How Samsung Missed the AI Boom: The Day It Said “No” to Nvidia
In an industry racing toward artificial intelligence supremacy, few moments have proved as fateful—or as overlooked—as one pivotal meeting in 2018. The location: Samsung headquarters in Korea. The visitor: Jensen Huang, the charismatic CEO of Nvidia. The purpose: to forge a collaboration that, if realized, could have changed the power balance of the global AI hardware market.
Instead, Samsung chose a different path.
A Meeting That Shaped the Future
Back in 2018, Nvidia was hungry to secure the next leap in AI computational power. At the heart of this quest was HBM (High-Bandwidth Memory), a critical technology for the world’s fastest AI chips. Jensen Huang, already renowned for his hands-on approach and vision, flew to Korea personally, hoping to convince Samsung—a global leader in memory and semiconductors—to partner more closely on:
-
Joint HBM production for Nvidia’s future AI chips
-
Advanced foundry collaboration, challenging TSMC’s dominance
-
Deeper software ties through CUDA, Nvidia’s foundational AI platform.
According to semiconductor insiders, Huang saw this as a win-win and a chance for Samsung to ride the swelling wave of AI demand. However, to his dismay, the proposals were summarily turned down—reportedly, without even the chance for a high-level strategic discussion. Huang later lamented, “There was no one at Samsung to discuss long-term strategy with me”.
Samsung’s Lost Opportunity
Samsung’s rejection wasn’t just about missing one contract. Without its support, Nvidia pivoted quickly to other partners:
-
SK Hynix seized the opportunity, rapidly becoming Nvidia’s preferred source for HBM and reaping the rewards of the AI computing boom.
-
TSMC became the undisputed foundry of choice, handling the manufacturing of Nvidia’s most advanced chips.
-
Samsung’s foundry ambitions stalled, and the company watched rivals—including SK Hynix and TSMC—leapfrog ahead in AI-ready technologies.
The AI boom, fueled by the explosive demand for Nvidia’s powerful chips, meant that Samsung missed out on supplying the very memory and technology powering the generative AI revolution—the biggest growth story in modern tech.
The inside story hints at broader difficulties. At the time, Samsung was reportedly rocked by internal turmoil, with the company’s chairman under investigation—a factor that may have left strategic decision-making in limbo. Sources say that Huang couldn’t even secure a meeting with the Samsung chairman, leaving him frustrated and disappointed.
Today, Samsung is trying to catch up. While it continues to develop advanced memory and chip technologies, reports suggest it remains behind SK Hynix in meeting Nvidia’s technical and qualification demands for next-generation HBM chips.
Meanwhile, Nvidia and SK Hynix now dominate the AI hardware race, thanks in large part to a door that closed in a Seoul boardroom seven years ago. For Samsung, the lesson may be simple but sobering: In technology, even giants cannot afford to turn away visionaries at the door.
