“Unprecedented” – that’s the word frequently used by venture capitalist Mary Meeker—once known as the ‘Queen of the Internet’—in her latest trends report on artificial intelligence (AI) development and adoption.
The 340-page report, titled ‘Trends — Artificial Intelligence,’ charts out the speed at which costs of usage are dropping, and how its adoption curve is unlike any tech disruption of the past. “The pace and scope of change related to the artificial intelligence technology evolution is indeed unprecedented…” Meeker writes in her report, her first major trends report since 2019.
While largely upbeat about AI’s disruptive promise, the report also outlines cautions against well-known pitfalls including hallucinations, biases, misinformation and slow moving regulation. It also said that while AI platforms have racked up the user-base, revenue per user is still quite low for most of them, with a median of $23.
The adoption of AI platforms has been unlike anything that has come before it, the report said. For instance, it took the likes of Instagram, WhatsApp, and YouTube between 2-4 years to reach 100 million users, but for ChatGPT, it took less than 3 months.
The report also speculated, based on data from Morgan Stanley, that while it took between 6-12 years for 50% households in the US to have access to mobile and desktop internet, it will take only 3 years for the same number of households to become users of AI platforms.
Owing to its large demography and internet penetration, India has been a key user-base market for AI companies, the report said. It is the second largest market for ChatGPT, and contributes the highest percentage of its mobile app users (13.5%), ahead of countries like the US (8.9%), and Germany (3%).
India is also the third-largest user base (6.9%) for China’s homegrown platform DeepSeek, and is behind only China (33.9%) and Russia (9.2%). However, the thing to note here is that ChatGPT, one of DeepSeek’s main rivals, is banned in both China and Russia. Indians therefore contribute a substantial user base to DeepSeek, despite the availability of its Western rivals.
The report said that two different philosophies in shipping AI models are playing out in parallel – closed and open source.
Closed models follow a centralised, capital-intensive arc. These models – like OpenAI’s GPT-4 or Anthropic’s Claude – are trained within proprietary systems on massive proprietary datasets, requiring months of compute time and millions in spending, it said. They often deliver more capable performance and easier usability, and thus are preferred by enterprises and consumers, and – increasingly – governments. However, the tradeoff is opacity: no access to weights, training data, or fine-tuning methods, the report added.
Meanwhile, platforms like Hugging Face have made it frictionless to download open source models like Meta’s Llama or Mistral’s Mixtral, giving startups, academics, and governments access to frontier-level AI without billion-dollar budgets. “And China (as of Q2:25) – based on the number of large-scale AI models released – is leading the open-source race, with three large-scale models released in 2025 – DeepSeek-R1, Alibaba Qwen-32B and Baidu Ernie 4.5,” it said.
“The split has consequences. Open-source is fueling sovereign AI initiatives, local language models, and community-led innovation. Closed models, meanwhile, are dominating consumer market share and large enterprise adoption. We’re watching two philosophies unfold in parallel – freedom vs. control, speed vs. safety, openness vs. optimization – each shaping not just how AI works, but who gets to wield it,” Meeker said in her report.
