Applied Materials Bets $500M on Singapore as AI Chip Demand Redraws Asia's Supply Map

Applied Materials, the largest maker of chipmaking equipment in the United States, is putting $500 million into a new manufacturing campus in Singapore, and the number behind that number tells the real story. The company plans to nearly double its global production capacity, and President and CEO Gary Dickerson told Nikkei Asia the chip business will grow more than 30% this year, with similar growth expected in the years that follow.

That is not a company hedging. That is a company building ahead of demand it believes is structural rather than cyclical.

What Applied Materials is actually doing

The Singapore campus will expand from one facility to three. When the buildout is complete, the site will house nearly 50% of Applied Materials' total production capacity, turning a Southeast Asian city-state into the operational center of gravity for a company headquartered in Santa Clara, California. The expansion comes with roughly 1,000 new hires in Southeast Asia, concentrated in Singapore.

For context, Applied Materials sells the tools that fabs from TSMC to Samsung use to etch, deposit, and inspect the layers that make up a modern chip. When AI demand pushes foundries to add wafer capacity, the equipment order book is the leading indicator. Dickerson's 30% growth figure is a read on how much new fab capacity his customers intend to bring online, and the Singapore investment is his way of making sure he can build the machines fast enough to fill those orders.

Locating that capacity in Singapore rather than expanding at home is a deliberate supply chain decision. The city offers proximity to the Asian foundry base, a deep pool of precision manufacturing labor, and a politically stable jurisdiction at a moment when chip supply chains are being pulled in opposite directions by US export controls and Chinese self-sufficiency drives. Putting half your capacity in one place is a concentration bet, and Applied Materials is making it on Singapore.

The demand engine sits in Taiwan

The scale of money moving through the AI hardware chain came into sharp focus at Computex and Nvidia's GTC Taipei this month. Nvidia CEO Jensen Huang said his company is spending $150 billion a year on Taiwan's supply chain, a figure that on its own reorders how you think about where value accrues in the AI buildout. Huang also used the Taipei stage to push Nvidia's Vera CPUs, a direct challenge to Intel and AMD in the server processor market, and to show off co-packaged optics switch technology that swaps copper and pluggable transceivers for optical connections inside the rack.

Each of those moves expands the addressable market for upstream equipment suppliers. More CPUs, more optical components, and denser systems all translate into more advanced packaging and more wafer starts, which is precisely the demand Applied Materials is building Singapore capacity to serve.

SK Hynix raises the memory stakes

The memory side of the chain is escalating just as fast. SK Group Chairman Chey Tae-won, making his first public appearance in Taiwan in years, told Nikkei Asia that SK Hynix aims to triple its wafer capacity by 2034. That came barely a week after he said the company would double capacity. SK Hynix is the key supplier of high-bandwidth memory for Nvidia's accelerators, a position that has carried it to record profits and a trillion-dollar valuation.

Chey's willingness to commit to a tripling on a roughly eight-year horizon is a statement about how durable he expects AI infrastructure spending to be. He did acknowledge that more stock market volatility is likely, and SK Hynix shares, like Samsung's, were hit by recent selling. He also floated the idea of an "economic community" between South Korea and Japan, a sign that the chip race is reshaping regional diplomacy alongside corporate strategy.

Japan rewires the subsea cable model

The AI boom is reaching infrastructure that sits far below the data center. Japanese companies are pushing for a larger role in Asia's subsea cable market, where demand is transforming how these systems get built and owned. Construction costs are roughly doubling, timelines are stretching under regulatory and geopolitical pressure, and the traditional large consortium model is straining.

NTT Data Group, one of the world's leading undersea cable builders, has responded by forming a dedicated venture with a small group of Japanese partners rather than a sprawling consortium to construct the Intra-Asia Marine cable. The system links Japan and South Korea with Malaysia and Singapore, with branches to Taiwan and the Philippines. The leaner structure is meant to move faster, and it positions Japan as a hub for the digital infrastructure carrying AI-era traffic. The strategic logic mirrors what Applied Materials is doing on land: when demand accelerates, control over capacity and a streamlined build process become competitive weapons.

The market being left behind

The clearest evidence that AI is sorting winners from losers comes from where the money is not going. India's stock market has lost ground to Asian rivals this year because investors chasing AI cannot find local companies positioned to capitalize on it. Foreign investors pulled a net $30 billion from Indian stocks this year, including a record $2.3 billion on June 1 alone, according to securities depository NSDL. The market slipped to seventh place globally by value, down from fifth just a week earlier.

The value of Indian equities held by foreigners fell to a 10-year low of 6.9 trillion rupees, or about $72 billion, on June 9. Taiwan and then South Korea overtook India's market capitalization. Eighteen months ago, Indian stocks were worth more than double Taiwan's and roughly 3.5 times South Korea's, a lead that analysts at Bernstein say evaporated within five months of 2026. "If you're trying to play AI as a theme, there is literally nothing to play in India," said Arvind Chari of Quantum Advisors India.

Foreign investors have also cut exposure to India's IT services sector by more than a third since January, wary of how AI could disrupt the country's largest white-collar employer. That is the flip side of the Applied Materials story. The same force adding capacity in Singapore, profits in Icheon, and cables under the Pacific is draining capital from economies that build software for humans rather than hardware for machines.

What it means

Put the pieces together and a pattern emerges. Capital, capacity, and strategic attention are consolidating into a tight geographic band running from Taiwan through South Korea, Japan, and Singapore, the places that make the chips, the memory, the equipment, and the cables that AI runs on. Applied Materials concentrating half its capacity in one city, NTT shrinking its consortium to move faster, and SK Hynix committing to a tripling are all responses to the same signal: demand strong enough to justify multi-year, multi-billion-dollar commitments to physical infrastructure.

The risk in that consolidation is the obvious one. Concentration cuts both ways, and a company with half its capacity in a single location or an economy levered to one demand theme is exposed if the AI spending curve flattens. Chey's warning about volatility and the recent selling in memory names are reminders that the market is pricing this buildout in real time. For now, the builders are betting the curve keeps climbing, and they are spending accordingly.