# The Biggest Investment Cycle in History Has Already Begun | Raoul Pal The Journey Man

https://www.youtube.com/watch?v=3n2UpyKETZw

[00:00] This is the most powerful technology the world has ever seen.
[00:02] It's the greatest discovery in all of humanity and we're not going to allow it to go bust.
[00:08] We can't.
[00:08] As Lynn Alden [music] would say, nothing stops this train.
[00:10] China and the US are now locked into a race where nobody [music] can stop spending and nobody can stop accelerating.
[00:18] If one country has a recession and can't spend on intelligence and that lasts [music] 18 months, they're going to be more than 3 years behind.
[00:25] You can't be left behind in the great race.
[00:28] If you fear the robots from that AI are going to take over your jobs, [music] you might as well invest in them.
[00:33] This super cycle means we won't have a bear market or a recession or something [music] in 2027 or 28.
[00:40] Hi everyone.
[00:41] I'm Raoul Pal and welcome to my show, The Journey Man, where I travel to that nexus of understanding between macro, crypto and the exponential age of technology.
[00:51] There is a lot going on as these forces of technology that I call the exponential age.
[01:01] Meet the liquidity flow from the everything code.
[01:07] And it's a point of hyper-acceleration.
[01:09] It's also a point where people are struggling with is this all going to blow up?
[01:15] Is something going to go horribly wrong here and it's going to be the dot-com boom and bust all over again?
[01:21] Or is there a chance that this is a super cycle?
[01:24] And while I'm open to the idea that problems can form, my general view is this is likely to be a super cycle.
[01:33] And I'm going to talk you through today the reasons why using my frameworks of the everything code, uh the exponential age, the economic singularity and even the universal code.
[01:43] Hopefully, it'll give you something to think about, a new way of interpreting this and a new way of seeing the world.
[01:49] Whether it's going to be a full super cycle or not, it's definitely never a certainty, but for me, the probability is rising.
[01:55] So, let's go through a presentation that I think will help you understand where we are.
[02:04] So, where are we going?
[02:06] Well, the everything code is the world we're leaving.
[02:08] That's where liquidity papers over the cracks and the debt rollover that takes the funding mechanism.
[02:15] That's what we've been seeing.
[02:16] That's the debasement of currency, the everything code that optically pushes up the price of assets, and we found there were two asset classes over time that outperform everything else.
[02:26] One is technology, and the other's crypto, which is basically a technology.
[02:30] The world we're entering is the exponential age.
[02:33] When I first started talking about this, I don't know, in 2021 and 22, people thought I was insane.
[02:38] I'm like, everything is about to go exponential, um and everything will change forever, and that has become more and more true.
[02:46] That's the world we're standing inside now, where the output of intelligence, the economy's most important input, is bending vertically off the page.
[02:56] Again, intelligence is going exponential.
[02:57] Humanity's never gone through anything like this before.
[03:02] And where we're going is the economic singularity.
[03:05] And that's where the system has to reorganize around a new substrate of intelligence altogether, and that's new systems, new economies, everything is new.
[03:14] That's where we're going, and that's the idea behind the you have 5 years to unfuck your future.
[03:21] It's now coming to about 4 years, uh and I stand by the 2030 idea that that is likely to be happening by then, when things change dramatically.
[03:29] So, let's go to the everything code.
[03:32] Here's the formula that you all know, which is GDP growth equals population growth plus productivity growth plus debt growth.
[03:40] Those are the three inputs to it all, and that is the world that we've been living in for a very long time now.
[03:47] This is where we're leaving.
[03:49] We're not going to stop debasing in but we're going to change the mechanism by which GDP grows faster than debt and debt to GDP falls, which is what we did in the 1950s.
[03:59] That's the big bets that Scott Bessent, Walsh, Trump, they're all.
[04:05] Taking that this is the thing that's about to happen.
[04:09] But let's go through the everything code again because it's very important for you to understand this.
[04:13] Demographics are destiny.
