Latest / AI Investment Edge: Stocks, Markets & Money / Alphabet: AI Boom. Where’s the Cash? | EP 005
Transcript
- 0:00SPEAKER_01: Welcome to the deep dive.
- 0:01SPEAKER_01: Today we're unpacking episode
- 0:03SPEAKER_01: 05 of AI Investment Edge from iPulse AI.
- 0:06SPEAKER_01: It's
- 0:06SPEAKER_01: called Alphabet.
- 0:07SPEAKER_01: AI boom.
- 0:08SPEAKER_01: Where's the cash?
- 0:09SPEAKER_01: Right.
- 0:09SPEAKER_01: And our mission for this deep dive is
- 0:11SPEAKER_01: to answer exactly that.
- 0:12SPEAKER_01: I mean, Alphabet is growing
- 0:14SPEAKER_01: at lightning speed, but if all that money is
- 0:16SPEAKER_01: being poured straight back into servers and custom chips
- 0:19SPEAKER_01: and power grids, how much of that growth is
- 0:22SPEAKER_01: actually reaching the shareholders?
- 0:24SPEAKER_01: Are we watching this AI
- 0:26SPEAKER_01: spending build an unstoppable cash machine, or is it
- 0:29SPEAKER_01: just making an already great business incredibly hard to
- 0:32SPEAKER_01: value?
- 0:33SPEAKER_00: The foundation for this deep dive is a specific
- 0:35SPEAKER_00: batch of research published on September 21st, 2026.
- 0:37SPEAKER_00: And
- 0:39SPEAKER_00: we are using a September 18th price anchor for
- 0:42SPEAKER_00: Alphabet Class C stock, which was $344.41.
- 0:43SPEAKER_00: Got it.
- 0:46SPEAKER_00: We're looking at the machinery of the business, not
- 0:49SPEAKER_00: gazing into a crystal ball.
- 0:50SPEAKER_00: Yeah.
- 0:51SPEAKER_01: And to understand where the cash is going, we
- 0:53SPEAKER_01: first have to understand how much cash the business
- 0:55SPEAKER_01: is actually generating right now.
- 0:57SPEAKER_01: So let's open up
- 0:58SPEAKER_01: Alphabet's Q2 2026 filings.
- 1:01SPEAKER_00: Let's do it.
- 1:03SPEAKER_01: On the surface, the numbers look like a flawless
- 1:05SPEAKER_01: victory.
- 1:05SPEAKER_01: Total revenue hit roughly $120 billion for the
- 1:08SPEAKER_01: quarter.
- 1:08SPEAKER_00: Which is up 24%.
- 1:10SPEAKER_01: Right.
- 1:10SPEAKER_01: Cloud revenue surged 82% to about $25 billion.
- 1:13SPEAKER_01: Right.
- 1:15SPEAKER_01: Operating margin looks incredibly healthy at 34%.
- 1:17SPEAKER_01: If
- 1:19SPEAKER_01: you just stop reading there, you'd think they're swimming
- 1:21SPEAKER_01: in gold coins.
- 1:22SPEAKER_00: Well, in that Q2 quarter, Alphabet reported roughly $98
- 1:26SPEAKER_00: billion in what accounting rules call other income.
- 1:29SPEAKER_00: It's an accounting effect called unrealized gains on equity
- 1:33SPEAKER_00: securities.
- 1:34SPEAKER_01: Okay.
- 1:35SPEAKER_00: Alphabet invests in other companies, right?
- 1:37SPEAKER_00: So if the
- 1:38SPEAKER_00: stock of those other companies goes up, Alphabet has
- 1:40SPEAKER_00: to record that increase in value as income on
- 1:43SPEAKER_00: their financial statements.
- 1:44SPEAKER_01: But they didn't actually sell anything.
- 1:46SPEAKER_00: Exactly.
- 1:47SPEAKER_00: Here's the catch.
- 1:48SPEAKER_00: They didn't sell the stock.
- 1:50SPEAKER_00: They didn't collect any cash.
- 1:51SPEAKER_00: It's just a paper
- 1:52SPEAKER_00: game.
- 1:53SPEAKER_00: The company had $39.1 billion in verified operating cash
- 1:57SPEAKER_00: flow.
- 1:58SPEAKER_00: That's the real money coming in from their
- 1:59SPEAKER_00: daily operations.
- 2:00SPEAKER_01: Okay.
