The Rundown – Deep Dive: Can AMD Beat Nvidia at Its Own Game?

Host: Zaid Admani
Date: October 18, 2025
Duration: ≈10 minutes

Episode Overview

This weekend’s deep dive explores the intensifying rivalry between AMD and Nvidia, focusing on whether AMD’s recent moves and key partnerships can seriously threaten Nvidia’s current dominance in the AI-focused GPU market. Host Zaid Admani breaks down AMD’s evolving business, analyzes landmark deals with OpenAI and Oracle, discusses the hardware vs. software battle, and evaluates AMD’s prospects and risks for investors.

Key Discussion Points & Insights

1. AMD’s Business Breakdown

(00:41–02:12)

• Segments:
  • Embedded (11% of revenue): Chips for smart sensors, robotics, and cars; $824M Q2 2025 revenue.
  • Client and Gaming (47%): Consumer CPUs for Windows PCs, PS5 & Xbox custom chips; $3.6B Q2 revenue.
  • Data Center (AI Focus, 42%): CPUs & AI GPUs (Instinct line: Mi350, Mi450 chips); $3.2B Q2 revenue, expected rapid growth.

"If you recently bought a Windows computer, there is a decent chance that it's being powered by an AMD Ryzen processor." — Zaid (01:37)

2. The AI Race: AMD’s Instinct Chips vs. Nvidia Blackwell

(02:13–03:12)

• Product Focus: AMD's Instinct Mi350 and Mi450 chips directly challenge Nvidia’s Blackwell chips.
• Market Traction: 7 out of the top 10 AI builders (Microsoft, Meta, Oracle, Tesla) are now using AMD.
• Wall Street Buzz: Recent wins have increased investor attention.

3. The OpenAI Mega-Deal: A Turning Point?

(03:13–05:10)

• Historic Purchase: On October 6, OpenAI committed to buy 6 gigawatts of AMD AI GPUs (Mi450 chips), a multi-billion dollar order.
• Financial Incentives:
  • OpenAI gains options for up to 160 million AMD shares (10% of the company).
  • Shares vest in tranches tied to chip deployment and AMD hitting a $600 share price.
• Revenue Impact: AMD expects "double digit billions of annual incremental data center AI revenues" starting H2 2026.

"OpenAI is financially incentivized for AMD stock price to keep going up if they actually want to get the stake in the company. So it's a pretty clever deal and I think it's a win-win for both sides." — Zaid (04:39)

4. Oracle Deal and Industry Trends

(05:11–05:46)

• Oracle Partnership: 53,000 Mi450 chips scheduled for late 2026.
• Industry Implication: Major AI infrastructure builders diversifying away from Nvidia; AMD emerges as a viable alternative.

5. The Nvidia Moat: Hardware & CUDA Software

(05:47–07:36)

• Hardware: AMD CPUs have beaten Intel, and new AI chips (Mi450) are closing in on Nvidia.
• Software Dominance:
  • Nvidia CUDA: The “iOS of AI”—ubiquitous, mature, and central to developer workflows.
  • AMD ROCm: AMD’s open-source alternative struggles with bugs, poor support. Still not developer-friendly or industry-ready.
• Switching Costs: Deep CUDA ecosystem makes moving to AMD tough, despite hardware advances.

"Nvidia's real monopoly comes from their software, which is called CUDA... all the most important AI tools are built with CUDA. Now AMD also has a competing software called ROCm, but from everything I've read, nobody really likes it." — Zaid (06:45)

6. Bull Case vs. Bear Case for AMD

(07:37–08:58)

• Bull Case:

  • OpenAI partnership could spur ROCm maturity and developer adoption.
  • CEO Lisa Su’s proven track record in breaking monopolies (e.g., overcoming Intel in CPUs).
  • Remarkable recovery from near-bankruptcy to PC leadership.
  • Personal note: Lisa Su and Nvidia CEO Jensen Huang are cousins, highlighting a family rivalry.

• Bear Case & Risks:

  • OpenAI deal assumes it can afford $100B+ in chips despite ongoing massive losses (~$20B/year against only ~$13B revenue).
  • Possible AI “bubble burst” may shrink the total market.
  • AMD’s forward P/E ratio (43x) is now higher than Nvidia (33x), making it pricier given their relative size.

"AMD stock has almost doubled in 2025... AMD is more expensive when looking at future earnings than even Nvidia, which is the undeniable market leader." — Zaid (08:47)

7. Nvidia’s Continued Dominance

(08:59–09:36)

• Current Scale: Q2 data center revenues—Nvidia: $41B vs. AMD: $3.2B.
• Margins: Nvidia ~70%, AMD <50%.
• Summary: AMD has momentum, but the gap in revenue, margins, and software keeps Nvidia far ahead.

8. Conclusion: Can AMD Topple Nvidia?

(09:37–10:18)

• Short Answer: No—Nvidia’s lead is massive and entrenched.
• However: AMD doesn’t need to “win” outright; if it grabs 15–20% market share in a fast-growing market, it can become one of the world’s most valuable companies.
• Investor Wisdom: Duopolies offer strong investment potential—even “just” being second best can mean outsized returns.

"As an investor, monopolies are great, but duopolies aren't so bad either, and that seems to be AMD's best case scenario." — Zaid (10:13)

Notable Quotes & Memorable Moments

• “This OpenAI deal… is one of the biggest GPU orders in history.” (04:00)
• “The best way I can describe CUDA is that it’s like the iOS of AI… [but] from everything that I’ve read, nobody really likes [AMD’s] ROCm.” (06:33)
• “Lisa Su… has a history of taking down monopolies… Now she’s trying it again, but this time she’s facing Nvidia—who is led by her cousin Jensen Huang. Still blows my mind that those two are cousins.” (07:55)
• “If all the predictions about AI are true… even if AMD only takes 15 to 20% of the market, that should still make them one of the most valuable companies in the world.” (09:58)

Important Timestamps

• 00:41 — Breakdown of AMD’s three business segments
• 02:13 — Data center/AI chips and competing with Nvidia
• 03:13 — Details of OpenAI’s mega-purchase and incentive deal
• 05:11 — Oracle partnership and industry diversification
• 05:47 — The hardware war and the software “moat” (CUDA vs. ROCm)
• 07:37 — AMD’s CEO history and breaking monopolies
• 08:24 — Valuation concerns and OpenAI deal risks
• 08:59 — Nvidia’s continued performance edge and revenue dominance
• 09:37 — Final analysis: duopoly, market share, and investor implication

Takeaway

Nvidia's dominance in the AI GPU market remains unchallenged for now, largely due to its software ecosystem (CUDA) and outsized scale. However, AMD’s new hardware, strategic partnerships (OpenAI, Oracle), and proven leadership under Lisa Su position it as a credible contender. Even without overtaking Nvidia, capturing a minority share of the exploding AI hardware market could make AMD an immense success—highlighting the power of duopolies for investors.