Latest / The PhilStockWorld Investing Podcast / Thursday Thrust: One Stock to Rule Us All (NVDA, of course!)

Transcript

Thursday Thrust: One Stock to Rule Us All (NVDA, of course!) · The PhilStockWorld Investing Podcast

• 0:00
  Roy: So it is Thursday, 02/26/2026.
• 0:04
  Penny: Yes, it is.
• 0:05
  Roy: And, if you've been glued to the terminals for the last twenty four hours just watching those tickers flash red and green
• 0:12
  Penny: Oh, I have been.
• 0:13
  Roy: You might be feeling a sense of, let's call it cognitive dissonance today.
• 0:17
  Penny: Cognitive dissonance is probably the the polite way to put it. I mean, call it whiplash, honestly.
• 0:22
  Roy: Whiplash. Yeah. That's fair.
• 0:23
  Penny: Because we are looking at a market that is simultaneously euphoric about the future and absolutely terrified of the present reality.
• 0:33
  Roy: It really is a tale of two markets right now. Because on one hand you have Nvidia.
• 0:38
  Penny: The king.
• 0:38
  Roy: The absolute king. The emperor of this AI revolution. And they report earnings that didn't just beat expectations, they absolutely crushed them. Again
• 0:47
  Penny: Like they always do.
• 0:48
  Roy: Right. But the stock reaction, it was it
• 0:51
  Penny: was muted. Almost a shrug from Wall Street.
• 0:53
  Roy: Exactly. Almost a shrug. Meanwhile, you look at the other side of the screen and you have stalwarts like IBM dropping off a cliff.
• 0:59
  Penny: Yeah. Worst trading day in twenty five years for them.
• 1:02
  Roy: Which is wild. Yeah. And suddenly, everyone on Wall Street is talking about utility poles and copper wires. Like, it's the most exciting trade in tech.
• 1:10
  Penny: It really feels like we passed an inflection point overnight. Yeah. The narrative has completely shifted from, you know, look at this magical technology to, wait, what does this actually break?
• 1:22
  Roy: Exactly. And that is what we are doing on today's deep dive. We're gonna unpack this strange moment.
• 1:27
  Penny: It's a lot to unpack.
• 1:28
  Roy: It is. We've gone through a massive stack of reports from the AGI Roundtable Consulting Group, specifically their breakdown of Nvidia's earnings and this broader sort of market schizophrenia we're seeing.
• 1:39
  Penny: Right.
• 1:39
  Roy: So the mission today for you listening is to figure out why a massive earnings beat resulted in a collective 'meh'. Why the market is suddenly terrified of something called the saspocalypse.
• 1:51
  Penny: Great word by the way.
• 1:52
  Roy: It's terrifying but catchy. We need to have a very serious conversation about the physical wall that might stop this whole party in its tracks. We're talking the Matrix economy versus the Adams economy.
• 2:02
  Penny: And just to set the stage here because this context is crucial, the buzzword of the week coming straight from Nvidia CEO Jensen Huang is Egenetic AI.
• 2:12
  Roy: Egenetic AI. We've heard this term bubbling up for a while now.
• 2:15
  Penny: We have, but Hong really planted the flag this week.
• 2:19
  Roy: He did.
• 2:19
  Penny: And this is the key to understanding the numbers. We aren't talking about chatbots anymore.
• 2:24
  Roy: Right. No more just typing a prompt.
• 2:26
  Penny: Exactly. We aren't talking about typing a prompt into a box and getting a funny poem or a recipe for chili. We are talking about software that actually does things.
• 2:36
  Roy: Agents that negotiate code plan execute.
• 2:39
  Penny: And correct their own errors. And that shift from chat to agent is what is driving every single financial metric we are going to look at today.
• 2:46
  Roy: Okay, let's unpack those numbers first. Because even if the market reaction was lukewarm, the numbers themselves are, well they're staggering.
• 2:56
  Penny: They are defying gravity at this point.
• 2:58
  Roy: I feel like we're running out of superlatives for Nvidia quarters.
• 3:00
  Penny: We really are. So Nvidia reported revenue of $68,100,000,000 for the quarter. That is up 73% year over year. But the number that really matters, the one you need to circle, is the data center revenue.
• 3:13
  Roy: Which is the chips and networking for AI.
• 3:15
  Penny: Exactly. That alone was $62,300,000,000
• 3:19
  Roy: $62,000,000,000 in three months.
