As AI Makes Software Cheaper to Build, Data and Distribution Become the Moat

Anthropic's own enterprise usage data shows that its Claude Code tool now costs companies a median of about $13 per developer per active day, with most users spending under $30 per day. That's remarkably cheap to produce the sorts of results that used to require a team of engineers and months of work. When writing software gets this inexpensive, the code itself stops being the valuable part. What's left to fight over is what a company already owns that artificial intelligence (AI) can't generate from scratch: proprietary data, and a position so embedded inside its customers' operations that ripping it out becomes its own project.

Image source: Getty Images.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

Salesforce (NYSE: CRM) just gave investors a live look at how seriously this is being taken, rolling out a pricing model in June that charges customers $2 per resolved customer service case, and nothing if the AI agent fails, a direct acknowledgment that charging per human seat stops making sense once software does the seat's job.

Where the moat is widening

Palantir (NASDAQ: PLTR) built its entire platform around what it calls "ontologies" -- structured maps connecting a company's data to its real-world operations, so an AI agent can understand, for example, that a specific supplier links to a specific shipping route and a specific customer order, not just scattered text. CrowdStrike (NASDAQ: CRWD) disclosed in its 2026 Threat Hunting Report that it now analyzes more than 7 trillion security events every day across its customer base, narrowing that down to roughly 14 million daily detection leads. That's a level of telemetry scale that a competitor can't shortcut simply by building better AI.

ServiceNow (NYSE: NOW) runs more than 100 billion workflows a year across over 8,800 enterprise customers, and a company that has spent a decade routing its information technology tickets and human resources requests through that platform isn't likely to jump ship and start over with a new provider just because a cheaper AI tool exists somewhere else.

Here's a dumbed-down version of why that matters for anyone putting money into AI stocks right now: A lot of investors are still pricing software companies as if the app or the algorithm is the hard part. It isn't. The hard part is the years of data the incumbents possess and the painful switching costs that businesses have to pay to leave them. That's not "I like this company because it has good technology." It's closer to "This company has something a rival literally cannot buy or build quickly," which tends to be a sturdier reason to own a stock.

Where the case comes with real caveats

AppLovin (NASDAQ: APP) grew revenue 53% last quarter, and that gain was powered by its advertising engine, which learns which ads convert best across its network of mobile apps. This provides it with a genuine data advantage. The caveat is real, though: AppLovin's own filings flag its dependence on other distribution platforms, meaning the house that it has built sits on land it doesn't own.

Adobe (NASDAQ: ADBE) is racing to embed its Firefly AI directly into Photoshop, Premiere, and Illustrator, expanding what it calls a Creative Agent across its whole suite this year. Its installed base of working creative professionals is a real moat, but generative AI tools are commoditizing the raw act of creating images and video outside Adobe's ecosystem entirely, a tension Adobe hasn't fully resolved yet.

This is a part of the broader AI conversation that is easy to miss. Having an advantage and owning that advantage outright aren't the same thing. AppLovin's data about which ads work well is valuable, as we've seen in its revenue growth, but it's built on top of others' app stores, which means larger companies have a say in how AppLovin operates at any time. The millions of working designers and editors using Adobe's software provide it with a real head start, but that head start can hold up only if Adobe's AI tools stay noticeably better than the numerous free alternatives that are showing up.

Where the moats seem to be narrowing

Salesforce is in an odd position: The same per-resolution pricing shift that proves AI agents work is also an admission that its legacy per-seat CRM licenses are exactly what those agents are designed to replace. Snowflake (NYSE: SNOW) management is guiding for its product revenue growth to decelerate from 30% to 27% for the coming fiscal year, even as it expects its margins to hold up. That's a sign that data warehousing is becoming easier to replicate as AI tools make querying and managing data less dependent on Snowflake specifically.

These two examples show that having a big customer base doesn't automatically provide a business with protection from being disrupted by AI-based competition. It can just as easily be the thing AI is about to disrupt. Salesforce built its fortune charging companies based on how many of their employees logged into its software; once an AI agent does that employee's job instead, charging per seat stops making sense, and Salesforce knows it, which is why it's rewriting its own pricing model before someone else forces the issue. Snowflake's slowing growth rate is quieter evidence of something similar: AI tools are making it easier for customers to manage their own data without leaning so heavily on data cloud providers.

Micah Zimmerman has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Adobe, CrowdStrike, Palantir Technologies, Salesforce, ServiceNow, and Snowflake. The Motley Fool recommends the following options: long January 2028 $330 calls on Adobe and short January 2028 $340 calls on Adobe. The Motley Fool has a disclosure policy.