The Bull Case for OpenAI

Building THE AI platform

Thus far most of the discussion about AI has been centered around AI as a productivity enhancer. Tools like GitHub Copilot and Salesforce Einstein have helped shape this discussion as they highlight just how much more productive workers can be when they leverage AI. These enhancements are awesome and exciting, but if I’m being honest they haven’t struck me as revolutionary to the scale of the internet or even emblematic of a true platform shift. Rather, in the classical Clayton Christensen sense, these tools feel like sustaining innovations – improvements that better serve existing customers for existing use cases, but do not serve fundamentally new use cases. It’s not surprising we would see sustaining innovations coming out of GitHub and Microsoft – they are, after all, incumbents – but it does mean we’ll likely need to look elsewhere for the true platform shift of AI. For a while I didn’t know where that real shift would come from, but with recent leaks that OpenAI is in discussion with former Apple Chief Design Officer Jony Ive to build the “iPhone of Artificial Intelligence”, I think I finally see it. Let’s dive in. 

What is a platform?

If we’re going to talk about platform shifts, it’s useful to first define what a platform is. To do this I’ll borrow from the book The Business of Platforms which defines a platform as “technology that brings together individuals or organizations so they can innovate or interact in ways not otherwise possible, with the potential for nonlinear increases in utility and value.” Platforms come in two basic types:

1. Transaction Platforms serve as intermediaries for direct exchange – think eBay enabling commerce.

2. Innovation Platforms serve as foundations upon which other technology can be built – think the Sony Playstation enabling game development.

When we talk about platforms in the sense of a platform shift, we’re talking about innovation platforms. The internet, cloud, and mobile were each technological advances that allowed myriad new innovations to be built on top of them, and those innovations collectively created tremendous value.

How big is a platform opportunity?

When we zoom in to the individual company level, it’s clear the businesses that build these innovation platforms capture an outsized share of the value created. Chrome is the base layer of how we interact with the internet, and Google is a $1.5 trillion business. AWS is the dominant cloud infrastructure provider, and analysts attribute the majority of Amazon’s $1.3 trillion valuation to cloud. The iPhone is synonymous with the shift to Mobile, and today Apple is worth $2.6 trillion.

What’s striking, beyond just the massive scale of these businesses, is that Apple’s market value is roughly equal to that of Google and Amazon combined. While these companies are extremely well covered and there’s a lot of nuance to explain their valuations, for the purposes of this post I’d like to focus on one key difference: Apple built THE mobile platform (in the US) while AWS built one of a few cloud platforms and Google was late to the battle for internet browser dominance. Clearly if you can create the base level innovation for a platform shift and then leverage network effects to come to define that platform shift yourself, you can build a multi-trillion-dollar business. This, I believe, is the opportunity for OpenAI.

Apple as a blueprint for OpenAI

To better understand how Apple built the world’s most valuable business and what OpenAI can learn from their success, its useful to study the iPhone as a platform. The iPhone can be thought of as a multi-sided platform with four primary non-Apple constituents:

1. Users. Everyday people like you and me who buy iPhones and use them daily.

2. 3^rd Party Application Developers. Individuals or companies who build the non-Apple applications we use on our iPhones.

3. Accessory creators. Businesses that make the non-Apple cases, chargers, headphones, etc. that we use with the iPhone.

4. Advertisers. Companies who want to advertise a product or service to iPhone users.

Each of these constituents have a symbiotic relationship with Apple whereby they use the iPhone and App Store as a platform to connect with each other and enjoy increasing value as the platform scales. To illustrate this dynamic, consider the figure below:

I won’t explain each of the individual connections here, but what’s clear from this diagram is Apple enjoys A LOT of benefit from being at the center of it all. Just look at all the green arrows flowing money into Apple! While this true for many platform businesses, it’s especially true for Apple for two main reasons:

1. In the iPhone and the App Store Apple has both the defining innovation platform and the defining transaction platform of mobile. Their dominance in devices ensures new applications are created on iOS, and their App Store dominance then ensures they capture a significant share of the value those applications create.

2. Apple established market leadership quickly and has not faltered. The iPhone was launched in 2007 and by 2009 Apple already had over 50% smartphone market share in the US. They’ve consistently retained or added to that share in the years since, and as a result have been able to capture so much of the value of mobile over time.

OpenAI’s potential

If we imagine a future with a next generation device that allows us to more easily interact with AI applications across domains, it’s not hard to see OpenAI at the center of that picture as a multi-trillion-dollar business. What’s more, but with GPT OpenAI has the potential to one-up Apple and create three platforms in one:

• An “AI-phone” and “AI-os” could serve as the base innovation platform for developing AI-enabled applications.

• GPT could then serve as a distinct but complementary innovation platform that actually powers the AI within those applications, allowing developers to leverage OpenAI’s LLM ratably. OpenAI could develop this as an IaaS business model like the hyperscalers model with cloud.

• An “AI store” could then act as a transaction platform connecting users and application developers, of course with OpenAI taking a cut.

Not only does using Apple as a blueprint help OpenAI develop a potentially all-time great business model, but following in Apple’s footsteps actually helps solve the most pressing challenge I see for OpenAI today: usage.

While ChatGPT exploded on the scene to the tune of 1 billion monthly page visits by February, visits peaked at 1.9bn in May of this year and have been declining since. While it’s difficult to be sure what’s driving this, based on my own experience I would argue it’s because AI and ChatGPT are inconvenient to use. While ChatGPT does have a mobile application, most of the time I want to use GPT I’m on my computer which means I have to search for ChatGPT online. If I then want to use a different GPT-powered application I’m again forced to perform a different search, which when compared to the ease of toggling between apps on my iPhone is annoyingly time-consuming.

If OpenAI wants to continue to grow usage and make the most of this innovation platform they’ve already built, they need to make interacting with ChatGPT and other GPT-powered applications as frictionless as possible. I believe following in Apple’s footsteps and developing an “AI-phone” to centralize AI interactions is the right way to do this, and if the Jony Ive rumors are true I’m betting Sam Altman believes this as well. My hope for OpenAI is it’s not just Apple’s hardware genius Altman is looking to replicate, but rather their multi-platform business model as well. If Altman and co. get this right I think there’s a viable path to OpenAI becoming the most valuable business in the world, which is plenty of reason to be bullish in spite of their (rumored) latest $80bn valuation.