• 3:23
  Penny: To put that in perspective, that's larger than the annual GDP of some mid sized nations.
• 3:27
  Roy: Generated in ninety days.
• 3:29
  Penny: Essentially by selling SAN that has been taught to think.
• 3:32
  Roy: That is a crazy way to put it, but it's true. And they aren't slowing down, are
• 3:36
  Penny: they? At all. They projected $78,000,000,000 for the 2027.
• 3:42
  Roy: But the thing that really stood out to me in the AGI Roundtable report was the margins.
• 3:46
  Penny: Oh, the margins are the whole story.
• 3:48
  Roy: Because usually when you make hardware, you have massive physical costs.
• 3:51
  Penny: Right. If you're Apple or making cars or refrigerators, you pray for 40% gross margins, maybe 50% if you are incredibly efficient.
• 4:00
  Roy: Right.
• 4:00
  Penny: NVIDIA's gross margins are holding at roughly 75%.
• 4:04
  Roy: 75%. On hardware.
• 4:06
  Penny: It is a software margin on a hardware product.
• 4:09
  Roy: Which means they have absolute pricing power.
• 4:11
  Penny: Exactly. There is no competition that matters right now.
• 4:14
  Roy: Right.
• 4:14
  Penny: They are selling the oxygen for the new economy and they can charge literally whatever they want.
• 4:20
  Roy: So here's the mystery then. Usually numbers like that would send a stock to the moon. Right. We've seen it before. Nvidia Beats stock goes up 10% everyone cheers.
• 4:29
  This time the reaction was flat. Why? Is the market just tired?
• 4:35
  Penny: It's not fatigue. It's about the price of perfection.
• 4:38
  Roy: Okay. Explain that.
• 4:39
  Penny: The market is priced for absolute perfection right now. When you are valued this highly great isn't good enough. You need impossible.
• 4:47
  Roy: Right. The bar is just too high.
• 4:48
  Penny: But I think there's a deeper narrative shift happening that investors are still trying to wrap their heads around. It's the shift from training inference.
• 4:55
  Roy: Help us distinguish those two. Because to, you know, a lay person, it's all just AI chips. Right?
• 5:01
  Penny: Sure. So for the last three years, the story has been everyone needs chips to train these massive models.
• 5:06
  Roy: Like building GPT four or Gemini.
• 5:09
  Penny: Exactly. That's building the brain. It's capital intensive. It's huge. But theoretically, you only do it once per model version.
• 5:14
  Alright. But Jensen Huang said something critical on the earnings call. He said, Compute is revenue.
• 5:20
  Roy: Compute is revenue. It sounds like a marketing slogan. What does it actually mean operationally?
• 5:25
  Penny: It means we are fully moving into the inference economy. When you use an agentic AI to write code or analyze a 50 page legal contract.
• 5:35
  Roy: Or run a biological simulation for a new drug.
• 5:38
  Penny: Exactly. That takes computing power. It generates tokens, pieces of text or data. In this new economy, generating those tokens is directly tied to revenue generation for the companies using them.
• 5:50
  Roy: Okay. So it's the difference between building a factory and running the factory.
• 5:54
  Penny: Precisely. If you build a factory, you pay for the steel once.
• 5:57
  Roy: Right.
• 5:57
  Penny: But if you run the factory, you pay for the electricity every single second. Inference is the electricity.
• 6:03
  Roy: Ah, I see. So if compute is revenue, there's basically no ceiling on demand.
• 6:07
  Penny: None. As long as the AI can generate a dollar and 1¢ of value for a dollar of compute, you will keep buying chips forever.
• 6:15
  Roy: And that explains why they mentioned that even their older stuff, the six year old chips, are totally sold out.
• 6:20
  Penny: Completely sold out in the cloud. They are transitioning from the Blackwell architecture to the new Rubin chips, but they literally cannot make them fast enough.
• 6:29
  Roy: So the demand isn't just for the newest, shiniest thing.
• 6:31
  Penny: No. It's for any capacity that can run these agents.
• 6:34
  Roy: Now there was one specific red flag in the report though, or maybe a yellow flag depending on your risk tolerance.
• 6:41
  Penny: Let me guess China.
• 6:42
  Roy: The China wild card. This part is fascinating. Nvidia actually has approval to ship some limited ships the H200s to China right?
