Microsoft and AWS commit $3.5B to AI deployment: integration becomes the battleground

On 2 July, Microsoft announced Frontier Company, an operating business backed by $2.5B and 6,000 industry and engineering experts, dedicated to deploying AI directly inside client organisations — with the London Stock Exchange Group, Unilever and Accenture as early partners. Two days earlier, AWS committed $1B to an equivalent Forward-Deployed Engineer model, following the joint ventures already launched by OpenAI and Anthropic this spring.

What Cardan-AI takes from this: when the four largest AI technology vendors pour billions into consulting and integration, they are publicly acknowledging that technology alone does not produce outcomes. The bottleneck is no longer the model — it is process redesign, change management and ROI measurement. That is the core business of consulting, not of software licensing.

The blind spot in these offerings: they target the Fortune 500 and will, by construction, sell their parent company's technology stack. For a European mid-cap or SME, two direct consequences: these armies of experts will not reach you for a long time, and when they do, their diagnosis will always end with the same prescription — their own platform.

Cardan-AI recommendation: the mid-market needs vendor-independent partners able to compare scenarios (proprietary API, open-source, on-premise) without conflicts of interest. Before any commitment, require your integrator to document the alternatives it discarded and the reversibility criteria. An advisor that can only recommend one stack is not an advisor — it is a sales force.