Latest / Deep Values: Value Investing Deep Dives / Alphabet: 2026 Q2 AI Infrastructure and Financial Earnings Analysis
Transcript
- 0:00Welcome to this deep dive. So I want to start off today by
- 0:04just asking you a question. What happens when you know one
- 0:07of the most profitable cash printing software businesses in
- 0:11human history suddenly decides to just completely rewrite its
- 0:14own DNA? Yeah, it's a.
- 0:16Massive. Shift, right.
- 0:17I mean think about a company that practically invented the
- 0:20modern asset light digital monopoly and then virtually
- 0:25overnight they decided to become the most capital heavy concrete
- 0:28pouring infrastructure company on the entire planet.
- 0:31It is a massive historic pivot. It really is.
- 0:34It's unprecedented. But before we jump into the
- 0:36weeds on how they are actually pulling this off, I have some
- 0:38quick housekeeping for you to access more advanced researches,
- 0:41applications and our latest personal Deep Value agents.
- 0:44Welcome to explore and join Deep Values APP to track total market
- 0:49temperature. Welcome to Search and download
- 0:50the iOS App Buffett Indicator from the Apple App Store or
- 0:54Bubble Watch from Google Play. Essential tools right now,
- 0:57honestly. Definitely.
- 0:58And you know, we have to do the legal guardrails.
- 1:01All analysis is based on publicly available sources and
- 1:04may be subject to revisions or differing interpretations.
- 1:06Yeah. All contents from this channel
- 1:08is for research and entertainment purposes only and
- 1:11does not constitute any investment or financial advice
- 1:14or service. Conduct your own due diligence
- 1:17before making investment or financial decisions.
- 1:19Perfect, got to keep the lawyers happy.
- 1:21Always OK, with that out of the way, let's look at the actual
- 1:24numbers because our mission today is to unpack a
- 1:27comprehensive mid 2026 financial research, analysis and valuation
- 1:32report on Alphabet Inc, you know, the parent company of
- 1:34Google. We are breaking down this
- 1:36transition from a nimble search engine to a Hege AI utility and,
- 1:40well, what it means for anyone holding tech in their portfolio.
- 1:43Which is pretty much everyone at this point.
- 1:44Exactly. So to understand the sheer force
- 1:47behind this pivot, we have to look at the first half of 2026.
- 1:50First quarter revenues hit a staggering $109.9 billion.
- 1:55That's a 22% increase year over year.
- 1:57Then Q2 comes in at 119.8 billion, a 24% increase.
- 2:02The growth isn't just steady, it's it's actively accelerating.
- 2:05Well, it is accelerating and the underlying mechanism driving
- 2:09that speed is really Google Cloud.
- 2:11I mean, if you isolate the cloud division, the numbers are
- 2:14frankly astonishing. Oh yeah, they're off the church.
- 2:17Right. In the first quarter, cloud
- 2:18revenue surged 63% to $20 billion and then in the second
- 2:23quarter, it expanded by 82% to 24.8 billion.
- 2:27What we're seeing in the data is enterprise AI adoption basically
- 2:31hitting a critical mass. Companies aren't just, you know,
- 2:34experimenting anymore. They are embedding these tools
- 2:37into their daily operations. OK, let's unpack this because it
- 2:40feels like Google Cloud is going through this crazy teenage
- 2:44growth spurt where it's just aggressively outgrowing all its
- 2:46old clothes. But hold on, let me stop you
- 2:48there, because when I look at the Q2 earnings in this report,
- 2:51there is a massive glaring anomaly that feels like kind of
- 2:55a trap for anyone just skimming the financial headlines.
- 2:58Oh, the net income number. Yes, net income for Q2
- 3:02skyrocketed 298 percent, $212.1 billion.
- 3:08That yielded an earnings per share of $9.11.
- 3:11I mean, if I'm just a casual investor reading that, my jaw
- 3:14drops. I'm thinking Google has somehow
- 3:17figured out how to quadruple its actual day-to-day business
- 3:20profitability in a single quarter, but I know that can't
- 3:23be right. What is actually going on under
- 3:25the hood of that $112 billion number?
- 3:28Yeah, that is a crucial call out because you are looking at a
- 3:30classic accounting distortion, that 298% spike that was not
- 3:35driven by them selling more ads or more cloud subscriptions.
