When the experienced engineer and executive Pat Gelsinger assumed the role of CEO at Intel in 2021, the renowned chipmaker was facing challenges. Having failed to adapt to the mobile era and lagging in cutting-edge microprocessor manufacturing, Intel was also struggling to meet the increasing demand for chips in the tech industry’s quest for artificial intelligence.
With unwavering optimism, Gelsinger pledged to orchestrate a remarkable resurgence at Intel. He committed to revitalizing its stagnant corporate culture, prioritizing core engineering, and unveiling an enhanced manufacturing strategy to rival industry giants like TSMC and Samsung.
Recently, Gelsinger proudly announced that Intel’s revival strategy was progressing smoothly. He introduced a rebrand of the company’s “foundry” division, responsible for producing chips for other companies, highlighting that Intel’s latest manufacturing advancements would soon yield silicon chips comparable in efficiency and capability to those from TSMC. Securing Microsoft as the first major client for this innovative chipmaking technology marked a significant milestone for Intel as it strives to showcase its competitiveness in the AI era.
In a conversation with WIRED senior writer Will Knight, conducted over Zoom from his residence in Santa Clara, California, Pat Gelsinger shed light on Intel’s AI-focused transformation. Here is a condensed version of their dialogue:
Will Knight: This week, you unveiled Intel’s reimagined chip manufacturing business as an “AI-era system foundry.” Can you elaborate on this concept?
Pat Gelsinger: Two years ago, I initiated Intel’s strategic vision, emphasizing the unexpected growth in generative AI. While Nvidia has dominated this space, Intel stands uniquely positioned to capture 100% of the AI market. Our expertise in integrating networks, memory, and supply chains has generated immense enthusiasm among customers seeking our comprehensive solutions.
Regarding the surge in AI development, how do you perceive reports indicating OpenAI’s ambition to raise $7 trillion for chip development to advance AI progress?
Initially, the $7 trillion figure seemed staggering. However, upon closer examination, current AI models rely on about 10,000 GPUs, with projections suggesting a future demand of 10 million GPUs for larger models. This budget encompasses expenses for training advanced models, power consumption, and data centers.
You mentioned Intel’s progress towards the “18A” manufacturing process to rival TSMC’s offerings. What additional strategies are you implementing to regain a competitive edge?
The industry is focused on advancing transistor technology, particularly with ribbonFET. Notably, Intel’s innovation in backside power delivery enhances chip performance and density, enabling the production of more chips per wafer—a significant value proposition.
Intel recently secured Microsoft as a foundry customer. How do you plan to assure potential clients of Intel’s improved capabilities in this domain?
Our goal is to become the preferred foundry for all clients. Whether it’s Nvidia’s AI chips or Alphabet’s TPUs, we aim to be the go-to provider. By investing billions in the 18A process, we are committed to ensuring high performance and scalability, with separate financial reporting for the foundry business starting next quarter.
Intel’s access to next-generation lithography machines from ASML is crucial for competing with TSMC. How pivotal is this in gaining a competitive advantage?
Lithography remains integral to chip production, and Intel’s commitment to cutting-edge technology, including extreme ultraviolet lithography, signifies a strategic advantage. Our proactive approach to adopting advanced techniques reflects our determination to lead in process technology.
As Intel stands to benefit from the CHIPS Act, addressing talent shortages in the US chip industry is crucial. How do you propose addressing this challenge?
The CHIPS Act’s workforce development funding is a positive step towards resolving talent shortages. While engineering capacity is sufficient, the scarcity lies in skilled trades like construction workers and precision technicians. By focusing on workforce training, we can bridge this gap effectively.
Lastly, is that a diagram of the iconic 486 microprocessor next to you?
Yes, it is the 486 microprocessor, a design that has even been featured in the Museum of Modern Art. Reflecting on its artistic significance at a chip art exhibit was a surreal experience, highlighting the intersection of technology and creativity in my career journey.
