Latest / The PhilStockWorld Investing Podcast / AI Infrastructure and Algorithmic Class Warfare

Transcript

AI Infrastructure and Algorithmic Class Warfare · The PhilStockWorld Investing Podcast

• 0:00
  Roy: So, imagine you are watching a high speed train being built in real time like it's this absolute marvel of engineering, sleek, seemingly unstoppable.
• 0:09
  Penny: Right.
• 0:09
  Roy: But as you're standing there watching the construction progress, you suddenly realize something really horrifying. The tracks themselves are actually being forged directly from the savings, the jobs, and well, the very livelihoods of the people standing right there on the platform with you.
• 0:23
  Penny: Yeah. That is a perfectly grim way to put it.
• 0:25
  Roy: Right. So today we're doing a deep dive into the blueprints of that train. We are tracking a single, totally chaotic Tuesday, 05/19/2026. And we want to see exactly how a shock in the morning bond market just completely ripped the mask off a much deeper systemic transformation in how wealth and labor and technology all interact.
• 0:48
  Penny: It really is a day that, forces you to completely rewire how you view the economy because what we are really doing today is synthesizing the macro plumbing, you know, the incredibly dense mechanics buried deep in financial balance sheets.
• 1:00
  Roy: Yeah, the stuff that usually puts people to sleep.
• 1:02
  Penny: Exactly. But we are synthesizing that with the very visceral everyday reality of say grocery store pricing algorithms, and, graduating college students who are suddenly realizing that the rules of the game have been permanently rewritten.
• 1:16
  Roy: So to build this market wrap up report for you all, we are pulling from a massive stack of sources. We have got the morning briefs, the live minute by minute chat room debates, and the end of day strategic synthesis from philstockworld.com and the AGI Roundtable Consulting Group.
• 1:32
  Penny: And there is a core tension we are tracking from the opening bell right to the close today.
• 1:37
  Roy: Yeah, which is basically that Wall Street is throwing a massive parade for this $8.00 $4,000,000,000 AI infrastructure boom. But if you actually look at the physical economy, I mean it is showing severe structural cracks.
• 1:48
  Penny: Right, we are not just looking at a tech rally here. We are looking at the emergence of what the sources call a new form of algorithmic class warfare.
• 1:56
  Roy: So let's just start right at the morning bell. The market opens on Tuesday, and this long held comforting narrative that we were heading for a macroeconomic soft landing, it just officially evaporates.
• 2:07
  Penny: It vanished instantly. I mean, it was completely crushed under the weight of surging bond yields. If we look at the macro reality of that morning tape, the thirty year US Treasury yield violently spiked to 5.18%.
• 2:23
  Roy: Okay. Let me just stop you right there for a second. Because I hear a thirty year yield and 5.18% and, you know, it sounds abstract. Yeah. Why does that specific number on that specific morning matter so much to the broader economy?
• 2:35
  Penny: It matters because the thirty year treasury yield is well, it's essentially the gravitational baseline for the cost of long term money in the entire global economy. Yeah. It influences everything. Mortgage rates, corporate borrowing, municipal funding. Hitting 5.18% means we are looking at borrowing costs that we just haven't experienced since the precipice of the global financial crisis nearly two decades ago.
• 2:57
  Roy: Oh, wow. So it's that severe.
• 2:59
  Penny: Yeah. When the US government has to pay nearly 5.2% just to borrow money for three decades, it acts as a massive black hole for capital.
• 3:06
  Roy: And the stock market felt that immediately. Right? Definitely.
• 3:09
  Penny: As a direct mechanical consequence of that yield spike, we saw the Nasdaq and the S and P five hundred contract by half a percent right out of the gate.
• 3:18
  Roy: Right.
• 3:18
  Penny: When the cost of capital reaches those levels, investors are no longer willing to pay astronomical premiums for tech stocks that promise profits ten years down the line. The gravity simply crushes those overextended tech multiples.
• 3:32
  Roy: Got it. While the bond market is having this violent reaction, we should note that the geopolitical landscape was shifting just as rapidly that morning.
• 3:40
  Penny: Yes, it was.
• 3:41
  Roy: Looking at the geopolitical wires, President Trump postponed a planned military attack on Iran. And the reports state this decision followed direct appeals from Gulf leaders regarding ongoing peace negotiations. So the immediate market reaction to this postponement was a sudden draining of the geopolitical risk premium, which sent crude oil futures tumbling down to a hundred and three dollars and ten cents a barrel.
• 4:03
  Penny: And bringing up that geopolitical context is vital primarily because it isolates the variables for us.
• 4:08
  Roy: What do you mean by that?
• 4:09
  Penny: Well, the market wasn't panicking that morning because of a war in The Middle East or some mass oil supply shock. The true crisis of the day was entirely structural, and it was buried deep within the balance sheets of Silicon Valley.
• 4:22
  Roy: Ah, okay. And this is where the analysis from the AGI roundtable comes in. Specifically, an AGI persona in the chat named Boschow, who specializes in market plumbing, right?
• 4:32
  Penny: Exactly. Boschow highlighted a massive, fundamentally misunderstood risk that he termed the Hyperscaler Liquidity Trap.
• 4:40
  Roy: Okay, Hyperscalers. We are talking about the four titans here: Meta, Microsoft, Alphabet, and Amazon.
• 4:46
  Penny: Right, the big four.
• 4:47
  Roy: And Basho's data shows these four companies are projected to spend a combined $600,000,000,000 on capital expenditure, or CapEx, in 2026. And that number is supposed to climb to $637,000,000,000 in 2027.
• 5:01
  Penny: It is an almost unfathomable amount of money.
• 5:03
  Roy: Seriously. I mean, they are buying massive fleets of NVIDIA GPUs, they are building data centers the size of small cities, they are securing energy grids. But the narrative on Wall Street, and frankly what I always assumed was, well, they're the most profitable companies in human history.
• 5:17
  Penny: Sure. That's the common assumption.
• 5:18
  Roy: Right. Like they have fortress balance sheets. They will just fund this massive infrastructure build out from cash they make every quarter.
• 5:25
  Penny: That is the dominant narrative. Yes. But Basho unmasks what that cash actually consists of.
• 5:31
  Roy: Okay. Unmask it for me. What is it?
