This article puts together the annual AI Trends Report from Bond Capital, a unit of Mary Meeker, the queen of the web, to paint an in-depth picture of how AI technology is reshaping the world at an astonishing rate. (Synopsis: Hon Hai internal experiment: AI can replace 80% of jobs, chairman Liu Yangwei exposed the trinity blueprint of the future factory) (Background supplement: Podcast essence: AI and robots cover the cryptocurrency boom, the next era of micro-entrepreneurship is the strongest) Mary Meeker, who has the reputation of "Queen of the Internet" (Queen of the Internet), The company he founded, Bond Capital, is known worldwide for publishing online trend reports over the years. BOND's latest "Trends – Artificial Intelligence" report recently painted a 340-page deep picture of how AI technology is reshaping the world at an astonishing rate. From disruptive changes in user habits, unprecedented scale of capital expenditures, to intense business models and far-reaching implications for the physical world and the future of work, the astonishing acceleration of AI development is revealed. Due to the large amount of content in the report, this is the first of three articles that will be published after focusing on AI for you. A new "year of the dog" in AI technology, and even faster The Bond report opens with a compelling tone, based on Venter Cerf (Vint Cerf)'s observation in 1999 that "the year of the internet industry is like the year of the dog", BOND further states that the rate of evolution of AI "is climbing dramatically at an even faster rate... And the capabilities of the machines themselves are surpassing us." Bond said that this unprecedented pace of development is not empty talk, but is supported by a large number of user, usage and revenue charts that continue to grow to the upper right corner, as well as spending charts that also grow to the upper right corner. This is not only the accumulation of quantity, but also a qualitative leap, indicating that a new technological paradigm is taking shape. The ubiquity of the global network infrastructure (reaching 5.5 billion citizens), the massive digital data set accumulated over decades, and the emergence of breakthrough large-scale language model (LLM) – especially OpenAI's ChatGPT has rapidly gained popularity with its "extremely easy to use and responsive user interface" since its launch in November 2022. The BOND report highlights that the world is undergoing change at an unprecedented rate, and AI is the core engine of this change. Hailed as "the most successful overnight hit in history (nine years after its founding)," ChatGPT's influence, unlike the gradual proliferation of the early Internet, swept the world almost instantaneously. This phenomenon has ignited a fierce competition between incumbents and emerging challengers as they scramble to build next-generation AI infrastructure, including more autonomous agent interfaces and sovereign AI models. The interweaving of technological progress and geopolitical forces is an important theme throughout the report. Is change happening faster than ever? The answer is yes. The Compounding Effect of Technology Millennium Dimension: Looking back at global GDP growth over the past thousand years, it is clear that there is a strong correlation between major technological advances such as printing, the steam engine, electrification, networking, and even cloud computing, and significant economic expansion. Each wave of technology lays the foundation for subsequent development and shows an accelerated trend. Scale continues to expand: The development cycle of computers continues to shorten, but the scale continues to expand. This exponential growth is clearly visible from the early million-scale mainframes to the potential market of "tens of billions of devices" targeted by today's AI era. The technical cornerstone of this process has also evolved from the initial CPU to the combination of big data and cloud computing, and then to the current era of computing power with GPUs as the core. Multi-dimensional explosion of AI: Training data scale: Over the past fifteen years, the amount of text data used to train key AI models has grown at an astonishing annual average of 260%. This means that the model is able to learn from a broader, deeper knowledge, which improves its ability to understand and generate. Training Computing Power (FLOPs): What grows in tandem with the scale of data is the computing power required for training. Over the past fifteen years, the computing power (measured in FLOPs) used to train key AI models has grown by an average of 360% per year, opening up the possibility of larger, more complex models. Algorithm improvement: In addition to the improvement of data and computing power, the progress of the algorithm itself is also crucial. Over the past nine years, smarter algorithms have led to an average annual increase in effective computing power of 200%, which means the same or even better results can be achieved with fewer resources. AI supercomputer performance: The performance of dedicated AI supercomputers is also