The Zhitong Finance App learned that Orient Securities released a research report saying that the commercial value of enterprise AI delivery capabilities has been shown, and leading model manufacturers have incorporated deployment capabilities into the core part of commercializing enterprise AI. FDE is expected to become a key delivery layer between models, cloud platforms, and enterprise applications. In the future, as large model capabilities continue to increase and agents continue to develop, FDE demand is also expected to accelerate. The bank believes that in the process of continuous development of large models and agents, FDE has become a key link for enterprises to implement AI capabilities, and demand will also accelerate.
Orient Securities's main views are as follows:
The commercial value of enterprise AI delivery capabilities has been shown in Palantir's results
FDE, or Frontier Deployment Engineer, was pioneered by Palantir. Palantir's AIP platform can be connected to models from many manufacturers such as OpenAI, Anthropic, and Google. Its competitiveness comes more from the combination of the Ontology platform and the FDE delivery system: FDE penetrates customer business, connects the model to enterprise data, business logic, permission systems and operation processes, and completes development, testing and continuous optimization around specific business goals. Recently, Palantir's earnings report had a simultaneous high increase in revenue, providing circumstantial proof that the commercial value of this model has been implemented.
OpenAI establishes a deployment company, and the leading model manufacturer incorporates deployment capabilities into the core part of enterprise AI commercialization
In May 2026, OpenAI announced the establishment of a deployment company controlled by OpenAI and reached an acquisition agreement with Tomoro; the completion of the transaction still requires regulatory approval and other conditions, and is expected to bring in about 150 FDE and deployment experts. The company received an initial investment of more than US$4 billion and collaborated with 19 investment institutions, consulting firms and system integrators, and its partners covered more than 2,000 investee companies. The specific project will begin with high-value workflow screening, and FDE will complete system design, development, testing and deployment within the customer organization to connect the OpenAI model with the customer's data, tools, permission control and core business processes. Model manufacturers invest large-scale capital and directly allocate delivery teams, reflecting that enterprise customers' concerns about AI have covered system integration, daily operation, and quantifiable business results. The number of token purchases and PoCs is no longer sufficient to measure the actual degree of enterprise AI adoption, and deployment and delivery has become a core card point.
FDE is expected to become a key delivery layer between models, cloud platforms, and enterprise applications. The commercial value is driven by both deployment revenue and resource consumption
Enterprise AI production systems need to simultaneously handle private data access, model evaluation, system integration, security compliance, authority management, and organizational collaboration. Improving model capabilities cannot automatically solve these problems. Through on-site development and continuous feedback, FDE condenses customer requirements into connectors, evaluation sets, industry templates, tool interfaces, and platform functions; these reusable assets help shorten subsequent deployment cycles and drive continuous consumption of model calls, cloud resources, databases, and security governance services. The business model can also cover project delivery fees, software subscriptions, pay-as-you-go calls, and pricing linked to business results. In the future, as large model capabilities continue to increase and agents continue to develop, FDE demand is also expected to accelerate.
Risk Alerts
Competition in the industry intensified; commercialization fell short of expectations.