Back to Growth Lab

Stripe just agreed to buy OpenRouter for more than $8 billion.

The obvious interpretation is:

Stripe wants to own more of AI.

I think the more interesting interpretation is:

Stripe believes the value in AI may move away from the models themselves.

OpenRouter doesn't build GPT.

It doesn't build Claude.

It doesn't build Gemini.

Instead, it sits above them.

One API.

More than 400 models.

More than 70 providers.

And increasingly, software can use OpenRouter to decide which underlying model should handle which task.

That distinction matters.

Because whenever an industry produces an explosion of supply, value often starts moving somewhere else.

At first, scarcity lives in the product.

Then everybody builds the product.

The products get better.

Differences narrow.

Switching becomes easier.

And suddenly the question changes from:

“Can I get this?”

to:

“Which one should I choose?”

That second question creates an entirely new market.

Stripe appears willing to pay more than $8 billion to own a bigger piece of that market. OpenRouter had reportedly been valued at just $1.3 billion in May.

That is the part chiropractors should pay attention to.

Because chiropractic may be approaching its own version of this problem.

Chiropractic doesn't have a supply problem

There are chiropractors everywhere.

Thousands of websites.

Thousands of offices.

Hundreds of techniques.

Millions of reviews.

Everyone claims to provide personalized care.

Everyone says they get great results.

Everyone says they treat the root cause.

Everyone has testimonials.

Everyone can buy Facebook ads.

Everyone can build a decent website.

Everyone can post educational videos.

And the average patient has almost no ability to meaningfully distinguish between them.

That creates an interesting inversion.

For decades, chiropractors believed the scarce asset was the care itself.

Become a better adjuster.

Learn another technique.

Buy another piece of equipment.

Improve your report of findings.

Build a prettier office.

Those things matter.

But from the market's perspective, access to chiropractic isn't particularly scarce anymore.

Certainty is scarce.

The patient isn't asking:

Can someone adjust me?

They're asking:

Who should I trust with my problem?

Those are very different businesses.

One is the business of delivering care.

The other is the business of reducing uncertainty.

And I increasingly believe the second is where an enormous amount of value will accrue.

Better chiropractors may actually make this problem worse

Here is the counterintuitive part.

Suppose chiropractic gets substantially better over the next decade.

Education improves.

Technology improves.

Business systems improve.

Clinical knowledge spreads faster.

Bad practices disappear.

Good techniques become more widely adopted.

AI helps doctors make better decisions.

The average chiropractor becomes more sophisticated.

You might assume that makes it easier for great chiropractors to differentiate.

I think there's a good chance the opposite happens.

When everybody gets better, quality becomes harder to use as a differentiator.

This happens in technology constantly.

At first, one company having a certain feature is extraordinary.

Eventually everybody has it.

Then the feature stops being a competitive advantage and becomes table stakes.

The same thing happens with services.

“Great care.”

Table stakes.

“Friendly staff.”

Table stakes.

“Advanced technology.”

Eventually table stakes.

“Personalized plans.”

Table stakes.

The more competent the underlying supply becomes, the less useful generic claims of competence become.

Which means something else becomes more valuable.

The ability to prove who is unusually good at what.

That's a very different asset.

The opportunity isn't another directory

Think about the information available today when someone needs a chiropractor.

Google knows proximity.

Google knows reviews.

Google knows website authority.

Meta knows interests.

Yelp knows ratings.

A clinic website knows what the doctor claims to treat.

But almost nobody has a structured answer to:

Which doctor consistently gets meaningful outcomes for which type of patient with which condition?

That's an information gap.

And information gaps create businesses.

Imagine a system containing millions of relationships like:

Condition → Patient → Provider → Intervention → Outcome → Location

Now layer on:

severity,

duration,

demographics,

patient-reported improvement,

verified reviews,

treatment history,

confidence,

and volume.

Eventually you don't just have a directory of chiropractors.

You have a results graph.

And the results graph becomes interesting because it can answer questions the underlying clinics cannot answer individually.

Not:

Is Dr. Smith a good chiropractor?

But:

Among chiropractors within 15 miles, who has the strongest body of verified outcomes for patients like me dealing with chronic sciatica?

That's a dramatically more valuable question.

This is where AI changes the game

For years, online marketing rewarded whoever was best at attracting attention.

Rank higher.

Get more clicks.

Run more ads.

Create more content.

But AI interfaces are beginning to compress the search process.

A patient doesn't necessarily need ten blue links anymore.

They can ask:

Who should I see?

And receive an answer.

That's an enormous change.

Because when search becomes an answer engine, the economic prize shifts.

You don't merely want traffic.

You want to influence the answer.

And answer engines need underlying evidence.

They need entities.

Relationships.

Corroboration.

Structured information.

Third-party validation.

Evidence that helps them discriminate between ten businesses making almost identical claims.

In other words:

AI makes the decision layer more important.

The interesting part about OpenRouter is that it embraces the same idea from the opposite direction.

There are hundreds of models, so OpenRouter helps software navigate that complexity. Its own routing tools can classify requests and choose models based on the task.

More supply didn't eliminate the need for OpenRouter.

More supply created the need for OpenRouter.

That is the insight.

Now apply that to chiropractic

Imagine there are 50 competent chiropractors in a metro.

If one is obviously superior, selection is relatively easy.

But imagine 40 of them are pretty good.

Now the selection problem gets harder.

More competent supply creates greater demand for trustworthy differentiation.

