AI and the Pace of Change

The Distinctive Nature of the AI Revolution

The current wave of artificial intelligence adoption presents a unique shift compared to previous technological revolutions, such as the rise of mobile technology, social media, or cloud computing.

Mary Meeker's Comprehensive Report

A recent 340-page report released by venture capitalist Mary Meeker highlights the exceptional speed of AI development, implementation, investment, and utilization. The term “unprecedented” appears 51 times within the report, underscoring this rapid progression.

Meeker’s report, titled “Trends — Artificial Intelligence,” asserts that the scale and velocity of change associated with AI technology are truly without parallel, as evidenced by extensive data analysis.

A Historical Perspective

This report carries particular weight given Meeker’s background. As the founder and general partner at Bond, and formerly known as the “Queen of the Internet” for her influential Internet Trends reports, she possesses a unique vantage point.

Prior to Bond, she led Kleiner Perkins’ growth practice from 2010 to 2019, providing backing to prominent companies including Facebook, Spotify, Ring, and Block (formerly Square).

Having paused her annual trends reports since 2019, Meeker has revisited this format to meticulously document how AI adoption is surpassing all prior technological advancements in history.

Unprecedented Growth Metrics

The rapid user acquisition of ChatGPT – reaching 800 million users within just 17 months – is described as unprecedented.

Similarly, the number of companies achieving substantial annual recurring revenue, and the rate at which they are doing so, is also considered unprecedented.

Declining Costs and Increasing Competition

The speed at which the cost of utilizing AI is decreasing is remarkable. While model training costs can reach up to $1 billion, inference costs – the expense of using the technology – have already fallen by 99% over two years, measured per 1 million tokens, according to Stanford research.

The swift pace at which competitors are matching features, often at reduced costs, including the emergence of open-source alternatives, particularly from Chinese models, is further described as unprecedented.

For instance, Nvidia’s 2024 Blackwell GPU demonstrates a 105,000-fold reduction in energy consumption per token compared to its 2014 Kepler GPU.

Infrastructure Development

Furthermore, companies like Google, with its TPU (tensor processing unit), and Amazon, with its Trainium chips, are actively developing large-scale infrastructure for their cloud services, demonstrating a rapid pace of innovation.

Meeker emphasizes that these initiatives are not merely experimental projects, but rather represent significant, long-term investments.

Financial Implications

One area where AI hasn’t yet surpassed all previous tech revolutions is in immediate financial returns.

While venture capitalists are heavily investing in AI, both AI companies and cloud service providers are experiencing substantial cash burn due to the massive infrastructure requirements.

This situation benefits consumers and businesses through rapid improvements and lowered costs, as competition intensifies.

However, it remains uncertain which of the current companies will ultimately achieve long-term profitability and establish themselves as the next generation of tech leaders. Meeker notes that “Only time will tell which side of the money-making equation the current AI aspirants will land.”

For the broader public, the implications suggest a period of significant and rapid change.