Nvidia Poised to Overtake Apple as TSMC’s Top Client

When Jensen Huang, CEO of Nvidia, first met Morris Chang, the founder of Taiwan Semiconductor Manufacturing Co. (TSMC), decades ago, he famously declared that Nvidia would one day become the chip foundry’s largest customer. This ambitious prediction is now on the cusp of becoming a reality, according to industry analysts and Huang himself. Nvidia is set to overtake Apple as TSMC’s top client this year, a significant shift that underscores the explosive growth of artificial intelligence and its foundational role in the semiconductor landscape.

This transition marks a pivotal moment in the semiconductor industry, reflecting Nvidia’s ascendancy as a critical enabler of the global AI infrastructure build-out. Huang himself confirmed this development in a recent podcast, stating, “Morris will be happy to know Nvidia is TSMC’s largest customer now.”

Ben Bajarin, principal analyst at Creative Strategies, projects that Nvidia will generate approximately $33 billion in revenue for TSMC this year, accounting for about 22% of the foundry’s total revenue. In comparison, Apple, currently believed to be TSMC’s largest customer, is expected to contribute around $27 billion, or roughly 18% of the total. “The scale of this has drastically changed,” Bajarin noted. “A couple of years ago, you could just see how much more capacity Nvidia was demanding from TSMC.”

While both Nvidia and TSMC have declined to comment, and Apple did not respond to a request for comment, TSMC’s own financial disclosures provide further evidence of this trend. The company reported that its top 10 customers accounted for 76% of its net revenue, with its largest customer at the time representing 22% and the second-largest 12%.

The impact of Nvidia’s burgeoning demand is profoundly evident in TSMC’s financial performance. High-performance computing (HPC) sales, which encompass Nvidia’s AI chips, constituted 55% of net revenue in the fourth quarter, a substantial increase from 40% in 2022, the year the AI boom truly ignited with the launch of OpenAI’s ChatGPT. AI accelerators, a segment dominated by Nvidia, represented a significant “high-teens” percentage of TSMC’s total sales in 2025.

Nvidia’s sales trajectory is outpacing that of Apple. The company is anticipated to report 66% growth in fiscal year 2026, reaching $213 billion in sales. This contrasts with Apple’s 6.4% revenue growth in its fiscal year 2025. Furthermore, Nvidia’s AI chips are more complex and larger than those manufactured for Apple’s consumer devices, resulting in a higher cost per unit.

This industry realignment highlights TSMC’s indispensable role as the world’s leading contract chip manufacturer, serving a vast array of processor designers including Advanced Micro Devices (AMD), Intel, Broadcom, and Qualcomm. Market researcher TrendForce estimates that TSMC commands approximately 70% of the global chip manufacturing revenue market. Meanwhile, rival Intel, despite its ambitions to produce leading-edge chips in the U.S., is still seeking a major anchor customer, and recently faced production concerns and a significant stock drop after issuing soft quarterly guidance.

Huang’s deep understanding of TSMC’s critical role is evident in his consistent engagement. He made five visits to Taiwan last year, notably attending TSMC’s annual sports day, even donning the same red shirt as the employees. During one such visit, he reportedly toured a fab utilizing the 3-nanometer technology essential for Nvidia’s Rubin chip, which is now in full production and slated for shipment later this year.

TSMC’s recent quarterly results further illustrate the transformative impact of AI chip demand on its business. The company reported $33.73 billion in net revenue for the December quarter, a 21% year-over-year increase, and projected 30% sales growth for the current year. This robust outlook is largely fueled by the anticipated compound annual growth rate of “mid-to-high fifties” for AI chips through 2029.

“Our customers’ customer, who are mainly the cloud service providers, are also providing strong signals and reaching out directly to request the capacity to support their business,” stated TSMC CEO C.C. Wei, underscoring the company’s confidence in the long-term AI trend. “Thus, our conviction in the multi-year AI mega-trend remains strong.”

TSMC plans to invest between $52 billion and $56 billion in capital expenditures this year, with further increases anticipated to meet AI demand. These investments are expected to come online by 2028. However, the company maintains a degree of conservatism regarding long-term projections, with Wei expressing a measured concern about the potential for an AI bubble, given the substantial capital outlay required.

Historically, TSMC’s close relationship with Apple, driven by the high volume of iPhone chips, allowed the foundry to make aggressive investments in cutting-edge “leading node” technologies. The advanced nodes enable greater power efficiency, a critical factor for extending battery life in devices. This pursuit of power efficiency is equally vital for Nvidia, as energy consumption directly impacts the return on investment for its AI systems.

Bajarin notes that this dynamic shift changes the calculus for TSMC. “What was the driving force for TSMC — Apple — now shifts to Nvidia, and to some degree AMD, which is sort of the guarantee-scale customer that helps you justify the increase in CapEx to each new node,” he explained.

While Apple will continue to be a major customer, the emphasis on Nvidia is clear. During TSMC’s earnings call, Nvidia was the only customer whose product lineage was specifically mentioned, with Wei highlighting the performance leaps from Hopper to Blackwell to Rubin. From Wei’s perspective, the primary bottleneck in the AI industry remains “TSMC’s wafer supply.”