Mary Meeker's Tech Report Shows Pace, Scope Of AI Adoption

Forbes

14 hours ago

SAN FRANCISCO, CA - OCTOBER 19: Partner at KPCB, Mary Meeker, (L) and special correspondent for ... More Vanity Fair, Nick Bilton, speak onstage during "The State of the Valley: Where’s the Juice?" at the Vanity Fair New Establishment Summit at Yerba Buena Center for the Arts on October 19, 2016 in San Francisco, California. (Photo byfor Vanity Fair)
Tech media is looking hard at a new report from Mary Meeker indicating that ChatGPT's search volume over three years is faster than that of Google search, which is now the dominant means of traditional hyperlink-based Internet use.
It's the first report of its kind in six years, and weighs in at around 360 pages, from someone with bona fides in the tech space.
'Her previous annual reports have been pivotal in shaping understanding and investment within Silicon Valley and the broader technology sector,' writes an anonymous author at Ikala. 'This AI-centric report is a natural extension of her influential work, now directed at what many consider the most transformative technology of our time.'
So this powerhouse of a report, with so much detail, comes from someone with an established reputation: Meeker now runs the Bond fund, with almost $6 billion in investment capital. At Kleiner Perkins, she helped lead the charge to invest in some of the biggest tech companies, like Facebook and Spotify. (She was also famously involved in the IPOs of both Netscape, and later, Google.) This is one reason that so many people with close ties to the industry are paying attention as the report is released.
Some of what Meeker found was 260% annual growth in AI training model data since 2010, and 360% annual growth in training compute over the same time frame.
In addition, she noted, user and subscriber and revenue growth for ChatGPT shows its popularity with today's Internet user.
You can see visuals on this with charts from the report's slides, like one labeled: 'Seem Like Change Happening Faster Than Ever? Yes, It Is,' with its familiar hockey-stick curve, or this interesting combination of tags: 'AI Monetization Threats + Rising Competition + Open-Source Momentum + China's Rise.'
And, as noted by some readers, Meeker uses the word 'unprecedented' quite a bit, as in this introductory remark:
'To say the world is changing at unprecedented rates is an understatement. Rapid and transformative technology innovation / adoption represent key underpinnings of these changes. As does leadership evolution for the global powers.'
And…
'OpenAI's ChatGPT – based on user / usage / monetization metrics – is history's biggest 'overnight' success… AI usage is surging among consumers, developers, enterprises and governments. And unlike the Internet 1.0 revolution – where technology started in the USA and steadily diffused globally –ChatGPT hit the world stage all at once, growing in most global regions simultaneously.'
I think that's a pretty good thesis of the report as a whole. Let's break some of this down.
Charts Paint Thousands of Words
Meeker also found that the Google ecosystem, used as a baseline for the AI dev community in general, developed fivefold from May of last year until now, from 1.4 million developers up to 7 million.
She looked at work on image, audio and more, and the upshot of it all is that developers are swarming to this new area of AI research. Just look at what's happening on Github, or ask a corporate team what their focus is.
In addition, Meeker delved into how companies will be deploying all of the fruits of their developer teams' AI labor, for one, in interacting with a target audience. She quotes Decagon Co-Founder & CEO Jesse Zhang in comments last year, saying:
'In a few years, every company will have AI agents running their customer experiences. Customer support staff are no longer fielding routine tasks; they are now becoming AI managers – configuring, training and overseeing the AI agents that handle repetitive work.'
That's just one of many major use cases that will change the face of business soon.
Tech Compounding Over Fifty-Plus Years
Then there's the lifespan of tech adoption for various past revolutions, where Meeker measured AI against other innovations.
The personal computer, she said, took 20 years to reach half of all households.
Desktop Internet took 12 years; mobile Internet took six years.
AI is seeming to take about three years, for a kind of Moore's law advance in adoption speed, and then there's the evolution, correspondingly rapid, of AI's various forms. Meeker writes:
'Multimodal AI models … embed text, pictures, sound, and video into a shared representation and generate outputs in any of those formats.…the path to this capability unfolded stepwise: OpenAI's CLIP paired vision and language in 2021; Meta followed with ImageBind in 2023 and Chameleon in 2024; and by 2024-2025, frontier systems such as GPT-4o, Claude 3, and Chameleon had become fully multimodal.'
150% Annual Growth Over Six Years of Performance Gains from Better AI Supercomputers
CEOs are paying attention, too. Meeker's work shows how AI is being embraced by business – quickly.
There's a Morgan Stanley adoption survey where a full 75% of global CMOs who responded were experimenting with AI. And again, the numbers bear out that any forward-thinking company is putting significant chips, no pun intended, on AI work.
Here's another business statistic: Meeker observed a 21% gain in related capex, and 28% in data spend, as well as 37% in annual cloud revenue growth on the part of hyperscalers.
Then there's the issue of cost: with Meeker noting 'cost deflation' and 10x increases annually in what a dollar will buy for AI inference. I wrote about this after attending a conference earlier this year, and it's remarkable how much more your money gets you in LLM performance compared to just months ago.
Another thing that comes through from Meeker's report is how new users will start from zero with an AI-native environment. They won't be encumbered with the obsolescence of traditional Google search or sitting at a terminal. The machine will meet them in their native language, verbally.
'Imagine, for a moment, how different your next week would look if there were no internet,' Meeker writes. 'Every facet of modern life – how we work, how we communicate, how we govern, and more – would likely be turned on its head. The internet has been woven into so many facets of life, big and small, that – for many – it is difficult to imagine a world without it. In the next decade or two, imagining a world without AI will likely feel the same.'
Meeker also takes on numbers from around the world, with top ChatGPT use in India at 13.5%, with the U.S. at around 9%, and Indonesia and Brazil both at 5%. Check the report for more.
There's also a nod to U.S./China competition, with Chinese LLM performance taking the lead.
'Chinese AI capabilities now underpin nationally strategic areas such as battlefield logistics, target recognition, cyber operations, and autonomous decision-making platforms,' Meeker writes. 'In 2025, Chinese state media highlighted the integration of AI into non-combat support functions (e.g., military hospitals), while the Ministry of Science and Technology reinforced its commitment to 'indigenous innovation' in strategic technologies. … The implications of Chinese AI supremacy would be profound.'
That dominance is likely to be a significant feather in China's cap as we get closer to new interface models where we're engaging with humanoid entities endowed with AI.
In other words, that's the next step – step one is interfaces that move from a text base to verbal. Step two is where those verbal interfaces are tied into a body that can approach yours, to pursue you in terms of interaction, to seek you out and to combine that with task-based functionalities, a.k.a. a robot butler or other sort of utility.
All of this bears a lot of thinking as we move through 2025, and it's a good example of a comprehensive survey of where we're at with artificial intelligence right now.