---
title: "The Bacon \u0026 the Skillet: When Does the AI Market Congeal?"
description: "The AI market is sizzling with innovation, but when will market share stabilize? An exploration of switching costs, model improvements, \u0026 the moment the fat congeals."
categories: ["AI","strategy"]
keywords: ["AI market share","switching costs","AI model innovation","Gemini 3 vs GPT-4","S-curve","software moats","AI strategy","market consolidation","activation energy","category creation"]
ai_summary: "The AI market is currently fluid with rapid innovation and shifting leadership, but this \"sizzling\" phase will eventually cool as model improvements follow an S-curve. When performance gains become incremental, switching costs and ecosystem lock-in (\"congealing fat\") will replace raw model performance as the primary competitive moat."
date: 2025-11-21
lastmod: 2026-07-28
canonical_url: https://www.tomtunguz.com/the-bacon-and-the-skillet-when-ai-market-share-congeals/
author: "Tomasz Tunguz"
---


The AI market today is bacon in a hot skillet. Everything is sizzling, moving, & changing at an incredible pace. We're all watching it closely.

Market share is fluid because no one yet knows what AI can do & the second we think have grasped it, models improve. The [Nvidia chip performance & the launch of Gemini 3](https://tomtunguz.com/gemini-3-proves-pretraining-scaling-laws-intact/) the biggest gain ever in Google model performance suggest no simmering ahead.

As long as the underlying models hurtle towards PhD level performance, people will continue to test. How much better is Gemini 3 at coding? tool calling? writing?

If the progress is material, then the benefit of switching is worth [the activation energy](https://tomtunguz.com/activation-energy-switching-costs/).

{{< email_image src="activation_energy" alt="Activation energy diagram showing the effort required to switch between AI tools" width="540" height="360" >}}

Today, startups, incumbent software companies, cloud providers & AI labs all are competing. First the model, then infrastructure (memory & retrieval), then tools, then applications. Will the foundational models play at the application layer? Or will the applications differentiate themselves enough to overcome model differences?

Who can take advantage of the next big leap in model performance fastest? Which sales team can reach the target customers first & write the RFP?

This is the [Great Game of Risk in Category Creation](https://tomtunguz.com/category-creation-speed-strategy/) & aggression wins.

But this era of fluidity won't last forever. The rate of improvement in AI models will eventually attenuate. When the performance gap between the best model & the second-best model shrinks, the incentive to switch evaporates.

Switching costs will start to matter more than marginal performance gains. The custom tools I’ve built, the muscle memory I’ve developed, the integrations my company has deployed, the enterprise contracts signed, all inertia.

At that point, the fat begins to congeal.

The winners will be those who use the sizzling phase to build fat worth congealing around.

This fun analogy came up during my conversation with Harry, Jason, & Rory. {{< youtube sWkpLLH_8jQ >}}