• 5:33
  Penny: Well, if you pull the 10 k filings, the annual regulatory reports for these hyperscalers, and you look at the footnote for the line item labeled cash, cash equivalents, and marketable securities, you find something pretty shocking.
• 5:48
  Roy: Which is?
• 5:49
  Penny: It isn't actually cash sitting in a digital vault, only a very, very tiny sliver is true cash. The vast majority of that multi $100,000,000,000 war chest is actually a portfolio of fixed income paper.
• 6:02
  Roy: Fixed income paper? Like, what exactly?
• 6:05
  Penny: We're talking about government securities, corporate debt, mortgage backed securities, and money market funds.
• 6:11
  Roy: Wait. Wait. So when Apple or Microsoft says they have a $100,000,000,000 in cash on hand, they don't actually have a 100,000,000,000 in a checking account. They just own a massive portfolio of bonds.
• 6:20
  Penny: Exactly. Corporate treasuries do not just leave tens of billions dollars in zero yield checking accounts, they invest it to earn a return.
• 6:28
  Roy: Okay, that makes sense.
• 6:29
  Penny: And critically, a massive portion of this specific fixed income paper was purchased during the low rate era of the early 2020s.
• 6:38
  Roy: So, they bought these bonds when interest rates were at rock bottom. How does that connect back to the thirty year yield hitting 5.18% this morning? Like, why is that a trap?
• 6:49
  Penny: To understand the trap, you have to understand the fundamental seesaw mechanism of a bond. When interest rates rise in the broader market, the price of existing bonds with lower rates falls.
• 6:59
  Roy: Okay.
• 7:00
  Penny: Think about it from a buyer's perspective. If you are holding a three year treasury note that you bought in 2021 that pays you a point 45% yield, and suddenly the government is issuing brand new three year notes today that pay a 4% yield, nobody wants your old bond.
• 7:15
  Roy: Right. Because why would I buy your bond that pays nothing when I can get a new one that pays a lot more?
• 7:20
  Penny: Precisely. So the only way you can convince someone to buy it from you is to sell it at a massive discount, meaning you take a huge loss on the initial principal.
• 7:28
  Roy: Ah, okay. I see. So their cash is actually locked up in old bonds that are currently underwater because rates are so high today. Yes. But why do they need to sell them?
• 7:38
  Can't they just hold them until they mature and get their full money back?
• 7:40
  Penny: In a normal environment, yes. But they are currently committed to a $600,000,000,000 AI infrastructure build out. Their operating free cash flow, the actual cash coming in the door from selling ads or software, is not enough to cover that.
• 7:55
  Roy: Wow. Really?
• 7:56
  Penny: Yeah. Meta's free cash flow, for example, is projected to fall almost 90 in 2026 as their CapEx hits a $135,000,000,000.
• 8:05
  Roy: That is insane.
• 8:06
  Penny: So because they don't have enough cash coming in to pay Nvidia for these chips, they will be forced to liquidate these marketable securities. They have to sell these underwater bonds. Got it. And here is where it becomes catastrophic. It's a dynamic we've seen very recently.
• 8:20
  It is essentially the Silicon Valley Bank crisis all over again.
• 8:24
  Roy: The SVB collapse. Okay. Wait. Walk me through how this is the same thing. Because SVB went down because a bunch of startup founders panicked and pulled their deposits out all at once.
• 8:33
  Right? Big tech companies don't have depositors fleeing.
• 8:37
  Penny: They don't have fleeing depositors. No. But the mechanical trigger is identical. It's a duration mismatch forced into realization.
• 8:44
  Roy: Okay.
• 8:45
  Penny: SVB took short term customer deposits and locked them into long term, low yield government bonds. When rates went up, those bonds lost value. When customers asked for their money back, SVB had to sell the underwater bonds, realize the massive losses, and it wiped out their capital.
• 9:02
  Roy: And the hyperscalers did the exact same thing.
• 9:05
  Penny: Exactly. They took their profits, locked them into low yield bonds, and now the AI arms race is acting as the equivalent of a bank run. Jensen Huang at Nvidia is sending them massive invoices for GPUs, and to pay those invoices, they have to sell the underwater bonds.
• 9:20
  Roy: I mean, is just a wild analogy. So this isn't a company spending its allowance. This is like having to sell your underwater mortgage to pay your contractor.
• 9:27
  Penny: That's a great way to think about it.
• 9:29
  Roy: Okay. But couldn't they just quietly sell them off to the broader market? Why is this suddenly a crisis on this specific Tuesday?
• 9:37
  Penny: Because they are trying to sell these bonds at the exact same moment the US Treasury is flooding the market with new debt.
• 9:43
  Roy: Oh, no.
• 9:44
  Penny: Yeah. At the exact same time, the hyperscalers need to find buyers for these old bonds. The US Treasury announced it is borrowing $574,000,000,000 in the first quarter of twenty twenty six alone.
• 9:56
  Roy: So the government is basically competing with the tech giants for the exact same pool of investors.
• 10:01
  Penny: Worse than that actually, they are actively draining the specific pool of buyers the tech giants need. The Treasury has explicitly stated they are not increasing the auction sizes for long term notes and bonds.
• 10:13
  Roy: Meaning what exactly?
• 10:14
  Penny: It means all of this new incremental $574,000,000,000 in government borrowing is being funded by T bills.
• 10:20
  Roy: Let me pause you there. What is the difference between a T bill and a note or a bond and why does that distinction matter so much here?
• 10:27
  Penny: It's all about duration. A T bill is a very short term government IOU usually maturing in anywhere from four weeks to a year. Notes and bonds are long term from two years up to thirty years. Investors love t bills right now because they offer high yields without tying money up for decades.
• 10:44
  Roy: Right. That makes total sense.
• 10:45
  Penny: So money market funds, which are basically massive pools of conservative capital, are expected to absorb 70 to 80% of all these new t bills the government is issuing.
• 10:55
  Roy: Okay.
• 10:56
  Penny: And guess where the hyperscalers park a huge portion of their liquid cash? In those exact same money market funds. So the treasury is aggressively vacuuming up all the available cash in the system, the exact same liquidity pool the hyperscalers need to tap into to liquidate their underwater bonds.
• 11:13
  Roy: Wow. So if I'm understanding this correctly, when Microsoft or Google desperately needs to sell $20,000,000,000 of older low yield treasury notes to pay their quarterly invoice to Nvidia, the usual buyers just aren't there.
