Latest / AI Investment Edge: Stocks, Markets & Money / Amazon: Bigger Cloud Profits. A Bigger Cash Test. | EP028
Transcript
- 0:00SPEAKER_00: Imagine an airline buying a massive brand new fleet of 747s in pure cash today.
- 0:08SPEAKER_01: Right.
- 0:08SPEAKER_01: Just a massive upfront expense.
- 0:10SPEAKER_00: Exactly.
- 0:11SPEAKER_00: And, you know,
- 0:11SPEAKER_00: the planes look incredible on the balance sheet and
- 0:13SPEAKER_00: the daily ticket sales look like pure profit.
- 0:17SPEAKER_00: But behind the scenes,
- 0:18SPEAKER_00: it takes a decade of flying those planes at full capacity just to earn back
- 0:22SPEAKER_00: the cash they originally burned to buy them.
- 0:24SPEAKER_01: Right.
- 0:25SPEAKER_01: The payback period is huge.
- 0:26SPEAKER_00: Yeah, and that is the exact tension we are exploring today as we dive
- 0:30SPEAKER_00: into Amazon's latest financials for AI Investment Edge, Amazon Episode 28.
- 0:35SPEAKER_00: It
- 0:35SPEAKER_01: really is.
- 0:37SPEAKER_00: Because on paper, Amazon is generating massive operating profit.
- 0:41SPEAKER_00: But, and this is wild, its company-defined trailing free cash flow is actually negative.
- 0:46SPEAKER_01: Yeah, it's a huge paradox.
- 0:48SPEAKER_00: So the big question we're answering for you today...
- 0:51SPEAKER_00: Can today's incredibly expensive cloud build out actually produce tomorrow's spendable cash?
- 0:57SPEAKER_00: So let's unpack this.
- 0:59SPEAKER_00: Welcome to the deep dive.
- 1:00SPEAKER_01: Thanks.
- 1:01SPEAKER_01: I mean, it really is the defining question for the company right now.
- 1:05SPEAKER_01: And just to set our baseline for everyone listening,
- 1:08SPEAKER_01: we are looking at dated September iPulse AI research.
- 1:11SPEAKER_00: Right.
- 1:12SPEAKER_01: And I want to emphasize right up front, this uses a September price basis.
- 1:16SPEAKER_01: So it's not some fresh October rescore.
- 1:18SPEAKER_00: Okay.
- 1:18SPEAKER_00: That's important context.
- 1:19SPEAKER_01: Right.
- 1:20SPEAKER_01: And the model is assigned a neutral one-year label and a buy five-year label.
- 1:25SPEAKER_00: Okay.
- 1:26SPEAKER_00: And just so you listening have the map, we are looking at fixed endpoints here, right?
- 1:30SPEAKER_01: Exactly.
- 1:31SPEAKER_01: Those fixed endpoints fall in the third quarter of 2027 and the third quarter of 2031.
- 1:36SPEAKER_01: Okay.
- 1:36SPEAKER_01: And what's fascinating here is that
- 1:40SPEAKER_01: a strong long-term buy label doesn't actually mean every single entry price is
- 1:45SPEAKER_01: attractive right now.
- 1:46SPEAKER_00: Right.
- 1:46SPEAKER_00: Because things fluctuate.
- 1:48SPEAKER_01: Exactly.
- 1:48SPEAKER_01: So to really understand why the short term looks neutral,
- 1:51SPEAKER_01: but the long term is a buy,
- 1:53SPEAKER_01: we have to look under the hood at the four major financial realities of
- 1:57SPEAKER_01: the business.
- 1:57SPEAKER_00: Yeah.
- 1:58SPEAKER_00: Let's briefly map out that ecosystem so we know exactly what we were evaluating.
- 2:02SPEAKER_01: Sure.
- 2:03SPEAKER_00: Because when we say Amazon today,
- 2:05SPEAKER_00: you're looking at online stores,
- 2:06SPEAKER_00: third-party seller services,
- 2:08SPEAKER_00: the massive logistics network,
- 2:10SPEAKER_00: prime subscriptions, advertising, and obviously Amazon Web Services or AWS.
- 2:14SPEAKER_01: Right.
- 2:15SPEAKER_01: That's the whole package.
- 2:16SPEAKER_00: Right.
