Google CEO Sundar Pichai has identified a critical economic problem threatening enterprise AI adoption: companies worldwide are depleting their entire annual artificial intelligence budgets by May due to skyrocketing token consumption. At Google's I/O 2026 developer conference, Pichai shifted the conversation from AI capabilities to AI economics, warning that the rapid rise of AI agents has created unprecedented costs for businesses. To address this crisis, Google unveiled Gemini 3.5 Flash, a new model designed to deliver high performance while dramatically reducing expenses.

Why Are Companies Burning Through AI Budgets So Quickly?

The explosion of AI agent usage has caught enterprises off guard. Tokens are the basic units that AI models process; each word or piece of data a model reads or generates consumes tokens. As companies deploy more AI agents to automate workflows, their token consumption has skyrocketed, causing them to exhaust budgets meant to last the entire year.

"Companies are already blowing through their annual token budgets and it's only May," said Sundar Pichai, CEO of Google.

Sundar Pichai, CEO of Google

To illustrate the scale of the problem, Pichai noted that top companies are processing approximately 1 trillion tokens per day. This massive volume, multiplied across thousands of enterprises, represents billions of dollars in AI infrastructure costs that companies didn't anticipate when they set their annual budgets.

How Can Companies Reduce AI Spending Without Sacrificing Performance?

• Shift to Efficient Models: Companies can migrate 80 percent of their workloads from expensive frontier models to a combination of Gemini 3.5 Flash and other optimized models, potentially saving over $1 billion annually for large enterprises.
• Leverage Flash's Coding Capabilities: Gemini 3.5 Flash performs better than its predecessor, Gemini 3.1 Pro, on coding and agentic benchmarks, making it suitable for complex development tasks without premium pricing.
• Enable Multi-Step Automation: The model can handle long-horizon workflows, executing multi-step tasks such as application development, code maintenance, and document preparation in a single session.
• Utilize Multimodal Features: Gemini 3.5 Flash can generate interactive web interfaces, graphics, and animations while supporting complex reasoning tasks, expanding its use cases beyond text processing.

Pichai's proposal is straightforward: if enterprises shifted 80 percent of their workloads from other frontier models to Gemini 3.5 Flash, they could save more than $1 billion dollars annually. As Pichai explained, "That is real savings they can pour back into their company".

As Pichai

What Makes Gemini 3.5 Flash Different?

Google unveiled Gemini 3.5 Flash as the first release in the new Gemini 3.5 family at I/O 2026. The model is designed to deliver improved performance for coding, agentic AI tasks, and multimodal understanding while maintaining the faster response speeds associated with the Flash series. According to Google, the model now performs better than Gemini 3.1 Pro across several coding and agentic benchmarks, including Terminal-Bench 2.1, GDPval-AA, and MCP Atlas.

Availability is immediate and broad. Gemini 3.5 Flash is now accessible globally through the Gemini app, AI Mode in Search, Google AI Studio, Android Studio, and enterprise platforms. The model also powers new AI experiences across Google products, including the upcoming Gemini Spark personal AI agent. Google has also positioned the model to work with its updated Antigravity platform, allowing multiple AI subagents to collaborate on larger workflows.

Why Does Google Have a Cost Advantage Over Competitors?

Google's ability to offer lower-cost models stems from a structural advantage that competitors cannot easily replicate. Google owns the full technology stack, including custom chips, data centers, cloud infrastructure, AI models, and consumer applications. This vertical integration allows Google to control costs at every layer.

Analysts estimate that Google pays 50 to 75 percent less for AI compute than its rivals, thanks to its custom TPU (Tensor Processing Unit) chips and direct sourcing of computing resources. By contrast, competitors like OpenAI rely on third-party infrastructure providers including Microsoft, Oracle, and Nvidia, paying margins at every layer of the technology stack. This cost structure makes it difficult for competitors to match Google's pricing on enterprise AI services.

The timing of Pichai's announcement is significant. With enterprises already facing budget crises mid-year, Google is positioning Gemini 3.5 Flash as the practical solution to a problem that will only intensify as AI adoption accelerates across industries. For companies struggling to justify AI spending to their boards, the promise of $1 billion in annual savings represents a compelling business case to adopt Google's latest technology.