The 'Imaginary Customer' Problem: Why ICP Often Misleads•
The AI Revenue Illusion: GMV vs ARR•
The 'Global Market' Myth: Why Local Wins First•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
The Hidden Cost of 'Friendly' Feedback•
GTM 101: Startup Revenue Architecture•
Agile: The Secret Behind Success vs. The Reason We Fail•
The Hidden Crisis: How Silent Quitting Reveals a Broken Feedback Loop in Modern Organizations•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
Get Off the Sidelines: It's Your Turn to Play•
The PM Role is Evolving. Enter the Producer.•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
Why GTM Is Suddenly Everywhere (The Shift from PLG)•
Pricing is Power: If You Can't Raise It, You Don't Have a Business•
Agile: The Secret Behind Success vs. The Reason We Fail•
The 'Imaginary Customer' Problem: Why ICP Often Misleads•
GTM 101: Startup Revenue Architecture•
Why GTM Is Suddenly Everywhere (The Shift from PLG)•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
Get Off the Sidelines: It's Your Turn to Play•
The Hidden Cost of 'Friendly' Feedback•
The 'Global Market' Myth: Why Local Wins First•
The AI Revenue Illusion: GMV vs ARR•
The PM Role is Evolving. Enter the Producer.•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
Pricing is Power: If You Can't Raise It, You Don't Have a Business•
The Hidden Crisis: How Silent Quitting Reveals a Broken Feedback Loop in Modern Organizations•
The Hidden Cost of 'Friendly' Feedback•
Agile: The Secret Behind Success vs. The Reason We Fail•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
The 'Global Market' Myth: Why Local Wins First•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
The Hidden Crisis: How Silent Quitting Reveals a Broken Feedback Loop in Modern Organizations•
The PM Role is Evolving. Enter the Producer.•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
Why GTM Is Suddenly Everywhere (The Shift from PLG)•
Get Off the Sidelines: It's Your Turn to Play•
Pricing is Power: If You Can't Raise It, You Don't Have a Business•
GTM 101: Startup Revenue Architecture•
The 'Imaginary Customer' Problem: Why ICP Often Misleads•
The AI Revenue Illusion: GMV vs ARR•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
GTM 101: Startup Revenue Architecture•
The AI Revenue Illusion: GMV vs ARR•
Get Off the Sidelines: It's Your Turn to Play•
Agile: The Secret Behind Success vs. The Reason We Fail•
The 'Global Market' Myth: Why Local Wins First•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
The 'Imaginary Customer' Problem: Why ICP Often Misleads•
The AI Revenue Illusion: GMV vs ARR•
The 'Global Market' Myth: Why Local Wins First•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
The Hidden Cost of 'Friendly' Feedback•
GTM 101: Startup Revenue Architecture•
Agile: The Secret Behind Success vs. The Reason We Fail•
The Hidden Crisis: How Silent Quitting Reveals a Broken Feedback Loop in Modern Organizations•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
Get Off the Sidelines: It's Your Turn to Play•
The PM Role is Evolving. Enter the Producer.•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
Why GTM Is Suddenly Everywhere (The Shift from PLG)•
Pricing is Power: If You Can't Raise It, You Don't Have a Business•
Agile: The Secret Behind Success vs. The Reason We Fail•
The 'Imaginary Customer' Problem: Why ICP Often Misleads•
GTM 101: Startup Revenue Architecture•
Why GTM Is Suddenly Everywhere (The Shift from PLG)•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
Get Off the Sidelines: It's Your Turn to Play•
The Hidden Cost of 'Friendly' Feedback•
The 'Global Market' Myth: Why Local Wins First•
The AI Revenue Illusion: GMV vs ARR•
The PM Role is Evolving. Enter the Producer.•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
Pricing is Power: If You Can't Raise It, You Don't Have a Business•
The Hidden Crisis: How Silent Quitting Reveals a Broken Feedback Loop in Modern Organizations•
The Hidden Cost of 'Friendly' Feedback•
Agile: The Secret Behind Success vs. The Reason We Fail•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
The 'Global Market' Myth: Why Local Wins First•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
The Hidden Crisis: How Silent Quitting Reveals a Broken Feedback Loop in Modern Organizations•
The PM Role is Evolving. Enter the Producer.•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
Why GTM Is Suddenly Everywhere (The Shift from PLG)•
Get Off the Sidelines: It's Your Turn to Play•
Pricing is Power: If You Can't Raise It, You Don't Have a Business•
GTM 101: Startup Revenue Architecture•
The 'Imaginary Customer' Problem: Why ICP Often Misleads•
The AI Revenue Illusion: GMV vs ARR•
The OKR Paradox: Why Google's Goal-Setting Framework Became Its Own Enemy•
The 'Glass Slipper' Trap: Why Early Retention Metrics Mislead•
GTM 101: Startup Revenue Architecture•
The AI Revenue Illusion: GMV vs ARR•
Get Off the Sidelines: It's Your Turn to Play•
Agile: The Secret Behind Success vs. The Reason We Fail•
The 'Global Market' Myth: Why Local Wins First•
Your Dashboard Is Lying to You (Stop Hiding Behind Data)•
OBF
KR
AboutWhy OBFInsightServicesContact Us
Back to Insights
The AI Revenue Illusion: GMV vs ARR