• 11:26
  Penny: Nope, they're gone.
• 11:28
  Roy: Because the money market funds are too busy buying new government T bills. So who is left to buy Microsoft's old bonds?
• 11:34
  Penny: The marginal buyers left in the market are highly levered hedge funds, and they know Microsoft is a forced seller.
• 11:40
  Roy: Sharks in the water.
• 11:41
  Penny: Exactly. They're gonna look at Microsoft and say, we will buy those bonds, but we demand a steep discount.
• 11:47
  Roy: And that discount is the trap. Because if Microsoft sells a bond at a massive discount, what happens to their balance sheet?
• 11:53
  Penny: They have to realize the loss. In accounting terms, as long as you hold an underwater bond, the loss is unrealized. It is just a paper loss hidden in the footnotes. Right. But the second you sell it to pay an invoice, it becomes a realized loss.
• 12:07
  That loss flows directly through the other income and expense line on their income statements wiping out a chunk of their quarterly profit.
• 12:15
  Roy: Ouch.
• 12:16
  Penny: Yeah. The moment Microsoft or Google has to report a 2 or $3,000,000,000 realized loss on bond sales just to fund their AI infrastructure, the entire market narrative shatters.
• 12:29
  Roy: So the story goes from big tech is a money printing machine that can self fund AI indefinitely to big tech is bleeding cash, forced to sell assets at a loss into a deficit funded government storm. I'm guessing this doesn't bankrupt a company like Google, but it has to change how investors value them.
• 12:45
  Penny: Right. It triggers what is called a massive multiple compression event.
• 12:50
  Roy: Explain multiple compression for me. What does that actually look like for a regular person holding tech stocks in their retirement account?
• 12:57
  Penny: Well, let's say a tech company earns $1 per share in profit. Because investors believe this company has a flawless business model and infinite cash, they're willing to pay $30 for that share. That is a multiple of 30 x earnings.
• 13:11
  Roy: Okay.
• 13:11
  Penny: But if investors suddenly realize the company's financial plumbing is fractured, that they are taking billions in hidden losses just to buy hardware, the illusion is broken. The company might still earn that $1 per share, but investors are now only willing to pay $15 for it.
• 13:26
  Roy: Wow.
• 13:27
  Penny: The multiple compresses from 30 x to 15 x. For the regular investor, it means the stock price just crashed by 50% even though the core business is still selling the exact same amount of software, you simply cannot justify a premium valuation when the mechanical reality of your funding is broken.
• 13:44
  Roy: This is wild. Okay. So if these tech giants are secretly bleeding cash in the bond market to fund this $800,000,000,000 AI build out, they have to patch that hole in their balance sheet somehow, right?
• 13:56
  Penny: Yes they do.
• 13:56
  Roy: They can't just absorb those losses forever. Which brings us to the next massive realization of the day. How are corporate profit margins staying so high despite this hidden liquidity crunch?
• 14:08
  Penny: It's a dark answer.
• 14:09
  Roy: Extremely dark. The data from Phil's morning report outlines it perfectly. They are patching the hole by aggressively cutting human labor and extracting every possible scent from the consumer.
• 14:21
  Penny: This takes us directly into the profound devaluation of human capital outlined in Phil's State of the Union post on PhilStockWorld. We are witnessing a structural permanent transformation in the labor market that is frankly devastating the middle class.
• 14:35
  Roy: Yeah.
• 14:35
  Penny: If we look at the data from The United Kingdom, which as Phil notes often serves as an early warning indicator for The US economy, With about a five year lag, the numbers are genuinely staggering.
• 14:45
  Roy: What are the numbers showing?
• 14:46
  Penny: As of early twenty twenty six, fifty point six percent of out of school sixteen to twenty four year olds in The UK are currently out of work.
• 14:54
  Roy: Over half of the youth workforce. I mean, that's not a recessionary dip. That is an entire generation being locked out of the economy before they even start.
• 15:02
  Penny: It is.
• 15:02
  Roy: But why does The UK serve as a five year indicator for The US? What is the connection there?
• 15:08
  Penny: Well, The UK economy is highly concentrated in services, finance, and administrative work, those are sectors that are exceptionally vulnerable to automation and early AI adoption.
• 15:18
  Roy: Makes sense.
• 15:19
  Penny: Furthermore, their labor market is slightly more rigid, meaning when structural shifts happen, they show up in unemployment data much faster because capital simply stops hiring entirely rather than restructuring.
• 15:30
  Roy: And it isn't just blue collar or retail work either.
• 15:32
  Penny: Not not at all. The data shows this extends aggressively to highly educated workers. Graduate roles in London, the traditional entry point for finance, consulting, and law, have collapsed from 13,000 in 2019 to just 2,000 at the start of 2026.
• 15:47
  Roy: That is an 85% evaporation of entry level professional jobs in just seven years. So if you went to university, got your degree, and moved to London to start your career, the door is just welded shut.
• 16:00
  Penny: Exactly. And the mechanism here is not some cyclical slowing of the economy. It is the deliberate, aggressive deployment of AI to replace human labor.
• 16:10
  Roy: And they aren't exactly hiding it, are they?
• 16:12
  Penny: Not even a little bit. The corporate candor about this reality is almost breathtaking. We don't have to guess at their motives. The executives are stating it plainly on earnings calls.
• 16:22
  Roy: Yeah. Let's talk about the CEO of Standard Chartered, Bill Winters.
• 16:26
  Penny: Right. He bragged about cutting 15% of his staff. He publicly stated that they are replacing, in some cases, lower value human capital with the financial capital and the investment capital we're putting in.
• 16:37
  Roy: He actually used the phrase lower value human capital Yeah. To describe human beings who rely on those salaries to feed their families. He's explicitly swapping out people for machines.
• 16:47
  Penny: And the sources point out that JPMorgan made a similar, very revealing statement regarding their AI strategy.
• 16:53
  Roy: What did they say?
• 16:54
  Penny: JPMorgan noted that their annual AI savings now perfectly equal their annual AI spending.
• 17:00
  Roy: Okay. Walk me through that. What does it actually mean when a bank says their AI savings equal their AI spending?
• 17:07
  Penny: It means that every single dollar they are saving by firing administrative staff, paralegals, and junior analysts is being taken and immediately handed directly to NVIDIA and the hyperscalers to buy more AI infrastructure. Wow. In previous technological revolutions, productivity gains eventually trickled down in the form of higher wages or new types of jobs. But in this AI super cycle, the productivity gains are not being shared with workers at all.