[04:16] And once you understand that demographics are baked in the cake and they extend forward, you can see the world for what it is.
[04:26] It's incredibly useful to understand why things are happening.
[04:29] So here's the main chart that shows the births rate of the US.
[04:35] And you can see how it's been falling and falling over time.
[04:39] The trend rate of growth is collapsing.
[04:41] In most developed nations now, it's gone negative.
[04:43] So populations are starting to shrink.
[04:45] The US is slow but not as slow as others.
[04:49] But the point being is population growth tells you in the future where your labor force participation rate is.
[04:57] The economic participants in the economy, what are they likely to do?
[05:01] Well, they're likely to collapse.
[05:03] As the baby boomer cohort shrinks over time, we'll.
[05:06] Find that there's less people in the workforce.
[05:08] So that obviously leads to a slowing economy.
[05:11] And the slowing economy is what we use debt for to paper over the cracks.
[05:18] So as I mean this chart is I call the greatest chart in all of macro and again the one that most people don't understand and certainly haven't copied, which is bizarre cuz they copied most of my charts, is it's the labor force participation rate put up against debt to GDP on an inverted scale.
[05:32] They're the same thing.
[05:34] So as the labor force shrinks, the economy slows and we use debt as a way to paper over the cracks to keep growth growing.
[05:44] Now, the issue is is productivity there is the output, you know, per hour for workers.
[05:50] And the older the population gets, the less productive they get.
[05:55] So, that whole economic formula is not working.
[05:57] And productivity over the time has been falling.
[06:01] And that's again exactly in.
[06:06] Line with this aging population idea.
[06:08] And that has led to the slowing of trend rate of GDP.
[06:11] And it's been measurable.
[06:13] And what the bet is is whether we can change this picture in the future by AI intelligence, robots, and technology overall.
[06:22] So, as I said, debt is the fuel for the economy right now.
[06:25] And that has been the issue.
[06:27] So, debt to GDP keeps rising, total debt keeps rising, and the trend doesn't stop.
[06:31] It just keeps going up, and we're going back close to the final years of COVID.
[06:36] After we, you know, corrected debt to GDP for a little bit, it's going back up again cuz it can't stop because the shrinking population.
[06:45] There's no way around it.
[06:46] As Lyn Alden would say, nothing stops this train.
[06:48] You can't stop this until you grow GDP.
[06:51] And so, how do you pay for all of this debt?
[06:55] Well, it's liquidity.
[06:57] And this is the secret of the everything code.
[06:59] Is like nobody's able to default.
[07:01] You don't want to default in this.
[07:05] So,
[07:07] You fund the debt by debasing your currency.
[07:13] It's basically borrowing money on your credit card to pay off a credit card.
[07:14] You just keep extending the debt to do it.
[07:18] And that's the interest payments.
[07:19] That's where they start compounding over time as the debt grows.
[07:25] And so, somehow you have to pay for it, and you pay for it by the debasement of currency.
[07:29] So, here's this chart, which is US total liquidity.
[07:31] Um that's including bank credit and bank lending versus uh government debt as percentage of GDP.
[07:39] And you can see the relationship here.
[07:42] It's very tight.
[07:45] It's telling you that you have to increase liquidity to fund debts.
[07:50] And that's been the story of our times.
[07:52] And that's been the story that makes the rich richer, the poor poorer because earnings don't go up as much as assets.
[08:01] And the majority of the population have earnings.
[08:04] And they've got poorer because even in real terms, they've.
[08:09] Really struggled.
[08:12] But in asset terms they've lost out massively.
[08:14] Those who can own assets have increased wealth, particularly if you owned technology.
[08:19] But those who don't just get poorer every day.
[08:23] People don't understand this mechanism, but this is the mechanism of debasement.
[08:25] So the interest payments drive the liquidity.
[08:27] So the total liquidity we can look forwards because we know the size of the interest payments to come.
[08:33] It's like a 36-month lead on what liquidity has to do.