- 3:37Right, it couldn't be. No, it was primarily driven by a
- 3:40$99 billion net unrealized gain on non marketable equity
- 3:43securities. OK, wait, unpack net unrealized
- 3:46game for the listener who isn't an accountant.
- 3:48What does that actually mean in plain English?
- 3:50It basically means paper wealth. So Alphabet invests heavily in
- 3:53other startups and tech companies, AI labs for example.
- 3:57When the estimated value of those specific startups goes up,
- 4:00accounting rules actually require Alphabet to record that
- 4:04increase in value on their income statement.
- 4:06Even if they didn't sell anything.
- 4:07Exactly, even if they haven't actually sold the shares for
- 4:10cash, no actual cash traded hands.
- 4:13It is purely an accounting elevation in the other income
- 4:16column. Wow.
- 4:18So it's kind of an illusion of operational success.
- 4:20If you strip away that $99 billion paper game, what does
- 4:24the actual core business look like?
- 4:26Well, if you remove that paper game, their actual operating
- 4:29margin, you know the real profit they make from their core
- 4:31day-to-day business actually contracted slightly to 34% in
- 4:35Q2. Oh wow, so it shrank.
- 4:37It didn't explode. It shrank primarily because they
- 4:40are spending so much to support this new infrastructure.
- 4:44But you know, that doesn't mean the core business is weak.
- 4:47The reason? Cloud who is jumping 82% is due
- 4:50to systemic integration. By the end of 2025, nearly 75%
- 4:54of Google Cloud customers were actively utilizing Alphabets AI
- 4:58tools. What does that mean in terms of
- 4:59revenue though? Like how does someone just using
- 5:02an AI tool translate into a massive spike in cloud revenue?
- 5:05It fundamentally changes the billing model.
- 5:08In the past a company used Google Cloud mostly for static
- 5:12storage just holding data. Now they are using consumption
- 5:15based billing for generative AIAPIS.
- 5:18Oh, so it's a pay as you go? Thing exactly.
- 5:20Every time a company's customer service bot generates a response
- 5:24using Google's AI or an enterprise tool summarizes a
- 5:27massive document, they are alphabet servers.
- 5:30They pay per computation. It is a highly active,
- 5:34constantly ticking meter rather than just a flat storage fee.
- 5:37That makes total sense and it seems like they are successfully
- 5:40pushing that model outside of just the enterprise space too.
- 5:43The report highlights this massive growth in non
- 5:46advertising consumer products. By early 2026, they had 350
- 5:50million paid consumer subscriptions.
- 5:51Plus YouTube Premium and Google One, Yeah.
- 5:53Right. And the Gemini app hit 750
- 5:55million monthly active users. And Waymo.
- 5:58Waymo, which honestly felt like this weird sci-fi side project
- 6:01for the last 10 years, is now doing over 500,000 paid
- 6:04autonomous rides every single week.
- 6:06Which is exactly why this company is at such a fascinating
- 6:09inflection point right now. They have these incredibly
- 6:12powerful diversified revenue streams, but the computing power
- 6:16required to run Gemini for 750 million users or you know,
- 6:20process the real time sensor data for half a million Waymo
- 6:23rides is astronomically. High right?
- 6:25It's not cheap. Not at all.
- 6:27Sustaining those models requires a physical infrastructure that
- 6:30fundamentally changes how Alphabet manages its money.
- 6:33And see, that brings us to the biggest shock in this entire
- 6:35report. With all that cash coming in
- 6:38over 119 billion in a single quarter, you would naturally
- 6:41assume Alphabet is just raining cash down on his shareholders.
- 6:45For the last 10 years, they were the undisputed kings of the
- 6:48share buy. Oh.
- 6:49Absolute. Kings, Yeah, in 2023 alone, they
- 6:52bought back and retired $62.1 billion of their own stock,
- 6:56which, you know, artificially boosts the value of the
- 6:58remaining shares. But in Q 12026, buybacks dropped
- 7:02to absolute zero. They completely suspended A10
- 7:04year capital return program. Yeah.
- 7:06And not only did they stop returning capital, they actively
- 7:09started diluting shareholders. They authorized an $80 billion
- 7:13equity raise. That's 40 billion at the market
- 7:16and a 10 billion private placement.
