Antonio Linares
The AMD Era Has Begun.
10/9/2025, 1:30:05 PM
Economic Summary
- AMD's chiplet-based approach lets it produce specialized, high-memory chips at marginal cost, lowering total cost of ownership for customers and attracting hyperscaler deals (e.g., ORCL, META, OPENAI).
- AI adoption is still early; only a small percentage of companies and consumers use it today, implying substantial future CapEx and compute demand beyond any single buyer.
- Inference is becoming a major, distributed workload—from core data centers to edge devices—creating fractal demand that favors architectures optimized for on-chip memory and specialization (benefiting AMD).
- Macro risks include possible breakdowns in AI scaling laws and large energy/infrastructure bottlenecks that could materially limit the projected exponential compute growth.
Bullish
- AMD's chiplet platform enables personalized, cost-effective chips for specialized workloads.
- Major customers (Oracle, Meta, OpenAI) validating AMD technology and driving backlog.
- Distributed inference and edge AI create large, growing markets favoring AMD's architecture.
- AMD can pursue multiple multi-trillion-dollar AI opportunities simultaneously at marginal cost.
Bearish
- AI scaling laws may not persist, reducing long-term demand for specialized AI compute.
- Energy and infrastructure bottlenecks could constrain massive AI compute growth.
- Large incumbents could eventually replicate chiplet advantages and erode AMD's lead.
- OpenAI deals concentrated and somewhat uncertain, creating deal-concentration risk.
Bullish tickers
AMDORCLMETAOPENAI
Bearish tickers
AMDOPENAINVDA
AMD
Bullish
Chiplet platform enables personalized, memory-heavy chips with TCO advantage; validated by deals from Oracle, Meta, and OpenAI; well-positioned for distributed inference and edge AI.
Bearish
Risks include a potential slowdown in AI scaling laws, energy/CapEx bottlenecks, concentrated deals, and eventual competitive catch-up.
NVDA
Bullish
Established GPU leader with strong software and ecosystem advantages, representing competitive pressure to AMD.
Bearish
May lose share in some inference workloads where AMD's chiplet memory advantage wins; pivoting to chiplets could be difficult.
ORCL
Bullish
Leaning into AMD for distributed inference; rising backlog validated demand for AMD-powered datacenter workloads.
META
Bullish
Chose AMD MI-series for large-scale Llama inference, signaling AMD competitiveness on major AI workloads.
OPENAI
Bullish
Public validation by leaning into AMD signals strong product-market fit for AMD's AI chips.
Bearish
Deal concentration and uncertain long-term commitments create execution and demand risk.
PLTR
Bullish
Mentioned as an example of the innovator's dilemma; used comparably to illustrate how platforms become hard to replicate.
People mentioned
JeremyMark PapermasterLisa SuSam Altman