[08:38] And it's going to have to keep rising.
[08:40] And at first it used to be the central bank balance sheets.
[08:43] And then it included Fed net liquidity.
[08:46] And now it's total liquidity which includes the banking system.
[08:50] Cuz the banks and the central bank or the treasury are the only people who can create money.
[08:56] So they're creating money via loans and leverage.
[08:59] And that comes into the system.
[09:02] And that is the way of funding the interest payments.
[09:04] And this debasement or liquidity if you want the nice term.
[09:09] For it is the thing that drives every asset.
[09:13] It is the it is the movements of the denominator.
[09:18] The denominator falls because you're devaluing the currency and the scarce asset rises.
[09:22] Incomes don't rise.
[09:25] This is why price earnings keep going up is the price goes up but earnings are more consistent with GDP growth over time.
[09:33] Um but there's only two assets that beat this and that's Nasdaq and and Bitcoin or crypto.
[09:42] And they're all the same thing really.
[09:45] So, liquidity keeps going up as we're aware of for a new all-time highs in total liquidity.
[09:49] This is global total liquidity.
[09:53] And you can see the Nasdaq mirrors that very nicely.
[09:55] It's been accelerating because of what's been going on with the intelligence side of the equation.
[10:02] And Bitcoin slightly underperformed it in recent times, but generally speaking, these correlations are super high over long terms.
[10:08] They shift and morph in the.
[10:11] Short term, so they're not for trading, they're for the direction of the overall market over time.
[10:17] So, they tell you that over time as you'd base the currency, the value of Bitcoin and crypto goes up, the value of Nasdaq goes up, and they're both in secular uptrends because they're technologies driven by adoption, so they outperform anything else.
[10:34] No other assets really matter.
[10:35] And partly because these assets are based around the new intelligence layer for the global economy.
[10:42] So, now we understand liquidity is everything.
[10:43] This is the everything code and the rollover of the debt is the cyclicality.
[10:47] Now, we don't know how cyclical it is going to be going forwards because so much debt is being issued at the front end.
[10:57] So, if you issue a lot of debt at the front end, it removes a lot of the cyclicality.
[11:02] And issuing debt at the front end is stimulus in itself.
[11:07] So, that's the pattern we're seeing from the US government now is a reticence to.
[11:12] Issue debt further out the curve because rates are too high, so they keep piling in the short-term debt, which is ongoing stimulus to the economy.
[11:20] So, we get less of that big cyclicality and that's going to be very important if we're talking about a super cycle.
[11:28] But that's what I think is coming here.
[11:31] So, let's talk about the exponential age.
[11:33] [snorts]
[11:34] The exponential age is where we are now, when everything is starting to go exponential.
[11:40] And I developed a whole thesis around this called the universal code, and it has four laws, and I'll do a piece on this um soon so you have a better understanding for it.
[11:50] But, the first law, the most important thing to understand is the entire system.
[11:56] And this is the universal system.
[11:56] This is every single observable factor in the universe organizes around intelligence per unit of energy.
[12:07] And particularly the economic system um too.
[12:10] So, energy turns into compute.
[12:13] Compute becomes intelligence.
[12:15] [snorts]
[12:15] And we see this everywhere in physics, biology, markets, mycelial networks, everything.
[12:20] It's just running at different scales.
[12:22] We see it at the universal scale with the boss solar system imagery of networks.
[12:25] We see it at the smallest micro level um observable in all of physics.
[12:32] So, we're seeing this process run everywhere.
[12:35] Everything, every process is increasing intelligence per unit of energy.
[12:40] It seems to be the purpose of the universe itself.
[12:43] And when the economy starts hyper-focusing on this one metric, this is what it feels like.
[12:52] It's the exponential age.
[12:54] It's when everything starts going vertical.
[12:56] And if you think about it in this bizarre world that we live in, we're running electricity through silicon, which is the second most popular substance on Earth after carbon, and we are turning it into intelligence.