• 6:51
  Penny: Right. These are the compliant versions Kind of dumbed down a bit to meet US export controls.
• 6:56
  Roy: But Nvidia reported zero revenue from those shipments so far.
• 7:00
  Penny: Zero?
• 7:01
  Roy: Zero. Not just low but actually zero.
• 7:03
  Penny: Flat zero. It is a massive mark and the second biggest economy in the world sitting at zero in their guidance.
• 7:09
  Roy: Because of the geopolitical uncertainty and import restrictions.
• 7:12
  Penny: Exactly.
• 7:12
  Roy: And yet, and this is the crazy part, they still beat guidance by billions.
• 7:17
  Penny: That tells you how insanely strong the demand is everywhere else. They don't even need China to crush the numbers. But it also highlights the risk. If the geopolitical winds shift or if The US tightens restrictions further, that zero stays zero. It's a massive chunk of the global market that is currently just inaccessible.
• 7:35
  Roy: Okay. So the demand is there, the chips are sold out, the margins are astronomical, but here is where the AGI roundtable report pivots and frankly it gets a little scary.
• 7:45
  Penny: Yes it does.
• 7:46
  Roy: We have the chips but do we have the plugs?
• 7:48
  Penny: Yes, the physical wall.
• 7:50
  Roy: The report calls this the clash between the matrix
• 7:53
  Penny: The software, the AI, the dream.
• 7:55
  Roy: Right. And the atoms economy. The physical stuff like copper, steel, and electricity.
• 7:59
  Penny: And electricity is the hard bottleneck here. You cannot code your way out of a lack of power.
• 8:04
  Roy: No. You really can't.
• 8:05
  Penny: We are talking about a grid that needs roughly $2,000,000,000,000 in upgrades just to handle this AI boom, and we simply aren't building it fast enough.
• 8:13
  Roy: And this week, we saw what the roundtable calls political theater regarding this exact issue, the rate payer protection pledge.
• 8:21
  Penny: It sounds nice, doesn't it?
• 8:22
  Roy: It sounds wonderful. We promise not to hurt the rate payers.
• 8:25
  Penny: The idea is that the hyperscalers, so Amazon, Microsoft, Google, they pledged to bring their own generation.
• 8:31
  Roy: Basically saying, don't worry, regular citizens. We will build our own power plants so we don't crash your grid or jack up your utility rates.
• 8:38
  Penny: Right. But Roundtail analysts, specifically the one profiled as Hunter, called this a scam.
• 8:44
  Roy: That is very strong language. Why a scam?
• 8:46
  Penny: Because of the basic physics of the power grid. You can't just plug a data center into a private nuclear plant and cut the cord to the rest of the world.
• 8:54
  Roy: Because data centers require, what, 99.999% uptime?
• 8:58
  Penny: Exactly. Nuclear plants go down for maintenance. Solar obviously doesn't work at night. Wind doesn't work when it's calm.
• 9:04
  Roy: So they still need the public grid no matter what?
• 9:06
  Penny: They absolutely need the public grid for backup for frequency regulation for overall stability.
• 9:12
  Roy: So what are they actually doing then?
• 9:14
  Penny: What they are doing is buying up all the best baseload power. Look at Amazon's deal for the nuclear output from Susquehanna.
• 9:21
  Roy: Right.
• 9:22
  Penny: They took that clean, reliable, twenty four seven energy right off the public market for themselves. That is what the roundtable calls the privatization of reliability.
• 9:31
  Roy: So they take the filet mignon power?
• 9:32
  Penny: And the public gets the hamburger helper.
• 9:34
  Roy: Wow.
• 9:35
  Penny: We get what the report calls the socialization of decay. We get the older, doodier, less reliable plants to keep our lights on, and we get stuck with the massive bill for the transmission upgrades needed to support these data centers.
• 9:49
  Roy: Data centers acting essentially as parasites on the grid.
• 9:52
  Penny: Their words not mine, but the logic holds up.
• 9:55
  Roy: There is also a timeline issue here that frankly blew my mind when I read it. We just said NVIDIA chips are sold out, but they become obsolete in what, eighteen months?
• 10:04
  Penny: Right. An h 100 or a Blackwell chip is old news in two years tops.
• 10:08
  Roy: Right.
• 10:08
  Penny: You depreciate that asset incredibly fast. The tech moves at the speed of light. But a power plant Mhmm. A natural gas peaker plant, or a nuclear facility.
