The semiconductor shortage during the pandemic proved to be a significant challenge, leading to prices that made them nearly inaccessible. Although that period is now in the past, OpenAI CEO Sam Altman is adopting a forward-looking strategy. A recent Bloomberg report discloses Altman’s plan to secure “billions“ in funding for the establishment of a series of AI chip-manufacturing facilities.
Altman’s perspective is rooted in the belief that even leading foundries like TSMC, Intel, and Samsung may have limitations in producing future AI-focused chipsets. To overcome these limitations, he envisions the creation of a ‘network of AI chip factories,’ as outlined in the report.
Indeed, entering the domain of AI chip manufacturing would position OpenAI in direct competition with industry giants such as Google, Microsoft, Amazon, and others. These major players have been developing their custom chipsets instead of opting for off-the-shelf processors, showcasing the evolving landscape of the AI industry.
Notably, companies like Google have taken the bespoke route for their chipsets, signaling the continuous advancement of the AI sector. While Google initially utilized Samsung’s foundry for constructing Tensor chipsets, recent developments suggest a shift toward TSMC for the next-gen Tensor G5.
However, Sam Altman’s approach appears distinctive. The report indicates that he aims to secure funding to establish an independent AI chip-manufacturing powerhouse, minimizing dependence on external sources. Altman envisions AI technologies becoming increasingly self-sufficient, establishing their semiconductor supply chain from scratch.
Altman Engages in Active Talks with Investors
According to the report, Altman is actively engaging in discussions with investors, including G42 based in Abu Dhabi and SoftBank Group. The primary focus of these discussions is not only to address the current AI chip shortage but also to establish capabilities to meet future requirements.
The scale of this project is noteworthy, with funding discussions, particularly with G42, reportedly exceeding $8 billion. The ongoing trend of decreasing fabrication nodes for chip architectures contributes to the substantial costs involved. Mass-producing a 2nm or 3nm class node could presently incur an estimated cost of around $30 billion.
The underlying reality is that the technology essential for modern fabrications comes with a high price tag, with individual tools costing well over $100 million today. The question arises: Will this strategic move position Altman as the AI chip visionary, or will Nvidia maintain its supremacy? Share your thoughts in the comments below.
0 Comments