- 2:17SPEAKER_00: A quick note for you as you track this, advertising is a revenue category here.
- 2:22SPEAKER_00: It is not a separately reported operating profit segment.
- 2:26SPEAKER_01: Yeah, that's a good distinction.
- 2:28SPEAKER_01: Which brings us to our first major finding, earnings.
- 2:33SPEAKER_00: Okay, hit me with the numbers.
- 2:35SPEAKER_01: So in the second quarter, sales rose 20% to about $201 billion.
- 2:39SPEAKER_00: Wow, $201 billion.
- 2:41SPEAKER_01: Yeah.
- 2:42SPEAKER_01: And operating income reached roughly $27 billion.
- 2:46SPEAKER_01: But then you look at net income and it was suddenly about 63 billion.
- 2:50SPEAKER_00: Wait, hold on.
- 2:52SPEAKER_00: 63 billion?
- 2:53SPEAKER_01: Yeah.
- 2:54SPEAKER_00: That net income is massive compared to the operating income.
- 2:57SPEAKER_00: I mean, is that all real profit just from selling physical goods and renting out cloud space?
- 3:02SPEAKER_01: No.
- 3:03SPEAKER_01: And this is where you have to be incredibly careful reading headline numbers.
- 3:07SPEAKER_00: Okay.
- 3:07SPEAKER_01: You really have to judge repeatable operating performance because
- 3:11SPEAKER_01: that 63 billion includes large private investment revaluations.
- 3:16SPEAKER_00: I see.
- 3:16SPEAKER_01: Specifically,
- 3:18SPEAKER_01: an anthropic valuation gain of about $50 billion pre-tax,
- 3:22SPEAKER_01: which is just tucked into other income.
- 3:24SPEAKER_00: So wait, so it's just a paper gain.
- 3:26SPEAKER_00: It's like, I don't know, your house going up in value.
- 3:29SPEAKER_00: You don't actually have that money in your checking account to go buy groceries.
- 3:32SPEAKER_01: Precisely.
- 3:33SPEAKER_01: It is not AWS sales and it is definitely not cash received.
- 3:39SPEAKER_00: Right.
- 3:39SPEAKER_01: And a critical accounting note for anyone building models out there.
- 3:43SPEAKER_01: You cannot just subtract a pre-tax gain directly from after-tax net income to try
- 3:49SPEAKER_01: and like manufacture an adjusted profit metric.
- 3:52SPEAKER_00: Because it just distorts the reality of the business.
- 3:55SPEAKER_01: Exactly.
- 3:57SPEAKER_00: Okay.
- 3:57SPEAKER_00: So if that massive net income is heavily influenced by paper gains,
- 4:00SPEAKER_00: let's look at the actual operations, which is finding number two, profitability.
- 4:05SPEAKER_01: Right.
- 4:06SPEAKER_00: So AWS generated about $17 billion of operating income on $42 billion of sales.
- 4:12SPEAKER_00: That is a reported operating margin of roughly 39%, while the whole group margin is about 14%.
- 4:19SPEAKER_01: Yeah, quite a gap.
- 4:21SPEAKER_00: But 39% is incredible.
- 4:22SPEAKER_00: Does that mean these AI servers are just like minting pure profit?
- 4:26SPEAKER_01: I have to pause you there because the reality of that 39% is a bit more complex.
- 4:32SPEAKER_00: Oh, really?
- 4:33SPEAKER_01: Yeah.
- 4:34SPEAKER_01: That margin actually includes a roughly half billion dollar unrealized energy contract benefit.
- 4:39SPEAKER_00: Okay, so more accounting quirks?
- 4:41SPEAKER_01: Exactly.
- 4:42SPEAKER_01: It's a limited component of their reported profit, but it is an unrealized gain.
- 4:46SPEAKER_01: It is not realized energy savings lowering their daily power bill.
- 4:50SPEAKER_00: Got it.
- 4:51SPEAKER_01: So the accounting is doing some heavy lifting again to make the margins look a bit sweeter.
- 4:55SPEAKER_00: Yes.
- 4:56SPEAKER_01: And you also have to understand how depreciation works here.
- 4:59SPEAKER_00: Right.
- 4:59SPEAKER_01: Because hardware depreciation is an accounting expense that is recognized slowly as
- 5:04SPEAKER_01: the equipment is used over the years.
