Latest / AI Investment Edge: Stocks, Markets & Money / EP 45 | Adobe Stock Analysis: Can Firefly Pay Off?
Transcript
- 0:00SPEAKER_00: I want to start today with a scenario that is probably playing out
- 0:04SPEAKER_00: in corporate offices all over the world right now.
- 0:06SPEAKER_01: Oh, absolutely. It's everywhere.
- 0:08SPEAKER_00: Right. So imagine you manage a creative team
- 0:11SPEAKER_00: and suddenly you give them artificial intelligence tools that allow them
- 0:15SPEAKER_00: to produce like double the amount of work in half the time.
- 0:18SPEAKER_01: Yeah. Which sounds great on paper.
- 0:19SPEAKER_00: Exactly. But the question is, do you still need to pay for the exact same number
- 0:24SPEAKER_00: of software licenses if your team shrinks?
- 0:26SPEAKER_00: I mean, if a creative team does more with fewer people,
- 0:29SPEAKER_00: does Adobe just lose those subscription seats or, you know,
- 0:32SPEAKER_00: can they find a clever way to charge for all that extra output?
- 0:35SPEAKER_01: That is, well, that's really the defining question for the entire software industry right now.
- 0:39SPEAKER_00: Yeah.
- 0:40SPEAKER_01: It completely reframes how we think about productivity and, you know,
- 0:44SPEAKER_01: monetization in the AI era.
- 0:46SPEAKER_01: Because if you're just making your users faster
- 0:49SPEAKER_01: so that they can eventually cancel half their subscriptions.
- 0:52SPEAKER_00: You're just cannibalizing your own business.
- 0:54SPEAKER_01: Right. You aren't innovating.
- 0:55SPEAKER_01: You're just eating your own revenue.
- 0:57SPEAKER_00: Exactly. And the market is completely torn on how this plays out.
- 1:01SPEAKER_00: So welcome to this deep dive.
- 1:02SPEAKER_00: Today, we are looking at the AI investment edge from iPulse AI.
- 1:07SPEAKER_00: Specifically, this is episode 45 with research dated November 16th, 2026.
- 1:12SPEAKER_01: It's a really thorough report.
- 1:14SPEAKER_00: It is. And before we get into the actual mechanics of how Adobe might pull this off,
- 1:18SPEAKER_00: let's just establish the baseline.
- 1:20SPEAKER_00: The source has published a strong buy rating at both the short and long fixed horizons.
- 1:25SPEAKER_00: Plus, there's a separate current action label of strong buy and keep.
- 1:28SPEAKER_01: Right.
- 1:28SPEAKER_00: And just to anchor our conversation today,
- 1:30SPEAKER_00: this research uses the November 13th close of about $243 a share.
- 1:34SPEAKER_01: And we should clarify right up front that these are research labels.
- 1:38SPEAKER_00: Yeah. Very important point.
- 1:39SPEAKER_01: They're based on specific analytical frameworks.
- 1:41SPEAKER_01: They are not financial guarantees and they definitely aren't promises
- 1:45SPEAKER_01: of future stock performance.
- 1:46SPEAKER_00: OK, so let's really unpack this.
- 1:48SPEAKER_01: Right.
- 1:48SPEAKER_00: Because I'm looking at these strong buy ratings,
- 1:50SPEAKER_00: but I am also looking at this massive looming threat of AI efficiency.
- 1:55SPEAKER_01: The seat shrinkage.
- 1:56SPEAKER_00: Right.
- 1:57SPEAKER_00: If companies are getting more output per employee and thus hiring fewer employees,
- 2:02SPEAKER_00: it just seems obvious that seat counts will drop.
- 2:05SPEAKER_00: So if teams shrink, does efficiency actually hurt Adobe's bottom line?
- 2:10SPEAKER_01: To answer if efficiency hurts the bottom line,
- 2:12SPEAKER_01: we really have to look closely at how Adobe's revenue actually flows.
- 2:17SPEAKER_01: You have to strip it down to the bare mechanics.
- 2:18SPEAKER_00: Okay, let's do that.
