Three semiconductor and equipment companies are emerging as the biggest beneficiaries of an expected surge in artificial intelligence infrastructure spending, even as investors overlook them in favor of more famous names. Goldman Sachs recently forecast that global AI infrastructure capital expenditures could climb to between $920 billion and $1.4 trillion in 2027, up from more than $700 billion expected this year. While chip designers like Nvidia grab headlines, the real money may flow to the companies that manufacture chips, design custom processors, and build the machines that make advanced semiconductors possible.

Which Companies Control the AI Chip Supply Chain?

The semiconductor industry operates in layers, and each layer has dominant players. At the manufacturing level, Taiwan Semiconductor Manufacturing Company (TSMC) holds an almost unshakeable position. The company produces advanced logic chips for virtually every major chip designer, from Nvidia to Broadcom to Alphabet's custom processors. TSMC's advantage is not just scale but also expertise; it is currently the only foundry capable of producing advanced logic chips in massive quantities with high yields, meaning few defects.

This dominance gives TSMC tremendous pricing power. Recent reports indicate the company will raise prices on its newer 3-nanometer chips by 15 percent later this year, a move that underscores its leverage as demand for AI chips intensifies. As more chip companies compete to secure manufacturing capacity, TSMC benefits from both higher volumes and higher margins.

Broadcom, meanwhile, has transformed itself into a diversified powerhouse by acquiring other chipmakers and infrastructure software companies over the past decade. The company now designs custom application-specific integrated circuits (ASICs) for hyperscalers like Google and Amazon. Unlike Nvidia's general-purpose graphics processing units (GPUs), Broadcom's custom chips are optimized for specific AI workloads, allowing customers to handle AI tasks more cost-efficiently at scale.

In fiscal 2025, Broadcom's AI chip sales surged 65 percent to $20 billion, representing 31 percent of its total revenue. The company expects AI chip sales to grow roughly fivefold to over $100 billion by fiscal 2027, accounting for at least 58 percent of its projected $171.5 billion in total revenue. From fiscal 2025 to fiscal 2028, analysts expect Broadcom's revenue and earnings per share to grow at compound annual growth rates of 53 percent and 66 percent, respectively.

What Role Does Equipment Manufacturing Play in the AI Boom?

Behind every advanced chip lies a machine. ASML, a Dutch semiconductor equipment manufacturer, produces the only extreme ultraviolet (EUV) lithography systems in the world. These machines optically etch circuit patterns onto silicon wafers with precision so fine that they enable the creation of the world's smallest, densest, and most power-efficient chips. Without EUV technology, the AI infrastructure boom would not be possible.

ASML's EUV machines cost up to $400 million each and require multiple planes to ship. All of the most advanced chip foundries, including TSMC, Samsung, and Intel, rely on these systems to manufacture cutting-edge chips for fabless chipmakers like Broadcom and Nvidia. This monopoly on critical technology gives ASML extraordinary pricing power and makes it a linchpin of the entire semiconductor market.

The company is experiencing robust demand from both foundries and memory makers. ASML's EUV machines are used not only to manufacture advanced logic chips but also to produce high bandwidth memory (HBM), a critical component in AI accelerators. From 2025 to 2028, analysts expect ASML's revenue and earnings per share to grow at compound annual growth rates of 17 percent and 26 percent, respectively.

How to Understand the Hidden Winners in AI Infrastructure Spending

• Manufacturing Bottleneck: TSMC controls the only foundry with the scale and expertise to produce advanced logic chips in mass quantities, giving it pricing power as demand surges and competitors fight for capacity.
• Custom Chip Advantage: Broadcom's custom ASICs for hyperscalers offer cost-efficiency gains over general-purpose GPUs, with AI chip sales expected to grow from $20 billion to over $100 billion by fiscal 2027.
• Equipment Monopoly: ASML's exclusive control of EUV lithography technology makes it indispensable to every advanced chip manufacturer, positioning it to benefit from sustained demand regardless of which chip designer wins market share.

The broader context matters here. Alphabet, one of the largest AI infrastructure spenders, is investing heavily in its own tensor processing units (TPUs) rather than relying solely on Nvidia's GPUs. This diversification of chip types benefits TSMC, which manufactures all of them. Alphabet's TPUs have become so well regarded that the company now allows select customers like Anthropic to order directly from Broadcom, adding another high-margin revenue stream for Broadcom.

Recent market volatility has tested investor confidence in semiconductor stocks. In early June, tech stocks across Asia experienced a sharp sell-off after Broadcom's fiscal second-quarter revenue missed market estimates. The VanEck Semiconductor Exchange-Traded Fund (SMH) lost over 9 percent in a single day, and TSMC fell 2.96 percent, while ASML initially declined before recovering to close 3.6 percent higher.

However, the underlying fundamentals remain intact. Goldman Sachs' forecast of $920 billion to $1.4 trillion in AI infrastructure spending by 2027 suggests that the demand driving these companies' growth is structural, not cyclical. As long as hyperscalers continue building data centers and training larger AI models, they will need more chips, more manufacturing capacity, and more equipment to produce those chips.

The lesson for investors and observers is that the most profitable companies in a boom are often not the most visible ones. While Nvidia dominates headlines and SpaceX captures investor imagination with record valuations, the companies that manufacture chips, design custom processors, and build the machines that make advanced semiconductors possible may deliver more sustainable returns. TSMC, Broadcom, and ASML occupy positions of structural advantage in the AI supply chain, and their pricing power suggests they will capture significant value as the $1 trillion AI infrastructure spending wave unfolds.