- 5:07SPEAKER_00: Okay.
- 5:08SPEAKER_01: That is completely different from actually repaying the debt used to buy that server.
- 5:13SPEAKER_01: That repayment is a financing cash flow that actually leaves your bank account today.
- 5:17SPEAKER_00: Ah, that makes sense.
- 5:18SPEAKER_01: Right.
- 5:19SPEAKER_01: And while advertising and fulfillment efficiency certainly help the bottom line,
- 5:23SPEAKER_01: margins are never immune to the brutal reality of competition.
- 5:27SPEAKER_00: Yeah, which naturally leads us to finding three, the balance sheet.
- 5:32SPEAKER_00: If depreciation is just an accounting expense, let's look at the actual cash.
- 5:37SPEAKER_00: So cash and marketable securities were about $123 billion at the end of June.
- 5:43SPEAKER_00: But that sits against roughly $129 billion of long term debt plus current debt
- 5:49SPEAKER_00: and lease obligations.
- 5:50SPEAKER_01: Yeah.
- 5:50SPEAKER_01: So the big takeaway here is that this $123 billion is financial capacity.
- 5:56SPEAKER_01: But it has massive commitments attached to it.
- 5:58SPEAKER_00: Right.
- 5:58SPEAKER_00: It's already spoken for.
- 5:59SPEAKER_01: Exactly.
- 6:00SPEAKER_01: This is not a debt-free company.
- 6:02SPEAKER_01: Those private investment paper gains we just talked about,
- 6:05SPEAKER_01: they do not pay suppliers who are building your data centers.
- 6:08SPEAKER_00: Yeah.
- 6:08SPEAKER_00: Suppliers want actual cash.
- 6:10SPEAKER_01: Exactly.
- 6:11SPEAKER_01: And you can never assume that refinancing your debt is guaranteed in the future.
- 6:15SPEAKER_00: Right.
- 6:16SPEAKER_00: So if the paper gains aren't real cash and the debt is very real,
- 6:21SPEAKER_00: we have to look at Finding Four, the real heavy hitter.
- 6:24SPEAKER_00: Investment intensity.
- 6:26SPEAKER_01: Yeah, this is the big one.
- 6:28SPEAKER_00: Let's walk through the specific cash bridge.
- 6:30SPEAKER_00: Walk me through that.
- 6:31SPEAKER_01: Sure.
- 6:31SPEAKER_01: So this is where the rubber meets the road.
- 6:34SPEAKER_01: You take $161 billion of operating cash.
- 6:37SPEAKER_00: Okay, $161 billion.
- 6:39SPEAKER_01: Then you subtract $169 billion of net property and equipment purchases.
- 6:44SPEAKER_00: Oh, wow.
- 6:44SPEAKER_01: Right.
- 6:45SPEAKER_01: That equals about $7.6 billion of company-defined free cash outflow.
- 6:50SPEAKER_01: And that's on a trailing 12-month basis.
- 6:52SPEAKER_00: Outflow.
- 6:53SPEAKER_00: Right.
- 6:54SPEAKER_00: Meaning they are bleeding $7.6 billion in free cash because the infrastructure is so expensive.
- 6:59SPEAKER_01: Exactly.
- 7:00SPEAKER_01: Investment is growing faster than incoming operating cash.
- 7:03SPEAKER_01: The implication is stark.
- 7:05SPEAKER_01: The ultimate return on this massive growth still needs to be earned.
- 7:09SPEAKER_00: Right.
- 7:10SPEAKER_01: Now, these purchases support general technology, AWS, and fulfillment.
- 7:14SPEAKER_01: Amazon does not give us a neat maintenance-only split,
- 7:18SPEAKER_01: nor is there any claim that all of this capital expenditure is purely AI.
- 7:23SPEAKER_00: Okay, that's good to know.
- 7:24SPEAKER_01: And remember, having a massive customer backlog of future orders is not cash in the bank today.
- 7:30SPEAKER_00: Yeah, you know,
- 7:31SPEAKER_00: this entire dynamic really reminds me of our earlier deep dive into Alphabet,
- 7:34SPEAKER_00: the one called AI Boom, Where's the Cash?
- 7:36SPEAKER_01: Oh, absolutely.
