Google has released two new hardware products designed to cut the latency and token costs associated with running autonomous software agents inside a production environment. The Gemini 3.6 Flash, for example, is described as a more energy-efficient alternative to previous models, potentially reducing the costs of powering agent tokens.
The economics of running autonomous software agents in a production environment are complex, requiring a model to reason through multi-step tasks competently. However, the current cost equation few vendors advertise directly is not ideal for many organizations. This can make it difficult to justify the high costs of deploying enterprise AI agents into these environments.
Gemini 3.6 Flash targets reducing token costs, and by extension latency, while also offering improved energy efficiency. These products are designed to provide a more robust and scalable alternative to previous hardware offerings from Google. By cutting through some of the complexity in running autonomous software agents, Gemini 3.6 Flash is poised to revolutionize how these models operate in enterprise environments.