How Facebook's Open Compute Project became a major force in data center hardware, with hundreds of companies, including HP, Foxconn, and Goldman Sachs on board
How Facebook is eating the $140 billion hardware market — It started out as a controversial idea inside Facebook.
Context & Ripple Effects
The Open Compute Project began as a contested idea inside Facebook and has since pulled in hundreds of members — HP and Foxconn on the supply side, Goldman Sachs on the buyer side — turning hyperscale server designs into a shared specification anyone can build against. The arc runs through Google's decision to join the project Facebook founded, which ended the era of rival hyperscalers treating infrastructure design as pure competitive secrecy.
The significance today is that OCP is no longer a cost-saving club but the template Facebook keeps replicating: the Telecom Infra Project applies the same open-specification playbook to telecom equipment, and Facebook has continued feeding it donations such as its Fabric Aggregation Layer networking architecture. A Wired retrospective on Facebook's open-source portfolio from Cassandra and GraphQL to React places OCP alongside software projects as part of one strategy — commoditize the layers below you.
First-order effects
- HP and Foxconn now compete inside a framework where the buyer (Facebook) writes the specifications, shifting margin pressure from system integration to manufacturing scale — Foxconn's cloud-and-networking business becoming its majority revenue contributor reflects exactly this positioning.
- Goldman Sachs' membership shows even non-tech buyers adopting open hardware for their own data centers, expanding OCP beyond web companies into financial services.
Second-order effects
- Google joining rather than fighting OCP forces other large operators to choose: adopt the specs or bear higher costs alone — the same dynamic that pushed telecom gear into Facebook's Telecom Infra Project two years later.
- Traditional enterprise server vendors lose pricing power on commodity racks, pushing them toward services and differentiated components instead of closed systems.
Third-order effects
- If hyperscaler-written open specifications keep winning, infrastructure value concentrates at the layer above — the platforms and chips — while hardware becomes standardized capacity that financiers like Goldman Sachs can underwrite at scale, a pattern visible in the bank's later role alongside Nvidia and private-credit firms backing AI data-center funding.
- OCP establishes the governance model where the largest consumer of hardware effectively sets the industry standard through a foundation it founded — regulatory and antitrust attention to that gatekeeping remains an unresolved question.
The trend: Hyperscale operators are converting proprietary infrastructure designs into open standards they govern, first in servers via Open Compute and then across adjacent markets like telecom equipment.