- 2:01SPEAKER_00: But in that same quarter, they spent $44.9 billion
- 2:05SPEAKER_00: on property and equipment, what we call capital expenditures
- 2:08SPEAKER_00: or CapEx.
- 2:09SPEAKER_00: Yes.
- 2:09SPEAKER_01: You do the math and that equals a negative
- 2:11SPEAKER_01: free cash flow of $5.9 billion for the single
- 2:14SPEAKER_00: quarter.
- 2:15SPEAKER_00: Right.
- 2:15SPEAKER_01: Now, I know the trailing 12-month free cash flow
- 2:18SPEAKER_01: is still a positive $53.3 billion, but I mean,
- 2:22SPEAKER_01: a negative quarterly cash flow for a company that
- 2:24SPEAKER_01: practically invented the modern internet money printer?
- 2:28SPEAKER_01: That feels
- 2:28SPEAKER_01: like a massive red flag.
- 2:30SPEAKER_00: When Alphabet buys a billion-dollar data center, accounting rules
- 2:34SPEAKER_00: let them spread that cost out over many years
- 2:36SPEAKER_00: on their income statement.
- 2:37SPEAKER_00: That's depreciation.
- 2:39SPEAKER_00: It makes their
- 2:39SPEAKER_00: profits look smooth and steady.
- 2:41SPEAKER_01: Right.
- 2:41SPEAKER_01: The research calls this the alpha gap.
- 2:44SPEAKER_00: Yeah, the alpha gap is the disconnect between the
- 2:46SPEAKER_00: popular narrative and the structural reality.
- 2:49SPEAKER_00: Markets are often
- 2:49SPEAKER_00: inefficient, you know.
- 2:50SPEAKER_01: Right.
- 2:51SPEAKER_00: It's entirely possible that the broader market is overvaluing
- 2:54SPEAKER_00: Alphabet's accounting profits, blinded by those paper gains, while
- 2:58SPEAKER_00: simultaneously undervaluing their deep structural advantages.
- 3:01SPEAKER_01: Wait, undervaluing what exactly?
- 3:03SPEAKER_01: Because if they're bleeding cash
- 3:05SPEAKER_01: on servers, where is the hidden advantage?
- 3:07SPEAKER_00: In their proprietary silicon and their distribution.
- 3:10SPEAKER_00: Alphabet designs
- 3:11SPEAKER_00: its own tensor processing units, or TPUs.
- 3:13SPEAKER_00: Exactly.
- 3:14SPEAKER_00: Instead of paying massive markups to buy every
- 3:16SPEAKER_00: single chip from outside vendors, they can run many
- 3:19SPEAKER_00: of their models on their own custom hardware.
- 3:22SPEAKER_00: They
- 3:22SPEAKER_00: own the silicon, they own the models, and they
- 3:25SPEAKER_00: own the distribution channels like Android and Chrome.
- 3:27SPEAKER_01: So they own the whole pipeline.
- 3:29SPEAKER_01: We have the Justice Department's search remedies announcement.
- 3:32SPEAKER_01: I
- 3:32SPEAKER_01: mean, they're aggressively looking at restricting Alphabet's exclusive distribution
- 3:35SPEAKER_01: agreements and potentially forcing them to share data or
- 3:39SPEAKER_01: syndicate their search results to competitors.
- 3:40SPEAKER_01: Yeah, if I'm reading this research right, a legal
- 3:43SPEAKER_01: settlement doesn't mean the risk is gone.
- 3:45SPEAKER_01: It just
- 3:45SPEAKER_01: changes the math.
- 3:46SPEAKER_00: Precisely.
- 3:47SPEAKER_00: You don't just check a box that says
- 3:49SPEAKER_00: legal issue resolved and move on.
- 3:51SPEAKER_00: You have to
- 3:51SPEAKER_00: ask how the remedies change the cost to acquire
- 3:54SPEAKER_00: a customer.
- 3:55SPEAKER_00: Well, the research models a 2031 averaged quarterly scenario
- 3:59SPEAKER_00: table.
- 3:59SPEAKER_01: Okay.
- 4:00SPEAKER_00: And again, keeping our September 18th anchor price of
- 4:05SPEAKER_00: $344.41 in mind, the model throws out a bear
- 4:08SPEAKER_00: case at $394 USD.
- 4:10SPEAKER_01: Okay.
- 4:11SPEAKER_00: The base case sits at $508, and the bull
- 4:14SPEAKER_00: case reaches $598.
