Microsoft said it is funding a new enterprise AI deployment business with $2.5 billion, turning Microsoft Frontier Company into a services push for large customers trying to run more than one AI model inside their own systems.

The Microsoft Frontier Company enterprise AI deployments plan starts with Unilever and Novo Nordisk as initial clients, while Rodrigo Kede Lima, who has led Microsoft’s Asia business, will serve as president.

Microsoft said the group will embed 6,000 industry and engineering specialists with customers.

Microsoft Frontier Company Starts With $2.5 Billion Funding

Microsoft said Frontier Company will be a separate operating entity focused on AI implementation for enterprise customers.

The company said the new group will combine forward deployed engineers, technical consultants, support staff, sales employees and industry specialists.

Microsoft said the 6,000-person group will work directly with customers to design, deploy and improve AI systems.

The company said the teams will bring industry knowledge, change management, continuous improvement work and enterprise AI engineering into customer projects.

Microsoft said Lima has spent the past six years at Microsoft and has 30 years of industry experience.

The company said he previously led enterprise transformation work across the Americas and Asia before taking the Frontier Company role.

Customers Keep Output From Frontier Company Work

Microsoft said Frontier Company will advise customers on Microsoft AI tools and external providers.

The company said the service will help connect those tools with internal customer data while allowing customers to retain the output of the work.

Microsoft said its arrangement leaves control of customer data, intellectual property and competitive information with the customer.

Under the arrangement, customers keep Frontier Company project results rather than returning them to Microsoft.

Microsoft listed OpenAI, Anthropic, Microsoft AI, open-source providers and industry-specific models as supported options for the platform.

The company said that structure lets customers match different models to different workloads instead of relying on one provider across all use cases.

Multi-Model AI Raises Cost And Governance Work

Microsoft said Frontier Company will help customers evaluate models, integrate them with existing systems and switch between them when required.

The company said the platform will include tools to observe, govern, manage and secure AI systems, including FinOps practices to assess returns on AI investment.

Judson Althoff, CEO of Microsoft Commercial Business, said Microsoft’s own Copilot experience shaped the new company.

Althoff said Microsoft initially made a mistake by binding Copilot only to OpenAI models, and said customers wanted model swappability for state-of-the-art systems and fine-tuning.

Althoff said customers placed more value on their own data combined with the right models than on any single model.

He also said businesses needed the ability to move between models quickly as AI projects moved closer to business processes and operations.

LSEG And Land O’Lakes Are Named Customer Examples

Microsoft named London Stock Exchange Group as one customer example.

Microsoft said its teams added AI features to LSEG Workspace for searches across structured and unstructured financial content.

Microsoft said the LSEG work is being refined through client feedback and real-time user testing.

The company also named Land O’Lakes among the customers working with Frontier Company teams.

The enterprise AI services market is already crowded.

Microsoft said Palantir helped popularise the forward deployed engineering job title, while the article said AWS has launched a $1 billion embedded-engineer unit and OpenAI launched its own deployment company in May.

Microsoft said it has forward deployed engineering partnerships with Accenture, Capgemini, EY, KPMG and PwC.

CNBC put Microsoft’s enterprise and partner services revenue at about $2.1 billion in the March quarter, a 2.5% year-on-year increase.

Microsoft has not disclosed customer contract values, pricing for Frontier Company work, measured return on investment from the named deployments, or whether Unilever, Novo Nordisk, LSEG and Land O’Lakes have moved the systems into full production.