• 10:18
  Roy: That takes thirty years to pay off. Exactly. So the round table asks a very uncomfortable question. If this AI bubble bursts or even just slows down in five years, who pays for the remaining twenty five years on that power plant?
• 10:32
  Penny: Let me guess. It's not Microsoft.
• 10:34
  Roy: No, it is the ratepayer. Grandma in Ohio.
• 10:37
  Penny: Spot on. The tech giants will have moved on to the next efficient jurisdiction or the next big technology, and the public will be paying for a natural gas plant that was built for an AI agent that doesn't even exist anymore.
• 10:49
  Roy: That is the physical wall right there.
• 10:51
  Penny: It is. It's not just about running out of power. It's about the financial toxicity of trying to build physical infrastructure at the speed of software.
• 10:58
  Roy: And this tension between the infinite speed of AI and the slow expensive reality of the physical world is really the defining conflict of this market right now.
• 11:10
  Penny: It completely explains why boring utility companies are suddenly trading like hot tech stocks.
• 11:16
  Roy: It really does. Investors are waking up to the fact that the matrix cannot exist without the atoms.
• 11:21
  Penny: Exactly.
• 11:21
  Roy: Speaking of conflict, we have to talk about the other panic that set in this week. The susbocalypse.
• 11:27
  Penny: I really do love that term. Sus, of course, standing for Software as a Service.
• 11:31
  Roy: The subscription model we all know and, let's be honest, kind of hate.
• 11:34
  Penny: We all hate it. And Apocalypse standing for, well, the end of the world for those companies.
• 11:39
  Roy: Because Nvidia's earnings were good, but if you looked at the software stocks on Thursday, yikes. Brutal. IBM had its worst day in twenty five years, dropped about 13%. Salesforce service now, all of them taking huge hits. What exactly happened on Thursday to trigger that?
• 11:57
  Penny: It was triggered by a single blog post.
• 11:59
  Roy: Wait, a blog post wiped billions off the market cap of IBM?
• 12:03
  Penny: It did. Anthropic, the creators of the Cloud AI model, released a tool. And it wasn't even a full product launch, it was just a demonstration.
• 12:12
  Roy: Okay. What did it demonstrate?
• 12:13
  Penny: They showed an agent that can successfully modernize COBOL code.
• 12:17
  Roy: Okay. For those who aren't computer historians, remind us why COBOL matters today.
• 12:21
  Penny: COBOL is this ancient coding language that essentially runs the entire global banking system.
• 12:27
  Roy: Right.
• 12:27
  Penny: It's clunky. It's old. And the people who actually know how to write it are all retiring or already retired.
• 12:32
  Roy: So usually, you need very expensive human consultants often from IBM to maintain it or translate it for new systems.
• 12:40
  Penny: Exactly. It is a massive, reliable cash cow for IBM's consulting business.
• 12:44
  Roy: And Anthropic just casually said, Our AI can do that now.
• 12:48
  Penny: Pretty much. They showed that their AI agent can ingest the old code, understand the business logic, and rewrite it in modern languages efficiently, quickly, and incredibly cheap.
• 12:59
  Roy: And the market just did the math instantly.
• 13:01
  Penny: Instantly. If AI can do the consulting and coding work, what happens to IBM's billable hours?
• 13:08
  Roy: They just evaporate.
• 13:09
  Penny: Right. That 13% drop wasn't just a bad trading day. It was the market fundamentally re pricing the value of human labor in the software sector.
• 13:18
  Roy: Wow.
• 13:19
  Penny: It's the sudden realization that routine maintenance is no longer a viable business mode.
• 13:23
  Roy: This connects perfectly to another report mentioned in our sources the Citrini Research Note. Now this was technically presented as a thought experiment about the year 2028 but the market clearly treated it like a prophecy this week.
• 13:36
  Penny: You're talking about the concept of ghost GDP.
• 13:38
  Roy: Ghost GDP. It sounds spooky.
• 13:40
  Penny: It is a little spooky. The idea is that these AI agents, this agentic AI we keep talking about, will make corporations incredibly productive.
• 13:47
  Roy: Right. They generate huge amounts of revenue and output.
• 13:50
  Penny: So on paper, GDP goes up. The economy looks like it is booming. Yeah. But they don't need human workers to do it. So overall, employment and middle class wages actually go down while GDP goes up.