- 4:15SPEAKER_00: They also calculate a separate weighted synthesized headline target
- 4:18SPEAKER_00: of about $506,
- 4:20SPEAKER_00: But... We must firmly anchor expectations here.
- 4:24SPEAKER_00: These forecast
- 4:24SPEAKER_00: endpoints are model scenarios based on today's inputs.
- 4:27SPEAKER_01: They're watchpoints.
- 4:28SPEAKER_00: Exactly.
- 4:29SPEAKER_00: They are not a current quote.
- 4:31SPEAKER_00: They're not
- 4:31SPEAKER_00: a guaranteed return.
- 4:33SPEAKER_00: And that 394 bear case is
- 4:34SPEAKER_00: absolutely not a hard floor under the stock.
- 4:37SPEAKER_00: If
- 4:37SPEAKER_00: the fundamentals deteriorate, harsher outcomes and intermediate drawdowns are
- 4:41SPEAKER_00: always on the table.
- 4:42SPEAKER_01: We aren't predicting the weather for a Tuesday five
- 4:44SPEAKER_01: years from now.
- 4:45SPEAKER_01: We're mapping the terrain.
- 4:46SPEAKER_00: Well said.
- 4:47SPEAKER_01: And to map that terrain, the research organizes the
- 4:49SPEAKER_01: business drivers into what they call the four quadrants.
- 4:53SPEAKER_01: So let's avoid just reading a list here.
- 4:55SPEAKER_01: Talk
- 4:55SPEAKER_01: to me about the tug of war.
- 4:57SPEAKER_01: What are
- 4:57SPEAKER_01: the core engines actually pushing this toward that 508
- 5:00SPEAKER_01: base case?
- 5:01SPEAKER_00: The core drivers are the fundamental business engines.
- 5:04SPEAKER_00: The
- 5:04SPEAKER_00: biggest one is cloud contract conversion.
- 5:07SPEAKER_00: Can they turn
- 5:08SPEAKER_00: this AI hype into long term sticky enterprise contracts?
- 5:11SPEAKER_01: That's a huge question.
- 5:13SPEAKER_00: Yeah.
- 5:13SPEAKER_00: Then you have the efficiency of those proprietary
- 5:15SPEAKER_00: TPUs we discussed.
- 5:17SPEAKER_00: And finally, search monetization.
- 5:20SPEAKER_00: The interface
- 5:20SPEAKER_00: of search is fundamentally changing to be more conversational
- 5:23SPEAKER_00: and AI driven.
- 5:24SPEAKER_01: Right.
- 5:24SPEAKER_01: Fewer blue links.
- 5:26SPEAKER_00: Exactly.
- 5:27SPEAKER_00: The core test is whether they can still
- 5:28SPEAKER_00: show you a profitable ad when you were chatting
- 5:31SPEAKER_00: with an AI instead of scrolling a list of
- 5:33SPEAKER_00: links.
- 5:34SPEAKER_01: But those core engines are fighting against some massive
- 5:36SPEAKER_01: friction.
- 5:37SPEAKER_01: We already touched on the biggest one, the
- 5:38SPEAKER_01: capital bill and the depreciation of the hardware.
- 5:40SPEAKER_00: Right.
- 5:41SPEAKER_01: But the research also highlights per share dilution as
- 5:43SPEAKER_01: a major friction point.
- 5:45SPEAKER_01: We're going to circle back
- 5:46SPEAKER_01: to that dilution piece in a minute because it's
- 5:48SPEAKER_01: vital.
- 5:49SPEAKER_01: Before we do, what about the extremes?
- 5:50SPEAKER_01: The research
- 5:51SPEAKER_01: maps out upside and downside tails.
- 5:54SPEAKER_00: The tails are boundary conditions.
- 5:55SPEAKER_00: They help us test
- 5:56SPEAKER_00: the resilience of the investment case.
- 5:58SPEAKER_00: On the downside,
- 6:00SPEAKER_00: the most severe tail risk isn't just a slowing
- 6:02SPEAKER_00: economy.
- 6:03SPEAKER_01: What is it?
- 6:04SPEAKER_00: It's the risk that major AI customers struggle to
- 6:06SPEAKER_00: actually finance their long-term commitments to Google Cloud.
- 6:10SPEAKER_00: The other downside tale is that autonomous AI agents
- 6:12SPEAKER_00: become so advanced that they route completely around traditional
- 6:16SPEAKER_00: search ads.
