Mary Meeker’s AI Trend-Analysis

Dear Readers,

There are moments when numbers make history. 800 million weekly users in just 17 months - with this sentence, Mary Meeker not only describes the meteoric rise of ChatGPT, but also a fundamental change: we are not experiencing another tech hype, but the development of a new digital infrastructure that is having a profound impact on the economy, politics and society.

In her current 340-page analysis, Meeker outlines a world in which speed has become the currency, developers the strategic capital and efficiency the sharpest weapon in global competition. As investments in data centers and chips pile up, a new phase begins: not of the biggest, but of the smartest models - adapted, targeted, resource-efficient.

This DeepDive takes you through five key trends that make Meeker's report not just a data firework, but a political wake-up call: those who don't understand how power, productivity and innovation are reorganizing now will not be carried by the wave, but overrun by it.

All the best,

Mary Meeker’s AI Trend-Analysis

A single sentence from Mary Meeker's latest “Trends - Artificial Intelligence” report sums up the drama of the moment: “ChatGPT reached 800 million weekly users in just 17 months - it took Netflix more than a decade to gain an eighth of that.”

Rarely before has a technology shown such a steep diffusion curve; the word “unprecedented” appears 51 times in Meeker's 340-page set of slides. But speed alone does not explain the hype. At the same time, the cost of training large models is falling rapidly, while a new generation of highly efficient special models is being launched. Society, the economy and geopolitics are simultaneously coming under pressure to change.

Against this backdrop, the key question arises: Is the current wave of AI adoption, cost reduction and global competition creating a new basic infrastructure in which it is not size but efficiency, access to talent and domain-specific orientation that determine success and the balance of power?

Automate Prospecting Local Businesses With Our AI BDR

Struggling to identify local prospects? Our AI BDR Ava taps into a database of 200M+ local Google businesses and does fully autonomous outreach—so you can focus on closing deals, not chasing leads.

Ava operates within the Artisan platform, which consolidates every tool you need for outbound:

• 300M+ High-Quality B2B Prospects

• Automated Lead Enrichment With 10+ Data Sources Included

• Full Email Deliverability Management

• Personalization Waterfall using LinkedIn, Twitter, Web Scraping & More

Book a demo and supercharge your sales team

1. Rapid Adoption - more than just an erratic trend

The historically unprecedented user explosion of ChatGPT is not an outlier. According to Meeker, 90 percent of ChatGPT sessions now come from outside North America, and generative AI tools are now reaching the 100 million mark in months rather than years. This is accompanied by a boom in new models: the number of high-performance systems has increased by an average of 167 percent per year since 2020, with large-scale models even increasing by 1 150 percent between 2022 and 2024.

2. Cost Pressure as a Market Driver

The flip side of hyper-adoption is huge training budgets. OpenAI, xAI and Anthropic together received around 95 billion US dollars in capital, but have only generated 12 billion to date. Meeker warns that cheaper alternatives - such as China's DeepSeek - could undercut the top dogs in the medium term. Advances in hardware and better algorithms continue to drive down inference costs; those who fail to scale and optimize risk being relegated to a commodity business.

3. Infrastructure Race - chips, clouds, clusters

Behind the scenes, the race for computing power is picking up speed. The report shows an average performance increase of 150 percent per year for AI supercomputers, driven by chips-per-cluster and performance-per-chip. While Nvidia GPUs are reaching their supply limits, specialized ASIC solutions are emerging from Google, Tenstorrent & Co. The result: a global battle for manufacturing capacities and supply chains that is increasingly charged with security issues.

4. Human Capital - developers as a strategic resource

Meeker quotes Steve Ballmer's mantra “Developers, developers, developers” and links it to the stricter US immigration rules. Europe could benefit here if it integrates talent more quickly. The fact that 75 percent of global CMOs are already testing or using AI tools also illustrates how widely specialists will need AI skills in the future.

5. Practical Examples and Early Movers

J.P. Morgan reports a 65% increase in value between 2023 and 2025 through AI-supported applications in fraud prevention, marketing and internal processes. At the same time, a recent BTOS study shows that only around seven percent of all US companies are using AI productively, but the trend is rising sharply. This illustrates a paradox: operational penetration lags well behind public perception, but has enormous potential to catch up.

Interim result: 2025 marks the transition from experiment-driven size to competition-critical efficiency. Companies that drastically reduce costs and computing requirements while building strong developer ecosystems and focused domain solutions will set the rules of the game. With artificial intelligence, we are seeing the rise of a technology at an unprecedented speed. The adaptation, distribution and increase in efficiency is almost unparalleled. And this trend is likely to continue.

The analysis shows:

• The speed and global breadth of AI adoption go beyond familiar diffusion models.

• Costs and economies of scale are shifting, which is why efficiency strategists can operate on an equal footing with capital colossi.

• Talent is becoming a key resource; migration and education policies will influence market shares in the future.

• AlthoughPraxisimpact is growing rapidly, it is still in the early stages overall - the greatest productivity leverage is yet to come.

This answers the question formulated at the beginning: yes, we are witnessing the emergence of a new basic infrastructure whose success factors are not solely based on gigantic amounts of data. The decisive factors will be how cheaply and purposefully AI performance can be provided, how open platforms are for developers and how cleverly politicians set the framework conditions.

Outlook: Those who invest in specialized, energy-efficient models today, promote open ecosystems and attract talent globally will secure a key role for themselves tomorrow. It remains to be seen whether states and companies can agree on a common value and regulatory framework - or whether fragmented standards will slow down the full potential. The answer to this question will determine whether AI becomes the biggest booster of prosperity since electrification - or just the next fast-moving tech fad.

Sources:

🔗 https://arxiv.org/pdf/2505.14721

🔗 https://www.ndtvprofit.com/technology/mary-meeker-ai-trends-report-mind-boggling-numbers-paint-ais-massive-growth-picture

🔗 https://www.saastr.com/the-top-10-learnings-from-mary-meekers-latest-report-on-ai/?utm_source=chatgpt.com 

🔗 https://www.bondcap.com/reports/tai

As explained in the last DeepDive, we are living in a revolutionary era. Dario Amodei never tires of emphasizing how much we are heading towards a changing world, with major implications for the labour market.

Mary Meeker's data seems to back this up. ChatGPT is the fastest adapting technology ever. Within a few days, millions of people have tried ChatGPT for the first time and now, according to the latest surveys, AI is used in almost 50% of all work activities in the world of work.

AI is here to stay and is constantly evolving. It is a revolution - empirically verifiable. The only question is what conclusions we as a society draw from it for our coexistence.