November 11, 2026
Taiwan Semiconductor Manufacturing Company (TSMC), the world’s largest semiconductor foundry, has approved strategic investments totaling up to US$1.5 billion, marking a significant expansion of its involvement in the artificial intelligence industry. The investments include up to US$500 million each in OpenAI and Anthropic, alongside another US$500 million to establish an investment company focused on the semiconductor industry and its supporting ecosystem.
The decisions, approved during TSMC’s board meeting on November 10, 2026, represent an important development in the company’s long-term strategy. While TSMC has traditionally focused on manufacturing advanced semiconductors for technology companies, the new investment plans extend its financial interests into the AI software companies that are helping drive global semiconductor demand.
The board approvals establish maximum investment amounts rather than confirming that all transactions have been completed. Nevertheless, they indicate TSMC’s intention to strengthen its position across the expanding AI computing ecosystem, from advanced chip manufacturing and packaging to the developers of the AI models that require increasingly powerful computing infrastructure.
Under the approved investment plans, TSMC intends to acquire up to US$500 million in common shares of Anthropic, the developer of the Claude AI models. The investment will be made at Anthropic’s initial public offering (IPO) price. Separately, TSMC has approved an investment of up to US$500 million in preferred shares of OpenAI Group PBC, the company behind ChatGPT and other artificial intelligence technologies.
These investments represent a strategic expansion beyond TSMC’s traditional semiconductor manufacturing activities. Although OpenAI and Anthropic are primarily AI model developers rather than conventional foundry customers, their growing computing requirements have a substantial influence on the semiconductor industry. Training advanced AI models and operating large-scale inference systems requires enormous computing resources, creating demand for graphics processing units (GPUs), application-specific integrated circuits (ASICs), high-bandwidth memory (HBM), networking semiconductors and advanced packaging technologies.
As AI models become more sophisticated and their adoption expands across enterprise and consumer markets, cloud service providers are investing heavily in computing infrastructure. Companies developing AI accelerators must continuously improve processing performance, memory bandwidth and energy efficiency to meet these requirements. This creates opportunities for semiconductor foundries capable of manufacturing advanced logic chips and supporting increasingly complex chip integration technologies.
For TSMC, investing directly in AI model developers provides financial exposure to a part of the industry that helps determine future computing requirements. Rather than relying exclusively on semiconductor manufacturing revenue, the company is positioning itself to participate in the broader commercial growth of artificial intelligence.
One of the most important semiconductor industry developments associated with the rapid expansion of artificial intelligence is the growing interest in custom AI accelerators. While GPUs remain central to AI training and inference infrastructure, custom ASICs can provide advantages for specific workloads, particularly when computing performance, memory bandwidth, power consumption and operating costs must be optimized for large-scale deployment.
Cloud service providers and AI infrastructure companies are increasingly evaluating purpose-built semiconductor architectures rather than relying entirely on general-purpose processors. Custom ASIC development allows companies to optimize processing engines, memory interfaces, interconnects and power management for particular AI applications. These designs often require advanced semiconductor manufacturing processes, sophisticated electronic design automation tools, semiconductor intellectual property and advanced packaging solutions.
TSMC is positioned to benefit from this trend through its advanced manufacturing capabilities and established semiconductor design ecosystem. Growing demand for AI computing can create business opportunities not only for TSMC but also for ASIC design service providers, semiconductor IP vendors, EDA companies and packaging specialists supporting the development of next-generation AI processors.
The investments in OpenAI and Anthropic do not, by themselves, establish new chip supply agreements or manufacturing commitments. However, they illustrate the increasingly close relationship between AI software development, computing infrastructure requirements and semiconductor technology.
In addition to its investments in AI companies, TSMC’s board approved up to US$500 million to establish a dedicated investment company responsible for making strategic investments in the semiconductor industry and related ecosystem.
The new entity could provide TSMC with additional flexibility to invest in technologies and businesses relevant to its long-term manufacturing roadmap. Advanced semiconductor production depends on an extensive network of equipment suppliers, materials manufacturers, photomask technology providers, semiconductor IP developers and testing companies. As semiconductor manufacturing progresses toward 2nm and subsequent technology generations, cooperation across this supply chain is becoming increasingly important.
Advanced lithography equipment, photomask manufacturing, specialty materials, process control and semiconductor testing all play critical roles in achieving the performance, yield and reliability required for leading-edge semiconductor production. Strategic investments in selected companies could help strengthen relationships with technology suppliers, support innovation and improve access to technologies that are essential for future manufacturing processes.
The proposed investment company also reflects the changing nature of semiconductor manufacturing, where technological competitiveness increasingly depends on the capabilities of the entire ecosystem rather than the wafer fabrication process alone.
TSMC has previously used minority investments to participate in important developments across the semiconductor industry. Two notable examples are its investments in Arm Holdings and IMS Nanofabrication.
In September 2023, TSMC invested approximately US$100 million in Arm during the company’s initial public offering, acquiring a small equity position in one of the semiconductor industry’s most important processor architecture and intellectual property providers. Arm-based architectures are widely used across mobile devices, embedded systems, automotive electronics and increasingly in data center computing applications.
Following the appreciation of Arm’s share price, TSMC gradually reduced its ownership position. The company sold approximately 850,000 Arm shares in 2024 and completed the sale of its remaining approximately 1.11 million shares in April 2026. The transactions generated substantial proceeds compared with its original investment, demonstrating that strategic semiconductor investments can also deliver significant financial returns.
Another important investment occurred in September 2023, when TSMC agreed to acquire approximately 10% of Austria-based IMS Nanofabrication from Intel. The transaction valued IMS at approximately US$4.3 billion.
IMS specializes in multi-beam electron-beam mask-writing equipment used in the manufacture of advanced semiconductor photomasks. These technologies are particularly important for extreme ultraviolet (EUV) lithography, which enables the production of highly complex semiconductor patterns required for leading-edge manufacturing processes.
By investing in IMS, TSMC strengthened its relationship with a critical supplier of advanced photomask technology. Unlike its investment in Arm, which provided exposure to semiconductor intellectual property, the IMS investment was directly connected to manufacturing technologies needed for continued semiconductor process development.
The proposed investments in OpenAI and Anthropic extend this strategy further, allowing TSMC to participate financially in the AI applications and computing demand that ultimately influence semiconductor manufacturing requirements.
Alongside the three strategic investment plans, TSMC’s board approved approximately US$40.93 billion in capital appropriations to support its long-term manufacturing capacity and technology development roadmap.
The approved capital will primarily support the installation and expansion of advanced semiconductor manufacturing capacity, advanced packaging operations, mature and specialty technology manufacturing, fabrication plant construction and facility infrastructure. It also includes capital investments associated with research and development activities in 2027, recurring capital expenditures and capitalized leased assets.
The approval demonstrates the scale of TSMC’s ongoing investment in semiconductor manufacturing infrastructure as demand grows for advanced logic processes and increasingly complex semiconductor packaging solutions. Importantly, the US$40.93 billion represents newly approved capital appropriations and should not be interpreted as TSMC’s total annual capital expenditure or as spending already completed.
Expansion of advanced process capacity will be particularly important for semiconductor companies developing high-performance processors, AI accelerators and custom ASICs. At the same time, additional advanced packaging capabilities will help support semiconductor architectures that combine multiple logic dies, memory devices and other components within integrated systems.
