AI Eats The World

Platform Shifts

Every 10–15 years, a platform shift reshapes technology. We are currently living through the deployment phase of Generative AI, marked by massive capital expenditure, uncertainty, and the search for value.

I'm riffing on former a16z Partner Benedict Evans' 2025 update, AI eats the world.

The Next Platform Shift

We have seen this pattern before. Mainframes, PCs, the Web, and Smartphones each represented a sigmoid curve of adoption. Now, Generative AI attempts to start a new curve. However, unlike previous shifts where we knew the physical limits (screen size, bandwidth), we do not yet know the ceiling of LLM capability.

What happens in a platform shift?

Who is affected, and how much?

Dominance in one era does not guarantee survival in the next. Microsoft dominated the PC era with a 90%+ share of computing, but when the center of gravity shifted to smartphones, its share of total computing devices collapsed.

Microsoft OS Share

Furthermore, early leaders often disappear. The "first" to market was not the winner in PCs, browsers, search, social, or smartphones. We should expect similar turbulence in the AI model layer.

PC Market Share

How will the new thing work? We don't know

For every new platform, we forget how many ideas failed and how unclear everything was.

Similar to how the internet eventually became the default medium for meeting partners, AI adoption may follow a slow-then-explosive trajectory.

This often leads to bubbles.

But when the dust settles, the world has changed

US heterosexual couples who met online, by year of meeting

The internet has gone from the New Thing to a basic part of daily life.

New platforms mean new tools (and new revenue)

SaaS means the typical large enterprise in the USA now uses 4-500 apps

One way this platform shift is different, though

For PCs, the web or smartphones, we knew the physical limits of what could happen next year

With LLMs, we don’t know how much better this could get 🤯

Another platform shift, or more?

We know this will get better, but we don’t know how much

So how will the new thing turn out?

If this is ‘only’ as big as mobile or the internet, that seems like enough

2. The Capex Boom

The race for AGI has triggered an investment cycle without precedent. The "Big 4" (Microsoft, Meta, Alphabet, Amazon) are projected to spend over $330 billion in 2025 alone on AI infrastructure. This level of spending is driven by FOMO and the existential threat of missing the next platform.

"The risk of under-investing is significantly greater than the risk of over-investing."

— Sundar Pichai, Q2 2024

FOMO drives a capex surge

~$400bn in 2025 for the big four alone (for comparison, global telecoms is ~$300bn).

Big Tech Capex

"The very worst case would be that we have just pre-built for a couple of years."

— Mark Zuckerberg, Q3 2025

This spending is reshaping the physical world. In the US, construction spending on data centers is rapidly approaching spending on office buildings.

A new investment cycle

US data centre construction is overtaking office construction.

US Construction Spending

Nvidia can’t keep up

Nvidia vs Intel

This demand flows to one bottleneck: Nvidia. While Intel struggles to maintain relevance, Nvidia's revenue has exploded, decoupling entirely from traditional semiconductor cycles.

TSMC can't keep up either

TSMC unwilling/unable to expand capacity fast enough to meet Nvidia’s book

Quarterly revenue ($bn)

The Constraints

We are seeing wild estimates for future capacity. Some forecasts suggest we will need to triple global data center capacity by 2030, costing trillions of dollars. However, real-world constraints like power, chips, and permits are becoming major bottlenecks.

Utility power availability is now the primary bottleneck in the US, where power demand growth has historically been flat (around 2%) but is now spiking due to AI.

US power backlogs becoming a major issue

US power demand growth is ~2%, and AI might add 1% that’s hard to build fast (this is not an issue in China)

Main constraints to data centre construction, USA (February 2025)

"It’s been almost impossible to build capacity fast enough since ChatGPT launched"

— Kevin Scott, Microsoft CTO

Data Center Capacity

“Three trillion dollars!”

Annualised AI capex aspirations are a similar magnitude to mature global capital-intensive industries

The hyperscalers can afford it…

The hyperscalers have massive cash flows, which generally cover their capital expenditures.

Cash Flow vs Capex

… up to a point

However, when we include leases, the total burden is much higher, especially for Amazon and Microsoft.

Annual Capex and Free Cash Flow, 2010 to 2025e ($bn)

Hyperscalers add leases and debt, while some analysts suggest Oracle’s cloud capex might be >100% of revenue.

Circular Revenue

Without its own cashflows, OpenAI partners with Nvidia, Oracle, Softbank, and petrodollars.

OpenAI is buying Nvidia chips with Nvidia’s cashflow…

Which comes from the hyperscalers…

and using Nvidia’s cash to turn AMD into an Nvidia competitor, and pay Broadcom to make its own chips…

Yes, but where has all this money got us?

After three years, lots more science and engineering, but no real clarity on the shape of the market

Model Proliferation

Meanwhile, the models themselves are converging. On general benchmarks, the gap between the leading model and the followers is shrinking, suggesting that raw intelligence might become a commodity.

Models converge and leaders change weekly

Dozens of (saturated) benchmarks to choose from, but on the most general, the leaders are very close.