- 7:37SPEAKER_00: The thematic comparison there is incredibly strong because both
- 7:41SPEAKER_00: of these tech giants are asking investors to essentially take a leap of faith
- 7:45SPEAKER_00: to assess what this huge,
- 7:46SPEAKER_00: expensive AI investment will eventually earn years down the road.
- 7:50SPEAKER_01: It's the exact same fundamental question.
- 7:52SPEAKER_00: So with this severe cash intensity in mind,
- 7:56SPEAKER_00: how does the evidence weight the actual share price over our fixed horizons?
- 8:02SPEAKER_01: Let's set our current reference first.
- 8:04SPEAKER_01: We are using a price of about $250 per ordinary share.
- 8:08SPEAKER_01: And just for clarity,
- 8:09SPEAKER_01: that is based on an IEX trade dated October 22, just before 4 p.m. in New York.
- 8:15SPEAKER_01: It is not a certified consolidated closing price.
- 8:18SPEAKER_00: Okay, so $250 is our starting line.
- 8:21SPEAKER_00: Where does the base case take us from there?
- 8:23SPEAKER_01: From that reference,
- 8:24SPEAKER_01: the supported evidence-weighted base endpoints are about $279 in the third quarter
- 8:28SPEAKER_01: of 2027 and $446 in the third quarter of 2031.
- 8:29SPEAKER_01: Wait, so we are looking at cumulative price returns of roughly 11% by 2027 and 78% by 2031?
- 8:32SPEAKER_01: Correct.
- 8:43SPEAKER_00: And to be crystal clear for you listening, these are cumulative returns, not annual returns.
- 8:49SPEAKER_00: Plus we are assuming zero future ordinary dividend cash.
- 8:52SPEAKER_00: So there is no dividend uplift baked into those numbers.
- 8:55SPEAKER_01: Right.
- 8:55SPEAKER_01: Zero dividends.
- 8:56SPEAKER_00: But I have to push back here.
- 8:59SPEAKER_00: 78% upside over five years?
- 9:01SPEAKER_00: I mean, how is that physically possible when they are currently in
- 9:04SPEAKER_00: a $7.6 billion free cash outflow?
- 9:07SPEAKER_00: What has to go right for them to pull that off?
- 9:10SPEAKER_00: And obviously we don't promise actual returns, but based on the model, what's the logic?
- 9:14SPEAKER_01: Yeah, so that is the qualitative bull case right there.
- 9:18SPEAKER_01: For that upside to happen,
- 9:19SPEAKER_01: billable AWS workloads,
- 9:21SPEAKER_01: advertising, and better logistics have to grow much faster than their capital needs.
- 9:25SPEAKER_00: Yeah, I see.
- 9:26SPEAKER_01: Basically,
- 9:27SPEAKER_01: the cash conversion improves drastically,
- 9:30SPEAKER_01: and the returns finally justify today's massive build-out.
- 9:33SPEAKER_00: But the bear case is the exact opposite, right?
- 9:35SPEAKER_01: Exactly.
- 9:36SPEAKER_00: Customers use less capacity than expected,
- 9:39SPEAKER_00: replacing and depreciating the hardware stays incredibly expensive,
- 9:42SPEAKER_00: and competitors force Amazon to lower their prices just to keep market share.
- 9:47SPEAKER_01: That's the classic infrastructure cycle risk.
- 9:49SPEAKER_00: Yeah, we actually talked about this exact dynamic from the supplier's angle
- 9:52SPEAKER_00: in our NVIDIA deep dive.
- 9:53SPEAKER_01: Right, right.
- 9:55SPEAKER_01: It's the same cycle viewed from two different ends of the supply chain.
- 9:58SPEAKER_01: Amazon buys and develops its own custom silicon, while Nvidia sells the computing platforms.
- 10:03SPEAKER_00: Okay, so if the path to hitting that 78% long-term return relies
- 10:07SPEAKER_00: on specific drivers winning out against distinct frictions,
- 10:11SPEAKER_00: let's look at the four quadrants of that battle.
- 10:14SPEAKER_01: Let's break it down.
- 10:15SPEAKER_01: The ordinary drivers are straightforward,
- 10:17SPEAKER_01: cloud demand successfully translating into billable consumption,
- 10:21SPEAKER_01: highly useful advertising placed right beside shopping intent,
- 10:25SPEAKER_01: the successful development of custom silicon,
- 10:28SPEAKER_01: and increasingly efficient fulfillment.