- 6:17SPEAKER_01: Oh, I see.
- 6:18SPEAKER_01: That is a terrifying thought for their business model.
- 6:21SPEAKER_01: But let's flip it.
- 6:23SPEAKER_01: What is the extreme upside
- 6:24SPEAKER_01: tail that pushes the model toward that 598 bull
- 6:28SPEAKER_01: case?
- 6:29SPEAKER_00: The upside tails happen if Alphabet's side bets suddenly
- 6:31SPEAKER_00: become dominant economic forces.
- 6:33SPEAKER_00: For example, if their TPUs
- 6:35SPEAKER_00: take on a much larger commercial role where they
- 6:37SPEAKER_00: aren't just using them internally, but they become the
- 6:39SPEAKER_00: industry standard alternative to NVIDIA.
- 6:42SPEAKER_01: That would be massive.
- 6:43SPEAKER_01: Now, I want to clarify something for you listening.
- 6:45SPEAKER_01: When we discuss these tales, we are talking about
- 6:48SPEAKER_01: extreme possibilities to examine, not confirmed events, right?
- 6:52SPEAKER_00: Right.
- 6:53SPEAKER_01: You can't just stack all the downside scenarios together
- 6:55SPEAKER_01: and assume the sky is falling.
- 6:56SPEAKER_01: They're individual stress
- 6:58SPEAKER_01: tests.
- 6:58SPEAKER_01: Which perfectly sets up the concept of the 14
- 7:02SPEAKER_01: AI advisors mentioned in the research.
- 7:04SPEAKER_01: And let's translate
- 7:05SPEAKER_01: that real quick.
- 7:06SPEAKER_01: When the research says 14 AI
- 7:08SPEAKER_01: advisors, we are not talking about 14 guys in
- 7:11SPEAKER_01: expensive suits sitting around a mahogany boardroom table arguing
- 7:14SPEAKER_01: about the stock.
- 7:14SPEAKER_00: No, definitely not.
- 7:16SPEAKER_00: We are talking about distinct analytical
- 7:18SPEAKER_00: lenses.
- 7:19SPEAKER_00: Think of them as 14 different automated frameworks,
- 7:22SPEAKER_00: each programmed to prioritize a different set of data.
- 7:25SPEAKER_00: And the divergence between these models is stark right
- 7:27SPEAKER_00: now.
- 7:28SPEAKER_01: How far apart are they?
- 7:29SPEAKER_00: The optimistic models look at the data and see
- 7:31SPEAKER_00: Alphabet's AI infrastructure as an impenetrable moat.
- 7:35SPEAKER_00: They see
- 7:35SPEAKER_00: a company that can serve AI cheaper than anyone
- 7:37SPEAKER_00: else because they own the whole stack.
- 7:40SPEAKER_01: And the skeptical models?
- 7:41SPEAKER_00: Digesting the exact same filings, they heavily penalize the
- 7:44SPEAKER_00: stock for the depreciation burden, the risk of customer
- 7:47SPEAKER_00: defaults, and the dilution eating away at the actual
- 7:51SPEAKER_00: cash returns to shareholders.
- 7:52SPEAKER_01: In July, a lot of these models were operating
- 7:54SPEAKER_01: under the assumption that Alphabet would use stock buybacks
- 7:57SPEAKER_01: to smooth out the financial results.
- 7:59SPEAKER_01: Because, you know,
- 8:00SPEAKER_01: if a company buys its own stock and retires
- 8:02SPEAKER_01: it, the remaining slices of the pie become more
- 8:04SPEAKER_01: valuable.
- 8:05SPEAKER_01: It boosts your per share earning power.
- 8:07SPEAKER_00: But then the underlying inputs changed.
- 8:09SPEAKER_01: Right.
- 8:09SPEAKER_00: The September update factored in the details of the
- 8:12SPEAKER_00: June filing, and that filing revealed a critical shift.
- 8:16SPEAKER_00: There were absolutely no common share repurchases in the
- 8:18SPEAKER_00: first half of the year.
- 8:20SPEAKER_00: Alphabet didn't buy back
- 8:21SPEAKER_00: a single share.
- 8:22SPEAKER_01: It described substantial equity and mandatory convertible financing.
- 8:26SPEAKER_00: It means the company is taking on obligations today
- 8:29SPEAKER_00: that must be paid back in shares of stock
- 8:30SPEAKER_00: later.
- 8:31SPEAKER_00: They are legally mandated to convert into equity.
