Brian Balfour predicts a new distribution platform will emerge within about six months, most likely centred on ChatGPT, and startups should move onto it early.
Balfour's four-step cycle says platforms open to third parties for distribution, then close for control and monetisation, as Facebook's 2007 developer platform did before cutting off organic reach.
Balfour says ChatGPT's moat is context and memory, and he advises startups to make one focused bet and plan their exit early by owning part of the user experience.
Why ChatGPT may be the next big growth platform
Distribution, not product, now separates winners, and platforms repeat the same cycle of opening to developers for growth and then closing to monetise. Brian Balfour believes a new platform will emerge within about six months, most likely ChatGPT, because its context and memory moat and rising retention show escape velocity. Startups should pick one platform, bet on it early, and plan their exit from the start.
Takeaways
Distribution now matters more than product quality. A great product is necessary but not sufficient. Following Alex Rampell, startups win by getting distribution before incumbents can copy them. That is harder now because incumbents copy faster, organic reach on LinkedIn and X has shrunk, and AI makes software easier to build.
Platforms follow a predictable four-step cycle. A category forms, a platform builds its moat with developers, opens to third parties for distribution, then closes to monetise and control. Facebook in 2007, Google search, Apple's 30 percent App Store cut and LinkedIn all followed it, and cycles are getting shorter.
ChatGPT is the likeliest next distribution platform. The moat is context and memory, since model outputs are similar. ChatGPT's retention curves level off high and rise over time, a smile curve. It has at least ten times Claude's users, and OpenAI is hiring for agent-platform roles.
Expect a preferred-partner phase, then an opening. Brian expects an initial 10 to 20 partner companies, then a wider developer platform, shopping and new monetisation to fund free usage. Apple is best placed but least active, Google is second, and Anthropic is focusing Claude on developer tools and coding.
Startups should make one focused bet. Late-stage companies can spread bets, but startups must pick one platform and go all in. Judge platforms on retention and engagement first, then user quality and monetisation, value exchange for developers, and scale. Android has about 70 percent of devices but 30 percent of dollars.
Plan your exit from the platform early. Every platform eventually closes, so companies should own part of the user experience or build specialised data. Udemy's creator revenue share fell from around 80 percent to 25 to 30 percent. Those who fail to predict the last stage end up in hard positions.
Hard constraints drive AI adoption inside companies. Fewer than 10 percent of companies Reforge sees set firm limits, such as each function being one fifth of peer size, and they are farthest along. Executives are disconnected from usage, and slow functions like IT, legal and procurement set the pace.
How to evaluate and join a platform
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
Judge retention and depth of engagement first, not sign-ups.
Check user quality and monetisation, not just device share.
Look at the value exchange the platform offers developers.
Consider scale last.
Build relationships now with the platform's likely preferred-partner picks.
Plan your exit by owning part of the experience or accumulating specialised data.