I Can't Stay Quiet on OpenAI & AMD vs NVIDIA Stock (NVDA) Any Longer

Key Concepts

• AMD-OpenAI Partnership: A significant collaboration involving AMD supplying GPUs to OpenAI and OpenAI receiving warrants for AMD stock.
• GPU Deployment: The core of the partnership, focusing on the scale of AMD's Instinct GPUs to be deployed by OpenAI.
• Warrants and Stock Dilution: The mechanism by which OpenAI can acquire AMD shares at a low price, contingent on deployment milestones and stock performance, leading to potential shareholder dilution.
• AI Inference Demand: The exponential growth in the need for AI processing power, driven by multiple factors.
• Market Growth vs. Supply: The argument that AI market growth is outpacing chip production, meaning most available hardware will be deployed.
• AMD's Position in AI: The re-evaluation of AMD's role in the AI revolution, moving beyond direct competition with Nvidia.

AMD and OpenAI Partnership Details

The video details a monumental partnership between AMD and OpenAI, which is seen as a significant indicator of the future direction for AI stocks.

• Deal Scope: OpenAI plans to deploy 6 gigawatts (GW) of AMD's GPUs over the next few years.
• Stock Warrants: As part of the agreement, OpenAI receives warrants to purchase up to 160 million shares of AMD stock at a nominal price of 1 cent per share. These warrants vest upon the achievement of specific milestones.
• Milestone Examples:
  • The first tranche of shares unlocks upon the deployment of the first gigawatt of GPUs.
  • Subsequent tranches are tied to AMD's stock price reaching certain predetermined levels.

Technical Breakdown of GPU Deployment

The summary delves into the technical specifics of the GPU deployment and its implications for AMD's business.

• AMD Data Center GPUs and TDP: A table (not provided in transcript but referenced) would detail AMD's data center GPUs and their Thermal Design Power (TDP), which varies based on workload and cooling methods (air vs. liquid).
• Estimated GPU Quantity: Assuming liquid cooling for AMD's Instinct GPUs, the 6 GW deployment translates to approximately 4.3 million GPUs, depending on the specific mix of models.
• AMD's Current GPU Sales:
  • AMD's MI350X accelerators are estimated to cost around $25,000 each.
  • AMD's data center revenue was $3.2 billion last quarter, with approximately half ($1.6 billion) coming from their EPYC CPUs.
  • This implies roughly $1.6 billion in quarterly accelerator revenue.
  • At $25,000 per GPU, AMD sold approximately 64,000 GPUs last quarter.
  • Considering other revenue streams like FPGAs (from Xilinx acquisition) and DPUs (from Pensando acquisition), AMD's annual data center GPU shipments are estimated at around 200,000 units.
• Deal Impact on AMD's Shipments:
  • The OpenAI deal represents over 20 years of AMD's current annual data center GPU shipments.
  • Even the initial phase of supplying 1 GW of GPUs (starting in the second half of next year) equates to roughly 1 million GPUs, which is five times AMD's current annual sales.
• Financial Implications for AMD:
  • AMD generated approximately $1.6 billion in AI accelerator revenue in Q2, suggesting an annualized run rate of $7-8 billion.
  • This partnership is projected to more than double AMD's entire data center GPU business and increase their total revenues by about 40%, depending on the specific accelerators and pricing.

Comparison with Nvidia's Deal and Deal Structure

The video contrasts the AMD-OpenAI deal with OpenAI's previous engagement with Nvidia, highlighting key differences in structure and implications.

• Scale Relative to Nvidia: If Nvidia were to secure a deal of similar magnitude (20 years of data center GPU sales), it would represent a $2.3 trillion deal for approximately 70 million GPUs, equivalent to 120 GW of compute – the combined capacity of all existing data centers globally. This emphasizes the immense scale of the AMD deal relative to AMD's current size.
• Misconception about OpenAI's Stake: The common misunderstanding is that OpenAI is buying 10% of AMD. In reality, AMD is providing OpenAI with warrants to buy shares at a very low price, effectively financing OpenAI's infrastructure deployment based on performance milestones.
• Performance-Driven and Dilutive Nature: This is a high-stakes, performance-driven deal. The full 160 million shares vest only if OpenAI successfully deploys all 4.3 million AMD GPUs.
• Shareholder Dilution: The vesting of all 160 million shares would result in approximately 9% dilution for AMD shareholders, based on current outstanding shares.
• Contrast with Nvidia's Investment: This deal is distinct from Nvidia's recent $100 billion direct investment in OpenAI. The AMD deal is contingent on deployment and stock performance, offering significant upside only if milestones are met.
• Stock Price Milestone: A key vesting milestone requires AMD's stock to reach $600 per share, a roughly threefold increase from current prices. This would imply a $1 trillion valuation for AMD, making the equity granted to OpenAI worth approximately $100 billion, or 10% of the company.
• Immediate Benefits for AMD: Even before reaching the highest milestones, AMD secures a multi-billion dollar hardware order and a direct connection to a leading AI company, which is expected to attract other AI firms.

