Key Takeaways
- Nvidia’s revenue is soaring, significantly boosted by the rapid increase in AI “token” generation.
- Tech giants like Microsoft and Google report massive, even fifty-fold, annual increases in the tokens their AI systems process.
- Tokens are the fundamental units for AI inputs and outputs, representing everything from text segments to image pixels.
- This explosive token growth is driving immense demand for Nvidia’s chips, potentially outpacing current data center capacities.
- Analysts and tech leaders see rising token counts as a key indicator of AI’s practical value and adoption.
Nvidia’s financial performance is hitting new heights, with revenue climbing to $44.1 billion last quarter. But one of the most revealing metrics for the chip giant isn’t measured in dollars; it’s all about “tokens.”
These tokens are the basic building blocks of artificial intelligence. Whether it’s pixels in an image, segments of words, or bits of audio, all AI processes break down into these units. As AI tools become more sophisticated, they’re churning out tokens for AI-generated content, or inference, much faster than many anticipated.
Top tech executives, many of whom are Nvidia’s biggest customers, have been highlighting their own token growth. According to Business Insider, Nvidia’s CFO Colette Kress noted, “OpenAI, Microsoft, and Google are seeing a step function leap in token generation.” She added that Microsoft processed over 100 trillion tokens in the first quarter alone, a fivefold increase year-over-year.
Even Google CEO Sundar Pichai shared that monthly tokens produced across Google’s products had jumped by a factor of 50 in the last year, drawing gasps from the audience at Google I/O.
“Explosive token growth is what really matters, in the longer term,” Morgan Stanley analysts wrote before Nvidia’s recent earnings announcement. Following the call, Nvidia’s share price surged.
While tokens can vary in size depending on the content they represent, making direct comparisons tricky, Forrester analyst Alvin Nguyen explained they are “the closest to a standard that we have without needing more data and analytics.”
Nvidia CEO Jensen Huang views the increasing discussion of tokens by tech leaders as a clear sign that AI tools are delivering real value. “Very soon we’ll be talking about how many tokens we produce every hour, just as every single factory does,” Huang remarked at an event in Taiwan.
Given its dominant market position, Nvidia benefits directly from nearly all this token growth. Some analysts suggest this demand is growing so fast that current data center infrastructure is struggling to keep up.
“Every hyperscaler has reported unanticipated strong token growth,” the Morgan Stanley analysts noted, citing anecdotal evidence that demand for AI inference capabilities exceeds what existing systems can support.
The main challenge with tokens is the difficulty in tracking their growth unless companies voluntarily release these figures. However, it’s clear that whenever executives share any hint about token demand, investors will be listening intently for the foreseeable future.