[13:15] So, all of life as we know it, and all intelligence as we know it, was really putting energy through carbon.
[13:23] And now we're putting it through silicon, and it's million times faster.
[13:30] This is why it's so shocking for us is the speed of everything happening is because of the new substrate.
[13:34] It's so much faster than a synapse of human neuron.
[13:39] Yes, human neuron can parallel process more, but it's only a matter of time when the silicon substrate can do that.
[13:47] So, we were always limited by the speed of biology.
[13:49] Every system around us is built on the system of biology speed.
[13:52] And that is now changing.
[13:55] It's going to a new escape velocity and it's going exponential.
[13:59] So, let's talk about this idea of a super cycle.
[14:01] I'm not entirely convinced we're going to see another business cycle here.
[14:08] We've got the largest capital build out in human history.
[14:10] So, we are spending an extraordinary sum of money on financing.
[14:16] This.
[14:16] And that is going to be done by a number of ways, cash flows, debts, and other stuff.
[14:23] And we used to see things like semiconductors growing, you know, in a in a cycle, a very cyclical between down 20% up 40% a year growth that held for 35 years.
[14:33] And it it's well uh when this presentation was done, it was at 79%.
[14:37] It's now over 90% year-on-year growth.
[14:40] So, we've got something entirely new that we've not seen before.
[14:45] And the cynical mind might say, "You know what?"
[14:47] This is just another malinvestment boom.
[14:52] And we're going to talk about that because I don't think that's the case.
[14:55] You see, one of the issues here is that we've got a weird situation that nobody can defect.
[15:01] Nobody can fail here because everybody has to run flat out because intelligence is compounding too fast.
[15:12] It's compounding so fast that if anybody falls behind, they can never catch up.
[15:19] So, there's no equilibrium where it can slow because somebody will use the advantage to win.
[15:24] And if somebody wins, they get monopoly.
[15:29] So, CAPEX becomes a forced move and not a choice because if you don't, you're out of the race.
[15:33] And this is the biggest race of all time, the race for intelligence.
[15:36] It's a race of nations.
[15:39] It's a race of China versus the US.
[15:41] It's the race of model companies.
[15:43] It's the race between the frontier labs and the open source.
[15:46] It's a race that nobody can stop.
[15:48] And they will put more and more money into this.
[15:54] I even get to the point where people say, well, I'm going to come on to the debts in this sector.
[15:57] But let's say one of the big players went bust.
[16:00] Let's say it was OpenAI.
[16:04] And everyone says, well, this is a disaster.
[16:05] We'll just play through your mind what would happen if OpenAI went bust.
[16:11] Somebody is going to buy all the compute and double the compute.
[16:16] And then they would be the monopolistic power because more compute equals more.
[16:20] Intelligence, which means they scale faster.
[16:24] The US government I don't think could even allow that to happen because the lab that gets it becomes so powerful.
[16:31] So, it has to be kind of auctioned and split amongst others.
[16:34] So, I I find it hard to believe it's a malinvestment phase when somebody will find the money to buy any excess capacity because it's such a strategically important move cuz no lab can fail.
[16:47] So, let's look at the size of the CAPEX boom.
[16:49] So, it's really not that big.
[16:52] It's big versus the post-war boom.
[16:55] And this is projections going out to 2030, about 23% of GDP.
[17:00] Nowhere near what China did in the 2000s or Japan did in the '80s.
[17:06] So, we've got plenty of room to run.
[17:09] It's It's a big CAPEX boom.
[17:11] It's a very big CAPEX boom in the US history.
[17:13] But it's certainly not gigantic.
[17:16] We've seen much larger booms in the past.
[17:19] And also, when we look at the returns of the
[17:21] Cycle, we're early.
[17:24] So, so far, I've taken it from 2020, which is really the start of the of the exponential age.
[17:29] We could take it from 2022.
[17:32] It's even earlier.
[17:34] And the Nasdaq's done like 3x.