- 10:30SPEAKER_00: And the ordinary frictions fighting against those drivers?
- 10:33SPEAKER_01: The sheer cash required to physically build the capacity,
- 10:37SPEAKER_01: power and equipment availability bottlenecks,
- 10:40SPEAKER_01: hardware depreciation dragging on the books,
- 10:43SPEAKER_01: brutal competition, And,
- 10:45SPEAKER_01: you know,
- 10:45SPEAKER_01: the valuation sensitivity of relying on distant future cash flows instead
- 10:49SPEAKER_01: of money today.
- 10:51SPEAKER_00: Okay, those are the ordinary day-to-day battles the company fights.
- 10:55SPEAKER_00: What about the distinct tails, like the unexpected futures?
- 10:59SPEAKER_01: Well, the distinct upside tail is a conditional surprise where
- 11:03SPEAKER_01: the investment payback cycle is simply faster than expected.
- 11:06SPEAKER_00: Okay, so they make their money back quicker.
- 11:08SPEAKER_01: Exactly.
- 11:09SPEAKER_01: The massive capital spending stabilizes,
- 11:11SPEAKER_01: while the capacity they've already built starts generating substantial,
- 11:15SPEAKER_01: durable cash.
- 11:17SPEAKER_00: And the distinct downside tail?
- 11:19SPEAKER_01: That would be a broad AI spending retrenchment across the economy,
- 11:23SPEAKER_01: which leaves all this incredibly costly Amazon capacity sitting underused.
- 11:27SPEAKER_00: Ouch.
- 11:29SPEAKER_01: Or, alternatively,
- 11:30SPEAKER_01: a severe regulatory remedy that actively weakens their core marketplace model.
- 11:35SPEAKER_00: Wow, okay.
- 11:37SPEAKER_00: So we have all this data, the paper gains, the negative cash flow, the 78% potential upside.
- 11:44SPEAKER_00: Where is the market getting it wrong right now?
- 11:45SPEAKER_00: Like, where is the alpha gap?
- 11:48SPEAKER_01: So the core hypothesis here is that investors might be misreading the situation
- 11:52SPEAKER_01: in two completely different directions.
- 11:54SPEAKER_00: Oh, really?
- 11:55SPEAKER_01: Yeah.
- 11:56SPEAKER_01: First, they might be misreading the headline earnings,
- 11:59SPEAKER_01: assuming the company is wildly profitable,
- 12:01SPEAKER_01: without realizing those numbers are heavily flattered by investment revaluations.
- 12:05SPEAKER_00: Right, the paper gains.
- 12:06SPEAKER_01: Exactly.
- 12:08SPEAKER_01: And second,
- 12:08SPEAKER_01: they might be misreading the negative near term cash flow,
- 12:11SPEAKER_01: which is currently depressed by real physical build outs before the returns arrive.
- 12:16SPEAKER_00: So what does this all mean for you as you're analyzing this?
- 12:20SPEAKER_00: Here is the real paradox.
- 12:22SPEAKER_00: Negative cash flow alone doesn't prove that it's wasted investment.
- 12:26SPEAKER_00: You have to buy the plane before you can sell tickets.
- 12:29SPEAKER_01: Exactly.
- 12:30SPEAKER_00: But on the flip side,
- 12:31SPEAKER_00: strong operating profit alone doesn't prove the build out pays off
- 12:34SPEAKER_00: because accounting rules make margins look better than the actual cash in the bank.
- 12:38SPEAKER_01: Precisely.
- 12:39SPEAKER_01: The true test of this business right now is
- 12:42SPEAKER_01: whether this incredibly useful capacity can earn more than its financing,
- 12:46SPEAKER_01: its running costs and its eventual replacement costs.
- 12:49SPEAKER_00: But not everyone looking at the models agrees on when or even if
- 12:53SPEAKER_00: that capacity will ultimately cross that profitability finish line.
- 12:56SPEAKER_01: Now, there is real advisor disagreement here.
- 12:59SPEAKER_01: The central debate is all about timing and the actual return on invested capital.
- 13:04SPEAKER_01: The optimists are pointing to the custom chips and logistics efficiency as
- 13:08SPEAKER_01: a clear path to eventual cash conversion.