- 2:20SPEAKER_01: So historically, core products like Creative Cloud, Acrobat,
- 2:25SPEAKER_01: and their enterprise marketing tools, they rely on a strict paid subscription model.
- 2:29SPEAKER_00: Right, the classic software as a service thing.
- 2:31SPEAKER_00: You pay per seat, per user, every month.
- 2:34SPEAKER_01: Precisely.
- 2:35SPEAKER_01: But here's the causality with AI that people miss.
- 2:39SPEAKER_01: An AI generator can, you know, instantly create a stunning initial draft of a marketing image.
- 2:43SPEAKER_00: Yeah, it takes two seconds.
- 2:45SPEAKER_01: Right. But in a professional setting, that draft is never, ever the final product.
- 2:51SPEAKER_01: Human users still have to come in.
- 2:52SPEAKER_01: They have to do the layered editing,
- 2:54SPEAKER_01: apply strict corporate brand rules and route it through these complex approvals.
- 2:58SPEAKER_00: Oh, so it's not a finished piece.
- 3:00SPEAKER_01: Exactly.
- 3:01SPEAKER_01: Thus, usage-based credits, where you pay for the heavy processing of generating
- 3:05SPEAKER_01: that AI content, can complement those traditional paid seats.
- 3:10SPEAKER_01: They're betting that the sheer volume of AI generation will bridge the gap
- 3:13SPEAKER_01: and offset the loss of traditional subscriptions.
- 3:16SPEAKER_00: That makes sense in theory, but I really want to look at the hard financial evidence.
- 3:21SPEAKER_00: Like, is this actually happening?
- 3:23SPEAKER_01: We can definitely look at the numbers.
- 3:24SPEAKER_00: Yeah, because the research lays out four distinct financial findings.
- 3:28SPEAKER_00: Finding one is earnings and cash.
- 3:31SPEAKER_00: For the August fiscal third quarter, revenue was about $6.8 billion.
- 3:35SPEAKER_01: Which is up 13% year over year.
- 3:37SPEAKER_00: Right. And a stat that really jumps off the page.
- 3:40SPEAKER_00: Adobe noted that annualized recurring revenue for AI-first products grew more than 150%.
- 3:47SPEAKER_01: Now, I want to pause on that number because context is crucial here.
- 3:50SPEAKER_01: Growth of more than 150% is amazing, but we have to clarify what that actually is.
- 3:55SPEAKER_00: It's not the whole company.
- 3:56SPEAKER_01: No, it is a subset and it's an annualized metric.
- 3:59SPEAKER_01: It is not total company sales and it does not represent cash they've already received.
- 4:04SPEAKER_01: It's just projecting the current momentum forward.
- 4:06SPEAKER_00: That is a huge distinction so we don't get totally blinded by the hype.
- 4:10SPEAKER_00: Now, talking about actual cash,
- 4:12SPEAKER_00: the source frames the nine-month cash bridge using some really clear phrasing.
- 4:16SPEAKER_01: Walk us through it.
- 4:18SPEAKER_00: So they generated roughly $7.6 billion in operating cash.
- 4:22SPEAKER_00: You minus about $180 million in property and equipment purchases,
- 4:27SPEAKER_00: and that leaves about $7.5 billion of free cash flow.
- 4:31SPEAKER_01: It's a massive cash engine.
- 4:33SPEAKER_01: But we have to remember that acquisition and buyback cash are kept entirely separate
- 4:37SPEAKER_01: from that calculation.
- 4:38SPEAKER_00: Right. So that seven and a half billion is just the core engine.
- 4:41SPEAKER_00: But if they're generating that much cash, why is everyone so anxious about seat attrition?
- 4:47SPEAKER_01: Well, because free cash flow is backward looking.
- 4:50SPEAKER_01: The anxieties about what happens to profit margins two years from now
- 4:54SPEAKER_01: as AI usage explodes, which brings us to finding two margins.
- 4:59SPEAKER_00: Yeah, because generating all this AI content isn't free.
- 5:02SPEAKER_01: Exactly.
- 5:03SPEAKER_01: There is a fascinating contrast here.
- 5:05SPEAKER_01: Under standard accounting, their operating margin was about 35%,
- 5:09SPEAKER_01: which was actually lower than the year earlier.