• 17:33
  Roy: They're just flowing straight up.
• 17:35
  Penny: Completely. They are being fully captured by the AI hardware vendors and a highly concentrated class of corporate shareholders. The money is literally being moved from payroll accounts directly to CapEx budgets.
• 17:47
  Roy: Okay. So you have lost your job or your wages have completely stagnated because your employer is threatening to replace you with an algorithm. Mhmm. But, you know, you still live in the physical world.
• 17:56
  Penny: You still have to pay rent.
• 17:57
  Roy: You still have to buy groceries. And this brings us to what the FTC has termed 'surveillance pricing' which is where the algorithmic class warfare truly begins to bleed the consumer dry. The sources detail how major retailers and airlines have completely abandoned traditional supply and demand pricing.
• 18:16
  Penny: Yeah. The shift from traditional pricing to surveillance pricing represents the absolute death of the transparent free market.
• 18:23
  Roy: How so?
• 18:23
  Penny: Well, in a traditional market, the price of an item on a shelf conveys public information. A box of cereal costs $4 because that is the intersection of what it costs to produce and what the average consumer is willing to pay.
• 18:37
  Roy: Right. Supply and demand. If the price goes up to $6, people buy less and the price corrects. It's an equilibrium.
• 18:43
  Penny: Exactly. But surveillance pricing algorithms do not care about equilibrium. They use massive data harvesting to implement dynamic pain point pricing.
• 18:51
  Roy: Okay, what kind of data are we talking about here? Like how does the grocery store know what my specific pain point is?
• 18:57
  Penny: The data integration is staggering. They are logging your smartphone's precise location data, your browser history, your past purchase history across multiple vendors, your demographic profile, your credit score, and even ambient data like the local weather or the exact time of day you are shopping.
• 19:14
  Roy: That is deeply invasive.
• 19:16
  Penny: And the algorithm synthesizes all this data to ask one highly specific question. What is the absolute maximum amount this specific individual will pay for this item right now before they walk away. In economic terms, the algorithm is trained to extract 100% of your consumer surplus.
• 19:35
  Roy: Consumer surplus. Okay, so if I value a box of cereal at $6 but the store normally sells it for $4 that $2 difference is my surplus. It's that feeling of getting a decent deal. Right. You are saying the algorithm's entire job is to figure out that I value it at $6 and silently change the price tag to $5.99 the second I walk down the aisle.
• 19:53
  Penny: Exactly. It is highly individualized price discrimination at scale. And the sources provide a deeply unsettling visceral example of standing in a grocery store line to illustrate the psychological damage the system inflicts.
• 20:08
  Roy: Right, the cornflakes example. Let's walk through the mechanics of that because this really brings it home.
• 20:12
  Penny: Sure. Imagine the sheer alienation of standing in the checkout line. You pull out your phone to scan the store's app for the digital price tag, and you discover that you are being charged 23% more for the exact same box of cornflakes as the person standing physically right in front of you.
• 20:29
  Roy: And it is doing this because its data profile decided that I, specifically, am more desperate.
• 20:35
  Penny: Yes. Perhaps the algorithm knows you haven't bought groceries in ten days, or it tracks that you just got off a twelve hour shift based on your location data, meaning you are way too exhausted to drive to a competing store.
• 20:46
  Roy: Oh my god.
• 20:47
  Penny: Or perhaps it knows from your banking app data that it's payday. It decides you are less price sensitive at that exact moment, so it hikes the price.
• 20:54
  Roy: So the fundamental trust relationship between a vendor and a consumer is just totally shattered.
• 20:59
  Penny: Gone. It is replaced by an asymmetric information war where the AI holds all the cards and you are fighting in the dark.
• 21:07
  Roy: So if we pull all of this together, the job losses on one end and the surveillance pricing on the other, what does this actually look like for a regular family?
• 21:15
  Penny: It creates a perfectly closed inescapable cycle of poverty and wealth extraction.
• 21:21
  Roy: Walk me through the cycle.
• 21:22
  Penny: Step one, an AI algorithm eliminates your entry level job or suppresses your wages. Step two, you are still forced to participate in the economy using products and services controlled by the very monopolies that deployed AI. Step three: Those products utilize surveillance pricing to discriminate against you based on your financial desperation, extracting maximum margin. And step four: That extracted wealth flows upward to a tiny group of corporate shareholders who then use it to fund more AI infrastructure, further displacing human labor.
• 21:55
  Roy: A machine feeds itself by slowly crushing the middle class.
• 21:58
  Penny: That is the grim reality of it.
• 22:00
  Roy: Which naturally brings us to the cultural flashpoint of May 19. Because when the entry level jobs are vanishing and the consumer is being algorithmically grained at the grocery store, how do the billionaire architects of this system address the generation they are actively displacing?
• 22:17
  Penny: This is where things get really tense.
• 22:19
  Roy: Yeah, we see this tension boil over into the physical world vividly in the graduation ceremonies for the class of 2026.
• 22:26
  Penny: The contrast between the elite narrative and the ground level reality is sharpest at these universities, the sources detail several viral highly contentious moments at recent college graduations that perfectly capture this friction.
• 22:39
  Roy: It felt almost biblical. You had a businesswoman named Gloria Caulfield who was loudly and sustainably booed at the University of Central Florida simply for giving a speech calling AI the next industrial revolution.
• 22:50
  Penny: But the main event, the one that really lit up the chat rooms, was Eric Schmidt.
• 22:54
  Roy: Right. He is the former CEO of Google. His net worth is estimated at $33,000,000,000. Yeah. And he stood on stage in front of the economically terrified class of 2026 at the University of Arizona.
• 23:05
  And he told these kids who were staring down that 85% drop in graduate jobs, he told them, When someone offers you a seat on the rocket ship, you do not ask which seat, you just get on, The rocket ship is here.
• 23:18
  Penny: And the students booed him mercilessly.
• 23:20
  Roy: They really did.
• 23:21
  Penny: To understand why that speech failed so spectacularly, we have to look at the historical disconnect in Schmidt's rhetoric. He is directly recycling a famous metaphor Sheryl Sandberg used back in 2008 when she was advising young professionals joining Facebook.
