The Overlooked Geography Beneath the AI Economy

AI-native companies are growing faster than business executives of a previous generation could have dreamed. One executive at a prominent AI company estimated that his company could grow by 80 times this year. That growth doesn’t exempt them from a business practice as old as time—tax collection—and in the AI economy, tax compliance may be more complicated than ever.

In part, that’s because most of the world’s countries now abide by the taxation-by-destination rule, which requires businesses to collect the taxes of the country or state of the customer—not the country or state of the company.

In the US, there are over 16,000 combinations of tax rules and rates that can apply to a sale of goods or services. Tax jurisdictions do not nest neatly in cities or ZIP codes—and the rates they apply are constantly changing. Other countries use entirely separate systems, like the value-added tax (VAT) in Europe or the goods and services tax (GST) in Canada, complicating any company’s global expansion.

Without a systematic approach for managing tax compliance in a large number of jurisdictions, high-flying AI firms may expose themselves to legal risks or penalties for noncompliance. Or they hesitate to enter markets whose tax regimes they don’t fully understand. Tax compliance starts by understanding where the transaction is taxable, which is usually where the customer is located.

“That’s why it’s so important to know where your customers are,” says Aleksandra Bal, the global indirect tax technology lead at Stripe.

Stripe’s tax compliance product Stripe Tax uses sophisticated geographic information system (GIS) technology to instantly and accurately identify tax jurisdictions, saving clients significant time they might otherwise spend investigating local tax jurisdiction boundaries.

Taxing the Brave New World of the AI Economy

Stripe has processed online and in-store payments for 15 years. In 2025, the company reported that businesses running on Stripe generated $1.9 trillion in total volume, roughly 1.6 percent of global GDP. Since 2021, the company has also helped businesses collect taxes on those purchases, using location software to ensure that the right tax is applied to millions of transactions.

For AI companies focused on rapid scaling, tax management presents a steep learning curve. An enterprising AI company with 10 employees and customers in 100 countries likely doesn’t have the expertise to define each regional tax collection strategy. It might not even have a CFO.

“As they onboard, AI companies are immediately doing numbers of transactions where they can’t think about managing tax compliance manually,” says Erich Rentz, the tax geography lead at Stripe and the architect of the geospatial tools used by Stripe Tax.

Adding to the geographic complexities of tax collection are novelties specific to the AI economy, which lawmakers are only beginning to grapple with.

“Questions that were nonexistent a few years ago are appearing now,” Bal says. “There’s many new business models—for example: agentic commerce, usage-based billing, purchases made via LLM apps, micropayments.”

In the US, there is no uniform national approach to taxing AI services. States like New York treat digital automated services as taxable, while Indiana and Illinois do not consider the services of an AI chatbot subject to sales tax. Internationally, the rates applying to AI services vary and can be as high as 27 percent in Hungary or 25 percent in Sweden.

Most AI businesses are too busy to build the capability to manage tax internally—so Stripe Tax built it for them, using GIS technology as a powerful analytics engine.

Behind One-Tap Digital Purchases, a Universe of Tax Jurisdictions

To determine the right amount of tax, Stripe Tax starts by asking, Where? Understanding the geography of tax compliance in the US is managed by the company’s jurisdiction resolution system (JRS), which is broken into two parts: an offline map of all US tax jurisdictions and an online map that helps analyze transactions in real time.

Each time a purchase occurs, the online system taps a trove of offline geographic data and generates an accurate answer in milliseconds, rather than making several simultaneous analyses on the spot. Without that foundation of location data, even the most sophisticated AI models can’t reliably trace purchases to the right jurisdictions in every instance.

“Many vendors rely on ZIP codes, which we know are not reliable,” Bal says. “Our solution is very accurate, and that’s possible because we have the JRS.”

GIS technology organizes the precise locations of all US states and territories, over 3,200 counties, around 10,000 cities, and more than 3,000 districts. Bal, Rentz, and the Stripe Tax team track relevant legislation and update their maps throughout the year.

The explosion of special taxing districts—areas formed by local governments to levy taxes and help pay for public improvements like emergency services or water supply—has contributed to the complexity of tax allocation. New rules may also be forthcoming as governments seek ways to tax the windfall generated by the booming AI economy. For instance, beginning in 2027, California will impose new sales and use tax to prewritten software, SaaS, and other remotely accessed software.

A User-First Philosophy, Grounded in Location Insights

To ensure that rules are applied consistently, Rentz and colleagues invented the concept of a Stripe Place of Taxation (SPOT): areas where state, county, and local taxes align. They tend to be oddly shaped, with boundaries that might have hundreds or even thousands of sides.

GIS technology matches addresses to the correct SPOTs in milliseconds. Some places take slightly longer than others: Charleston, South Carolina, for instance, has a complicated boundary, and the city itself is composed of 70 discontinuous areas, resulting in a complex tax thumbprint.

At Stripe, GIS simplifies these byzantine distinctions. All that consumers see is a purchase confirmation a moment after tapping the buy button.

“We want to make sure that you can, from day one, be confident that you can sell into markets without being concerned that you don’t know the nexus rules,” Rentz says, referencing the laws that govern how out-of-state businesses collect sales tax. “Stripe is such a user-first company—we want to make sure we’re doing this right.”

Meeting a Global Need for Address Identification

AI users are rapidly multiplying worldwide: roughly 53 percent of the world’s population now uses the technology. Bal is eager to explore how JRS could be applied in markets beyond the US to help Stripe’s fastest-growing clients.

She sees potential, for instance, when analyzing transactions in places that might share European statehood but exist outside the continent. The island of Bonaire lies off the coast of South America but is part of the Netherlands, and yet sales there are governed by a different set of tax rules than those in Amsterdam or The Hague. Bonaire is especially interesting because it does not yet have postal codes that would assist in identifying the customer’s location.

“If we want to collect taxes correctly, we need to have more information than just the country code,” Bal says. “That’s the biggest way we’re trying to get the JRS abroad.”

Ultimately, in the AI economy, geography shouldn’t be a barrier to a company’s growth.

By harmonizing tax compliance across borders, Stripe and the GIS technology it relies on are giving a green light to AI-native businesses to scale faster than ever.