skyrocketing, with performance (measured in FLOP/s) increasing by an average of 150% per year over the past six years. Number of powerful AI models: The number of newly released large-scale AI models (those that require more than 10^23 FLOPs to train) has grown by an average of 167% per year over the past four years. ChatGPT: The epitome and catalyst for acceleration: The report cites the rise of ChatGPT as a landmark event in this round of AI acceleration, with impressive growth in user numbers (800 million weekly active users by April 2025), subscribers, and revenue. ChatGPT reached an average of 365 billion searches per year in just 2 years (by 2024), while Google reached this milestone in 11 years (by 2009), a full 5.5 times faster. This contrast highlights the rapid iteration and market penetration of AI technology. This fully illustrates how artificial intelligence can be used as a "compound interest engine on network infrastructure" to achieve explosive growth on the basis of existing technologies. The profound evolution of the way knowledge is disseminated: The report divides the history of knowledge dissemination into three stages: from the "static + physical dissemination" stage represented by printing from 1440 to 1992 to the "dynamic + digital dissemination" stage marked by the network from 1993 to 2021 and then to the present, a new era of "dynamic + digital + generative dissemination" represented by generative artificial intelligence such as ChatGPT. This process of evolution is not only an innovation in technological means, but also a fundamental change in the way information is obtained, processed and created. The report cites Martin M. H. Fischer (Martin H. Fischer) famously said: "Knowledge is the process of piling up facts; Wisdom lies in simplifying them", especially in the age of generative AI, which gives new life to knowledge by simplifying complex information. The evolution of artificial intelligence: important milestones and expanding capability boundaries: The report traces the Turing test proposed by Alan Turing in 1950 with a detailed "artificial intelligence milestone timeline" to the prediction of 2025. Covers topics such as the "AI Winter" (1967-1996), the triumph of the Deep Blue computer, the popularity of the Roomba robot vacuum, the birth of the Siri voice assistant, iterations of various GPT series models, and data from Anthropic, Meta, Google, and Ali...
The content is for reference only, not a solicitation or offer. No investment, tax, or legal advice provided. See Disclaimer for more risks disclosure.
Internet queen Mary Meeker 340-page AI Depth report ( on ): How artificial intelligence captures user mindshare at an unprecedented speed
This article puts together the annual AI Trends Report from Bond Capital, a unit of Mary Meeker, the queen of the web, to paint an in-depth picture of how AI technology is reshaping the world at an astonishing rate. (Synopsis: Hon Hai internal experiment: AI can replace 80% of jobs, chairman Liu Yangwei exposed the trinity blueprint of the future factory) (Background supplement: Podcast essence: AI and robots cover the cryptocurrency boom, the next era of micro-entrepreneurship is the strongest) Mary Meeker, who has the reputation of "Queen of the Internet" (Queen of the Internet), The company he founded, Bond Capital, is known worldwide for publishing online trend reports over the years. BOND's latest "Trends – Artificial Intelligence" report recently painted a 340-page deep picture of how AI technology is reshaping the world at an astonishing rate. From disruptive changes in user habits, unprecedented scale of capital expenditures, to intense business models and far-reaching implications for the physical world and the future of work, the astonishing acceleration of AI development is revealed. Due to the large amount of content in the report, this is the first of three articles that will be published after focusing on AI for you. A new "year of the dog" in AI technology, and even faster The Bond report opens with a compelling tone, based on Venter Cerf (Vint Cerf)'s observation in 1999 that "the year of the internet industry is like the year of the dog", BOND further states that the rate of evolution of AI "is climbing dramatically at an even faster rate... And the capabilities of the machines themselves are surpassing us." Bond said that this unprecedented pace of development is not empty talk, but is supported by a large number of user, usage and revenue charts that continue to grow to the upper right corner, as well as spending charts that also grow to the upper right corner. This is not only the accumulation of quantity, but also a qualitative leap, indicating that a new technological paradigm is taking shape. The ubiquity of the global network infrastructure (reaching 5.5 billion citizens), the massive digital data set accumulated over decades, and the emergence of breakthrough large-scale language model (LLM) – especially OpenAI's ChatGPT has rapidly gained popularity with its "extremely easy to use and responsive user interface" since its launch in November 2022. The BOND report