Company file
AMD
A fabless designer showing how long-lived supplier relationships and measured commitments turn product ambition into dependable access.
Structured snapshot available; the disclosed commitment category includes manufacturing and non-manufacturing items.
Operating profile
How AMD builds access
- Operating model
- Fabless CPU, GPU, and adaptive-computing designer
- Controls through
- Qualified suppliers, unconditional commitments, product planning, and ecosystem relationships
- Physical stack
- Wafers, substrates, advanced packaging, memory, servers, and software ecosystems
Structured snapshot
Reported exposure
Unconditional commitments, primarily wafers/substrates, cloud, software and technology
Due in fiscal 2026
FY2025 purchases of property and equipment
$12.166B ending balance ÷ $34.639B FY2025 revenue
Reading note. Commitment intensity is an ending balance divided by annual revenue, not CapEx, cash paid, asset value, or capacity controlled. Values are dated company records.
Comparability. Commitment category includes non-manufacturing items. Do not characterize the entire balance as wafer capacity.
A fabless model built on durable partnerships
AMD's history makes the distinction between ownership and access especially tangible. Separating from manufacturing ownership created a more focused design business, but it did not make physical production disappear. The company still needs qualified process technology, wafers, substrates, packaging, test, and the relationships that allow a product roadmap to become available systems at the moment customers need them.
Commitments reveal the planning horizon
AMD disclosed approximately $12.166 billion of unconditional commitments at the end of fiscal 2025, with $8.5 billion due during fiscal 2026. The filing category is broader than wafer capacity, which is precisely why it should be read carefully. It shows a company supporting a network of inputs and services, not a simple count of factory access or a substitute for owned capital equipment.
Products have to cross several physical layers
A modern AMD platform depends on more than a leading processor design. It has to be manufactured, packaged, paired with memory, integrated into a board and server, qualified by system partners, and supported by software that lets customers use it at scale. The company file is useful because it keeps those production layers visible even though the factories themselves sit with specialized manufacturing partners.
Access has value when it preserves choice
The strongest partner network is not merely a set of purchase obligations. It gives a designer credible routes through changing demand, product mix, and regional conditions. For AMD, the quality of supplier alignment, advanced-packaging access, memory availability, and customer readiness matters as much as the headline commitment amount. A resilient architecture is one that can keep its roadmap moving when a single layer becomes constrained.
Product mix determines which capacity matters
CPU, GPU, and adaptive-computing programs create different demands for wafers, packaging, memory, boards, and server integration. The operating task is to express that mix early enough for a partner network to plan around it, then keep learning as customer adoption changes. AMD's strategic leverage comes from making its technical roadmap a useful planning signal for every production layer that has to respond.
What to watch
AMD's manufacturing-control story is ultimately about translation: can product demand become a sufficiently clear signal for its suppliers, and can that signal be matched by qualified, deployable systems? The important distinction is constructive rather than binary. AMD does not need to own every factory to build durable access, but it does need a network of commitments and partners that can perform through the next product cycle.