AMD Ramps Supply to Meet Exploding AI Chip Demand as Stock Hits All-Time High

Supply Crunch Signals Surging AI Demand

Advanced Micro Devices faces a challenge most companies would envy: customers want to buy its chips faster than the company can produce them. CEO Lisa Su confirmed during a visit to Taipei on October 6 that demand currently exceeds AMD’s supply capacity, and the chipmaker plans to substantially increase production in 2027. The announcement drove AMD shares to a record close of $649.42, pushing the company’s market valuation above $1 trillion for the first time.

The supply shortage stems from unforecasted demand rather than production failures or structural constraints. During AMD’s second-quarter earnings call, an analyst asked Su directly whether the company faced shortages of server central processing units. Su confirmed that the server CPU supply chain has remained tight throughout the first half of the year because much of the demand arrived unexpectedly, outpacing AMD’s internal projections and forcing the company to accelerate capacity expansion plans.

The tight supply situation does not cap AMD’s ability to meet its own revenue guidance, according to Su, but it does limit how much the company can exceed those targets. She emphasized that 2027 demand is now better forecasted, and AMD has secured supply to more than meet the guidance management has provided to investors. The company is now planning its advanced wafer capacity three to five years ahead, a dramatic extension of its planning horizon that reflects the scale and persistence of AI-driven chip demand.

$30 Billion Commitment Reflects Aggressive Capacity Build

AMD has responded to surging demand with extraordinary financial commitments. The company’s unconditional obligations-primarily for wafers, substrates, and components, plus multiyear cloud service agreements-soared from approximately $12.2 billion at the end of 2025 to $25.7 billion by late March and $30.3 billion by late June. Over just six months, these commitments increased roughly 150%, signaling an aggressive push to secure manufacturing capacity across the supply chain.

The $30.3 billion figure is particularly striking because it approaches AMD’s entire 2025 revenue of $34.6 billion. This magnitude of commitment demonstrates management’s confidence that current demand trends will persist for years, justifying massive upfront investments in production capacity. Prepayments under supply deals also contributed to a $1 billion increase in AMD’s balance sheet obligations, further illustrating the company’s willingness to lock in capacity at premium prices.

Su’s visit to Taiwan included meetings with critical partners in AMD’s supply chain, most notably Taiwan Semiconductor Manufacturing Company (TSMC), which fabricates AMD’s most advanced chips. She also met with Foxconn and is coordinating with memory suppliers to secure sufficient components, particularly the high-bandwidth memory used in AI servers, which remains constrained across the entire semiconductor industry.

Data Center Revenue Drives Growth Trajectory

AMD’s latest financial results underscore the magnitude of AI-driven demand. The company’s second-quarter Data Center revenue reached $6.7 billion, representing a 107% year-over-year increase driven by robust sales of EPYC processors and Instinct GPUs. The Data Center segment now accounts for 58% of AMD’s total quarterly revenue of $11.5 billion, cementing its position as the company’s primary growth engine.

Management expects this momentum to accelerate as AMD’s Helios rack-scale AI systems ramp production. Helios was already in production according to the company’s July statement, with deployments to OpenAI expected to begin in the fourth quarter and accelerate throughout 2027. The Helios platform represents AMD’s integrated approach to AI infrastructure, combining compute, memory, and networking components in optimized rack configurations designed for large-scale AI workloads.

Wall Street Raises Targets on Agentic AI Potential

Investment analysts are responding to AMD’s supply expansion with substantially higher price targets. Citi raised its price target on AMD stock to $800 from $575, citing the emergence of always-on personal AI agents such as Meta’s Muse as a catalyst for sharply increased demand for computing, memory, and networking infrastructure. The firm now estimates that the CPU total addressable market could expand from $29 billion in 2025 to $300 billion by 2030, representing approximately a 60% compound annual growth rate.

“We’ve been able to increase our supply as we’ve gone through 2026,” Su said, adding that AMD would “substantially increase” supply in 2027, according to Reuters.

The agentic AI thesis rests on the observation that AI agents require substantially more compute than traditional chatbots because they operate continuously and execute tasks in the background rather than responding only to discrete user queries. This architectural shift could drive persistent, baseline demand for server infrastructure that dwarfs the capacity requirements of current AI applications, fundamentally expanding the market for AMD’s data center products.

Supply Strategy Positions AMD for Multi-Year Growth

AMD’s aggressive capacity investments reflect a strategic bet that AI infrastructure demand will remain elevated for at least the next three to five years. By extending planning horizons and locking in manufacturing capacity through massive prepayments, the company is positioning itself to capture share in what management expects will be a period of sustained industry expansion. The supply crunch, rather than representing a weakness, signals that AMD’s products have achieved product-market fit in the rapidly growing AI segment.

AMD shares traded more than 1.5% higher in premarket trading on Tuesday following Su’s comments, extending the stock’s recent rally. The milestone crossing of the $1 trillion valuation threshold places AMD among an elite group of technology companies and validates the market’s confidence in the company’s AI strategy. Investors are betting that AMD’s investments in capacity will translate into revenue growth that more than justifies the upfront costs, particularly as the company gains traction in both traditional data center markets and emerging AI workloads.

Memory supply remains a critical bottleneck across the semiconductor industry, particularly for the high-bandwidth memory required in AI servers. Su acknowledged these constraints and confirmed that AMD is working directly with memory suppliers to secure adequate component availability. Resolving memory bottlenecks will be essential to AMD’s ability to fully capitalize on the demand surge, as even abundant CPU and GPU capacity cannot address customer needs if memory components remain scarce.