AI Business Brief

In any product cycle, there are two curves to follow. The first is what your underlying technology can do. The second is what your users can actually do with it. They are not the same curve, and the gap between them — what the platform makes possible versus what the product surfaces — has a […]

For three years, the cleanest sentence in the AI economy has been: everyone pays NVIDIA — as explored in the economics of AI compute infrastructure — . Frontier Labs pays it. Hyperscalers pay it. Enterprises pay it through their cloud bills. Startups pay it through API costs that ultimately route to it. The single most

When Uber had to cap employee AI spending after burning through its budget in just four months, it exposed a brewing crisis in how companies are pricing internal AI adoption. The ride-sharing giant’s predicament reveals a fundamental shift in enterprise software economics that could reshape how tech companies structure their AI business models. Uber’s AI

For the past eighteen months, The Business Engineer has been making one consistent structural argument about the agentic AI race: in the next two to three years, the durable competitive moat will not be the frontier model — as explored in the intelligence factory race between AI labs — itself — it will be control

Microsoft’s Project Solara—an Android variant designed for AI agents rather than traditional apps—signals a fundamental shift in how Big Tech plans to monetize mobile computing. While Google and Apple extract 30% from every app store transaction, Microsoft is betting on a post-app future where AI agents handle tasks directly, bypassing app stores entirely. The $100

In late 2024, Boris Cherny — newly arrived at Anthropic from Meta — as explored in the interface layer wars reshaping consumer tech — — received an instruction from his manager, Ben Mann. Mann is an Anthropic co-founder, helped architect GPT-3 at OpenAI, and ran the company’s product engineering function at the time. The instruction

While tech headlines focus on Nvidia’s chip dominance, the real story is three radically different business models converging on the same $200 billion prize: AI agent computing. Nvidia’s hardware-centric approach, Microsoft’s platform strategy, and Google’s services model are about to collide in ways that will reshape how enterprise computing generates revenue. The Three Business Model

For three years, the frontier was a race. The lab that shipped the best model fastest captured the cycle. That game ended in April. Three governance postures have crystallized into commercial categories — and they do not converge. They separate. But the choice of posture is the surface of a deeper picture: three hidden games

Over the last year and a half, I’ve argued that the best mental model for understanding where AI is headed is to see the current wave of AI agents as a transitional phase. Today, agents are learning how to use computers built for humans. But that is not the end state. The real destination is

While Google’s Gemini Spark promises 24/7 AI assistance and startups offer “free” home cleaning in exchange for robot training data, a deeper business model war is emerging. The real question isn’t whether these AI services work—it’s whether companies can sustain giving away premium services to feed their data collection engines. Google’s Invisible Revenue Engine vs.

A decade ago, I started mapping the economics of AI on a blog called FourWeekMBA. Back then, the map was small, and the cycle was slow. Three years ago, when ChatGPT — as explored in the intelligence factory race between AI labs — crossed into the consumer mainstream, the map became a quarterly artifact. This