• 14:02
  Roy: So the stock market hits record highs, corporate profits are through the roof, but nobody has a job.
• 14:09
  Penny: Or at least the historical link between corporate success and human employment breaks. We've always assumed that if companies do really well, they hire more people.
• 14:17
  Roy: Right. That's the whole basis of the economy.
• 14:19
  Penny: Ghost GDP says not anymore.
• 14:21
  Roy: The Citrini report also highlighted something called friction businesses. This is a concept that I think every single listener needs to audit their own career against.
• 14:30
  Penny: Absolutely. A friction business is any company that makes money on the inefficiency of human transactions.
• 14:36
  Roy: So think about DoorDash or Visa or even a travel booking site.
• 14:40
  Penny: Right. I order a pizza, I pay a fee because I don't want to physically call the restaurant. I book a flight, I pay a fee because I don't wanna call the airline and coordinate it myself.
• 14:50
  Roy: It's a convenience tax.
• 14:51
  Penny: Exactly. Now imagine an AI agent doing that for you. An AI agent doesn't care about the pretty interface of the DoorDash app.
• 14:57
  Roy: It doesn't have brand loyalty.
• 14:59
  Penny: It doesn't get frustrated by hold times. It just ruthlessly scans every single available option on the Internet in milliseconds to find the absolute lowest price with zero friction.
• 15:11
  Roy: So the profit margin for the middleman?
• 15:13
  Penny: Disappears completely. Mhmm. If you are a middleman charging for friction, AI will optimize you existence.
• 15:18
  Roy: That is the real fear gripping the Matrix economy right now.
• 15:21
  Penny: Why pay Salesforce $300 a month for a software seat for a human sales rep if you don't actually employ the human sales rep anymore?
• 15:29
  Roy: That's the Spocalypse in a nutshell. The Pooh seat pricing model is dead if there are no seats to fill.
• 15:34
  Penny: The revenue model for the last twenty years of tech basically evaporated in the face of eugenic AI.
• 15:41
  Roy: So if software is considered dangerous right now and charging for friction is dying business model, Where is the smart money going? You mentioned earlier there was a massive rotation.
• 15:49
  Penny: Capital rotation, yes. That's when big funds move billions of dollars from one entire sector to another.
• 15:55
  Roy: Where are they putting it?
• 15:56
  Penny: They are rotating into halo assets.
• 15:58
  Roy: H A L O.
• 15:59
  Penny: Heavy assets, low obsolescence.
• 16:01
  Roy: Back to the Adams economy.
• 16:03
  Penny: Precisely. The roundtable points this out very clearly. Investors are fleeing the overvalued software stocks and moving that money into things AI absolutely cannot print.
• 16:14
  Roy: Like copper?
• 16:15
  Penny: Copper, energy infrastructure, actual physical builders, industrials, companies like Home Depot or Eaton or the big copper miners.
• 16:25
  Roy: Because you can't just type a prompt into an AI and ask it to build a high voltage transmission line.
• 16:31
  Penny: You cannot download a new power grid. You have to build it with steel and sweat and zoning permits.
• 16:36
  Roy: And that physical moat, the sheer mind numbing difficulty of doing things in the real physical world is becoming the most valuable asset in the market.
• 16:44
  Penny: Because it's the only thing that AI can't replicate instantly.
• 16:47
  Roy: It's so fascinating. It's almost like we are going back to a nineteenth century industrial boom just to power the twenty first century brain. We need coal, we need copper, we need massive tracts of land.
• 16:57
  Penny: We really are. It is a massive return to the tangible.
• 17:00
  Roy: And that actually brings us to the biggest risk of all. The one that keeps the geopolitical strategists up at night in these reports. We've talked about the companies Nvidia, Microsoft, Amazon. Let's zoom out for a second. Is it actually a good thing that three or four American tech companies basically own the brain of the global economy?
• 17:19
  Penny: The oligopoly of intelligence.
• 17:21
  Roy: Right.
• 17:21
  Penny: The roundtable cites Arthur Mench on this. He's the CEO of Mistral AI. That's the big European AI lab.
• 17:28
  Roy: And he calls this a civilization scale risk.
• 17:30
  Penny: That is heavy language.
• 17:31
  Roy: It is, but look at the sheer capital requirements. To compete in this agentic AI race, to buy the clusters that Nvidia is selling, you need a capital expenditure budget of $100,000,000,000 or more.