- 5:12SPEAKER_00: But Adobe's adjusted operating margin was about 44%.
- 5:16SPEAKER_01: Right. And we have to explain the causality there.
- 5:19SPEAKER_01: Why did the standard margin drop?
- 5:21SPEAKER_01: Because marketing and product development still cost money, even as subscriptions grow.
- 5:26SPEAKER_00: AI models are expensive to run.
- 5:27SPEAKER_01: Very. Every time a user generates an image, it costs Adobe compute power.
- 5:32SPEAKER_01: The sheer computing and labor required drag on those standard margins.
- 5:36SPEAKER_00: Which means they need a fortress of a balance sheet to weather this.
- 5:40SPEAKER_00: That's finding three. Looking at their balance sheet,
- 5:42SPEAKER_00: they have about $5.6 billion in cash and short term investments
- 5:47SPEAKER_00: against about $6.4 billion of debt.
- 5:49SPEAKER_01: With about $1.6 billion of that classified as current, meaning due soon.
- 5:54SPEAKER_00: Right.
- 5:55SPEAKER_00: And this is where that cash engine gets pulled in different directions because debt
- 5:59SPEAKER_00: and buybacks actively compete for that same cash.
- 6:02SPEAKER_01: And speaking of buybacks, we have to highlight this fascinating mathematical quark
- 6:06SPEAKER_01: from the report.
- 6:07SPEAKER_00: Oh, I loved this part.
- 6:08SPEAKER_01: Yeah. In the first nine months, lower buyback cash spending,
- 6:12SPEAKER_01: dropping from about $8.8 billion to about $6.8 billion,
- 6:16SPEAKER_01: actually accompanied more shares delivered to Adobe, not fewer.
- 6:20SPEAKER_00: Which sounds insane at first.
- 6:22SPEAKER_01: You spend $2 billion less, but get more stock.
- 6:25SPEAKER_00: Right.
- 6:26SPEAKER_01: But the math is simple.
- 6:27SPEAKER_00: Because the average price paid was lower during that period.
- 6:30SPEAKER_00: The stock dipped, so their cash just went further.
- 6:32SPEAKER_00: No share count needed to understand that.
- 6:34SPEAKER_01: It's just basic arithmetic.
- 6:36SPEAKER_00: It is.
- 6:37SPEAKER_01: But we should also explain the conditional EPS math here.
- 6:41SPEAKER_01: If you have an unchanged profit,
- 6:43SPEAKER_01: but you have fewer diluted shares because of buybacks, your EPS goes up.
- 6:47SPEAKER_00: It artificially boosts the metric.
- 6:49SPEAKER_01: Exactly. But it's conditional.
- 6:51SPEAKER_01: Changing earnings matter.
- 6:53SPEAKER_01: And employee stock issuance matters because they hand out new shares to engineers constantly.
- 6:58SPEAKER_01: And crucially, there is no buyback price floor.
- 7:01SPEAKER_00: Just because they're buying stock doesn't prevent it from falling.
- 7:04SPEAKER_00: Okay, let's tie this to finding four, investment intensity.
- 7:08SPEAKER_00: This part genuinely confused me at first.
- 7:11SPEAKER_01: The CapEx numbers?
- 7:12SPEAKER_00: Yeah. They spent less than $200 million on property and equipment
- 7:15SPEAKER_00: against nearly $20 billion in nine-month sales.
- 7:19SPEAKER_00: Meanwhile, quarterly R&D was about $1.3 billion.
- 7:22SPEAKER_01: That's a huge gap.
- 7:23SPEAKER_00: Right.
- 7:24SPEAKER_00: So wait, how can their AI investment be so massive if they are spending so little
- 7:28SPEAKER_00: on equipment purchases?
- 7:30SPEAKER_01: That is the exact right question.
- 7:32SPEAKER_01: And the answer is that rented compute is still a cash cost.
- 7:36SPEAKER_01: Adobe doesn't have to buy the servers.
- 7:38SPEAKER_01: They can pay another company to run compute on its behalf.
- 7:41SPEAKER_00: Ah, so it shows up as a hosting expense.