So the strange outcome may be:

The better chiropractors become, the more valuable the referee becomes.

That referee doesn't necessarily provide the care.

It provides confidence.

It organizes evidence.

It reduces uncertainty.

It helps demand find the appropriate supply.

That position can become extraordinarily powerful.

We've seen versions of it everywhere.

Google didn't need to write the world's websites.

Booking.com didn't need to build hotels.

Uber didn't manufacture cars.

Visa doesn't make the things you buy.

Their power came partly from sitting between fragmented supply and enormous demand.

Not because middlemen are inherently valuable.

Most aren't.

They're valuable when they solve an expensive coordination problem.

Chiropractic has a gigantic coordination problem:

There is a lot of supply and very little trusted infrastructure for determining who should receive the demand.

Which creates two different plays

For an individual chiropractor, the strategy is relatively straightforward.

Stop thinking only about acquiring leads.

Start accumulating reasons to be chosen.

Every meaningful patient outcome is an asset.

Every condition in which you develop unusually deep evidence is an asset.

Every physician or community relationship is an asset.

Every third-party mention is an asset.

Every article demonstrating expertise is an asset.

Every verified result is an asset.

Every association between your clinic and a particular problem strengthens the probability that some future decision system routes that patient toward you.

You are building your own little gravity well.

The question becomes less:

How many leads can I buy this month?

And more:

If someone — or some machine — had to decide who deserves this patient, how overwhelming is the evidence for us?

That's a better question.

But there is an even bigger company hiding above the clinic

The natural instinct is to use AI to make chiropractic marketing better.

That's useful.

But it may dramatically underestimate the opportunity.

The bigger company is not necessarily:

AI marketing software for chiropractors.

It may be:

the intelligence layer that understands chiropractic outcomes better than anyone else.

That changes everything.

Now your customer isn't merely the chiropractor.

Your underlying asset can eventually become useful to:

patients,

AI assistants,

search engines,

health systems,

employers,

payers,

researchers,

referral partners,

and other healthcare platforms.

Because all of them eventually encounter some version of the same question:

Where should this person go?

If you possess differentiated data capable of answering that question, you are no longer selling marketing.

You're selling — or powering — decision infrastructure.

That's a very different category of company.

The trap is confusing the wedge for the destination

Many great companies begin with something that looks much smaller than what they eventually become.

The initial product is often just the easiest way into the network.

For a results platform, the wedge might be:

Help chiropractors showcase outcomes.

Then:

Help them rank for conditions.

Then:

Help AI discover their evidence.

Then:

Help patients compare providers.

But none of those necessarily have to be the final destination.

Because every one of those activities can produce something more important:

better proprietary information about the relationship between patients, conditions, providers and outcomes.

Once that dataset becomes sufficiently deep, the product can flip.

Originally, chiropractors joined because they wanted visibility.

Eventually, visibility may come because chiropractors need to participate in the dataset.

That is when a product starts becoming infrastructure.

And infrastructure has a very different kind of moat.

The moat isn't having data

Everybody says they have a “data moat.”

Most don't.

A real data moat has a loop.

More chiropractors produce more outcomes.

More outcomes create better answers.

Better answers attract more patients, search engines and AI systems.

More demand creates more value for chiropractors.

More chiropractors participate.

More outcomes enter the graph.

And the product gets stronger.

That is the loop I'd be obsessed with.

Not:

How do we get another subscriber?

But:

Does each subscriber make the underlying network more valuable for every future participant?

If the answer becomes yes, you've stopped building ordinary SaaS.

You are beginning to build a network.

The market will eventually ask for an answer

For most of chiropractic's history, practices could control their own narratives.

You could say you were great.

Buy an advertisement.

Build a website.

Collect testimonials.

Tell your story.

But we're moving toward a world where software increasingly synthesizes the story for the patient.

The patient won't always read seven websites.

They'll ask.

And something will answer.

That means somebody — somewhere — will provide the evidence behind that answer.

Maybe Google assembles it.

Maybe OpenAI does.

Maybe healthcare platforms do.

Maybe several companies do.

But the important strategic question is:

Who supplies the truth layer?

Because the company that owns the model can change.

The interface can change.

ChatGPT can win.

Gemini can win.

Some company that doesn't exist yet can win.

It doesn't really matter.

If all of them need your underlying information to confidently understand the market, you occupy a much more durable position.

That's another reason the OpenRouter idea is useful.

Don't spend all your energy predicting which model wins.

Build something the winners need.

The play for chiropractors

So yes:

Become excellent clinically.

Build a remarkable office.

Market aggressively.

Generate leads.

Create content.

Rank on Google.

Learn AI.

Do all of it.

But understand the larger game underneath those tactics.

As supply expands and information explodes, being available becomes less valuable.

Being chosen becomes more valuable.

And being the system that helps determine who gets chosen may become more valuable still.

That is the progression:

Provide the service.

Then own demand.

Then own the evidence.

Then influence the decision.

Most chiropractors will spend the next decade fighting to become a slightly better option.

A smaller number will build such overwhelming evidence around a particular category that they become the obvious option.

And perhaps one or two companies will realize the biggest opportunity isn't becoming another option at all.

It's building the infrastructure that decides between them.

That's the real lesson hiding inside Stripe's OpenRouter acquisition.

When the product becomes abundant, don't just build a better product.

Own what becomes scarce next.

In chiropractic, I think that's certainty.