Weekly active users of generative AI tools as share of the population, June 2025

Adoption is also uneven globally. The US leads significantly in the usage of generative AI tools, while developed markets like Japan lag significantly behind.

3. Usage and Adoption

Despite the hype, we are early. ChatGPT has roughly 800 million weekly active users, but usage intensity varies. Surveys suggest that for many, AI is a "sometimes" tool rather than a daily habit.

ChatGPT Growth

When we look closer at engagement, we see a "gap." While weekly usage is climbing, daily usage remains a smaller fraction. The technology hasn't yet become invisible and essential for everyone.

Engagement Gap

Survey Meta-Analysis

When the dust settles, however, new technologies fundamentally change society. The internet went from a novelty to the primary way couples meet, displacing friends and family.

How Couples Meet

Vanity Metrics

Beware of big numbers. Companies love to report "Tokens Generated" to show growth. But this is like reporting "Bandwidth" in 1996. It's a cost, not a value.

If you have a loop that generates a million tokens of garbage, you haven't created value. You've just burned electricity.

Tokens vs Value

4. Automation and Unbundling

AI is often described as "infinite interns." This leads to the Jevons paradox: as the cost of intelligence drops, we will consume vastly more of it. We saw this with the steam engine, and we saw it with retail automation.

Steam Engine Labor

Automation also allows us to handle complexity. The introduction of barcodes in the 1970s allowed supermarkets to handle 5x more SKUs than before.

Average SKUs per supermarket, USA

We are also seeing the unbundling of software. Startups are flooding into every vertical, attempting to unbundle Excel, Google, and traditional SaaS using AI.

YC Batches

5. Enterprise Reality

How do you know what to automate? Step one: ask your systems integrator. Accenture and other consultancies are seeing an explosion in "Generative AI" bookings as enterprises struggle to figure out use cases.

Microsoft's capital intensity is spiking as it builds out the infrastructure for this new era.

Microsoft Capex Intensity

Accenture Bookings

Deployment takes time. A survey of CIOs shows that while many have launched "pilots," full production deployment is a multi-year journey. 40% don't plan significant deployment until 2026 or later.

CIO Deployment Timing

Usage varies significantly by function. IT, Knowledge Management, and Marketing are leading the way, while Manufacturing and Supply Chain lag behind.

AI Agent Adoption

This is normal. E-commerce took 25 years to reach ~30% of US core retail sales. The "old stuff" sticks around for a long time.

E-commerce Share

But the future takes time. Cloud computing is old and boring news, yet after more than a decade, only about 30% of enterprise workloads have actually moved to the public cloud.

Cloud Adoption

6. Value Capture

Where does the money go? Currently, brands spend a trillion dollars a year to talk to consumers. Search and Social dominate this, but AI agents promise to disrupt how products are found and bought.

Global Ad Revenue

We are already seeing generational shifts in how people find information. Younger users are much more likely to use Generative AI as a primary search tool compared to older generations who stick to traditional search.

Search Preference

And software companies are benefiting. Palantir has seen revenue re-accelerate as governments and enterprises scramble to deploy AI on their data.

Palantir Revenue

Physical AI

While we focus on LLMs in the browser, the physical world is also being eaten. After a decade of broken promises and billions in burnt capital, "automatic cars" (robotaxis) are finally starting to work at scale.

Monthly ‘robotaxi’ trips in California

But general-purpose robots are harder. This is "Moravec's Paradox": high-level reasoning (chess, poetry) requires very little computation, but low-level sensorimotor skills (walking, holding an egg) require enormous computational resources.

We are seeing progress in "Embodied AI"—robots that learn to move by themselves—but we are still far from a Rosie the Robot in every home.

7. Regulation & Society

It's not just about business. AI is colliding with society. We see a split between "Safety" (bias, hallucination, legal liability) and "Doomerism" (existential risk).

Evans dismisses the "Doomer" view—that AI will kill us all—as "childish," preferring to focus on the tangible, immediate risks of deployment.

Copyright & Legal

The legal battles are just beginning. The NYT vs OpenAI case is a bellwether for how we treat training data. Is it fair use? Or is it theft?

8. Models

The scaling laws hold. We are seeing exponential growth in model size and capability. The number of parameters in leading models has grown from 100 million to over a trillion in just a few years.

Model Parameters

This scale comes at a cost. Training costs are also rising exponentially. What cost $50k in 2018 now costs hundreds of millions, with billion-dollar training runs on the horizon.

Training Costs

9. Applications

What are we actually doing with this? While chat is the default interface, users are finding value in creative writing, coding assistance, and summarization.

GenAI Use Cases

New tools like Cursor are seeing explosive growth, challenging traditional IDEs.

Cursor Growth

Coding Assistant Share

10. So what?

We are in the deployment phase. The technology works, the capital is flowing, and the models are getting better. Now comes the hard work of figuring out the product-market fit, the business models, and the societal impact.

As with the internet and mobile, the real revolution isn't the technology itself, but what we build with it.

Consider the elevator operator. In 1950, there were over 100,000 elevator operators in the US. Automation (buttons) didn't just remove the job; it made the elevator invisible.

Elevator Operators

Otis launched the ‘Autotronic’ automatic elevator in 1950.