- 13:11SPEAKER_01: But the skeptics emphasize power bottlenecks,
- 13:13SPEAKER_01: the ongoing drag of depreciation, and intense pricing pressure from competitors.
- 13:18SPEAKER_00: And just to clarify the structure of the models in the research packet,
- 13:22SPEAKER_00: we are looking at 14 contributions across three model versions
- 13:25SPEAKER_00: with 12 sharing one specific version.
- 13:27SPEAKER_01: Right.
- 13:28SPEAKER_00: Just remember, these are correlated AI perspectives.
- 13:30SPEAKER_00: They are not 14 independent human analysts sitting in a room arguing.
- 13:34SPEAKER_01: No, definitely not.
- 13:35SPEAKER_00: So when will we get the next real world clue to settle this disagreement?
- 13:39SPEAKER_00: Let's trace the quarterly path.
- 13:41SPEAKER_01: Sure.
- 13:42SPEAKER_01: So if we look at the scenario checkpoints,
- 13:44SPEAKER_01: the fourth quarter of 2026 is about 3.7% above our reference price.
- 13:50SPEAKER_00: Okay.
- 13:50SPEAKER_01: The third quarter of 2027 is around 11%.
- 13:54SPEAKER_01: But the big cash conversion test hits around the second quarter of 2028,
- 14:00SPEAKER_01: where the weighted price path climbs to about 24% above the reference.
- 14:04SPEAKER_00: And our next confirmed event to test this path is Amazon's Q3 results call happening
- 14:09SPEAKER_00: on October 29 at 2 p.m. Pacific time.
- 14:11SPEAKER_01: Right.
- 14:12SPEAKER_00: And just to note, that's a future event, so no Dubai time zone conversion needed there.
- 14:17SPEAKER_00: What are we looking for in that call?
- 14:18SPEAKER_01: Well, back in July,
- 14:19SPEAKER_01: the guidance was about $197 to $202 billion in sales and roughly $23 to $27 billion
- 14:22SPEAKER_01: in operating income.
- 14:27SPEAKER_01: What you really need to watch is AWS growth, actual investment, and cash conversion altogether.
- 14:33SPEAKER_01: Do not just look at a single headline beat on earnings per share.
- 14:37SPEAKER_00: Yeah, it's a huge balancing act,
- 14:38SPEAKER_00: spending massive capital today against a very long operating payback.
- 14:43SPEAKER_00: It's actually a great time to mention our SpaceX deep dive as a perfect comparison
- 14:47SPEAKER_00: for exactly that kind of capital-intensive growth business.
- 14:50SPEAKER_01: Oh, absolutely.
- 14:51SPEAKER_00: You're building rockets,
- 14:52SPEAKER_00: or in this case,
- 14:53SPEAKER_00: data centers, hoping the long-term payoff dwarfs the short-term burn.
- 14:58SPEAKER_00: Speaking of how things evolve over time,
- 14:59SPEAKER_00: the underlying research thesis on Amazon has actually shifted significantly since
- 15:03SPEAKER_00: the summer.
- 15:04SPEAKER_01: It really has.
- 15:05SPEAKER_01: The long horizon label changed from neutral in July to buy in September.
- 15:09SPEAKER_00: Okay, why?
- 15:10SPEAKER_01: Well, back in July, the thesis emphasized resilient customer demand.
- 15:14SPEAKER_01: And to be clear,
- 15:15SPEAKER_01: that means defensive demand for their services,
- 15:17SPEAKER_01: not defensive spending by the company.
- 15:19SPEAKER_00: Got it.
- 15:20SPEAKER_00: And what changed in September to bump it to a buy?
- 15:23SPEAKER_01: September retained that defensive demand view, but it significantly sharpened the business test.
- 15:27SPEAKER_00: Okay.
- 15:28SPEAKER_01: The view now is that stronger AWS demand absolutely must become collected cash,
- 15:33SPEAKER_01: especially as mounting debt
- 15:34SPEAKER_01: and lease obligations narrow Amazon's financial flexibility.
- 15:38SPEAKER_00: Right.
- 15:38SPEAKER_00: You know, that is a totally different dynamic than our last public stock deep dive on Medtronic.
- 15:43SPEAKER_01: Oh, for sure.
- 15:44SPEAKER_00: They are in a completely different sector,
- 15:46SPEAKER_00: and their focus right now is simplifying their portfolio.