- 7:44SPEAKER_01: Exactly. It shows up as a hosting expense rather than a server purchase.
- 7:48SPEAKER_01: It still drags down margins, but it's not a capital expenditure.
- 7:51SPEAKER_01: Also, note that the entire R&D bill isn't just AI spending.
- 7:55SPEAKER_01: It includes heavy labor and training costs, too.
- 7:57SPEAKER_00: Okay, let's connect all these financial mechanics to the actual target dates in the research.
- 8:03SPEAKER_00: Let's talk valuation and the three scenarios.
- 8:05SPEAKER_01: We should start with a base scenario.
- 8:07SPEAKER_00: Right. So from the November 13th close of about $243,
- 8:11SPEAKER_00: the base scenario targets about $305 in the third quarter of 2027.
- 8:15SPEAKER_01: Which represents around 25% cumulative price upside.
- 8:18SPEAKER_00: Yes. And looking further, a target of about $576 in the third quarter
- 8:23SPEAKER_00: of 2031 means around 137% cumulative price upside.
- 8:28SPEAKER_01: And I really want to emphasize the parameters here.
- 8:30SPEAKER_01: These are fixed September dates.
- 8:32SPEAKER_01: They are not rolling targets, they are not annualized returns,
- 8:36SPEAKER_01: and they are definitely not promises.
- 8:38SPEAKER_00: Right. Zero dividends are assumed to.
- 8:40SPEAKER_01: Exactly. And no invented endpoints or probabilities should be given.
- 8:44SPEAKER_01: We're just looking at scenarios based on the data.
- 8:47SPEAKER_00: So what happens in that base scenario?
- 8:48SPEAKER_01: Well, in the base case, usage revenue and professional workflows offset some seat pressure.
- 8:54SPEAKER_01: It's a steady transition.
- 8:55SPEAKER_00: Okay. And then what about the bull scenario?
- 8:57SPEAKER_01: In the bull case, Firefly credits and GenStudio become regular paid production tools.
- 9:02SPEAKER_01: This drives bigger renewals, while model competition limits Adobe's compute costs.
- 9:06SPEAKER_00: So everything goes perfectly.
- 9:08SPEAKER_00: But the bare scenario is the opposite.
- 9:10SPEAKER_01: Right.
- 9:11SPEAKER_01: The bare scenario assumes companies run projects elsewhere or pay for fewer seats
- 9:16SPEAKER_01: before new usage catches up.
- 9:18SPEAKER_01: Compute costs are higher and the valuation is just lower.
- 9:22SPEAKER_00: So to understand which scenario plays out,
- 9:24SPEAKER_00: we need to map out the forces pushing and pulling Adobe's business
- 9:28SPEAKER_00: into four quadrants. Right.
- 9:29SPEAKER_01: This is a great way to visualize it.
- 9:31SPEAKER_00: Yeah. So quadrant one covers the core drivers, existing professional workflows,
- 9:36SPEAKER_00: documentation, enterprise distribution, sticky subscriptions,
- 9:40SPEAKER_00: and that strong operating cash.
- 9:42SPEAKER_01: Then quadrant two is the frictions, the headwinds.
- 9:45SPEAKER_01: You've got cheap and easy competing tools, the risk of fewer seats,
- 9:49SPEAKER_01: and the high cost of serving AI compute.
- 9:52SPEAKER_00: Quadrant three is the upside tail.
- 9:54SPEAKER_01: Right. Enterprise standardization.
- 9:57SPEAKER_01: Think of a massive brand standardizing on Adobe for repeatable campaign production
- 10:02SPEAKER_01: or tightening approval standards using Adobe's content provenance tools.
- 10:05SPEAKER_00: And finally, quadrant four is the downside tail.
- 10:08SPEAKER_00: This is where Arrival offers reliable end-to-end editing and approval
- 10:12SPEAKER_00: at a lower cost or where copyright liability constraints AI use.
- 10:16SPEAKER_01: And we should note on that point,
- 10:17SPEAKER_01: legal risk is conditional based on contract eligibility in terms.
- 10:20SPEAKER_01: There is no blanket legal guarantee there.
