Published on the 18/06/2025 | Written by Heather Wright

Meeker’s take: It’s reshaping our world forever…

Mary Meeker is optimistic about the future of the world under AI, which she says is already reshaping the world, but she does have a few notes of caution.

Meeker, dubbed ‘the Queen of the Internet’ and famed for her internet trend reports published between 1995 and 2019 which were essential reading for the tech sector, investors and policymakers, says it’s gametime for AI, it’s only going to get more intense and AI isn’t just another technology wave. Instead, she say it’s resetting business and society and touching on all aspects of life: technical, financial, social, physical and geopolitical.

“The value shifts from tools to outcomes – from CRMs to automated deal summaries, from service desks to AI-powered resolution flows.”

Her new weighty 340-page Trends – Artificial Intelligence report comes packed with data, as is usual for Meeker’s work: From the 800 million daily users of US based large language models recorded in April, to the 64 percent growth in capital expenditure, to US$212 billion, by the big six (Nvidia, Microsoft, Alphabet, AWS and Meta) in 2024.

Training costs for models have gone up, often exceeding US$100 million per model (with Anthropic boss Dario Amodei suggesting $10 billion models could be on their way this year).

ChatGPT meanwhile has garnered 800 million weekly active users, and while only 20 million of those have moved to paid subscriptions, the company is already approaching the US$4 billion revenue mark. The company hit 365 billion annual searches in its second year – a feat that took Google 11 years to achieve.

Microsoft Copilot, meanwhile, racked up 15 billion chats between August 2023 and December 2024.

AI use is ‘surging’ among consumers, developers, enterprises and governments, and while the Internet 1.0 revolution started in the US with steady global diffusion, ChatGPT hit the world stage everywhere, all at once growing in most global regions simultaneously.

AI, she says, is a compounder allowing for ‘wicked-fast adoption’ of easy-to-use broad-interest services.

Meeker now runs the US venture capital company Bond, which has almost US$6 billion in investment capital. She rose to prominence in the 1990s while at Morgan Stanley (where she played a key role in Netscape’s 1995 IPO) , and in the 2010s ran capital firm Kleiner Perkins’ growth practice, investing in future big name tech companies including Facebook, Spotify and Canva.

In her first report in six years, Meeker and co-authors, and colleagues at venture capital firm Bond, Jay Simons, Daegwon Chaie and Alexander Krey, talk about a tech revolution moving at ‘unprecedented’ pace and scale. Unprecedented, in fact, is a frequently used term in the report. There’s the unprecedented pace and scope of AI evolution, the unprecedented growth of AI use, the unprecedented capex growth… you get the picture.

“The pace of improvement in AI’s cognitive ability is astounding,” the report says. “In the three years since ChatGPT’s 11/22 public launch, we’ve gone from the reasoning capabilities of a high school student to those of a PhD candidate.”

Relatively new AI company founders have been especially aggressive about innovation, product releases, investments, acquisitions, cash burn and capital raises, the report notes, while more traditional tech companies, have increasingly directed more of their hefty free cash flows into AI to drive growth and fend off attackers.

In discussion of where the AI model development is going, the report notes that in the early days of modern machine learning most models, rooted in academic and collaborative traditions were open source. Roll on to 2019 and the parallel, closed source, movement emerged.

Trained within proprietary systems, on massive proprietary datasets, requiring months of compute time and millions in spending, closed source models such as ChatGPT and Anthropic’s Claude, deliver more capable performance and easier usability, making them popular with enterprises, consumers and increasingly governments. But the report notes that the trade-off is transparency.

Open source models, with lower costs, growing capabilities and broader accessibility for developers and enterprises alike, are now resurging as LLMs mature and competition intensifies.

Platforms like Hugging Face have made it frictionless to download models like Meta’s Llama or Mistral’s Mixtral giving startups, academics and governments access to ‘frontier-level AI without billion-dollar budgets’.

“Open source AI has become the garage lab of the modern tech era: Fast messy, global and fiercely collaborative.”

China is leading the open source race, with DeepSeek R1, Alibaba Qwen 32B and Baidu Ernie 4.5 all released this year.

“We are watching two philosophies unfold in parallel – freedom versus control, speed versus safety, openness versus optimisation – each shaping not just how AI works, but who gets to wield it.”

The report also sees a future of horizontal enterprise platforms, combining AI-native productivity, search, communication and knowledge management into a unified platform, and changing payment models.

“Think of it as Slack meets Notion meets ChatGPT, all in one platform.”

Vendors may charge not for siloed software licences, but for intelligence embedded throughout the stack.

“The value shifts from tools to outcomes – from CRMs to automated deal summaries, from service desks to AI-powered resolution flows.”

Incumbents and upstarts alike are already embracing the model: Microsoft is integrating Copilot across the stack, Zoom and Canva are layering GenAI into user-facing workflows and Databricks is infusing GenAI into its data and developer stack, while startups bet on AI-first workflows to challenge the suite model.

Specialist vendors, meanwhile are absorbing AI fast, embedding copilots, automating workflows and fine-tuning models on proprietary industry data.

With their workflows, trust and structured data already in place, they have a headstart in deploying domain-specific intelligence that goes far beyond summarising a meeting to instead flag regulatory risks, optimise pricing in real time or draft regulation compliant documentation.

“In many cases, their incumbency becomes their advantage: They can roll out AI as a feature, not a product, and monetise it without changing the buying motion.”

It may not be a winner-takes-all battle though, but instead convergence with horizontal platforms pushing breadth, stitching together knowledge across functions and specialists pushing depth, delivering AI that speaks the language of compliance, contracts and customer intent.

Meeker is clear the technology is changing the way we work and has widespread ramifications, with both physical automation such as robots and drones, and cognitive automation where AI systems can reason, create and solve problems.

But while some see an ‘agentic future’ where AI agents replace humans in many white-collar jobs, Meeker’s more positive saying while that is possible, history and pattern recognition suggest the role of humans is ‘enduring and compelling’.

“Technology-forward leaps have typically driven productivity and efficiency gains and more – but new – jobs. That said, this time it’s happening faster.”

In an extreme, entirely agentic future, humans would pivot towards oversight, guidance and training.

“Imagine facilities filled with humans teaching robots intricate movements or offices full of workers providing reinforcement learning human feedback to optimise algorithms. This is not conjecture. Companies like Physical Intelligence and Scale AI, respectively, are

building powerful businesses based on this view of the world.

“The idea of the human workforce re-configured to teach and refine machines as a primary function might sound dystopic, but it’s worth remembering historical parallels. Fifty years ago, this prospect of rows of cubicles and uniformed office workers sitting quietly in front of LED computers 10 hours a day likely sounded equally dystopic. Yet here we are.

“Technology has constantly redefined and evolved the nature of work and productivity… AI is no different.”

Meeker continues: “As investors, we always assume everything can go wrong, but the exciting part is the consideration of what can go right. Time and time again, the case for optimism is one of the best bets one can make.”