- 7:18Here's where it gets really interesting, because I really
- 7:21have to challenge this logic as an investor.
- 7:23This feels like a betrayal. The report shows Alphabet
- 7:26printed over $73 billion in free cash flow in 2025.
- 7:31If they are generating that much raw cash, why on earth do they
- 7:34need to issue new stock, dilute my shares and suspend buybacks?
- 7:38It's like a purely digital software company suddenly
- 7:41deciding they need to buy thousands of acres of land and
- 7:44pour physical concrete to build railroads.
- 7:46Well, what's fascinating here is that is exactly what they are
- 7:48doing. You really have to look at the
- 7:49scale of their new capital expenditures or CapEx.
- 7:52Historically, between 2021 and 2023, Alphabet spent about 32
- 7:56billion a year on infrastructure.
- 7:58OK, 32 billion. Right.
- 8:00But their guidance for 2020 26 is an astonishing 180 to $190
- 8:03billion in CapEx for a single year.
- 8:06Wait, what? So that 73 billion in free cash
- 8:08flow doesn't even cover half of the new construction bill?
- 8:12What exactly are they buying for $190 billion?
- 8:15They are building a physical capital Moat.
- 8:17Training foundational AI models is not like building a
- 8:20traditional software app where you know a few engineers write
- 8:24code on laptops in a coffee shop.
- 8:26Training the next generation of AI requires linking 10s of
- 8:29thousands of custom tensor processing units or TP use.
- 8:34These are specialized microchips designed specifically for the
- 8:37complex math of neural networks. You have to house those chips in
- 8:41massive data centers, secure specialized cooling systems so
- 8:44they literally don't meltdown, and sign multi decade energy
- 8:48contracts just to power them. Do you think industrial
- 8:50manufacturing process for intelligence?
- 8:52Yes, perfectly set, and this physical infrastructure spend is
- 8:57an existential necessity. The financial threshold required
- 9:00to operate at the cutting edge of AI is now so steep that
- 9:03software startups are completely priced out.
- 9:05So it's uncatchable. Exactly.
- 9:07By spending 190 billion, Alphabet is digging A Moat so
- 9:11wide and so deep that only a sovereign nation or a rival mega
- 9:15cap tech company could ever hope to cross it.
- 9:18They are taking on short term pain to guarantee long term
- 9:22dominance. OK, I understand the concept of
- 9:24the mode, but this transition raises a massive paradox for
- 9:28anyone trying to actually value this company given this
- 9:31staggering CapEx spend, the shareholder dilution and the
- 9:34pivot from a 90% margin software business to a heavy asset
- 9:38construction project. I mean, traditional value
- 9:41investors should be running away screaming.
- 9:42Oh. Absolutely.
- 9:43By strict traditional value standards, Alphabet absolutely
- 9:46fails the test. Right now they're trailing 12
- 9:49month. Price to earnings ratio is 26.1
- 9:52X. Their price to book ratio is 8.7
- 9:54X. And the discounted cash flow
- 9:56analysis in this report is wild for the listener.
- 9:59ADCF model basically tries to figure out what a company is
- 10:02worth to take based on all the cash it will theoretically
- 10:04generate in the. Future right Standard valuation
- 10:06metric. Standard stuff, but even if you
- 10:08assume a very generous 10% free cash flow growth rate every
- 10:12single year for the next decade, the intrinsic value of Alphabet
- 10:16comes out to about $168 per share.
- 10:19Yet the market price in mid 2026 is sitting around $345.
- 10:24You're paying more than double what the cash flows suggest it's
- 10:27actually worth. Yeah, it's steep.
- 10:29There is zero traditional margin of safety, yet the exact
- 10:32opposite of a value investor exodus is happening.
- 10:35Warren Buffett's Berkshire Hathaway, the undisputed king of
- 10:39strict value investing, bought 36.4 million shares in the open
- 10:43market. That's an $18.8 billion stake.
- 10:47And they were the ones who injected another ten billion
- 10:49directly into Alphabet through that private placement we just
- 10:51mentioned. I'm struggling to understand
- 10:54this. How does a stock with terrible
- 10:55value metrics, massive dilution and $190 billion construction
- 10:59bill become the darling of Warren Buffett?
- 11:01It requires a brilliant pivot in how you categorize the business.
- 11:05Berkshire Hathaway is not looking at Alphabet as a
- 11:08speculative technology company anymore.
