He sold his 2.2 billion fortune before the NFT crash and then quickly moved into the hottest AI track.

Author: Diya Fan

In 2022, a Stanford graduate’s net worth once exceeded $2 billion.

He created OpenSea, the world’s largest NFT marketplace, valued at $13.3 billion.

Just a few months before the NFT bubble burst, he made a more critical decision to leave.

Two years later, his new company grew 10 times in 7 months, secured investments from a16z, Sequoia, and Menlo, and reached a valuation of $500 million.

His name is Alex Atallah. His new company is called OpenRouter.

This is a story about timing and methodology replication.

Who is OpenRouter? What does it do?

If you’re an AI application developer, you’ve probably heard of OpenRouter. Its main function is to help developers solve the pain of model switching:

• Want to use Claude for coding but find it often lacks capacity

• Want to analyze with GPT but the cost makes you hesitate

• Want to try open-source models but realize you need to rewrite a new API integration

Each model provider’s API is different. Every time you switch models, you have to modify your code.

OpenRouter does something similar to Ctrip, bringing all airlines into one app.

One API, access to 300+ models. 60+ providers. Switch models? Change one line of code.

OpenRouter as a Multi-Model Aggregation Layer

Two startups, same methodology

Before starting his entrepreneurial journey, Alex Atallah already had a hardcore software background: Stanford Computer Science, Palantir engineer, co-founder and CTO of OpenSea…

OpenSea co-founder Alex Atallah (left) with Devin Finzer (right)

He explained the commonality of his two startups on a podcast:

“OpenSea organized this very heterogeneous inventory and put it together in one place… You see a lot of those similarities with how AI works today.” (OpenSea consolidated a very chaotic NFT inventory… AI is the same today.)

What is his methodology?

Identify a “fragmented ecosystem” and then build an “aggregation layer”.

• In the NFT era: metadata standards vary → OpenSea aggregates

• In the AI era: API standards vary → OpenRouter aggregates

Alex once said in a podcast that left a deep impression on me: If training a large AI model only costs $600, then in the future, there could be tens of thousands or even hundreds of thousands of models. At that point, they will need their own ‘market’.

Early 2023, this was an extremely contrarian view. The mainstream narrative then was: OpenAI is far ahead, and other models are just followers.

But Alex was right.

Today, there are over a thousand open-source models. Claude, Gemini, Llama, Mistral, DeepSeek… new players enter every few weeks.

In a world of explosive model growth, an “aggregation layer” is needed. That’s exactly where OpenRouter fits.

An underestimated huge market

Behind OpenRouter’s success lies the visible trend in the AI market: “Reasoning” will replace “training” as the main driver.

The difference between reasoning and training, and the future trend of this market, was clearly explained in Groq’s recent analysis—worth checking out.

COO Chris Clark’s view is worth noting:

“We believe that inference costs will eclipse salaries as the dominant operating expense for most knowledge-based companies over the next five to 10 years.” (We believe that in the next 5-10 years, AI inference costs will surpass salaries, becoming the largest operational expense for knowledge-based companies.)

This can actually be seen from OpenRouter’s own data.

OpenRouter’s token consumption approaches the 8 trillion mark

A well-known AI model “public review”

As one of the earliest players in this space, OpenRouter has an exclusive advantage: ranking list.

After processing over 100 trillion tokens, they know:

• Which model is best at coding

• Which model offers the best cost-performance ratio

• Which model suddenly excels at specific tasks

This ranking has become an important industry reference and is highly recognized in the developer community.

What’s even more incredible? In April 2025, a mysterious model called “Quasar Alpha” was launched on OpenRouter.

A few days later, everyone found out: It’s GPT-4.1, exclusively launched by OpenAI on OpenRouter.

Because OpenRouter has a killer asset: the largest multi-model usage dataset on the internet.

Every day, millions of developers call different models here. OpenRouter knows:

• Which model performs best on which task

• Which provider is most stable

• Which time period is cheapest

These data have created the industry’s most authoritative LLM ranking. According to Menlo Ventures, even Andrej Karpathy (former Tesla AI director, co-founder of OpenAI) has publicly recommended it.

Once the data flywheel starts turning, it’s hard for latecomers to catch up.

Andrej Karpathy mentions OpenRouter LLM rankings on X

How does OpenRouter make money?

OpenRouter’s business model is relatively simple: You spend $100 on models, they take $5.

Model providers set their prices, and OpenRouter charges a 5% cut. They earn “toll fees,” not “markup.”

This model aligns well with Western intermediary business practices:

2. Maintain neutrality: If OpenRouter has its own models, would you trust its rankings?

3. Grow naturally with the market: The bigger the AI market, the more revenue they generate

4. Network effects: More users → more data → more valuable rankings → more users

Alex’s words: “We want developers to not feel vendor lock-in. We want them to feel like they have choice and they can use the best intelligence, even if they didn’t before.” (We don’t want developers to be locked into vendors. We want them to have options and access to the best intelligence at any time.)

Financial data (disclosed)

8 people, with an annualized GMV of nearly $100 million.

This person’s efficiency is among the top in similar startups.

Large market, small space

After highlighting the advantages, it’s necessary to discuss some issues with this model:

OpenRouter’s core strengths are “data” and “community.” The flywheel has started turning (more users → better data → more valuable rankings), but this model also means its ecosystem’s prosperity depends heavily on the number of developers.

This business can’t thrive without more small and medium developers emerging, because they lack the time for aggregation development and the scale to negotiate prices with AI vendors, so they need an intermediary.

For big companies, this might be useful for testing, but once scaled, they will definitely bypass it.

In fact, even medium-sized projects with larger usage will want to avoid it, such as an open-source alternative called LiteLLM, which is free and self-deployable.

Cost-sensitive developers will ask: “Why give you 5%?”

If competition intensifies, this cut could drop to 3%, or even 2%.

Whether it can sustain the current disclosed 100x high valuation remains uncertain.

Of course, it’s still in early stages and will continue to grow rapidly. Its ceiling is a question to consider in analysis.

One-minute overview of OpenRouter

Q1: What is OpenRouter?

OpenRouter is a large language model (LLM) API aggregation platform. Through a single API interface, developers can access over 300 models (including GPT-4, Claude, Llama, etc.) without integrating each provider’s API separately.

Q2: How is OpenRouter different from LiteLLM?

Both provide LLM API aggregation, but with different models. OpenRouter is a managed SaaS charging a 5% fee; LiteLLM is open-source, deployable locally, and free. OpenRouter’s unique advantage is its public model ranking and broader provider coverage.

Q3: Who are the founders of OpenRouter?

Alex Atallah, Stanford Computer Science graduate, former co-founder and CTO of OpenSea (the world’s largest NFT marketplace). He left OpenSea in 2022 and founded OpenRouter in 2023. His personal net worth once exceeded $2 billion.

Q4: How much funding has OpenRouter raised?

In June 2025, OpenRouter completed a total of $40 million in funding (seed + Series A), led by a16z and Menlo Ventures, with Sequoia participating, and valued at about $500 million.

Q5: Why does OpenAI test new models on OpenRouter?

According to OpenRouter disclosures, OpenAI has tested new models anonymously on its platform to gather unbiased developer feedback. This indicates that the OpenRouter community has some influence in the industry.