highlights that the world is undergoing change at an unprecedented rate, and AI is the core engine of this change. Hailed as "the most successful overnight hit in history (nine years after its founding)," ChatGPT's influence, unlike the gradual proliferation of the early Internet, swept the world almost instantaneously. This phenomenon has ignited a fierce competition between incumbents and emerging challengers as they scramble to build next-generation AI infrastructure, including more autonomous agent interfaces and sovereign AI models. The interweaving of technological progress and geopolitical forces is an important theme throughout the report. Is change happening faster than ever? The answer is yes. The Compounding Effect of Technology Millennium Dimension: Looking back at global GDP growth over the past thousand years, it is clear that there is a strong correlation between major technological advances such as printing, the steam engine, electrification, networking, and even cloud computing, and significant economic expansion. Each wave of technology lays the foundation for subsequent development and shows an accelerated trend. Scale continues to expand: The development cycle of computers continues to shorten, but the scale continues to expand. This exponential growth is clearly visible from the early million-scale mainframes to the potential market of "tens of billions of devices" targeted by today's AI era. The technical cornerstone of this process has also evolved from the initial CPU to the combination of big data and cloud computing, and then to the current era of computing power with GPUs as the core. Multi-dimensional explosion of AI: Training data scale: Over the past fifteen years, the amount of text data used to train key AI models has grown at an astonishing annual average of 260%. This means that the model is able to learn from a broader, deeper knowledge, which improves its ability to understand and generate. Training Computing Power (FLOPs): What grows in tandem with the scale of data is the computing power required for training. Over the past fifteen years, the computing power (measured in FLOPs) used to train key AI models has grown by an average of 360% per year, opening up the possibility of larger, more complex models. Algorithm improvement: In addition to the improvement of data and computing power, the progress of the algorithm itself is also crucial. Over the past nine years, smarter algorithms have led to an average annual increase in effective computing power of 200%, which means the same or even better results can be achieved with fewer resources. AI supercomputer performance: The performance of dedicated AI supercomputers is also skyrocketing, with performance (measured in FLOP/s) increasing by an average of 150% per year over the past six years. Number of powerful AI models: The number of newly released large-scale AI models (those that require more than 10^23 FLOPs to train) has grown by an average of 167% per year over the past four years. ChatGPT: The epitome and catalyst for acceleration: The report cites the rise of ChatGPT as a landmark event in this round of AI acceleration, with impressive growth in user numbers (800 million weekly active users by April 2025), subscribers, and revenue. ChatGPT reached an average of 365 billion searches per year in just 2 years (by 2024), while Google reached this milestone in 11 years (by 2009), a full 5.5 times faster. This contrast highlights the rapid iteration and market penetration of AI technology. This fully illustrates how artificial intelligence can be used as a "compound interest engine on network infrastructure" to achieve explosive growth on the basis of existing technologies. The profound evolution of the way knowledge is disseminated: The report divides the history of knowledge dissemination into three stages: from the "static + physical dissemination" stage represented by printing from 1440 to 1992 to the "dynamic + digital dissemination" stage marked by the network from 1993 to 2021 and then to the present, a new era of "dynamic + digital + generative dissemination" represented by generative artificial intelligence such as ChatGPT. This process of evolution is not only an innovation in technological means, but also a fundamental change in the way information is obtained, processed and created. The report cites Martin M. H. Fischer (Martin H. Fischer) famously said: "Knowledge is the process of piling up facts; Wisdom lies in simplifying them", especially in the age of generative AI, which gives new life to knowledge by simplifying complex information. The evolution of artificial intelligence: important milestones and expanding capability boundaries: The report traces the Turing test proposed by Alan Turing in 1950 with a detailed "artificial intelligence milestone timeline" to the prediction of 2025. Covers topics such as the "AI Winter" (1967-1996), the triumph of the Deep Blue computer, the popularity of the Roomba robot vacuum, the birth of the Siri voice assistant, iterations of various GPT series models, and data from Anthropic, Meta, Google, and Ali...