Stripe Investor Letter: The Singularity Has Arrived, and the Strategic Logic Behind Acquiring OpenRouter

Original Author: Stripe

Compiled by: TechFlow

Intro: On August 19, Stripe sent investors a letter that was measured in tone but immense in ambition. On the surface, the letter announces the acquisition of OpenRouter — the largest in Stripe's history. But reading through it, what Stripe is really saying is this: In the AI era, every enterprise needs to manage two pipelines — one for money (revenue pipeline), and one for intelligence (inference pipeline).

Stripe already owns the first. Now it's going after the second. From Bridge to Privy, from Metronome to OpenRouter, this recent spate of acquisitions forms a complete mosaic built around "AI economic infrastructure." Stripe says that building economic infrastructure for the internet and building economic infrastructure for AI are essentially the same thing — and in doing so, it is also declaring that the boundaries of a payments company have disappeared.

Below is the full translation of the investor letter.

To our investors:

We're writing to share that we're announcing OpenRouter is joining Stripe, our largest acquisition to date. We've also acquired Bridge, Privy, and Metronome in recent months. We thought it would be valuable to share our thinking on the role these businesses play in Stripe's overall strategy.

Stripe's goal is to grow the GDP of the internet. When we think about a prosperous world, we focus on the underlying mechanisms that make everything possible: money, credit, monetary systems, legal structures, and risk management. We believe the world can and should be larger and more prosperous than it is today, and better economic infrastructure can help achieve that.

The Singularity

"The Singularity" is a vague, perhaps overused term, but we've decided to treat January 1 as the start of the Singularity and have been operating accordingly ever since. The Singularity is often compared to millenarian prophecies, but in our case, we simply observed a dramatic inflection point in long-term trends (such as the significant acceleration in the rate of new company formation) and decided to take this phase shift seriously.

As it turns out, optimizing for developers from day one is in many ways equivalent to optimizing for code integration and AI agents, because they all pursue programmability and frictionless onboarding. We're fortunate that a large portion of the world's AI companies have already chosen Stripe for this reason. This has also made it increasingly clear to us: Building economic infrastructure for the internet and building economic infrastructure for AI are essentially the same thing.

While the scale of transformation will obviously be enormous, no one can concretely predict how AI will reshape our world. Many predictions from smart people have already been thoroughly disproven. With humility in the face of uncertainty, we have two overarching goals.

First, we want to accelerate the diffusion of AI throughout the economy. As AI continues to push the boundaries of what's possible, we're seeing a wealth of delightful new products and services emerge (surely just a prelude to the astonishing creations yet to come), which require different, better-suited financial tools. The rise of AI in the economy is also creating new types of challenges, such as new forms of theft and fraud, requiring sophisticated technical measures to address effectively. Overall, the collective expectation and promise of AI is that it will bring greater material prosperity and abundance. We want to help make that happen.

Second, there is a concern that AI will lead to job displacement or power concentration — perhaps both. We believe AI must be deployed with a priority on enhancing human autonomy, and we want Stripe to play a role in protecting economic autonomy as a foundational element of a free society. Because of AI, we hope more companies will be founded, competing with established giants at higher efficiency. While early data supports this trend (as we've documented on platforms like the Stripe Economics Substack), nothing is guaranteed. As advocates for small businesses, we will do our utmost to keep the path clear.

We still believe there is no ceiling on the size of the global economy (already over $100 trillion). Though it may sound incredible at first, we think it's meaningful to consider a "quadrillion-dollar" world and enumerate the bottlenecks to achieving it. (If global per-capita GDP reached the level of the average Irish citizen — about $100,000 — we'd already be 80% of the way there.)

More than 5 million businesses are already on Stripe, and the flow of funds through Stripe represents roughly 2% of global GDP. We're pleased with this solid start, but we believe these numbers are negligible compared to what's possible in the future.

What Are We Building?

From a macro perspective, the global economy is clearly poised for substantial growth. From a micro perspective, the way commerce operates is changing rapidly. Incumbent companies are adjusting their business models (usage-based pricing is on the rise, while many traditional models are in decline) and rapidly mobilizing resources to launch new products and services (with speed and flexibility becoming core competitive advantages). The rate of new company formation is accelerating. AI agents are on the verge of becoming independent economic actors. Stablecoins are seeing rapid adoption, and adoption will accelerate further as they become the native currency of the AI economy. As tokens naturally have cross-border mobility, global reach is becoming more important than ever before.

We are building the economic tools this era requires as quickly as we can. The combination of native stablecoin support and reachability by AI agents is giving rise to a new set of foundational primitives:

• Discovery + Access: Stripe Projects (enabling AI agents to sign up and integrate with third-party services), Stripe Directory (product and service discovery for agents), Provisioning API (embedded onboarding)
• Usage Management: Metronome
• Payments: Bridge (stablecoin orchestration), Stripe's Agentic Commerce Suite, Tempo (blockchain payments for agents), MPP (Machine Payment Protocol)
• Capital Storage: Privy (crypto/stablecoin wallets), Open Standard (a new stablecoin)

Over time, we expect a structural mix shift, with the "AI economy" tech stack gradually taking share from the "pre-AI economy" tech stack. The adoption of these new products may not look dramatic on the surface: we're deeply integrating them into Stripe's existing products and platform, ensuring seamless adoption for any business.

To ensure Stripe is as useful as possible in this new phase, we're closely tracking our adoption among the world's fastest-growing and most important new companies. Today, 88% of companies on the Forbes AI 50 list (including OpenAI and Anthropic) are built on Stripe (with most of the remaining 12% still in the pre-monetization stage), and 100% of Brex's latest list of fastest-growing startups use Stripe. Most of these companies use more than ten Stripe products, and our revenue share from AI companies and crypto companies has been doubling year over year. We hope Stripe can, over time, serve as a barometer for tracking the overall growth of AI deployment.

OpenRouter

Zooming out, capital and intelligence are becoming the two digital streams underpinning every enterprise. To date, every developer has needed a simple, reliable way to manage their revenue pipeline, and serving that need gave birth to Stripe. Going forward, every developer will equally need a simple, reliable way to manage their intelligence pipeline.

This observation is what first led us to OpenRouter. OpenRouter has built the world's largest and most trusted token routing engine, supporting all major models and providers, and is deeply loved by users. Thanks to the utility of the product itself, the exceptional capabilities of the founders and the OpenRouter team, and an industry cadence where new models launch every week, its business has grown at an extraordinary pace (even by AI industry standards), with token consumption compounding at 9% per week since the start of the year.

OpenRouter is extraordinarily useful for any developer, and Stripe is one of the world's largest developer platforms. We therefore believe that bringing these two core needs together will generate substantial synergies and efficiencies.

However, we believe the case for integration goes beyond mere convenience. Through our work with customers, we've realized that "intelligence" is special: it's expensive, heterogeneous, and constantly changing. Just as enterprises must carefully manage the cost and return of every unit of financial capital, they must manage intelligence capital with equal rigor and precision. How valuable is this task? Which model should handle it? Who pays, when, and what's the time value of latency?

We've already seen the parallels between managing intelligence and managing capital directly in our own products. For example, Radar was originally designed to combat financial fraud, but it has proven equally effective at combating token fraud and is now used by many of the world's largest AI companies. Metronome (used by industry leaders such as Anthropic and Nvidia) is demonstrating that, in AI contexts, usage-based billing is inseparable from the serving and consumption of tokens themselves.

We expect the transaction to close in the coming weeks. We're excited to extend Stripe's financial capabilities into this new domain, helping businesses effectively allocate this new "intelligence currency."