Compute now costs money on every AI subscriber, squeezing margins. At MAU, Phil Carter argues web monetization's real value isn't just the App Store fee, it's the speed, cash flow, and conversion too.
The new unit economics of consumer subscription apps
I’ve been saying for a while that consumer subscription apps are easy to launch but hard to scale. The first part of that has become even more true recently – AI is bringing down the cost of software development, technical barriers to founding a company are eroding, and niche apps that would have been economically unviable a few years ago are now finding real audiences.
The second part, the scaling part, is becoming harder still. And a lot of that comes down to unit economics.
The marginal cost problem
For most of the history of the consumer subscription app, the marginal cost of an additional subscriber was effectively zero. Cloud infrastructure costs were negligible at a small scale, and fraud and chargeback exposure was manageable. The main costs were acquisition and the platform fee – both significant, but predictable.
That assumption has broken down for AI-native apps. If you’re serving AI-powered features, you’re paying for compute on every interaction. Every subscriber has a real marginal cost attached to them. And that changes the math across the entire business model.
It means LTV calculations need to account for costs that scale with engagement, not just with subscriber count. It means pricing needs to be more sophisticated – flat monthly fees may not accurately reflect value delivered or cost incurred. And it means every efficiency gain, whether in acquisition costs, platform fees, or monetization conversion, compounds in ways it simply didn’t before.
This is the context in which the conversation about web monetization needs to be understood. It’s not primarily about saving 30 cents on the dollar by bypassing the App Store. It’s about building a business with sustainable unit economics in an environment where margins are genuinely under pressure.
What web monetization actually unlocks
Beyond the fee saving, there are three underrated benefits of selling on the web that I think most app companies aren’t fully accounting for.
- Experimentation speed
Releasing a pricing change or paywall redesign through the App Store requires a new build and approval cycle. On the web, you can ship a change today and have data by tomorrow. For companies that take monetization experimentation seriously – and more should – this is a significant structural advantage. - Paid acquisition efficiency
If you’re running performance marketing on Meta or Google and your conversions happen inside the App Store, you’re passing probabilistic attribution data back to the networks. Web conversions allow you to pass deterministic data – actual purchase events – which tightens the feedback loop and, on the margins, makes your ad spend more efficient. In a world of rising CAC, those margins matter. - Cash flow
App Store payouts are slow – 60 days is typical, and in some cases, if chargeback or refund rates spike, payouts can be frozen entirely. I’ve seen companies unable to rely on their own revenue for months due to payout freezes. Web payments are faster and more predictable. For early-stage companies funding growth through paid acquisition rather than venture capital, getting paid faster directly reduces dilution.
The multi-channel future
Multi-channel monetization isn’t a fringe strategy. It’s the direction the market is heading. The companies building expertise here are, in most cases, the ones with the most sophisticated approach to monetization overall.
That matters because it’s not just a payments decision – it’s a competitive one. Companies that can iterate faster on pricing, that have better acquisition efficiency, that aren’t dependent on a single platform’s approval cycles and payout schedules, are going to compound advantages over time that aren’t visible in any single quarter’s P&L.
But multi-channel also introduces real complexity. Fraud management, chargeback exposure, tax compliance across jurisdictions. Then you’ve got subscriber lifecycle management across two channels with different billing mechanics, not forgetting alternative payment methods for global markets. These aren’t insurmountable, but they require thought, tooling, and, in many cases, the right infrastructure partners.
The CalAI lesson, properly understood
The Cal AI App Store removal generated a lot of anxiety in the industry – and some of it, I think, was misdirected.
The lesson from Cal AI isn’t “don’t do App2Web”. The courts have been unambiguous: App2Web is legal in the US, and Apple cannot prohibit it. The lesson is: if you push the boundaries in one area, you invite scrutiny across the board. CalAI’s actual violations, as Apple’s own statement made clear, were related to not offering in-app purchases (which is required) and misleading pricing mechanics – not the existence of a web checkout.
The bigger risk, honestly, is for smaller companies without the scale and relationships that a company like Cal AI has. A large, successful app can survive a few days off the App Store. For an early-stage startup, a months-long ban can be existential. If you’re going to implement App2Web flows, make sure everything else about your monetization experience is clean. You’re buying yourself more scrutiny, so give Apple nothing to find.
For most companies, though, the risk is in the other direction: being too conservative, missing the window, and watching competitors build a compounding advantage in monetization capabilities while you’re still treating web as an afterthought. I think full credit to CalAI for being willing to push the envelope. It's incumbent on some industry leaders to test the margins of the system – and the industry is clearer as a result.