- 10:22SPEAKER_00: That downside risk, the idea of a rival bypassing Adobe,
- 10:27SPEAKER_00: actually highlights a major misunderstanding in the market,
- 10:29SPEAKER_00: which the report calls the alpha gap.
- 10:32SPEAKER_01: The alpha gap is such a crucial concept.
- 10:34SPEAKER_00: It is.
- 10:35SPEAKER_00: So the concept is that the market might be pricing the creation of a first draft
- 10:41SPEAKER_00: as if it's the exact same as controlling everything that follows.
- 10:44SPEAKER_01: Right, but making a draft and finishing a project are totally different.
- 10:48SPEAKER_00: Completely. The real value and where Adobe actually competes is in getting paid
- 10:52SPEAKER_00: for repeatable approvals, version control, revisions, and final delivery.
- 10:57SPEAKER_00: It actually reminds me perfectly of the last published stock we discussed
- 11:01SPEAKER_00: in episode 44, Booking Holdings.
- 11:03SPEAKER_01: Oh, that's a great analogy.
- 11:05SPEAKER_00: Think about it. An AI assistant might help you search for a trip,
- 11:08SPEAKER_00: but booking earns its money when a booking is completed on its platform.
- 11:13SPEAKER_01: They're the toll booth at the end.
- 11:14SPEAKER_00: Exactly.
- 11:15SPEAKER_00: Similarly, Adobe earns its keep by keeping the editing and approval work inside its tools.
- 11:20SPEAKER_00: They get paid at the end of the corporate workflow.
- 11:23SPEAKER_01: And because of this gap between making a draft and finishing a project,
- 11:26SPEAKER_01: the analysts are really split on the timeline of Adobe's success.
- 11:29SPEAKER_00: Right.
- 11:30SPEAKER_00: We have different views from the iPulse AI advisors, and we should name them as such.
- 11:34SPEAKER_00: They are iPulse AI advisors, not just generic industry experts.
- 11:38SPEAKER_01: So you have the stronger view.
- 11:40SPEAKER_00: They expect enterprise workflow sales and credit packs to grow sooner than seat attrition.
- 11:45SPEAKER_01: Then you have the conservative view.
- 11:46SPEAKER_00: Yeah.
- 11:46SPEAKER_00: They see slower growth, with investors paying roughly the same valuation multiple
- 11:50SPEAKER_00: if contracts renew with smaller seat counts.
- 11:53SPEAKER_01: And finally, the skeptic.
- 11:55SPEAKER_00: They want to see disclosed paid conversion,
- 11:57SPEAKER_00: not just a free audience size playing with beta tools.
- 12:00SPEAKER_01: And my insight here is that these advisors share inputs and often the same underlying model.
- 12:05SPEAKER_01: So their consensus isn't independent proof.
- 12:08SPEAKER_01: The difference is purely in their assumptions on the timing and strength
- 12:11SPEAKER_01: of usage revenue versus losing seats.
- 12:14SPEAKER_00: Right. It's all about timing.
- 12:16SPEAKER_01: So to test whose timing assumption is right,
- 12:19SPEAKER_01: we have a series of specific quarterly watch windows.
- 12:22SPEAKER_00: Right. And we need to separate actual events from model watch windows.
- 12:26SPEAKER_01: Let's start with December 2026.
- 12:28SPEAKER_00: We watch guidance and the early priorities of new chief executive Anil Chakravarty,
- 12:32SPEAKER_00: who starts December 1st.
- 12:34SPEAKER_00: That's an actual event.
- 12:35SPEAKER_01: Then in the first quarter of 2027, look for the first evidence of upgrades and paid consumption.
- 12:41SPEAKER_00: And the second quarter of 2027?
- 12:43SPEAKER_01: Watch for repeat multinational GenStudio deployments.
- 12:47SPEAKER_00: By the third quarter of 2027,
- 12:49SPEAKER_00: we compare actual gross margin and renewal performance against serving costs
- 12:53SPEAKER_00: and seat pressure.
- 12:54SPEAKER_01: Exactly.
- 12:55SPEAKER_01: And later on, you watch the later cash per share test, sustained usage billing,
- 13:00SPEAKER_01: and whether larger accounts consolidate onto Adobe through 2031.