- 11:10They are viewing it as a global digital utility.
- 11:12Oh. Interesting.
- 11:13Think about Buffett's history. He loves railroads.
- 11:16He loves energy grids. He loves infrastructure that
- 11:20requires massive, painful up different capital expenditures,
- 11:25but then yields decades of predictable, unavoidable tolls.
- 11:29So the $10 billion private placement isn't really a tech
- 11:32gamble. It's Buffett acting as the
- 11:34ultimate underwriter for global infrastructure, just like he did
- 11:38when he injected capital into Goldman Sachs during the 2008
- 11:41financial crisis. Exactly that.
- 11:43Berkshire's bet is on alphabets data gravity.
- 11:46Alphabet has billions of users locked into Android, Chrome,
- 11:50YouTube, and Maps. If you combine that captured
- 11:53audience with this insurmountable physical
- 11:55infrastructure scale, Alphabet effectively gains an unregulated
- 11:58toll bridge pricing power over the future economy.
- 12:01Because everyone needs access to it.
- 12:02Right, They're laying down the power lines of the 21st century.
- 12:06If a business wants to plug into the modern AI economy, they have
- 12:09no choice but to pay off of its toll.
- 12:11That is a highly compelling thesis.
- 12:13But you know, if they are literally paving the globe with
- 12:16data centers, they are taking on an immense amount of physical
- 12:20and structural risk. What happens if this transition
- 12:23doesn't go smoothly? And that's the multibillion
- 12:26dollar question. Right.
- 12:27What are the actual forces that could derail this utility
- 12:30monopoly? Because we have to look at the
- 12:32structural threats and, crucially, who is actually
- 12:35steering the ship. Well, the most acute structural
- 12:38threat is to their profit margins, specifically the shift
- 12:42from traditional search to generative AI.
- 12:45We touched on this earlier, but let's break down the mechanics.
- 12:48Historically, traditional Google search was computationally very
- 12:51cheap, right? A user typed in a query, Google
- 12:55checked a pre index database gave you 10 blue links.
- 12:58You clicked one and they made ad revenue.
- 13:00The index was already built, they just had to retrieve the
- 13:02information. It was practically free.
- 13:04Correct, but generative AI changes the entire mechanism.
- 13:08When you ask an AIA complex question, it doesn't just
- 13:12retrieve link, it calculates the statistical probability of the
- 13:15next word in a sequence billions of times on the fly to generate
- 13:19a direct answer. Which takes massive power.
- 13:22Exactly. This is called inference.
- 13:24The inference costs for a generative AI query are
- 13:27substantially higher than traditional search indexing.
- 13:30So even if Alphabet maintains its absolute monopoly on user
- 13:33attention, the actual cost of serving that attention is rising
- 13:37permanently. This constantly threatens to
- 13:39compress their gross margins. And on top of the margin
- 13:41compression, they are fighting a multi front war.
- 13:43On the legal side, the antitrust scrutiny is intense right now.
- 13:48The European Union upheld a €2.4 billion fine for anti
- 13:52competitive practices. The US Department of Justice is
- 13:55actively threatening to break up their ad tech stack.
- 13:57Yeah, the regulatory pressure is definitely mounting.
- 14:00It is, and the DOJ is also targeting those massive default
- 14:03search agreements, like the billions Alphabet pays Apple to
- 14:07be the default search engine on the iPhone.
- 14:09If those distribution advantages are severed by a judge that told
- 14:14you, each suddenly has a massive detour.
- 14:17But what really stands out in the report is how the company
- 14:19reacts to these threats. Management routinely dismisses
- 14:23these existential legal battles as mere operational friction.
- 14:27Which is pretty. Gold very bold and the reason
- 14:30they can afford to dismiss outside pressure is because of
- 14:33their governance structure. So what does this all mean for
- 14:36the average shareholder? Well, the governance structure
- 14:38is absolute. The public markets provide the
- 14:40capital, but they provide 0 control.
- 14:42Right because Larry Page and Sergey Brin the founders whole
- 14:46Class B super voting shares those specific shares carry 10
- 14:50votes each compared to the one vote for regular shares.
- 14:53So being a shareholder in Alphabet right now is basically
- 14:55like riding in the back of a way MO economist car where the
- 14:58founders have permanently locked the steering wheel that's.