The AI Revenue Illusion: GMV vs ARR

4 min read

There is a misunderstanding circulating in the AI startup ecosystem.
It appears in pitch decks, on Twitter, and in VC meetings.

"We reached $1M ARR in 3 months."

It sounds impressive. It sounds like hyper-growth.
But in many cases, it is financial misrepresentation.

They are not reporting ARR (Annual Recurring Revenue) in the traditional sense.
They are reporting GMV (Gross Merchandise Value) and presenting it as SaaS revenue.

The Definition of ARR Matters

ARR is the gold standard metric for SaaS because it implies two things:

  1. High Margins (80%+)
  2. Low Marginal Cost of Replication (Selling the same code cost nothing)

When a traditional SaaS company sells a $100 subscription, it costs them $10 to service it. The $90 goes to R&D and Sales. This leverage is why SaaS valuations are high.

Now look at an AI "wrapper" startup.
They sell a $100 subscription.
But every time the user clicks a button, they pay OpenAI, Anthropic, or AWS $60.
Their gross margin is not 90%. It is 30%. Sometimes it is negative.

Calling this "ARR" is misleading.
You are not operating as a typical software company. You are operating more like a reseller.

GMV: The Flow of Money vs. The Ownership of Money

If you run a marketplace (like Uber or eBay), you track GMV.
GMV is the total money flowing through your system.
Revenue is the "take rate" you keep.

Uber does not claim the full fare as revenue. They claim their cut.

But AI startups are claiming the full subscription price as "ARR," ignoring the massive variable costs (inference) attached to every dollar.

This is the "Service Bureau" Trap.
In the 1990s, we had Service Bureaus. You paid them to run a job. They paid for the machines and labor. It was a low-margin, operationally heavy business.
Many AI startups risk becoming Service Bureaus with better UI.

The "Costs Will Drop" Fallacy

Founders argue: "But inference costs are dropping! Margins will improve!"

This is half-true and wholly dangerous.
Yes, the cost of GPT-4 will drop.
But user expectations will rise. They will demand GPT-5, video generation, and agentic workflows.
The demand for "smarter" (and more expensive) models will always outpace the cost reduction of older models.

You are on a treadmill. You are not building software leverage; you are renting compute arbitrage.

The Identity Crisis

You need to decide what you are.

Option A: Software Company

  • You own the IP.
  • Marginal cost is near zero.
  • You charge for value, not usage.

Option B: Tech-Enabled Service / Reseller

  • You aggregate third-party models.
  • You pay a tax on every user action.
  • You compete on operations, not just code.

Option B is a valid business. Consulting firms and resellers make money.
But they do not get 20x revenue multiples. They get 1x or 2x.

The problem is not the business model. The problem is the valuation arbitrage.
Founders want the valuation of Option A while running the P&L of Option B.

The "Day Trader" Perspective

I see founders celebrating "high volume" usage.
It resembles the perspective of a day trader bragging about moving $1 million in volume, only to realize he paid $10,000 in fees and made $50 in profit.

Volume can be misleading.
If your revenue scales linearly with your compute costs, you haven't built a product.
You've built a toll booth where you pay the toll for the customer.

Conclusion: Face the Reality

If you are an AI founder, look at your P&L.
Subtract your inference costs, your vector DB costs, your cloud GPU costs.

What is left?
That is your Revenue.
The rest was never yours. It was a pass-through cost to NVIDIA and Microsoft.

Build a business where the value accumulates to you, not your vendors.
Until then, reconsider using the term "ARR."
Say "Throughput." It's more honest.