If you want to know what's broken about billing infrastructure right now, don't read the press releases. Read the M&A activity.
In the last 24 months, two of the largest payments companies in the world have collectively spent who-knows-how-much acquiring three different billing and subscription specialists. Stripe picked up Lemon Squeezy in 2024 to add Merchant of Record capabilities. Stripe picked up Metronome earlier this year to add usage-based billing. Adyen picked up Orb today, almost certainly for the same reason.
This is not what a healthy stack looks like. This is what a stack looks like when the underlying architecture was built for a different era and the only way to stay current is to acquire the people who actually built for today.
The architecture problem
Subscription billing was a 2010s problem. Usage billing is a 2020s problem. The two are not the same, and the old stacks know it.
Stripe Billing was designed for the SaaS subscription era. Recurring monthly charges, simple tier upgrades, occasional add-ons. It worked beautifully for that. But the moment AI shipped and the industry pivoted to per-token, per-credit, per-call, per-minute, per-action models, the assumptions baked into those original systems started to crack. Hybrid pricing (a base subscription plus metered usage plus prepaid credits plus overage) is now table stakes for any AI-native product, and retrofitting it onto a system designed around flat monthly recurring revenue is a non-trivial engineering problem.
The same is true for Merchant of Record. Cross-border tax, VAT, sales tax in 50 US states, GST in Singapore, BTW in the Netherlands. All of this used to be optional sophistication that you grew into as you scaled. For an AI startup with global users from day one, it's a launch-day requirement. Bolting MoR onto a payments processor that was originally built as a US-first card acquirer is, again, a non-trivial engineering problem.
So instead of rebuilding, the giants are acquiring. Stripe + Metronome solves the usage problem. Stripe + Lemon Squeezy solves the MoR problem. Adyen + Orb solves the usage problem. Each acquisition is the same admission. The original architecture didn't account for this. The cleanest path forward is to buy the people who did.
What this means if you're a builder
You're going to be stitching for the next two years.
Pre-acquisition, you had to evaluate Stripe Billing AND Metronome AND Lemon Squeezy as separate products with separate docs, separate pricing pages, separate APIs, and separate teams. Post-acquisition, you'll evaluate "Stripe" but the underlying systems are still three different products that will take years to fully unify. You'll still be reading three sets of docs. You'll still be wiring up three sets of webhooks. Your accounting team will still be reconciling three different ledgers.
Worst case, the integration ends up where most ambitious M&A integrations end up. The larger product slowly absorbs the smaller one. The smaller product's roadmap stalls during the transition. The team you bought the company for leaves over the next 18 months. You end up with a half-finished Frankenstein that's worse than either piece was on its own.
In the meantime, you spend more time becoming an expert in someone else's product than you do building your own.
And here's the second-order effect that should sharpen the math on this. AI coding tools have made shipping product radically faster. Cursor, Claude Code, Windsurf, Cline, whatever your stack of choice. Developers are shipping in hours what used to take days, and in days what used to take sprints. Which means every hour you spend becoming an expert in someone else's billing API instead of building your own product is worth exponentially more than it was two years ago. The opportunity cost of distraction has gone up at exactly the same time the complexity of the billing stack has. Both curves are moving against you.
Why Subotiz exists
We built Subotiz because the bolt-on era should be ending, not starting.
Subscriptions, usage billing, hybrid pricing, credits, MoR, tax, fraud, retries. These aren't separate problems that should be solved by separate companies and stitched together later. They're the same problem from different angles, and they should be solved in a single AI-native stack that was designed for the modern category of company from day one.
If you're a builder, you shouldn't be evaluating Stripe Billing AND Metronome AND Lemon Squeezy. You shouldn't be evaluating Adyen AND Orb. You shouldn't be evaluating Paddle AND a tax engine AND a fraud tool AND a payment retry tool. You should be evaluating a single platform that handles all of it natively, exposes one API, has one set of docs, has one team behind it, and gets out of your way so you can ship product.
That's what we built. Subscriptions, usage, hybrid, MoR, tax, fraud, dunning. One stack. AI-native. Not a roll-up of acquired companies. A coherent product designed around how modern AI and SaaS businesses actually charge.
Live today on Product Hunt
We're on Product Hunt today. If the bolt-on era frustrates you as much as it frustrates us, we'd love your support. If you're knee-deep in evaluating billing infrastructure right now and want to see what an AI-native alternative looks like, come and talk to us.
The next decade of AI and SaaS won't run on the plumbing of the last one. Let's go.
Adam Mawdesley is CRO North America at Subotiz, an AI-native subscription and usage billing platform with Merchant of Record baked in. Twenty years across PayPal, Stripe, Klarna, and a few others have given him strong opinions on what this layer of the stack should and shouldn't look like in 2026.