- 15:50SPEAKER_00: Amazon is doing the exact opposite.
- 15:52SPEAKER_00: They are vastly expanding their capital commitments.
- 15:55SPEAKER_01: Yeah, completely different strategies.
- 15:57SPEAKER_00: It highlights two completely different capital allocation problems.
- 16:00SPEAKER_00: Mentronic is deciding what to keep versus build,
- 16:03SPEAKER_00: while Amazon is deciding whether
- 16:05SPEAKER_00: an incredibly expensive buildout actually improves their future cash returns.
- 16:09SPEAKER_01: Right.
- 16:10SPEAKER_01: Medtronic serves as a great baseline comparison for general capital allocation,
- 16:14SPEAKER_01: while Alphabet, NVIDIA,
- 16:17SPEAKER_01: and SpaceX are your core thematic companions
- 16:19SPEAKER_01: for understanding this specific massive infrastructure cycle.
- 16:23SPEAKER_00: Absolutely.
- 16:24SPEAKER_00: So with the cash test sharpened so dramatically, what would actually break this model?
- 16:29SPEAKER_00: Let's talk about invalidation.
- 16:31SPEAKER_00: How does the bull case fall apart?
- 16:33SPEAKER_01: To weaken the long-term bullish case,
- 16:35SPEAKER_01: you would need to see AWS demand or pricing fail
- 16:38SPEAKER_01: to justify all this new capacity they're building.
- 16:40SPEAKER_00: Makes sense.
- 16:41SPEAKER_01: Or if operating cash consistently lags behind their investment and replacement needs.
- 16:46SPEAKER_01: Or if their margins start relying on non-recurring gains to look healthy.
- 16:51SPEAKER_01: Ultimately, if they take on more debt without better cash conversion, the thesis breaks.
- 16:56SPEAKER_00: And what challenges the bear case?
- 16:58SPEAKER_00: What invalidates the skeptics?
- 17:00SPEAKER_01: Strong realized cash returns, disciplined investment, and highly durable customer demand.
- 17:06SPEAKER_01: If Amazon delivers those three things, the skeptics are invalidated.
- 17:10SPEAKER_00: Right.
- 17:11SPEAKER_00: So my advice to you, track the actual reported results.
- 17:15SPEAKER_00: Do not just watch to see if the share price magically follows a modeled calendar.
- 17:19SPEAKER_01: Definitely.
- 17:20SPEAKER_00: Because the real world doesn't care about a spreadsheet's timeline.
- 17:24SPEAKER_00: So let's bring all this evidence together for the final verdict.
- 17:27SPEAKER_01: Okay.
- 17:28SPEAKER_01: Synthesizing it all, the model's neutral over the short horizon and constructive over the long.
- 17:34SPEAKER_01: Amazon's underlying operating engine is undeniably stronger today,
- 17:38SPEAKER_01: but the cash and financing test remains completely open.
- 17:41SPEAKER_00: Yeah, it all comes down to that single decision hinge.
- 17:44SPEAKER_00: Whether this incredibly expensive capacity can eventually earn durable
- 17:48SPEAKER_00: after financing cash returns.
- 17:51SPEAKER_00: Can the fleet of 747s fly enough routes to finally pay for themselves?
- 17:55SPEAKER_01: That's the billion dollar question.
- 17:57SPEAKER_00: For the underlying evidence,
- 17:58SPEAKER_00: inspect the dated Amazon research and full quarterly forecast path on iPulse AI.
- 18:03SPEAKER_01: Yep.
- 18:03SPEAKER_00: We highly recommend you listen to or watch our earlier Alphabet, NVIDIA,
- 18:07SPEAKER_00: and SpaceX deep dives for companion perspectives
- 18:10SPEAKER_00: on this massive infrastructure cycle.
- 18:12SPEAKER_01: Definitely worth checking out.
- 18:13SPEAKER_00: Please like, comment, and subscribe.
- 18:15SPEAKER_01: Yes, please do.
- 18:16SPEAKER_00: We are publishing a new podcast every day of the week except Thursday.
- 18:19SPEAKER_00: Stay tuned for the next podcast on Micron technology.
- 18:22SPEAKER_01: Looking forward to it.
- 18:23SPEAKER_00: What happens when AI demand meets the memory cycle?