Intel's Missed Opportunity with OpenAI: A Strategic Misstep

Around 2017 and 2018, Intel had the opportunity to invest in OpenAI, a then-upcoming non-profit research organization focusing on generative AI. Discussions included Intel potentially buying a 15% stake for $1 billion and an additional 15% stake if Intel manufactured hardware for OpenAI at cost.

Intel's Decision

Intel ultimately declined the deal. Then-CEO Bob Swan was skeptical about the near-term market potential of generative AI models, questioning whether the investment would pay off. This decision significantly impacted Intel, a company that had once dominated the computer chip industry.

Why OpenAI Wanted Intel

OpenAI sought Intel's investment to reduce reliance on Nvidia's chips and build its own infrastructure. However, Intel's data center unit was unwilling to produce hardware at cost, leading to the collapse of the deal.

Consequences and Current State

The decision not to invest in OpenAI, which later launched ChatGPT in 2022 and is now valued at around $80 billion, has not been publicly discussed until now. Intel, once a leader in computer chips, has struggled in the AI era. Recent reports highlight Intel's strategic missteps and its fall behind competitors like Nvidia and AMD.

Stock Market Impact

Intel's recent second-quarter earnings report led to a significant drop in its stock price, the worst trading day for the company since 1974. For the first time in 30 years, Intel's market value fell below $100 billion. In contrast, Nvidia, with a market value of $2.6 trillion, has successfully pivoted to AI chips, leaving Intel far behind.

Intel’s Response

In response to questions about its AI progress, Intel pointed to its upcoming third-generation Gaudi AI chip, set to launch in the third quarter of this year, and its next-generation Falcon Shores AI chip, expected in late 2025. Intel CEO Pat Gelsinger expressed optimism about these products, indicating the company’s efforts to capture a larger share of the AI market.

Missed Potential and Strategic Failures

Microsoft's investment in OpenAI in 2019 propelled it to the forefront of the AI revolution, a position Intel could have secured. For over two decades, Intel believed that the CPU (central processing unit), the backbone of desktop and laptop computers, was the most effective tool for handling the processing tasks required to build and run AI models. This belief was deeply rooted in Intel's strategy, as confirmed by four former Intel executives with direct knowledge of the company's plans.

Intel engineers viewed the GPU (graphics processing unit) architecture, predominantly used for video gaming by rivals Nvidia and Advanced Micro Devices (AMD), as inferior and unsuitable for AI tasks. They considered GPUs to be "ugly" and inefficient compared to CPUs.

However, by the mid-2000s, researchers had discovered that GPUs were significantly more efficient than CPUs for the intensive data processing required to build and train large AI models. Designed for game graphics, GPUs could perform a vast number of calculations in parallel, making them ideal for AI applications. Nvidia's engineers recognized this potential early on and spent years refining GPU architecture specifically for AI uses. They also developed the necessary software to fully leverage these capabilities, creating a robust ecosystem for AI development.

"When AI hit ... Intel just didn't have the right processor at the right time," said Lou Miscioscia, an analyst at Japanese investment bank Daiwa. This misalignment was a critical misstep for Intel. While Nvidia and AMD capitalized on the growing demand for AI processing power, Intel clung to its CPU-centric approach, failing to pivot quickly enough to the burgeoning AI market.

Attempts to Catch Up

Intel's reluctance to embrace GPU technology for AI, combined with its slow response to market changes, led to a series of strategic misfortunes. Since 2010, Intel has made multiple attempts to produce a viable AI chip, including acquiring startups Nervana Systems in 2016 and Habana Labs in 2019, but these moves were too little, too late. Nvidia and AMD had already established themselves as leaders in AI processing, leaving Intel struggling to compete in a rapidly evolving industry.

Future Outlook

Intel's data center business, including its AI chips, is expected to generate $13.89 billion in sales this year, far behind Nvidia’s expected $105.9 billion. Despite the challenges, Intel remains committed to advancing its AI technology and capturing a greater share of the market in the future.