- 13:04SPEAKER_00: So why are we watching these specific metrics right now compared to previous reports?
- 13:08SPEAKER_00: Like what changed?
- 13:10SPEAKER_01: Well, we have to distinguish the previous batches.
- 13:12SPEAKER_01: The July batch was centered on established workflows, copyright, buybacks,
- 13:16SPEAKER_01: and if a usage model could ever replace seats.
- 13:19SPEAKER_00: Right.
- 13:20SPEAKER_01: Then the September batch focused
- 13:22SPEAKER_01: on if paid credit consumption would actually show up in enterprise renewals.
- 13:26SPEAKER_00: And if we do a callback to October's original Adobe episode two,
- 13:31SPEAKER_00: we covered the milestone of one billion monthly users.
- 13:33SPEAKER_00: But we have to clarify that measures combined activity across creativity and productivity tools.
- 13:40SPEAKER_01: Right. It is neither a free only count nor a paid subscriber count. It's blended.
- 13:44SPEAKER_00: Yeah. But what's really new is the November 10th product news at Adobe Max.
- 13:48SPEAKER_00: They introduced Firefly Graph in beta and AI assistance also in beta,
- 13:54SPEAKER_00: while Builder remains in private beta.
- 13:56SPEAKER_01: And we must remind the listener, a product launch does not equal paid conversion.
- 14:01SPEAKER_01: Just because it's announced doesn't mean it's making money.
- 14:03SPEAKER_00: True. But the tension between spending and cash here is fascinating.
- 14:07SPEAKER_00: It actually compares perfectly to our deep dive on Microsoft in episode 37.
- 14:11SPEAKER_01: Oh, absolutely.
- 14:12SPEAKER_00: Yeah, with Microsoft, we saw AI demand and spending moving in opposite directions
- 14:17SPEAKER_00: from free cash flow.
- 14:19SPEAKER_00: Though obviously Adobe's hardware ownership needs are different since they rent compute.
- 14:23SPEAKER_01: But the core dilemma of spending heavy now for future growth is the exact same.
- 14:28SPEAKER_00: Exactly.
- 14:29SPEAKER_00: We've brought up Adobe, Microsoft and Booking Holdings today,
- 14:32SPEAKER_00: all wrestling with different parts of this tech shift.
- 14:35SPEAKER_00: But before we wrap up, what could completely invalidate this strong buy thesis for Adobe?
- 14:40SPEAKER_01: The invalidation points are clear.
- 14:42SPEAKER_01: Weak renewals, seat loss before usage catches up,
- 14:46SPEAKER_01: gross margin pressure as AI scales
- 14:47SPEAKER_01: or weakening free cash flow per share could invalidate the thesis.
- 14:52SPEAKER_00: It's all about that margin.
- 14:53SPEAKER_01: Yes.
- 14:54SPEAKER_01: Conversely, paid renewals expanding while margins hold would answer the bear case.
- 14:58SPEAKER_01: And again, opening the Firefly beta isn't a test of cash on its own.
- 15:01SPEAKER_00: So to summarize, AI lets teams do more with less.
- 15:04SPEAKER_00: And Adobe's entire future rests on proving they can charge
- 15:07SPEAKER_00: for that extra output instead of just watching their seat counts evaporate.
- 15:11SPEAKER_00: It's a massive transition.
- 15:13SPEAKER_01: The investment case turns on converting a larger amount of creative work
- 15:16SPEAKER_01: into paid recurring demand without giving the margin away.
- 15:20SPEAKER_00: For the bigger picture, visit iPulse AI to explore our complete rankings
- 15:24SPEAKER_00: of covered assets, discover the strongest buys,
- 15:26SPEAKER_00: and see the research behind the ratings, all in one place.
- 15:29SPEAKER_01: Primary research conducted through the iPulse AI platform by Future Edge Group.
- 15:33SPEAKER_00: Like, comment, and subscribe.
- 15:34SPEAKER_01: We publish a new podcast every day.
- 15:36SPEAKER_00: Stay tuned for the next podcast on D-Wave Quantum.