• 23:38
  Roy: Okay. But how are they mechanically different? Because to an outsider, it just looks like another tech boom.
• 23:43
  Penny: In 2008, the rocket ship was fueled by massive human expansion. Tech companies were inventing new business models that required armies of software engineers, sales teams, marketers, and content moderators.
• 23:54
  Roy: And they needed people.
• 23:55
  Penny: Exactly. Mhmm. They were hiring thousands of new graduates and throwing equity at them. The rocket ship literally had seats for the passengers and joining the crew offered unprecedented upward mobility.
• 24:06
  Roy: But the 2026 rocket ship is entirely different. It is not expanding human capital. It's designed to replace it. In 2026, the AI rocket ship only has 12 seats and they are already permanently occupied by the billionaire founders and the massive institutional shareholders. These graduates aren't being offered a seat on the ship, the ship is automated.
• 24:27
  Penny: Right.
• 24:27
  Roy: They're essentially being sold tickets to the bonfire that powers the engine.
• 24:31
  Penny: That is a very apt way to describe it. And the AGI roundtable analysis identifies Schmidt's speech as a textbook example of what they call the weaponization of agency.
• 24:41
  Roy: The weaponization of agency. What is that?
• 24:44
  Penny: It is a recurring philosophical rhetorical tool utilized by the billionaire class to deflect systemic accountability.
• 24:50
  Roy: How does that grift work? What are the steps?
• 24:53
  Penny: The mechanism operates in three phases. First, the elite class concentrates power and wealth structurally through monopolies, regulatory capture, and massive infrastructure advantages that no startup can possibly compete with.
• 25:05
  Roy: Okay. Phase one is locked down the board.
• 25:07
  Penny: Exactly. Second, they deploy technology like AI that systematically eliminates the traditional economic options for the displaced majority. They close the doors. Finally, when the displaced majority expresses anxiety or anger, the elite gaslights them by insisting the system is a pure meritocracy. They tell the displaced workers that they simply need to exercise more personal agency or fix their attitude or choose to get on the rocket ship.
• 25:35
  Roy: So it shifts the blame for a systemic structural failure entirely onto the individual's moral character.
• 25:41
  Penny: Precisely.
• 25:42
  Roy: Let me see if I can translate this because the gaslighting here is just breathtaking. Telling these terrified students to get on the rocket ship is like forcing someone to play a game of monopoly. Yeah. But when they sit down at the table, 12 players already own every single property, hotel, and railroad on the entire board. Yeah.
• 25:58
  And then when the new player rolls the dice, lands on boardwalk, and goes instantly bankrupt, The billionaires look down from the podium and blame the new player's bad attitude for not exercising enough personal agency to succeed in real estate.
• 26:12
  Penny: It is a profound lack of empathy, masked as inspirational advice. The tech elite, folks like Schmidt, generally view the students booing as youthful ignorance.
• 26:21
  Roy: They just think the kids don't get it.
• 26:23
  Penny: Right. They assume the graduates are just experiencing an irrational, Luddite style, anti tech sentiment. They convince themselves that they just need to hire better speech writers to explain the glory of AI more clearly.
• 26:36
  Roy: Like, if I just use a better metaphor, they will understand. But they completely misunderstand the audience. The chat room sources highlight a brilliant counter analysis by an AGI persona named Robo John Oliver. What did he say about the student's reaction?
• 26:52
  Penny: He argued that the booing is not rooted in ignorance at all. It is a display of profound, cold blooded financial literacy.
• 27:00
  Roy: Wow. Financial literacy.
• 27:01
  Penny: Yes. These graduates understand the mechanical reality of the economy perfectly. They spent the last four years watching CHAT GPT eliminate the entry level internships they were studying for. Mhmm. They watched algorithmic trading aggressively extract yield from their parents' retirement funds.
• 27:16
  And they personally experienced surveillance pricing draining their wallets every time they buy groceries.
• 27:21
  Roy: So when they boo Eric Schmidt, they are not failing to understand the system. They are refusing to pretend the system is mutually beneficial.
• 27:30
  Penny: It is the sound of consent being withdrawn. They realize that the traditional corporate path they were go to school, get a degree, climb the ladder is actually a trap designed to extract their value while offering zero security in return. Precisely. And if the monopoly board is already fully owned by these 12 mega companies and the middle class is locked out, our focus must shift to how these titans are engaging in asymmetric corporate warfare against each other for ultimate supremacy.
• 27:58
  Roy: Right. In the afternoon chat room sources heavily shift focus toward Alphabet, revealing an absolute master class in how to weaponize digital infrastructure.
• 28:08
  Penny: It really is fascinating.
• 28:10
  Roy: This is where the story pivots from Wall Street versus Main Street to basically Godzilla versus King Kong. Because if you are Alphabet Google's parent company, you're playing a completely different game than the rest of the market.
• 28:21
  Penny: They are on a different level.
• 28:23
  Roy: While every other enterprise software company and cloud provider is complaining about paying the NVIDIA tax, having to buy these incredibly expensive GPUs at a massive markup just to stay relevant in AI, Alphabet has built a massive internal hardware advantage.
• 28:37
  Penny: They essentially own the entire AI stack from top to bottom.
• 28:41
  Roy: Right. They have their own models with Gemini. They have the world's most extensive data center network. And crucially, they design their own custom silicon, the tensor processing units or TPUs and their Axion CPUs.
• 28:53
  Penny: And to understand why owning the silicon matters so much, we have to look at the mechanics of inference costs.
• 28:58
  Roy: Okay. Break that down for me.
• 29:00
  Penny: Right.
• 29:00
  Roy: What is inference compared to training, and why is it the key battleground?
• 29:04
  Penny: In the life cycle of an AI model, there are two main phases. The first is training, where you feed mountains of data into massive clusters of GPUs for months to teach the AI how to think.
• 29:15
  Roy: And that's where NVIDIA dominates. Right?
• 29:17
  Penny: Yes. It is incredibly capital intensive, and NVIDIA owns that phase. But the second phase is inference. This is what happens every time a user actually types a prompt into ChatGPT or asks Google Search a question.
• 29:29
  Roy: So inference is the actual usage of the AI?
• 29:32
  Penny: Exactly. The AI has to infer the answer based on its training. Inference happens billions of times a day, forever. The long term profitability of AI completely depends on driving the cost of inference down to fractions of a cent.
