OpenAI Diversifies AI Chip Supply with Landmark AMD Deal, Boosting Samsung Electronics
Table of Contents
OpenAI is strategically expanding its access to cutting-edge AI processing power through a major agreement with AMD, a move poised to reshape the semiconductor landscape and deliver a significant windfall for Samsung Electronics.
OpenAI announced on June 6 that it has secured a contract with AMD for the delivery of AI accelerators totaling six gigawatts (GW) of capacity. This substantial investment underscores the growing demand for AI infrastructure and signals a deliberate effort to diversify away from sole reliance on current market leader NVIDIA. According to reports, 6GW is equivalent to the electricity consumption of approximately 5 million U.S. households, exceeding three times the output of the Hoover Dam.
A Strategic Alliance Beyond Supply
The agreement extends beyond a simple procurement deal. AMD will also receive 160 million shares of OpenAI common stock – roughly a 10% stake – exercisable at a price of one cent per share, contingent upon an increase in AMD’s stock price. “This contract will be able to generate tens of billions of dollars in sales,” stated Lisa Su, CEO of AMD, emphasizing the potential for a mutually beneficial partnership. AMD management anticipates exceeding $100 billion in sales from OpenAI and other customers over the next four years.
Industry observers characterize the collaboration as a “strategic alliance,” particularly in light of OpenAI’s recent $100 billion agreement with NVIDIA for AI server construction. This dual-pronged approach is widely seen as a calculated move to mitigate risks associated with relying on a single supplier and to strengthen OpenAI’s negotiating position in a rapidly evolving market. NVIDIA currently controls approximately 80% of the AI chip market, and concerns are mounting about its ability to meet escalating demand.
Samsung Electronics Positioned to Benefit
The deal’s ripple effects are expected to extend to Samsung Electronics, which is already a key supplier of components to AMD. The semiconductor industry is identifying Samsung as a “hidden beneficiary” due to its crucial role in producing High Bandwidth Memory (HBM) – a critical component that directly impacts AI chip performance.
While SK Hynix has previously dominated the supply of HBM3 and HBM3E to NVIDIA, Samsung has been steadily gaining ground. The company began supplying HBM3E 12-speed products to AMD in the latter half of the year and is actively targeting the delivery of HBM4 for AMD’s next-generation MI450 accelerator. “Samsung Mi450’s HBM4 will be supplied mainly by Samsung Electronics,” noted one semiconductor analyst, signaling a significant win for the South Korean tech giant. This partnership with AMD offers Samsung an opportunity to rebound from past performance challenges in the HBM market.
HBM Market Growth and Future Outlook
Despite potential short-term implications for SK Hynix, analysts predict that the overall growth of the HBM market will provide ample opportunities for all major players. The initial concerns surrounding an “AI giant theory” have subsided as global tech companies continue to aggressively invest in AI infrastructure, driving sustained demand for AI chips.
Investor attention is currently focused on Samsung Electronics’ third-quarter earnings, scheduled for release on the 14th. Securities firms are forecasting an operating profit exceeding 10 trillion won, a substantial improvement from the 4.6761 trillion won reported in the second quarter – a year-over-year decrease of 55.23%. According to Chae Min-sook, a researcher at Korea Investment & Securities, “The HBM3E 12-speed product certification for NVIDIA has been resolved, mitigating the HBM risk.”
