OpenAI's Partnership Blitz Is About Distribution, Not Innovation

OpenAI published a piece about their business model that most people glossed over. They explicitly listed their revenue streams: subscriptions, API, ads, commerce, and compute.

Ads and commerce. From an AI research lab.

The Instacart deal makes this concrete. They're building "Instant Checkout" directly into ChatGPT. You ask the AI what to make for dinner, it suggests a recipe, and you buy the ingredients without leaving the chat. OpenAI presumably takes a cut.

I'll be honest, this is clever. It's also a massive strategic pivot that nobody seems to be discussing. They're not just licensing technology anymore. They're becoming a transaction layer.

Is the technology actually new?

The ChatGPT for Excel announcement mentions GPT-5.4 for "regulated environments." That's a new model designation I hadn't seen before. Whether it's genuinely more capable or just a compliance-focused variant remains unclear.

The actual features (financial modeling, research acceleration, working with Bloomberg and FactSet data) sound useful. But useful for whom? Enterprise customers with existing data subscriptions. This isn't democratizing finance. It's making existing tools slightly more convenient for people who already pay for them.

Not that there's anything wrong with that. But let's call it what it is.

What about Apple?

The Apple partnership is the oldest of these deals and still the most opaque. ChatGPT integration across Apple devices sounds transformative until you remember that Apple has a long history of partnering with companies and then building competing features two years later.

I don't know what the contract terms look like. Nobody outside those two companies does. But if I were OpenAI, I'd be treating Apple as a distribution channel with an expiration date, not a long-term ally.

The real question nobody's asking

Here's what I keep coming back to. OpenAI has, by most accounts, the best large language models in the world. They have massive consumer adoption. They have enterprise customers paying serious money for API access.

So why the urgency? Why announce four major partnerships in quick succession? Why pivot toward ads and commerce?

Maybe it's just smart business diversification. Maybe the board wants multiple revenue streams before the next funding round. Or maybe they're seeing something in their internal metrics that suggests the moat isn't as deep as everyone assumes.

I called a friend who consults for one of their competitors last week. He wouldn't say much, but he did mention that enterprise customers are increasingly asking for "model-agnostic" architectures. They want to swap between OpenAI, Anthropic, and Google without rewriting their systems.

That's the commoditization pattern. That's what these partnerships are trying to prevent.

Whether it'll work? Too early to say. But I'd pay more attention to the deal structures than the feature announcements. The features are marketing. The deals are strategy.

James Chen · 3 hours ago · 5 min