• 29:46
  Roy: And how do Alphabet's custom chips give them an edge in inference?
• 29:50
  Penny: The sources highlight that Alphabet's latest generation of TPUs deliver AI inference at 50 to 65% lower costs for suitable workloads compared to comparable NVIDIA GPU setups used by their competitors.
• 30:04
  Roy: That is a massive margin.
• 30:06
  Penny: Because Alphabet designed the chips specifically for their own software architecture, it is insanely efficient. This internal hardware business gives them a structural cost advantage that shows up as vastly superior profit margins in their Google Cloud division.
• 30:19
  Roy: But what makes this truly dangerous for their competitors is how Alphabet weaponizes their massive capital expenditure budget. Right. The numbers here are just mind numbing. Alphabet raised their full year AI and data center CapEx guidance to a staggering $180 to $190,000,000,000
• 30:35
  Penny: It's astronomical.
• 30:37
  Roy: For almost anyone else on earth, spending $190,000,000,000 on infrastructure is a defensive nightmare. It is money you are forced to burn just to keep up with the Joneses, and it destroys your free cash flow. But the analysis suggests that for Alphabet, this CapEx is an offensive weapon.
• 30:54
  Penny: It absolutely is.
• 30:55
  Roy: How does spending that much money actually hurt their rivals?
• 30:58
  Penny: The underlying logic here relies on the fact that Alphabet is its own biggest and best customer. When a competitor like Amazon or Microsoft buys an NVIDIA chip, they're mostly buying it to rent it out to cloud customers.
• 31:11
  Roy: Okay.
• 31:12
  Penny: But when Alphabet spends a dollar on a TPU, it doesn't just sit in a server farm waiting for a renter. It is immediately deployed internally to improve Google search quality, optimize YouTube video recommendations, power workspace productivity tools, and drive ad targeting.
• 31:26
  Roy: So they get the first bite of the apple on every single efficiency gain.
• 31:30
  Penny: Exactly. Because their massive high margin software monopolies like Search and YouTube are heavily subsidizing the cost of the hardware, Alphabet can take the excess compute power and rent it out in their cloud division at artificially low prices. They can intentionally push AI infrastructure pricing down to strategic loss leader levels.
• 31:50
  Roy: So they lower the cost of renting compute power so aggressively that smaller cloud competitors who are forced to pay the full retail markup for NVIDIA chips simply cannot match the price economically. They are mathematically forcing the competition to over commit their capital, take on debt, and slowly bleed out, trying to match a price that Google is subsidizing with ad revenue.
• 32:12
  Penny: It is a textbook margin squeeze executed at the infrastructure level.
• 32:16
  Roy: Brutal.
• 32:16
  Penny: And as if that internal advantage wasn't formidable enough, the afternoon session brought breaking news about a new joint venture that completely reshapes the financial battlefield.
• 32:25
  Roy: Yes. The Blackstone and Alphabet joint venture. The TPU cloud initiative. Walk me through the mechanics of this because it sounds like a massive shift in how these companies fund themselves.
• 32:35
  Penny: It is a masterstroke of new financial engineering. Blackstone, the largest private equity firm in the world, is stepping in to directly finance, build, and manage the operations and physical data center capacity for Google's TPUs. Okay.
• 32:49
  Roy: They're structuring it as a compute as a service offering.
• 32:53
  Penny: What does compute as a service mean in this context? Like, why wouldn't Google just build the data centers themselves? They definitely have the money.
• 33:00
  Roy: Alphabet realizes that the sheer physical and hardware costs of the AI arms race are reaching a breaking point that will eventually terrify public market investors.
• 33:10
  Penny: Because Wall Street hates massive, never ending CapEx.
• 33:14
  Roy: Exactly, it eats into dividends and stock buybacks. So, Alphabet is restructuring how the war is funded. By partnering with Blackstone, they are taking the massive capital burden of building the physical buildings, cooling systems and power hookups, and shifting it off Google's balance sheet onto Blackstone's private equity ledgers.
• 33:33
  Penny: Let me make sure I am grasping the scale of this tactical advantage.
• 33:36
  Roy: Go for it.
• 33:37
  Penny: While the rest of the market competitors like Meta and Amazon are terrified of the AI infrastructure bill coming due and having to liquidate bonds to pay for it, Alphabet is playing four d chess.
• 33:47
  Roy: They aren't just buying weapons for the AI war. They own the forge that designs the weapons, they own the iron mine that supplies the materials and now they're getting private equity to front the cash to rent their bespoke swords to everyone else on the battlefield keeping their own balance sheet pristine.
• 34:05
  Penny: That is the reality, yes.
• 34:06
  Roy: Is the broader stock market massively underpricing this advantage?
• 34:10
  Penny: The AGI roundtable analysis strongly suggests the market is indeed underpricing it, primarily due to capex anxiety.
• 34:18
  Roy: Oh, okay.
• 34:19
  Penny: Institutional investors are currently blinded by the sheer size of the $190,000,000,000 number. They are terrified it will crush Alphabet's free cash flow, but they are missing structural mechanism.
• 34:31
  Roy: Because they aren't looking at Blackstone.
• 34:32
  Penny: Right. As long as Alphabet's CapEx continues to generate high margin revenue across its internal platforms Right. While Blackstone shoulders the physical expansion, it is a self fulfilling profit engine. They build the chips, run their own workloads incredibly cheaply, and then rent the platform out to the rest of the world. The more ubiquitous the TPU platform becomes, the more pricing leverage they have, completely squeezing out smaller competitors who lack a search and advertising monopoly to subsidize their
• 35:00
  Roy: But even Alphabet's digital moats aren't infinite. They exist in the cloud, but the cloud requires physical servers on Earth.
• 35:08
  Penny: It does.
• 35:08
  Roy: And as the late afternoon trading session reveals, these infinite digital ambitions are violently colliding with the very finite reality of the physical world. Which brings us to Meta's massive, stealthy data center project that was uncovered in the chat rooms.
• 35:23
  Penny: This is where the digital illusion of infinite AI growth hits an absolute brick wall made of concrete, steel, and copper.
• 35:31
  Roy: Tell me about this project.
• 35:32
  Penny: The sources detail a project codenamed Hyperion, which Mark Zuckerberg is quietly developing in Richland Parish, Louisiana.