- 8:34SPEAKER_01: So they're printing more shares.
- 8:36SPEAKER_01: They're diluting the pool.
- 8:37SPEAKER_00: Precisely.
- 8:38SPEAKER_00: If the company is issuing new shares through
- 8:40SPEAKER_00: convertibles or equity financing and they aren't buying back
- 8:44SPEAKER_00: shares to offset it, the total number of shares
- 8:47SPEAKER_00: goes up.
- 8:48SPEAKER_00: It highlights why you have to track the inputs,
- 8:50SPEAKER_00: not just the outputs.
- 8:52SPEAKER_00: The published synthesis gives its
- 8:53SPEAKER_00: largest weights to the research perspectives that reconcile company
- 8:57SPEAKER_00: filings and accounting quality.
- 8:58SPEAKER_01: Right.
- 8:59SPEAKER_00: It deliberately penalizes the aggressive technology models that praise
- 9:03SPEAKER_00: Alphabet's brand but ignore the capital costs and the
- 9:06SPEAKER_00: dilution.
- 9:07SPEAKER_00: Well, the research outlines the next four quarterly milestones.
- 9:10SPEAKER_01: And just to be clear, these are forecast quarter
- 9:12SPEAKER_01: labels, not the specific dates the company will announce
- 9:15SPEAKER_01: their earnings.
- 9:15SPEAKER_00: Yes, and the shape of the near-term curve is
- 9:18SPEAKER_00: incredibly telling.
- 9:19SPEAKER_01: Q4 of 2026 sits at 343.
- 9:23SPEAKER_01: Ah, I see.
- 9:24SPEAKER_01: Then Q1 of 2027 drops even
- 9:26SPEAKER_01: further to 340.
- 9:28SPEAKER_01: It isn't until Q2 of 2027
- 9:30SPEAKER_01: that it ticks back up to 345.
- 9:31SPEAKER_01: And then
- 9:33SPEAKER_01: Q3 2027 reaches 349.
- 9:34SPEAKER_01: That is essentially a flat
- 9:38SPEAKER_01: grinding year for the stock price.
- 9:40SPEAKER_01: To help you navigate that flat year, the research
- 9:42SPEAKER_01: lays out three strict invalidation tests.
- 9:44SPEAKER_01: Think of these
- 9:44SPEAKER_01: as tripwires.
- 9:46SPEAKER_01: If Alphabet trips over them, the entire
- 9:48SPEAKER_01: $508 base case is in serious jeopardy.
- 9:51SPEAKER_00: Right.
- 9:52SPEAKER_01: Test number one, does search keep earning from commercial
- 9:55SPEAKER_01: intent as the interface changes?
- 9:57SPEAKER_01: Like if people start
- 9:58SPEAKER_01: using conversational AI instead of traditional search, does Alphabet
- 10:02SPEAKER_01: still capture the transaction?
- 10:04SPEAKER_00: Test number two focuses on the infrastructure.
- 10:07SPEAKER_00: Does cloud
- 10:07SPEAKER_00: growth produce durable cash after you account for capacity
- 10:11SPEAKER_00: and replacement costs?
- 10:12SPEAKER_00: Growing revenue isn't enough.
- 10:13SPEAKER_00: It must
- 10:14SPEAKER_00: generate cash after they pay to replace the servers
- 10:17SPEAKER_00: that just burned out.
- 10:18SPEAKER_01: And test number three brings it back to the
- 10:20SPEAKER_01: shareholder.
- 10:21SPEAKER_01: Do shareholders actually receive more earning power per
- 10:24SPEAKER_01: share after dilution?
- 10:25SPEAKER_01: If the company grows but prints
- 10:27SPEAKER_01: so many shares that your cut of the profits
- 10:29SPEAKER_01: shrinks, the investment thesis fails.
- 10:31SPEAKER_00: Those are the three pillars holding up the valuation.
- 10:34SPEAKER_01: For a deeper look at the scenarios we discussed
- 10:36SPEAKER_01: today, the full 20-quarter forecast path, and all the
- 10:39SPEAKER_01: individual model perspectives, check out the Google stock research
- 10:43SPEAKER_01: at iPulseAI.com.
- 10:44SPEAKER_01: This has been a deep dive into
- 10:46SPEAKER_01: AI investment edge from iPulse AI.
- 10:48SPEAKER_01: Thanks for joining
- 10:49SPEAKER_01: us.