AMD's AI Growth Stalled: Is the Stock a Sell or a Buy?

AMD’s AI division has long been a focal point for investors, but 2025 has brought mixed signals. While the company’s overall revenue surged 32% year-over-year to $7.7 billion in Q2 2025, driven by robust performance in the Client and Gaming segments, its AI business faced headwinds from U.S. export restrictions and competitive pressures. The Data Center segment, which includes AI hardware, generated $3.2 billion in revenue—a 14% increase—but was marred by a $155 million operating loss due to halted shipments of MI308 GPUs to China [1]. This raises critical questions: Is AMD’s AI growth truly stalled, or is the stock undervalued amid a high-growth market?

Valuation: Overpriced Optimism or Justified Hype?

AMD’s stock currently trades at a Price-to-Earnings (PE) ratio of 99.7x, far exceeding the industry average of 30.1x and peer average of 34.1x [2]. A discounted cash flow (DCF) model estimates its fair value at $138.05, implying the stock is overvalued by 21.5% [2]. However, analysts like Morningstar have raised their fair value estimate to $155 per share, citing optimism around AMD’s AI GPU roadmap [5]. This divergence reflects a tug-of-war between short-term challenges and long-term potential.

The company’s gross margin of 54% remains stable, but export-related charges and inventory write-downs have dented profitability. For instance, the MI308 export ban led to an $800 million non-cash charge in Q2 2025 [1]. Despite this, AMD’s AI strategy is gaining traction: seven of the top ten global AI firms now deploy its Instinct chips, and the MI350 series offers 4x AI compute performance and 35x inference improvements over its predecessor [1].

Strategic Positioning: Innovation vs. Market Realities

AMD’s AI ambitions are anchored in next-generation hardware and ecosystem partnerships. The MI350 series, launched at the Advancing AI 2025 event, targets large language models (LLMs) with support for FP4 and FP6 datatypes [2]. Meanwhile, the company secured a $10 billion contract with a Saudi AI startup to build cloud infrastructure and expanded its AI ecosystem through partnerships with Red Hat, Oracle, and HUMAIN [1]. These moves signal a strategic pivot toward enterprise AI adoption.

However, AMD’s AI GPU market share remains below 10%, trailing NVIDIA’s dominance [4]. While the MI350’s performance metrics are compelling, the company faces hurdles in converting trial customers into long-term buyers. Seaport Research notes that AMD struggles to scale AI chip deployments beyond initial trials [6]. Additionally, NVIDIA’s CUDA ecosystem and Intel’s potential resurgence in AI chips pose credible threats [1].

Market and Regulatory Headwinds

The U.S. export restrictions have had a material impact. AMD estimates a $1.5 billion revenue hit in 2025 due to blocked MI308 sales to China [4]. A revenue-sharing agreement, requiring AMD to remit 15% of AI chip sales to China for export licenses, further pressures margins [4]. While the company is diversifying into the EU and India, these markets lack the scale of China’s AI demand.

Regulatory fragmentation in the U.S. also complicates AMD’s strategy. States like Colorado and Virginia have enacted AI-specific laws requiring transparency and impact assessments, raising compliance costs [5]. Globally, the EU’s General-Purpose AI Code of Practice and Singapore’s AI assurance initiatives add layers of complexity [1].

Risk vs. Opportunity: A Calculated Bet

AMD’s AI business is a double-edged sword. On one hand, the company is innovating rapidly, with the MI400 series expected to bolster its competitive edge in 2025 [3]. On the other, its valuation reflects aggressive growth assumptions. Analysts project AI GPU revenue could reach $7 billion in 2025, up from $5 billion in 2024, but this hinges on overcoming export and adoption challenges [5].

For investors, the key question is whether AMD can sustain its 32% revenue growth while navigating regulatory and competitive pressures. The stock’s high PE ratio suggests market confidence in its AI ambitions, but this optimism may not be justified if growth slows or export restrictions persist. Conversely, AMD’s strategic partnerships and product roadmap position it to capture a larger share of the $391 billion AI market, which is projected to grow at a 35.9% CAGR through 2030 [5].

Conclusion: A Buy for the Long-Term, a Sell for the Short-Term

AMD’s AI division is neither a complete stall nor a runaway success. The stock’s current valuation reflects both its high-growth potential and near-term risks. For long-term investors, AMD’s innovation in AI hardware and expanding ecosystem partnerships justify a cautious buy, particularly if the MI350 and MI400 series gain traction. However, short-term investors may want to wait for a pullback, as the stock’s 99.7x PE ratio implies significant margin for error.

In the end, AMD’s AI story is a bet on execution. If the company can navigate regulatory hurdles, scale its AI deployments, and maintain its gross margin above 50%, it could emerge as a major player in the AI era. But for now, the risks outweigh the rewards for those seeking immediate returns.

Source:
[1] AMD Misses on Earnings, but Delivers Strong Q3 Guidance,
https://leverageshares.com/en/insights/amd-misses-on-earnings-but-delivers-strong-q3-guidance/
[2] Is AMD's Surge Justified After AI Chip Momentum in 2025?
https://simplywall.st/stocks/us/semiconductors/nasdaq-amd/advanced-micro-devices/news/is-amds-surge-justified-after-ai-chip-momentum-in-2025
[3] AMD: The Next $1 Trillion AI Giant Set to Dominate Inference
https://mlq.ai/research/amd-advanced-micro-devices-inference/
[4] The AI Chip Market Explosion: Key Stats on Nvidia, AMD and Intel’s AI Dominance
https://patentpc.com/blog/the-ai-chip-market-explosion-key-stats-on-nvidia-amd-and-intels-ai-dominance
[5] 50 NEW Artificial Intelligence Statistics (July 2025)
https://explodingtopics.com/blog/ai-statistics