The Exponential Growth of AI Inference Demand

The video then shifts to the broader AI market and the factors driving unprecedented demand for AI processing.

• Global AI Market Growth: The global AI market is projected to grow 19x over the next nine years, with a compound annual growth rate (CAGR) of 38.5% through 2034.
• AI Capex Projections: Meta, Google, Amazon, and Microsoft are expected to spend a combined $410 billion in AI capital expenditures (capex) in 2026, up from $220 billion in 2024, indicating a similar annual growth rate of 37%. This growth rate is approximately three times faster than that of the S&P 500.
• Jensen Huang's "1 Billion X" Inference Demand: Jensen Huang, CEO of Nvidia, predicts a "1 billion x" increase in demand for AI inference. This is attributed to four multiplying factors:
  1. Generative AI Adoption: Rapid adoption of generative AI, surpassing the adoption rates of smartphones and tablets.
  2. Increased Token Usage per Prompt: Prompts now require significantly more tokens (hundreds of times more) as models analyze more sources, perform complex reasoning, and generate code or data tables.
  3. Expansion of AI Use Cases: AI applications are extending beyond text-based LLMs into diverse fields like material science, chip design, drug discovery, entertainment, and more. Evidence from NVIDIA GTC shows dramatic increases in patent filings and product prototypes after AI adoption.
  4. AI Agents: AI agents can now take actions on behalf of users, including initiating other AI processes. The analogy is made of augmenting human employees with AI, leading to significant productivity gains. The expectation is that individuals will utilize multiple AI agents for various tasks, similar to owning multiple computing devices.

Implications for the AI Stock Market

The exponential growth in demand has profound implications for the entire AI hardware and infrastructure sector.

• Supply vs. Demand Imbalance: With a projected billionfold increase in inference demand, supply is unlikely to keep pace.
• Ubiquitous Deployment of Hardware: In such a high-demand, supply-constrained market, every GPU from Nvidia, every ASIC from Broadcom, and every accelerator from AMD will be deployed.
• Skyrocketing Utilization Rates: This demand surge will lead to significantly higher utilization rates for AI infrastructure providers, including cloud giants like Amazon, Microsoft, and Google, as well as hardware and service companies like Coree, Dell, and Oracle.
• Early Stage of the AI Era: The speaker emphasizes that we are still in the very early stages of the AI revolution, underscoring the importance of understanding the underlying technology and market dynamics.

Author's Admission of Error Regarding AMD

The video concludes with the speaker admitting a misjudgment regarding AMD's role in the AI revolution.

• Underestimation of AMD's Potential: The speaker acknowledges underestimating AMD's position, failing to connect the massive growth in AI inference demand with AMD's ability to supply hardware.
• AMD's Dominance Outside Nvidia's Sphere: The speaker previously noted that AMD excels in markets where they don't directly compete with Nvidia. The OpenAI partnership highlights that the AI market is large enough for multiple players to thrive.
• Validation for AMD: The OpenAI deal serves as a strong validation for AMD, indicating OpenAI's confidence in AMD's ability to deliver advanced GPUs and their belief in AMD's potential for significant future growth (implied by the stock price milestone).
• Future Growth Potential for AMD: The partnership positions AMD favorably, as other AI companies are likely to adopt similar hardware, software, and development ecosystems as OpenAI. This makes AMD a potentially strong investment for long-term wealth creation, provided milestones are met.

Conclusion and Investment Outlook

The video concludes with a forward-looking perspective on AMD and the broader AI investment landscape.

• AMD as a "Get Rich Without Getting Lucky" Stock: The partnership, if milestones are achieved, positions AMD as a stock with significant upside potential, driven by fundamental growth rather than speculative hype.
• Call to Action: Viewers are encouraged to like and subscribe for more content.
• Final Investment Advice: The speaker reiterates the importance of investing in oneself and directs viewers to another video for insights into other stocks they are investing in for long-term wealth.