Stripe's $10B Acquisition of OpenRouter: Why Payment Giant Bets on AI Aggregation Layer

It's like a highway company suddenly deciding to acquire a toll booth network. Stripe's intent to acquire OpenRouter is surprising yet makes sense. Payments are the blood vessels of AI, but Stripe chose to buy the transfusion station directly.

Conclusion First

This deal isn't simple business complementarity, but a strategic positioning by Stripe in the "middle layer" of AI infrastructure. OpenRouter's valuation premium—close to $10 billion—reflects the market's strong expectation for an integrated platform for AI traffic distribution and payment settlement. Looking at industry cycles, AI is shifting from model-training-driven to application-driven, and API call volumes are about to see exponential explosion. Whoever controls the entry points for payments and aggregation controls the stickiest part of AI commercialization.

Data-Driven: Why is OpenRouter Worth $10 Billion?

OpenRouter is currently the world's largest AI model aggregation platform, providing API gateway services for over 200 LLMs, including mainstream models from OpenAI, Anthropic, Meta, etc. It is essentially an "AI router," allowing users to invoke multiple models with a single connection, choosing the optimal cost-performance ratio on demand, while obtaining unified billing and logging functions.

Revenue Structure Speculation: Based on public industry information, OpenRouter primarily profits from commissions on API call volumes (approx. 5%-10%) and premium feature subscriptions. Assuming its daily average API call volume exceeds 50 million, with an average charge of $0.002 per call, the annualized revenue is approximately $360 million. At a $10 billion valuation, the implied PS ratio is about 28x. Compared to similar AI infrastructure companies:

Company	Business Model	Valuation/Market Cap	Implied PS
OpenRouter	AI API Aggregation	~$10B	~28x
Together AI	Model Hosting + Inference	~$3B	Not Public
LangChain	Dev Framework + Platform	~$2B	~15x
Anthropic	Model Layer	~$60B	~40x

A 28x PS ratio isn't extreme in AI infrastructure, but consider this: OpenRouter's moat is network effects—the more users connect, the more models join, accelerating the data flywheel. Stripe's payment network can amplify this effect perfectly: Stripe's existing millions of merchants can directly invoke AI models via OpenRouter, while Stripe's payment infrastructure can seamlessly handle micro-transaction settlements for AI APIs.

Key Data Point: Total payment volume processed by Stripe in 2024 was approx. $1.4 trillion, with AI-related payments accounting for nearly 5%. If Stripe acquires OpenRouter, it can increase the proportion of AI payments to 15%-20%, meaning an additional $200-300 billion in payment processing volume, directly driving revenue growth.

Competitive Landscape: The "Three-Layer Game" of AI Infrastructure

Current AI infrastructure is differentiating into three layers:

• Model Layer: OpenAI, Anthropic, Google DeepMind, holding core intellectual assets
• Platform Layer: OpenRouter, LangChain, Replicate, providing model routing, orchestration, hosting
• Cloud Layer: AWS, GCP, Azure, providing underlying computing power

Stripe's acquisition implies a judgment: The model layer is commoditizing, while the platform layer will differentiate into high-barrier payment + distribution platforms. OpenRouter has already accumulated data advantages in model aggregation—knowing which model offers the best cost-performance for specific tasks, something the model layer itself cannot easily replicate.

Stripe's competitors are also moving. Adyen has launched AI payment plugins, PayPal is testing AI agent payments. But Stripe has a deeper developer ecosystem—Stripe's API covers over 60% of global SaaS enterprises, and this developer reach capability is a key lever for OpenRouter's expansion.

Industry Trends: Upgrading "Selling Shovels" Before the Explosion of AI Call Volumes

From the "AI Gold Rush" logic, early shovel sellers were GPU manufacturers (NVIDIA), mid-stage was cloud services (AWS/GCP), and now we enter the third stage—infrastructure services for AI applications. OpenRouter is that "smart shovel."

Key Assumption: By 2026, global AI API call volumes will grow 10x, reaching 100 billion calls per day. Each call requires payment and routing to the optimal model. The Stripe+OpenRouter combination will occupy 20%-30% of this market share, with annualized payment processing volume potentially reaching **$500-800

Original link: https://www.ithome.com/0/980/984.htm