AI Agents: Goldman Sachs Predicts Huge Tech Cash Flow Boost

The growth of agentic artificial intelligence is set to dramatically boost computing volume for companies running large language models. This will also increase cash flow for these big tech companies. Here's the thing: The use of tokens, which are units of text processed by LLMs, is expected to explode. Token consumption could multiply 24 times, reaching 120 quadrillion tokens per month between 2026 and 2030. This is happening as the cost of computing falls, positioning AI players for a period of "margin inflection." What's interesting is that agentic AI doesn't just respond to queries; it performs sequences of tasks. This requires many more tokens. For example, agentic AI will multiply consumer token consumption 12 times by 2030, covering things like online shopping. The bottom line is that this massive increase in token usage, combined with falling costs, could significantly improve the free cash flow of hyperscalers.