Adobe's AI Revenue Tripled Past Half a Billion Dollars. It Still Won't Tell You the One Thing That Matters.

The bet on Adobe is that Firefly wins new ground against free AI tools, not just defends the old. A tripling AI revenue line looks like a win.

But revenue cannot tell additive from defensive. The number that can, net new customers, is the one Adobe does not make easy to see.

Adobe did something this quarter that should have quieted the doubters. Its AI revenue tripled in a year and passed half a billion dollars. For a company the market spent two years calling the next casualty of artificial intelligence, that is a loud number.

And it still does not answer the only question that matters. Is that money Adobe winning new customers, or Adobe paying to keep the ones it already had from leaving.

The bet is additive. The fear is defensive.

When we ran Adobe through the Stock Story Firewall, the assumption underneath the stock came out clean.

The bet investors are making is that Adobe’s Firefly AI will accelerate net new revenue and prevent churn, rather than being neutralized by free or cheaper AI-native competitors.

Those are two very different jobs hiding inside one stock price. Accelerate net new revenue means Firefly is bringing in growth Adobe would not otherwise have. That is additive. Prevent churn means Firefly is holding customers who were being tempted away by Canva, Midjourney, and a wave of free tools. That is defensive.

Additive growth means the moat is widening. Defensive growth means the moat is being defended. The first is worth a premium. The second is worth a discount. And the tripling AI number cannot tell you which one you are looking at.

The moat, not the moment, is the question

Nobody serious doubts Adobe has a moat. The content library, the creator trust, the enterprise relationships, decades of being the default. That is not the debate.

The debate is whether generative AI is narrowing that moat faster than Adobe can monetize its own AI. Canva keeps climbing. Free image tools keep improving. Figma, the collaboration layer Adobe was blocked from buying, is now an independent competitor. The question is not whether Adobe is good. It is whether the ground is shifting under it.

So the supporting evidence and the weakening evidence point at the same spot from opposite sides. Supporting: AI revenue is now a disclosed and growing line, enterprise demand looks firm, and customers are upgrading to pricier tiers. Weakening: the things that would actually reveal erosion, net new subscriber counts, the margin cost of all that AI compute, real defection of casual creators to free tools, are exactly the things the headline numbers do not isolate.

The number that would settle it, and where it hides

Here is the mechanism that matters.

When an existing customer upgrades to a plan that includes AI credits, Adobe’s recurring revenue goes up. When a brand-new customer joins because of Firefly, Adobe’s recurring revenue also goes up. On the revenue line, those two events are identical. One is the moat expanding. One is the moat being defended at a higher price. You cannot tell them apart from dollars.

What separates them is net new subscribers. Not revenue. People. Are more customers choosing Adobe than before, or is a flat base simply paying more.

That is the figure Adobe does not put in front of you cleanly, especially after reshaping how it reports. We get the AI revenue. We do not get a clean count of whether the customer base is actually growing because of it.

The case that Firefly is genuinely winning

The bull case has real support, and it should be stated fairly.

AI revenue tripling past half a billion dollars in a year is not nothing, and Adobe chose to disclose it, which a company tends to do when the trend flatters it. Management raised its full-year targets. Enterprise demand for AI marketing tools is real and growing. And the higher-tier upgrades are themselves a sign customers value the AI enough to pay up for it.

If those upgrades are landing alongside a growing customer base, the additive story is true and the stock is cheap for what it will earn.

That is a real possibility. It is just not yet shown.

The test, in three questions

The Firewall left three questions, and they cut straight to it.

First, disclosure. Does Adobe ever break out AI revenue and customer growth cleanly enough to verify the monetization claim, or does it stay bundled in a way that conveniently cannot be checked.

Second, the buyback. Strip share repurchases out of Adobe’s earnings-per-share growth over the past two years. What is the organic rate underneath, and does it justify the multiple without assuming an AI acceleration that has not yet shown up in revenue.

Third, the bodies. Has the actual count of net new creative subscribers grown, stalled, or fallen over the last two years. That single trajectory tells you whether Firefly is recruiting or merely retaining.

What we are watching, and what it teaches

This is a Watch. Adobe is a genuinely high-quality business with a real, durable moat. But the story the stock is priced on, that AI is accelerating growth rather than defending it, has not produced the hard evidence yet. The numbers that would prove it are still missing or bundled.

And here is the idea to keep. A growing AI revenue line can mean two opposite things. It can mean a company is winning new ground, or it can mean a company is paying to hold ground it is quietly losing. Recurring revenue cannot tell those apart, because both show up as more dollars. Only net new customers can. So when a company shows you a beautiful AI number but not a clean count of new customers, do not assume you are seeing growth. You may be watching a very expensive defense.

Adobe may be winning. The moat may be widening. But until the customer count confirms it, the tripling number is a headline, not an answer.

We will be reading the subscriber and margin lines every quarter until the question is settled. Subscribe to get the read the day the evidence lands.

Not investment advice. The subscriber decides.