AMD Surpasses US$1 Trillion in Value as AI Demand Fuels Growth - Mexico Business News

AMD surpassed a US$1 trillion market capitalization for the first time on Sept. 21, 2026, as its shares rallied on growing demand for AI computing, expanding data center revenue, and investor expectations that the chipmaker can challenge Nvidia while the global AI chip market advances toward sustained growth through 2032.

Advanced Micro Devices (AMD) has crossed a market valuation milestone that reflects the growing financial importance of AI infrastructure. The chipmaker surpassed US$1 trillion in market capitalization, as investors continued to place capital behind semiconductor companies positioned to benefit from expanding AI computing demand.

AMD shares rose about 9.6% to US$613.31, reaching a record high during trading. The rally placed the California-based company among a limited group of semiconductor manufacturers to achieve the valuation threshold, reinforcing its position as a major competitor to Nvidia in the AI accelerator market.

The milestone follows a sustained increase in AMD's stock price, which has risen about 185% in 2026, significantly exceeding the Nasdaq's 15.8% gain over the same period. The company has also recorded five consecutive trading sessions of gains, during which its shares advanced approximately 24%. The market response illustrates how AI infrastructure spending continues to influence semiconductor valuations, even as investors assess broader economic conditions and the financial sustainability of AI-related capital expenditure.

AMD Expands Its Position in AI Infrastructure

AMD's growing market valuation is supported by developments in its data center business and its efforts to compete beyond individual processors. The company has accelerated its AI product launches and expanded its portfolio toward complete computing systems that combine processors, networking equipment, and related hardware.

This approach places AMD in more direct competition with Nvidia, whose position in AI computing extends across chips, networking, and integrated infrastructure. AMD has also benefited from demand for central processing units (CPUs) used alongside graphics processing units (GPUs) in servers supporting AI inference. This market has helped the company gain share from Intel in selected data center applications.

The company's second-quarter results provide a financial basis for the market's expectations. According to CNBC, AMD reported US$11.54 billion in revenue, a 50% increase from US$7.69 billion a year earlier. Its Data Center segment generated US$6.7 billion, representing 107% year-over-year growth.

AI chips contributed to this expansion, while the broader demand for data center infrastructure continued to support semiconductor investment. Lisa Su, CEO, AMD, said during the company's earnings call in August that AMD expected to double data center sales in 2027. The projection reflects the company's expectations for continued demand for computing infrastructure, although future performance will depend on customer investment, competitive conditions, and execution.

Despite reporting quarterly revenue above Wall Street estimates, AMD previously fell short of elevated investor expectations. The contrast between operational growth and market valuation highlights the pressure on semiconductor companies to demonstrate sustained returns from AI demand.

AI Investment Reshapes the Semiconductor Market

AMD's milestone comes as the broader AI chip industry enters a period of expansion and technological diversification.

Makreo Research projects that the global AI chip market will grow at approximately 22% annually between 2027 and 2032. The research attributes this expansion to hyperscaler capital expenditure, generative AI adoption, custom silicon development, and the expansion of AI data center infrastructure.

The report estimates that global investment in AI data center infrastructure will exceed US$6.7 trillion between 2025 and 2030. AI-specific data center power demand is also projected to increase from 44GW in 2025 to 155GW by 2030, highlighting the infrastructure requirements associated with expanding computing capacity. These projections indicate that semiconductor demand will be influenced not only by the number of AI models being developed but also by the physical infrastructure required to train, deploy, and operate them.

GPUs remain a dominant technology for large-scale AI training because of their parallel-processing capabilities. However, hyperscalers are increasingly developing proprietary accelerators to address performance requirements, supply constraints, and infrastructure costs.

Google, Amazon, and Apple are among the companies advancing custom silicon initiatives, including Google's TPU v5, Amazon's Trainium 2, and Apple's Neural Engine. These developments are contributing to a market in which general-purpose GPUs operate alongside application-specific integrated circuits (ASICs) and other specialized processors.

According to Makreo's analysis, Nvidia maintains an estimated market share exceeding 70% of the AI accelerator market. The company reported US$89 billion in Data Center revenue in its fiscal second quarter of 2027, representing 117% year-over-year growth and approximately 92% of total company revenue.

Nvidia's scale demonstrates the concentration of AI infrastructure demand, while the expansion of custom silicon introduces additional competitive dynamics across the semiconductor supply chain.

Inference Demand Creates New Competitive Pressures

The next stage of the AI chip market is increasingly connected to inference, the process through which trained models generate responses and perform tasks. While training large AI models requires substantial computing capacity, inference introduces different requirements related to efficiency, latency, and operating costs. This has encouraged chipmakers to develop specialized architectures and expand their commercial offerings.

Makreo identifies several transactions and partnerships in 2026 that reflect this direction. AMD agreed to acquire Toronto-based Taalas, a startup developing technology that hardwires AI model weights directly into silicon for high-speed inference.

Qualcomm established a multi-generation collaboration with Amazon Web Services focused on customized AI inference chips and advanced optical connectivity. Microchip Technology also agreed to acquire Israeli edge AI processor developer Hailo.

Together, these developments point toward a more heterogeneous AI infrastructure landscape, where GPUs, custom accelerators, and specialized inference processors operate within the same computing ecosystem.