• 17:45
  Penny: 100,000,000,000? There are entire sovereign countries that don't have that kind of money lying around.
• 17:49
  Roy: Yeah, exactly. OpenAI is reportedly raising 100,000,000,000. Microsoft and Google are spending nearly 200,000,000,000. Who else can play at that table?
• 17:57
  Penny: Nobody.
• 17:58
  Roy: Nobody. Not the startups and frankly not even most developed nations.
• 18:01
  Penny: So if you are France or India or Brazil, you are becoming a client state. You are essentially renting intelligence from a US oligopoly.
• 18:11
  Roy: Which is terrifying from a national security standpoint. Yeah. If the intelligence layer of your economy, your banking code, your government services, your military strategy is running on a server owned by a private company in Seattle or Santa Clara? Do you really have sovereignty anymore?
• 18:26
  Penny: That is exactly what Mench is asking. And what if those companies decide to turn off the tap? Or if they bake in cultural biases that hurt your local economy.
• 18:35
  Roy: Or if the US government simply steps in and says no AI for you due to some foreign policy dispute.
• 18:40
  Penny: That is the ultimate fear. Mensch argues this inevitably leads to national AI firewalls.
• 18:45
  Roy: Like digital borders.
• 18:47
  Penny: Yes. We might see a heavily fragmented global Internet where nations ban these US models to protect their own digital sovereignty, creating a sort of splintered splinter net.
• 18:57
  Roy: It is a genuinely terrifying thought. So just to recap the landscape we are looking at on 02/26/2026.
• 19:04
  Penny: We
• 19:05
  Roy: have the Matrix economy printing absolute fortunes via Nvidia, but hitting a massive physical wall of energy scarcity.
• 19:12
  Penny: We
• 19:13
  Roy: have the traditional software economy facing a brutal SaaS bocalypse because AI agents are simply too efficient for human pricing models. And we have a looming geopolitical crisis because only three companies on earth can afford to play the game.
• 19:28
  Penny: That is the board we are playing on right now.
• 19:30
  Roy: It is a lot to process. So for the listener tuning in, trying to figure out what to do with their portfolio or honestly just their own career, what is the actionable takeaway here?
• 19:39
  Penny: I think the most crucial takeaway is to ruthlessly audit your exposure to friction.
• 19:43
  Roy: Okay.
• 19:43
  Penny: Look at your investments, look at your business model, your daily tasks. Are you a middleman? Do you rely on the inefficiencies of others? Do you charge a premium for things that an AI agent negotiate down to zero in a millisecond?
• 19:57
  Roy: Because if you are, you are standing directly in the blast zone.
• 19:59
  Penny: You really are.
• 20:00
  Roy: Yeah.
• 20:00
  Penny: But if you are a builder, if you own the physical infrastructure, the energy, the literal atoms that the AI needs to survive, you are in the driver's seat.
• 20:09
  Roy: The AI needs you more than you need it. Precisely. That is a very powerful way to frame it. I want to leave you listening with one final thought from the roundtable report. Something that has honestly been stuck in my head all day.
• 20:20
  Penny: Oh, let's hear it.
• 20:21
  Roy: It's about the everyday consumer. They suggest we are rapidly moving toward an economy where AI agents optimize absolutely everything. We talked about DoorDash earlier, but apply that to your whole life. If I have a personal AI agent that automatically negotiates my cable bill, buys my groceries, trades my stocks and books my travel, does that agent care about brand loyalty?
• 20:44
  Penny: Not unless it is specifically programmed to care. It only cares about data.
• 20:48
  Roy: Exactly. Pure efficiency, it cares about price per ounce, it cares about nutritional density or historical returns. It does not care about a multi million dollar Super Bowl commercial.
• 20:59
  Penny: Right.
• 20:59
  Roy: So if consumption becomes completely automated, ruthless, and perfectly optimized, what happens to the profit margins of every single consumer brand of the S and P 500? If Coke and Pepsi are just variables in an optimization algorithm, where does all that brand value go?
• 21:14
  Penny: That is the multi trillion dollar question. If the consumer is an algorithm, advertising as we know it is totally dead.
• 21:21
  Roy: Something to really think about while you're charging your devices tonight. Thanks for diving deep with us. We will see you next time on the deep dive.