korea, and China spent decades opening
The Signal
Modi's semiconductor conference drew the right people to the room, but not necessarily the right commitments. Applied Materials and Lam Research make the machines that make the chips. They sell tools. If India builds fabs, these companies win regardless of whether those fabs ever ship competitive product. That's different from TSMC or Samsung committing to actual production capacity on Indian soil.
The $30 billion figure is India's total bet, subsidies included, not the pledges from US firms. India is trying to do in a decade what took Taiwan 40 years: build a semiconductor ecosystem from scratch. Taiwan started with packaging in the 1980s, moved to wafer fabrication in the 1990s, and became the world's chip factory by the 2010s. India is skipping steps.
"Equipment makers pledging billions sounds like momentum until you realize they're betting on subsidies, not on India's ability to run competitive fabs."
The timing matters. The US is reshoring chips through the CHIPS Act. Europe has its own semiconductor sovereignty plan. China is pouring money into self-sufficiency despite sanctions. India is entering a market where every major economy is already spending heavily to secure supply chains. The question isn't whether India can attract equipment makers—that's easy when you're writing checks—but whether it can attract the engineers, build the institutional knowledge, and maintain the capital discipline to actually compete.
Key gaps India faces:
- No deep bench of fab engineers with 10+ years of process experience
- No established supply chain for the 200+ inputs a modern fab needs
- No track record of execution at the scale required for sub-7nm production
The conference is part of Modi's broader pitch to make India a manufacturing hub, not just for services and software. But semiconductors aren't textiles or smartphones. You can't iterate your way to a working 5nm process. The capital requirements are staggering, the margins are thin until you reach massive scale, and the technology moves faster than most governments can write checks.
Applied Materials and Lam Research are making rational bets. India will build something, and when it does, it will need their equipment. Whether India builds fabs that matter in 2030—fabs that produce chips the world actually needs—is a separate question entirely.
The Implication
Watch where the actual fabrication commitments land, not the equipment deals. If TSMC, Samsung, or Intel announce production fabs in India with real capacity targets, that's signal. Until then, this is a subsidy play dressed up as industrial policy.
For anyone building in the agent economy, this matters in two ways. First, chip supply chains are infrastructure for AI. Where chips get made determines who controls the physical layer of the Fourth Web. Second, if India succeeds even partially, it shifts geopolitical leverage in AI compute. If it fails, it's a reminder that some industries have moats that money alone can't fill.