US tech companies are restructuring their engineering teams into compact “pods” of humans and AI agents working together.
A pod usually consists of only 1 to 8 people but still encompasses multiple roles such as engineers, designers, and applied scientists.
AI agents are currently taking on an increasing amount of software development work, including coding, testing, and product deployment.
Thanks to AI coding assistants, companies can build products with far fewer personnel than the traditional model required.
Coinbase recently laid off about 14% of its workforce to reorganize under the AI-native pods model.
CEO Brian Armstrong stated that the company will focus on hiring personnel capable of managing a “fleet of agents” instead of large teams.
Coinbase is also experimenting with “one-person teams,” where a single individual simultaneously takes on the roles of engineer, designer, and product manager with AI support.
Rob Witoff, Coinbase’s platform lead, mentioned that an AI adviser project now only requires a team of 3 people instead of the 10 to 15 people needed previously.
According to Witoff, small teams significantly reduce time spent on meetings, reviews, and internal alignment, thereby accelerating product development.
He believes this is precisely why startups usually operate faster than large enterprises—due to less overhead and a higher level of consensus.
An Ernst & Young expert noted that the pod is the next evolution of the scrum team model, which replaced the waterfall model years ago.
The difference is that pods are now supercharged by extremely powerful AI tools, causing the human scale to shrink even further.
Deepak Singh from AWS suggested that the more powerful AI becomes, the fewer personnel are needed in a team.
Amazon has long been famous for its “two-pizza team” model, meaning a team small enough to be fed by two pizzas to speed up decision-making.
According to Singh, a team of 16 people can now be split into 2 small pods of 8 people to focus on specific projects.
Singh observed that large teams often spend most of their time on internal communication rather than actually building products.
This trend shows that AI is not only changing work tools but also reshaping the organizational structure of tech enterprises.
📌 Conclusion: The “AI-native pod” model is becoming the new blueprint for the tech industry as AI agents grow powerful enough to replace most repetitive technical tasks. Companies now prioritize ultra-small yet multi-functional teams, where a few individuals can manage an entire “army of AI” to develop products at startup speed. This could reduce the demand for traditional engineering hiring while simultaneously elevating the value of professionals who know how to collaborate effectively with AI. In the future, the structure of tech enterprises may shift from hundreds of engineers to a network of hyper-lean pods powered by AI.
