From selling models to selling "co-running," cloud providers are collectively shifting— the final mile of AI still relies on human engineers to fill in the gaps with their own bodies. The title of front-line deployment engineer sounds like a technical position, but it’s actually a combination of senior customer service, project manager, and fire-fighter.

Google is recruiting hundreds of engineers to help clients implement AI, highlighting the practical challenges. Cloud has established a new "Frontier Deployment Engineer" team to provide more technical resources for clients. Industry trends also include OpenAI co-establishing deployment units with consulting/investment firms, and Anthropic forming a joint venture; Google is also negotiating with Blackstone, KKR, and EQT to provide AI model access for their portfolio companies. Despite the rapid growth in revenue for frontier AI companies, the number of human consultants is also increasing, indicating that deploying AI into production environments remains difficult.