AMD to invest more than $10 billion in Taiwan as the AI chip race with Nvidia heats up

AMD is making another move on the chessboard in the global race for leadership in AI infrastructure. This time the company has announced investments of more than $10 billion in partners in Taiwan, the world's main chipmaker.

A bet on local manufacturing

AMD is investing in developing and scaling production of both AI chips and software for cloud infrastructure. It is an ambitious move aimed at making AMD a serious competitor to Nvidia in the AI accelerator market.

Under Lisa Su, the company has chosen a partnership path in Taiwan: working with local manufacturers instead of building its own fabs. This should let it adapt faster to growing demand from cloud providers and enterprises.

The investments cover several key areas:

• Developing next-generation AI accelerators based on the CDNA architecture
• Scaling up chip production for global orders from cloud platforms
• Building a full software stack for AMD accelerators
• Strengthening the supply chain amid geopolitical risks
• Preparing for mass demand for AI infrastructure in 2026–2027

Why Taiwan

Taiwan is the heart of the global semiconductor industry. TSMC, the world's largest contract chipmaker, is based there. AMD's investments in partners in Taiwan mean the company wants to be closer to the manufacturing process, respond faster to market demand and minimize supply delays.

Geopolitical tension around Taiwan and possible trade restrictions are also pushing tech companies toward diversifying and localizing production. For AMD, a stronger presence in Taiwan is both a strategic cushion and a way to demonstrate its capabilities to large customers.

Competition with Nvidia is intensifying

Nvidia dominates the AI chip market with a share of around 85–90%, but its monopoly position is starting to erode. AMD is gaining momentum with its MI300X, as customers look for alternatives because of Nvidia chip shortages and price pressure. AMD's investments in Taiwan are a direct response to that demand and an attempt to narrow the gap.

Cloud giants such as AWS, Google Cloud, Meta and Microsoft Azure are expanding capacity for AI workloads. They have an interest in competition among chip suppliers, because it lowers prices and speeds up innovation. AMD is becoming an increasingly attractive option for them. Some customers are already openly saying they are ready to migrate to AMD chips if that gives them independence from Nvidia and better control over pricing.

Beyond chips

AMD's investments cover not only the chips themselves but also software, an important part of the strategy. AI hardware without optimized software runs inefficiently. AMD ROCm, an open platform for GPU accelerators, is becoming an increasingly mature rival to Nvidia's CUDA.

Local development in Taiwan will let AMD integrate hardware and software faster, improving performance and developer experience. For cloud providers, this means they can get a full turnkey solution from AMD instead of assembling a stack from different vendors. That speeds up adoption and lowers integration risk.

What this means

For the market, this is a signal that Nvidia's monopoly is starting to crack. AMD's investments show the company is serious about winning market share, not just testing the waters. For users and businesses, this may mean more choice, healthy price competition and faster innovation in AI infrastructure.

The AI chip market is entering a new phase: competition instead of monopoly. For AMD shareholders, these investments are a bet on exponential growth of the AI segment in the coming years. If the company can deliver on its promises and scale up production, revenue from AI chip sales could become the main driver of profit growth.

*Meta has been designated an extremist organization and is banned in Russia.