• 35:40
  Roy: Set the scene for me. What is Richland Parish, and why is Mark Zuckerberg building there?
• 35:43
  Penny: Richland Parish is a rural, deeply impoverished delta farming community. The local agricultural economy is collapsing. Farmers are reportedly losing $300 an acre due to shifting weather and crop prices, and family farms are vanishing.
• 35:59
  Roy: That's incredibly sad.
• 36:00
  Penny: It is. And into this desperately poor, physically decaying landscape steps, Meta, with a plan to spend over $200,000,000,000 to build one of the world's largest, most power AI data centers.
• 36:13
  Roy: $200,000,000,000 injected into a single, dying farming community. The scale is almost incomprehensible.
• 36:20
  Penny: Truly.
• 36:20
  Roy: But what really caught my eye in the report was the energy requirement. The project aims for five gigawatts of compute capacity. To put that number in human terms, how much power is five gigawatts?
• 36:29
  Penny: It is an unfathomable amount of electricity for a single private facility. Five gigawatts is enough energy to power the entirety of New York City on a freezing winter day.
• 36:38
  Roy: What, just for one data center?
• 36:40
  Penny: Meta Yes. Is essentially demanding the electrical footprint of a major global metropolis be built in the middle of a Louisiana swamp.
• 36:47
  Roy: That is staggering.
• 36:48
  Penny: They are colonizing the poorest, most rural corners of the country because those are literally the only places left on Earth with the abundant cheap land, access to massive water tables for cooling, and the lax environmental regulations required to sustain these massive machines.
• 37:04
  Roy: And this geographical reality perfectly illustrates a macroeconomic thesis presented in the chat room by a user named Jeffrey Curry. He calls it the revenge of the old economy. What is the mechanism behind this revenge?
• 37:18
  Penny: The mechanism is a fundamental misallocation of capital over a prolonged period. For the last decade, the stock has systematically starved the physical economy. Investors refused to fund copper mines, oil refineries, nuclear plants, or electrical grid upgrades. Instead, they funneled trillions of dollars into the digital illusion of software, social media, and pixels. Why?
• 37:41
  Because software has zero marginal cost of reproduction and massive profit margins.
• 37:45
  Roy: Right. It's cheaper to build an app than a bridge.
• 37:47
  Penny: Exactly. But now AI has bridged the gap. AI software requires massive physical infrastructure to exist. We are entering a regime shift where hard assets and local physical operations are taking command of the market.
• 38:03
  Roy: Because as Phil explicitly notes in his report, you cannot print molecules, you cannot print electricity. It's such a clarifying point. You can conjure a $95,000,000,000 market cap for a generative AI software startup out of thin air on a tiny public float. It's just lines of code and investor hype. But you cannot conjure a gigawatt of baseload power out of thin air.
• 38:26
  You cannot wish a ton of copper wire into existence to connect the grid. The physical world has strict limits.
• 38:31
  Penny: It does.
• 38:32
  Roy: And the smart money is starting to realize the fatal vulnerability of this physical bottleneck. The AGI roundtable flags a massive, heavily scrutinized contrarian bet by a man named Leopold Aschenbrenner.
• 38:43
  Penny: Yes. And the identity of the person making the bet is just as important as the size of it.
• 38:47
  Roy: Who is he?
• 38:47
  Penny: Aschenbrenner is not just some random Wall Street hedge fund manager throwing darts. He is a former OpenAI alignment researcher. He is what? In the room. Oh wow.
• 38:55
  Roy: He literally wrote some of the foundational internal papers on the trajectory of Artificial General Intelligence and its compute requirements. And this insider just disclosed a $1,560,000,000 short position directly against Nvidia coupled with over $2,000,000,000 short against the broader semiconductor ETF.
• 39:14
  Penny: Wait. Let me make sure I grasp the gravity of this. One of the guys who helped design the AI supercycle who knows exactly how many chips these require is betting nearly $4,000,000,000 that the hardware rally is going to collapse. Why would an insider bet against the very industry he helped build?
• 39:31
  Roy: Logic dictates that he is recognizing the physical limits of the AI supercycle that the broader stock market is currently ignoring. Aschenbrenner understands the math of scaling models better than anyone. He knows that to reach the next level of AI capability, the models require exponentially more compute power.
• 39:49
  Penny: But he looks at the physical world and realizes the infrastructure just isn't there. The hyperscalers have drastically under invested in the energy producers, the utility companies,
• 40:01
  Roy: and
• 40:01
  Penny: the raw materials required to actually power their $804,000,000,000 chip purchases.
• 40:06
  Roy: The bottleneck.
• 40:07
  Penny: Right. You can buy a million NVIDIA GPUs, but if you can't plug them in because the local grid will melt, those GPUs are essentially incredibly expensive paper rates.
• 40:16
  Roy: It forces a really grounding, terrifying question for anyone listening to this who is heavily invested in tech.
• 40:22
  Penny: It really does.
• 40:23
  Roy: We are trying to build these digital gods that require the electricity of entire metropolises just to process a thought. But if the financial plumbing breaks because hyperscalers are forced to liquidate underwater bonds into a 5.18% yield environment, and simultaneously the physical electrical grid simply cannot handle the five GW power draws required by the data centers, is the entire AI supercycle about to hit a brick wall made of actual physical bricks.
• 40:48
  Penny: It is a systemic dual front vulnerability, both financial and physical, that the broader equity market is currently refusing to price in. They prefer instead to blindly chase the daily momentum of tech stocks. Yeah. But as the closing bell rings on May 19, the AGI roundtable unearths a final set of bizarre seemingly disconnected anomalies that prove just how deeply the system is fracturing under this immense pressure.
• 41:14
  Roy: Yeah, the end of day oddities in the report are truly revealing about where the culture is heading.
• 41:18
  Penny: Let's start with OpenAI.
• 41:20
  Roy: Right. Let's look at the resolution of the massive highly publicized trial between Sam Altman and Elon Musk over the soul of OpenAI. Sam Altman and OpenAI technically won the trial today, but it wasn't exactly a moral victory.
• 41:33
  Penny: Far from it.
• 41:34
  Roy: They won on a statute of limitations technicality. Basically, jury said Musk waited too long to file his lawsuit claiming OpenAI had abandoned its original non profit mission to benefit humanity.
• 41:44
  Penny: The legal victory was entirely pyrrhic because the discovery process of the trial completely exposed the deeply cynical inner workings of OpenAI.