[17:37] The Roaring 20s did 4x in 7 years, but the '80s and 2000s did a 7x, the post-war boom is 6x.
[17:45] Um the industrial age was a 24x over an extended period of time.
[17:47] So, we're still early in this if it's a if it if it is a proper super cycle, which I think it is.
[17:57] I don't think it can't be more important than the industrial age or the post-war boom or the tech globalization boom.
[18:02] This is the most powerful technology the world has ever seen.
[18:08] It's the greatest discovery in all of humanity and will be the last discovery for all of humanity.
[18:13] And we're just going to think we're just going to allow it to go bust?
[18:17] We can't.
[18:19] What people are looking at is this.
[18:20] Is the
[18:22] Hyperscalers were pouring capital into this because they made more money than any other firms in existence.
[18:31] And then suddenly in 2026, the switch flipped.
[18:35] And now they start issuing debt because they want to go faster because of the story I told you before.
[18:40] Speed is everything.
[18:42] He who gets the most compute wins.
[18:45] So, what we're seeing is this hyperscaler race for compute.
[18:51] And they now spend more than their cash flow.
[18:55] And that sounds like it could be catastrophic.
[18:58] But really, these companies, they're extraordinary cash machines.
[19:00] And they can finance a lot.
[19:04] And over time, what you find is their investments in CapEx tend to lead to more profits.
[19:14] And when we look at the leverage of these firms, I'm always take the dumb idea of leverage, which is I just look at this a percentage of market cap.
[19:21] And percentage of market cap, these.
[19:23] Things are only like 4.5%.
[19:26] The telcos back in the 2000s were like 25%, 30% of market cap in debt.
[19:33] Uh with many of them at 100% and here we've got nothing.
[19:36] As a proportion of cash flows, it's nothing.
[19:38] It's just It's so low debt and these guys have such an ability to service debt.
[19:43] Now, even if the advertising market changes and some of their business model changes, the likelihood is that they find new sources of revenue, which is whether it's leasing out the compute or selling intelligence in different formats.
[19:58] So, I think that the game is still self-sustaining.
[19:59] Will they all sustain?
[20:02] I don't know.
[20:03] But over time, I think there's plenty of cash to do this.
[20:08] So, as I said, they built it with cash flow first.
[20:10] That was the first phase of this and now it's debt.
[20:15] Now, the debt era era is very interesting because debt, if we go back to our US total liquidity, is an increase in liquidity.
[20:23] It's a
[20:24] creation of money.
[20:27] So, we're creating more money
[20:30] into a CapEx boom that cannot stop. So,
[20:34] you're now adding liquidity
[20:36] into a CapEx boom, which increases GDP
[20:39] growth.
[20:41] And it's very hard to see that stopping.
[20:43] Meanwhile, the government is issuing at
[20:46] the short end
[20:47] to stamp out the the um
[20:50] cyclicality there, whilst
[20:54] the everything code showed you that
[20:56] we've got all of the payments to come.
[20:59] So, ongoing, we're starting to build
[21:03] a backdrop of nothing can stop this now.
[21:06] Now, when we look at total debt to GDP,
[21:09] can the economy take much more?
[21:11] And the answer is yes.
[21:13] We broke the economy twice on debt. Once
[21:16] was in the 1929, and once was in 2008.
[21:19] And it's roughly when it gets over,
[21:21] let's say, 165%
[21:23] of GDP.
[21:24] We're currently total debt in the US
[21:26] about 140%
[21:28] because we've been deleveraging. Now,
[21:30] the changes in things like the ESLR and
[21:34] the bank mandates that are coming force
[21:36] the banks to lever up, to create money,
[21:39] so that the public sector doesn't do it.
[21:42] So, as or it doesn't do it as much.
[21:44] So, they need the bank balance sheets in
[21:46] this, and they've got plenty of room to
[21:48] expand total non-financial debt to GDP.