- 15:00A great way to put it. You cannot change the
- 15:02destination. You just have to trust that
- 15:04their navigation system works. Look at the June 2020 sixth
- 15:08annual shareholder meeting. There were 14 independent
- 15:11proposals from outside shareholders.
- 15:14These weren't frivolous requests.
- 15:16They were demanding equal voting rights, AI safety oversight, and
- 15:20reports on the massive water usage needed to cool all these
- 15:23new data centers. And every single one of them was
- 15:26decisively crushed. Completely.
- 15:28Yeah, the proposal demanding equal shareholder voting
- 15:30actually received 3.8 billion votes in favor from the public,
- 15:34which shows a massive outside desire for reform.
- 15:37But the Founders simply deployed their 8.5 billion opposed votes
- 15:40and just ended the conversation. The board cannot overrule them.
- 15:44The shareholders cannot overrule.
- 15:45Them Which brings us to the ultimate question of risk.
- 15:49If you have founders with absolute, unchallengeable
- 15:51control who have decided to spend $190 billion on physical
- 15:55infrastructure, what happens to the business model if the
- 15:58economy takes a hit? If we connect this to the bigger
- 16:00picture, it creates a severe operational leverage risk.
- 16:04Let's look at how Alphabet used to survive economic downturns in
- 16:08the past. If a recession hit and AD
- 16:10revenue dropped, Alphabet's cost would automatically drop, too.
- 16:14Because it was variable. Right.
- 16:15A huge portion of their expenses were variable, like the traffic
- 16:19acquisition cost they pay to partners.
- 16:21If fewer ads were clicked, they paid out less money.
- 16:24Their margins had a natural shock absorber built in, but.
- 16:26You can't just lay off a concrete data center.
- 16:29No, you cannot. By front loading 190 billion
- 16:32into physical data centers, land and custom silicon, they are
- 16:36locking in massive fixed depreciation costs.
- 16:40If a macroeconomic ad recession hits in 2027, those newly
- 16:44established cooling bills, energy contracts and
- 16:46depreciation costs remain entirely immovable.
- 16:49They still have to pay the electric bill.
- 16:50Exactly. Their operating margins could
- 16:53suffer severe compression compared to historical norms.
- 16:56The founders have locked the steering wheel and pressed the
- 16:58accelerator directly into a high fixed cost future.
- 17:02They believe they can outrun the margin compression through
- 17:04enterprise cloud dominance. But if the AI adoption curve
- 17:08slows even slightly, the financial drag from that
- 17:10infrastructure will be immense. It is a high stakes gamble on an
- 17:15unprecedented scale. I mean, if we summarize what
- 17:17we've unpacked today, Alphabet is in the middle of a historic
- 17:20metamorphosis. They are actively sacrificing
- 17:23short term shareholder returns, diluting their own equity, and
- 17:27absorbing massive upfront costs, all to cement an insurmountable
- 17:31physical infrastructure advantage in the AI era.
- 17:34Completely transforming the business.
- 17:36They're transitioning from an agile software giant into a
- 17:39heavy capital intensive digital utility.
- 17:42And they're doing it with the explicit financial endorsement
- 17:44of traditional value heavyweights like Berkshire
- 17:46Hathaway, who clearly believe this infrastructure mode will
- 17:50become impenetrable. The transition from a search
- 17:52monopoly to an AI utility is virtually complete in terms of
- 17:56capital allocation. The question now is simply
- 17:59execution. Which leaves us with one final,
- 18:02unmentioned ripple effect for you to ponder.
- 18:05If Warren Buffett is right, and Alphabet successfully
- 18:09establishes itself as the indispensable monopolistic
- 18:12digital utility for the entire globes AI needs, how long will
- 18:17it be before governments around the world start treating them
- 18:19like 1? That's the real catch.
- 18:21Think about your local electric company or public water utility.
- 18:24They have monopolies, yes, but because they are essential to
- 18:27society, the government steps in to aggressively regulate their
- 18:30pricing, their access and their profit margins.
- 18:34Alphabet has perfectly engineered itself for survival
- 18:36today by building this massive infrastructure mode, but in
- 18:40doing so, they may have just written the blueprint for their
- 18:42own future price regulation. It's a fascinating trade off.
- 18:45Thank you for joining us on this deep dive.