If the rise of AI feels unlike past tech revolutions like mobile, social media, or cloud computing, that’s because it truly is different. Venture capitalist Mary Meeker just released a massive 340-page report highlighting this shift. The word “unprecedented” appears on 51 pages, emphasizing the rapid pace at which AI is being developed, adopted, invested in, and applied — all supported by a wealth of data and charts.
“The speed and scale of change brought on by artificial intelligence is truly without precedent, and the data proves it,” writes Mary Meeker in her latest report, Trends — Artificial Intelligence. It’s fitting that Meeker is the one sounding the alarm. Often called the “Queen of the Internet” for her influential annual Internet Trends reports, she’s now the founder and general partner at VC firm Bond. Before that, she led the growth fund at Kleiner Perkins from 2010 to 2019, where she invested in major names like Facebook, Spotify, Ring, and Block (formerly Square).
After a five-year hiatus from publishing these reports, Meeker has returned with sharp focus — meticulously laying out how AI is being adopted faster than any previous technology in history.
ChatGPT hitting 800 million users in just 17 months? Unprecedented. The explosion of companies reaching high annual recurring revenue at lightning speed? Also unprecedented. According to Meeker, the pace at which usage costs are plummeting is equally remarkable. While training a single AI model can now cost up to \$1 billion — itself an unprecedented figure — the cost of using these models has dropped sharply. In fact, inference costs (what users pay to run the models) have fallen by 99% in just two years, based on the cost per million tokens, as cited from Stanford research.
The speed at which AI competitors are replicating each other’s features — often at a fraction of the cost and increasingly through open-source efforts, especially from China — is unprecedented. As an example, Meeker highlights that Nvidia’s 2024 Blackwell GPU uses 105,000 times less energy per token than its 2014 Kepler predecessor. At the same time, major cloud providers like Google and Amazon are scaling up their own AI chips — Google’s TPU and Amazon’s Trainium — at a rapid pace. “These aren’t side projects,” Meeker notes, “they’re foundational bets.”
There is, however, one area where AI hasn’t surpassed past tech revolutions: financial returns. Despite a flood of VC funding, AI companies and cloud providers are burning through capital at a high rate. That’s because building and scaling AI requires enormous infrastructure investments.
All of this rapid progress is great news for consumers and businesses, who are benefiting from faster innovation and falling costs, Meeker notes. But despite the momentum, it’s still unclear which of today’s AI players will emerge as sustainable, long-term winners.
“Only time will tell which side of the money-making equation the current AI aspirants will land,” she writes — a reminder that while the technology is advancing at breakneck speed, profitability is still an open question.