AI Infrastructure Giants: A Look at Nvidia and ASML

The AI Revolution: A Look at Infrastructure Leaders

The artificial intelligence (AI) revolution has taken the world by storm and become a key catalyst for growth among some of the most valuable enterprises globally. According to a forecast by analysts at McKinsey & Company, global capital expenditures on data center infrastructure, excluding IT hardware, will surpass $1.7 trillion by 2030. These investments will be primarily driven by advances in AI technologies, and increasing demand for high-performance and edge computing.

Nvidia: The Backbone of AI Infrastructure

Nvidia has solidified its position as a crucial semiconductor player by providing key technology to underpin the world's AI infrastructure. The company's latest earnings results highlight its business momentum. In its fiscal 2026 second quarter, which ended July 27, revenues jumped by 56% year over year to $46.7 billion while net income surged by 59% to $26.4 billion. Record sales for its Blackwell GPUs were the primary driver of this growth.

Blackwell and NVLink 72: Cutting-Edge Technologies

The new Blackwell Ultra GB300 (the most advanced configuration of the Blackwell platform) is in high demand and has already generated tens of billions of dollars in revenue. Major cloud service providers are easily transitioning from the previous top-end GB200 Blackwell system to the new GB300 rack-based system without significant retooling, as both systems share the same architecture, software, and physical footprint. Nvidia is building 1,000 GB300 racks per week and expects to accelerate its output in the third quarter. The company anticipates full market availability of GB300 systems before the end of calendar 2025. Nvidia has also introduced NVLink 72, a high-speed connection technology, in its GB300 systems. By enabling low-latency and high-speed networking, NVLink 72 makes a complete rack (comprising several servers, networking solutions, and power systems) behave like a single computer. Compared to the previous Hopper architecture systems, GB300 NVLink 72 systems can generate 10 times the output using the same amount of energy. This should ensure strong demand for them from data centers, where energy efficiency is becoming a key operational challenge.

CUDA: A Strategic Competitive Advantage

Beyond its hardware and networking, Nvidia's proprietary Compute Unified Device Architecture (CUDA) parallel computing software platform has become a strategic moat for the company. CUDA has established itself as the preferred platform for AI training, inference (real-time model deployment), and development, and because it's only directly compatible with Nvidia hardware, that makes it difficult for clients to switch to competitors' AI chips. Nvidia has also committed to releasing new hardware architectures annually -- a major acceleration from its prior pace of once every two years. That, too, adds significantly to its competitive advantages. Overall, the company should capture a substantial share of what some forecast will add up to $3 trillion to $4 trillion in global AI infrastructure spending by the end of 2030.

ASML: The Cornerstone of Advanced Chip Manufacturing

Semiconductor equipment leader ASML plays a pivotal role in the design of advanced chips. Its deep ultraviolet (DUV) and more advanced extreme ultraviolet (EUV) lithography systems are used to etch the ultra-fine patterns of circuits on advanced logic and memory chips. It is the only company producing machines capable of manufacturing the chips that go into today's most powerful AI accelerators, high-bandwidth memory, and cutting-edge computing systems.

Strong Financial Performance and Expected Growth

ASML's second-quarter performance was impressive. Total net sales were 7.7 billion euros, near the top of guidance, while net income came in at 2.3 billion euros. Net bookings reached 5.5 billion euros, and the backlog was an exceptionally strong 33 billion euros, even after the cancellation of 1.4 billion euros' worth of orders from customers in China. It expects its advanced chipmaking customers to add 30% more EUV capacity in 2025 than they did in 2024. Chipmakers and memory players are expanding their most advanced production lines to build cutting-edge AI chips and faster memory chips for AI data centers. Rising chip complexity is further driving demand for EUV systems.

NXE:3800E and High-NA EUV: The Future of Lithography

ASML is also making rapid progress with its next-generation lithography machines. Its latest low-NA EUV system, called the NXE:3800E, has demonstrated high productivity, allowing clients to increase output with nearly the same number of systems. While this helps reduce costs for chipmakers, ASML will also benefit from higher average selling prices -- thereby boosting its gross margins. At the same time, ASML is also rolling out an even more advanced version of the lithography technology known as high-NA EUV. This system can print even smaller chip features in fewer steps, which makes production faster, cheaper, and more reliable. Its commitment to innovation has helped ASML build a strong moat and kept competitors from taking its market share.

Conclusion

Nvidia and ASML stand at the forefront of the AI revolution, leading innovation and driving growth in the AI infrastructure industry. With advanced technologies and strong financial performance, they are well-positioned to continue thriving in the years to come.