16 Billion Tokens Per Minute and What It Means for Content

Alphabet’s Q1 2026 earnings, released on April 29, included three data points that, taken together, describe a consumer AI ecosystem operating at a scale most content strategies have not caught up with. Sundar Pichai, CEO of Alphabet and Google, reported that Gemini is now processing more than 16 billion tokens per minute via direct API use by customers, up 60% from the previous quarter. Paid subscriptions across Google’s services reached 350 million, with the strongest quarter ever for consumer AI plans driven by the Gemini App. And Gemini Enterprise paid monthly active users grew 40% quarter over quarter.

Each of these numbers tells a piece of the same story. The consumer side has reached streaming-service scale. The enterprise side is consolidating major business workloads. The developer side, measured by API token throughput, indicates that Gemini is becoming foundational infrastructure for third-party applications. Content gets consumed across all three layers, and the brands showing up inside that consumption are the ones the underlying systems recognize as authoritative.

What 16 billion tokens per minute actually represents

Token throughput is the kind of number that sounds impressive without immediately communicating what it means in practice. A token is roughly three-quarters of an English word. Sixteen billion tokens per minute translates to approximately 12 billion words being processed every 60 seconds, or about 200 million words per second. That is the volume flowing through Gemini’s API alone, not counting the consumer-facing Gemini App, AI Overviews in Search, or Gemini features embedded in Gmail, Docs, and other Google products.

The “via direct API use by customers” qualifier is important. API usage means developers and companies building their own products on top of Gemini. Every customer support chatbot, every AI-powered search tool, every automated research assistant, every content summarization service that runs on Gemini’s API is consuming tokens at a rate that contributes to that 16 billion per minute figure. The number captures the ecosystem of AI-powered applications, not the consumer interface.

The 60% quarter-over-quarter growth rate matters as much as the absolute number. A system processing 10 billion tokens per minute last quarter now processing 16 billion is not a gradual increase. That rate of acceleration suggests more developers are building on Gemini, existing applications are scaling their usage, and the volume of content flowing through Gemini-powered products is expanding rapidly across industries and use cases.

Every one of those tokens represents content being retrieved, processed, synthesized, and presented to an end user. Some of that content comes from the open web. Some comes from enterprise data. Some comes from documents users upload directly. The portion coming from the open web passes through the same retrieval and citation logic that determines which brands, which pages, and which claims get surfaced in AI-generated answers. At 16 billion tokens per minute, the volume of content consumption happening through AI intermediaries is approaching a scale where traditional direct-to-reader content consumption looks like one channel among several rather than the default.

350 million paid subscriptions and the end of early-adopter framing

When commentators describe AI search as an “emerging” channel or frame AI visibility as “something to watch for the future,” the subscription data makes that framing difficult to maintain. Three hundred and fifty million people are paying for AI-powered services from Google. YouTube premium products and Google One drive the bulk, but Pichai specifically called out the Gemini App as having its “strongest quarter ever for consumer AI plans.”

For context, Netflix has approximately 300 million paid subscribers globally. Spotify has roughly 260 million premium subscribers. Disney+ has approximately 150 million. Google’s paid subscription base of 350 million, which includes AI-powered features across YouTube, Google One, and the Gemini App, now exceeds any single streaming platform.

The comparison is imperfect because Google bundles multiple products under its subscription umbrella. But the scale comparison makes a useful point. When people discuss Netflix’s influence on entertainment consumption, nobody describes it as “emerging” or “something to monitor.” Netflix at 300 million subscribers is an established, mainstream channel that entertainment companies build strategies around. Google at 350 million paid subscribers, with AI features increasingly integrated into those subscriptions, should receive the same treatment from content strategists.

A meaningful share of those 350 million subscribers are encountering AI-mediated content discovery as a standard part of their paid experience. The Gemini App summarizes articles, answers research questions, drafts content from source material, and generates recommendations based on user queries. Google One subscribers with Gemini integration get AI features baked into their everyday Google experience. These are not power users experimenting with a beta product. These are mainstream consumers whose daily interactions with content increasingly pass through an AI layer.

Gemini Enterprise at 40% quarterly growth changes the B2B visibility question

The 40% quarter-over-quarter growth in Gemini Enterprise paid monthly active users represents the enterprise side of the same adoption curve. Enterprise users interact with Gemini through their work accounts, using it to summarize documents, draft communications, analyze data, research markets, and evaluate vendors.

When an enterprise user asks Gemini to evaluate software options, compare service providers, or research a vendor, the response draws from the same pool of web content and citation signals that consumer Gemini uses. A brand absent from AI responses in the consumer context is also absent in the enterprise context, which means the sales team is invisible at exactly the moment a potential buyer is doing pre-purchase research inside the AI tool their company pays for.

The 40% quarterly growth rate means the enterprise audience for AI-mediated content discovery is expanding fast enough that any gap between a brand’s AI visibility and its competitor’s visibility compounds rapidly. A competitor showing up in Gemini Enterprise responses this quarter will be showing up for a 40% larger audience next quarter, and a further expanded one the quarter after that. The compounding works in both directions, which is what makes the timing question urgent rather than theoretical.

Three layers consuming content from the same trust signals

The consumer layer (350 million paid subscriptions, Gemini App), the enterprise layer (Gemini Enterprise, 40% QoQ growth), and the developer layer (16 billion tokens per minute via API) all consume content, and all rely on overlapping trust signals to determine which content to surface.

When the Gemini App answers a consumer’s question, it retrieves information based on search ranking, entity recognition, and third-party citation presence. When Gemini Enterprise answers a business user’s question, the retrieval logic is fundamentally similar, drawing from the same web data, the same knowledge graph, and the same authority signals. When a third-party application built on Gemini’s API processes a user query, it inherits Gemini’s retrieval architecture and its trust signal hierarchy.

A brand with strong third-party citation presence, consistent entity recognition, and structured, extractable content on authoritative domains is visible across all three layers simultaneously. A brand without those signals is invisible across all three. The work is the same regardless of which layer the end user happens to be on, because the underlying retrieval system is shared.

Link building and digital PR feed the trust signal layer that all three consumption channels draw from. Every editorial mention in a credible publication, every backlink from an authoritative domain, every guest post on a site with editorial standards contributes to the citation pool that Gemini, across all its deployment surfaces, uses to decide which brands deserve a mention in its responses. Link insertions into existing authoritative content put a brand inside pages that the retrieval system already trusts, across all three layers, without waiting for new content to earn its way into the pool.

People are already reading through AI instead of reading the page

The Q1 numbers describe a content consumption model that has already split into two parallel tracks. On one track, users visit web pages directly, read content, scroll, click, and convert. On the other track, users interact with an AI layer that retrieves content on their behalf, synthesizes it, and delivers an answer without requiring the user to visit the source page. Both tracks are active, both generate real business outcomes, and the second track is growing at 60% quarter over quarter in API volume alone.

Content strategies built exclusively for the first track (optimizing for ranking, click-through, and on-page engagement) are serving a channel that remains important but is no longer growing as fast as the AI-mediated channel. Content strategies that serve both tracks (ranking well for traditional search while also maintaining the trust signals, entity consistency, and structured content that AI retrieval systems favor) are the ones positioned to capture value from both.

The 16 billion tokens per minute figure is not a projection about where AI content consumption might go. It is a measurement of where it already is, growing 60% quarter over quarter, with 350 million paid subscribers on the consumer side and 40% quarterly growth on the enterprise side. The scale is here. The remaining question is whether a brand’s content strategy has caught up to it.