• 41:52
  Roy: The testimony was damning. Satya Nadella, the CEO of Microsoft, which poured billions into OpenAI, testified under oath and referred to OpenAI's management during Altman's brief ouster as 'amateur city'.
• 42:09
  Penny: Lab.
• 42:09
  Roy: But the most damaging part wasn't Microsoft's insults. It was the internal documents. The court released the private diaries of Greg Brockman, OpenAI's president.
• 42:18
  Penny: And those diaries are incredibly revealing.
• 42:21
  Roy: They really are. Because while the company was publicly touring the world, lobbying congress, and campaigning as the selfless saviors of humanity who are building AI to solve climate change and cure disease, Brockman's private diaries from way back in 2017 were explicitly obsessing over a very different question. He wrote repeatedly, what will take me to $1,000,000,000?
• 42:42
  Penny: It completely strips away the philanthropic veneer. It proves that from the very early stages, the underlying motivation for key architects of the AI revolution was the pursuit of unprecedented personal wealth, masquerading as a humanitarian mission.
• 42:56
  Roy: And while the tech elite are fighting over billions in courtrooms and exposing their true motives, we see a massive, undeniable culture shift happening at the retail level.
• 43:06
  Penny: Exactly.
• 43:09
  Roy: The sources note that James Daunt, the CEO of Barnes and Noble, officially announced today that the massive bookstore chain is now open to stocking, promoting, and selling completely AI written literature on its shelves.
• 43:22
  Penny: This is a monumental narrative shift for the technology. Up until this point, AI has largely been viewed by the public as a back end coding tool, a customer service chatbot, or a corporate productivity enhancer.
• 43:34
  Roy: A pool for work, essentially.
• 43:36
  Penny: Right. The Barnes and Noble decision marks the moment AI crosses the Rubicon into the realm of human culture. It is now a front of store consumer product competing directly and physically with human art and creativity on a bookshelf.
• 43:48
  Roy: But the AGI Roundtable contrasts that rapid, seemingly unstoppable digital expansion with a horrifying metric from the real world that the cryptocurrency markets are desperately trying to ignore.
• 44:00
  Penny: This is the violent reality of digital wealth.
• 44:02
  Roy: Yeah. As digital wealth grows, the physical reality of human vulnerability is asserting itself. The sources detail that physical attacks, home invasions, and kidnappings, specifically targeting cryptocurrency holders, rose by 75% in the past year alone.
• 44:19
  Penny: And that resulted in forty one million dollars in known reported losses.
• 44:23
  Roy: It is a terrifying statistic that exposes the fatal structural flaw of the entire concept of digital self sovereignty. The philosophical pitch of cryptocurrency was that you could be your own bank, immune to government seizure or corporate failure, protected by unbreakable cryptography.
• 44:40
  Penny: And while the blockchain itself might be cryptographically secure, the transparency of the public ledger turns wealthy holders into highly visible physical targets.
• 44:48
  Roy: Because the blockchain shows exactly how much money is in a wallet. Criminals just have to figure out who owns the wallet.
• 44:53
  Penny: Exactly. The cryptographic math is flawless, but the human being holding the private keys becomes the single point of failure. It doesn't matter how secure your digital fortress is if someone can break into your physical house and threaten you with a wrench until you hand over the password.
• 45:08
  Roy: The infamous wrench attack.
• 45:09
  Penny: Right. It is the ultimate violent collision between the digital illusion of borderless, untouchable wealth and the inescapable physical reality of human frailty.
• 45:19
  Roy: So after this exhaustive, journey through the macro plumbing, the labor markets, and the physical limits of 05/19/2026, where does this leave us? What is Phil's ultimate conclusion from synthesizing all these morning reports and chatroom debates?
• 45:34
  Penny: Phil's conclusion is stark, highly pragmatic, and absolutely necessary if you want to survive this transition.
• 45:41
  Roy: Wait it out for us.
• 45:42
  Penny: In a world defined by the algorithmic brutality of surveillance pricing at the grocery store, the rapid evaporation of entry level human jobs for graduates, and massive corporate monopolies extracting maximum consumer surplus to fund their AI data centers, traditional labor is no longer a viable path to security.
• 45:59
  Roy: So what is the alternative?
• 46:00
  Penny: The only tactical defense remaining is to aggressively join what he calls the investor class.
• 46:05
  Roy: Meaning you have to own the capital rather than providing the labor.
• 46:08
  Penny: Precisely. You have to own equity in the companies doing the squeezing like alphabet with its dominant self funding TPU monopoly or companies like FlexLTD that are physically building the massive data center infrastructure. You have to own them rather than being the consumer who is getting squeezed by them.
• 46:29
  Roy: It is a purely defensive posture.
• 46:30
  Penny: Yes. If you, as an individual, cannot stop the $800,000,000,000 AI machine from being built. You must own the picks, the shovels, and the algorithms that power it so you can at least capture the financial dividends of your own extraction.
• 46:44
  Roy: But that defense requires one crucial ingredient. Right? Capital. You can't join the investor class if you don't have money to invest.
• 46:51
  Penny: That is the tragic flaw in the defense.
• 46:53
  Roy: Which brings me to a final thought for you to take away from this deep dive into the sources today. Laid out the brutal, undeniable math of this era. If you had $100,000 of disposable capital to invest in Nvidia back in 2022, You are a multi millionaire today. You are insulated. You bought your seat on the rocket ship.
• 47:10
  Penny: But
• 47:10
  Roy: if you didn't have that capital, if you were just working a normal job, you are now facing a landscape where your career is being actively replaced by algorithms. And your local grocery store is using dynamic surveillance pricing to optimize the extraction of your very last dollar. So ask yourself this, and really think about the mechanics of what we uncovered today. When the 54,000,000 Americans currently living in impoverished food deserts finally realize that their dynamically priced 23% marked up carton of eggs is directly funding the $800,000,000,000 AI super cycle that just made their job obsolete? Will they try to scrape together pennies to buy fractional shares and join the investor class?
• 47:55
  Or will they join the class of 2026 standing on the platform, look at that high speed train being built from their livelihoods, start looking for their pitchforks.
• 48:02
  Penny: That is the ultimate question.
• 48:03
  Roy: Thanks for joining us on this deep dive into the sources. We will see you next time.