[21:52] So, when we look out in the future, we
[21:55] have the ability
[21:58] to add another 24 trillion of debt to
[22:01] 2030 before we would hit that upper
[22:05] envelope. And that what it does is
[22:08] assumes about 6% nominal GDP growth,
[22:11] which let's say is
[22:13] 3 or 4% 3% GDP and 3% 2%
[22:19] productivity, 1% inflation, whatever the
[22:20] number is. I think that's probably low.
[22:23] The faster GDP grows, the faster you can
[22:25] actually uh increase the amount of debt.
[22:28] Um I think GDP actually grows far faster
[22:30] than this, but I'm trying to be
[22:32] conservative.
[22:33] So, you've got huge amounts of debt that
[22:35] is able to be issued by the economy. The
[22:37] economy can manage higher levels of
[22:40] debt.
[22:41] But, what that means is we have the
[22:43] ability to triple the pace of debt
[22:46] growth that we've seen, and finance by a
[22:49] lot of it the hyperscaler revenues and
[22:52] others. The size of this
[22:55] new cash into the economy
[22:58] is more than a Marshall Plan every 6
[23:00] months. And it's all being rooted into
[23:03] one substrate, intelligence. It's an
[23:06] extraordinary concentration of capital
[23:09] into the biggest thing of all time.
[23:14] And we've already committed
[23:17] to the interest payments of 1.2
[23:19] trillion.
[23:20] So, financial conditions cannot be
[23:23] allowed to tighten
[23:25] because we've got to pay this. And if
[23:26] financial conditions tighten,
[23:29] interest rates go up.
[23:31] And this number keeps going up. And we
[23:33] have to pay for it by the basement or
[23:35] liquidity.
[23:37] So, the sovereign side is committed to
[23:39] this. The corporate side is committed to
[23:41] this. Everybody is committed to adding
[23:44] more liquidity and spending money in the
[23:46] economy. Everything
[23:49] is now based on this.
[23:50] So, a quick break in your regular
[23:52] programming. If you're serious about
[23:54] your future, grab my free report called
[23:57] Prepare for 2030.
[23:59] I think you've got 5 years to make as
[24:00] much money as possible. And this guide
[24:02] will help you navigate what's coming.
[24:04] The link is in the description. Download
[24:06] it now.
[24:08] So, just to recap,
[24:10] the debt the Treasury
[24:12] uses short-dated bills.
[24:15] That goes into the banking system and
[24:17] elsewhere, and it's a cheaper way of
[24:20] issuing and it means that there is less
[24:22] cyclicality. But, it is an injection of
[24:26] liquidity.
[24:27] The banks, well, over time, they absorb
[24:29] more of the Treasuries.
[24:32] And then they create lending on the back
[24:34] of it. They use it as collateral.
[24:37] So, we will see as the yield curve
[24:38] begins to resteepen,
[24:41] the banks picking up their ownership of
[24:42] Treasuries, which helps the Treasury out
[24:45] who have a lot of these bloody things.
[24:47] And then the other would be the Chinese,
[24:49] the Japanese, and others.
[24:51] And the Fed,
[24:53] now their main job is to get out of the
[24:55] way.
[24:56] Is to
[24:58] interest rates below
[25:00] GDP growth
[25:01] so they've got financial repression
[25:04] and then get regulatory clearance for
[25:06] the banks to absorb massive amounts of
[25:07] debt knowing that the economy can can
[25:10] take this
[25:11] and that's what the Fed are going to do.
[25:13] I think Wash cuts rates a couple of
[25:16] times and steps out of the way just like
[25:18] Greenspan did. Greenspan played the same
[25:21] playbook which is let productivity lead,
[25:23] let growth lead. That's the bet. If that
[25:26] happens, all of this works and the
[25:28] system is coalescing around this very
[25:30] big idea that the exponential age is
[25:33] going to drive productivity. If it
[25:34] drives productivity, then it drives GDP
[25:37] growth and lowers inflation. That lowers
[25:40] the interest rates
[25:42] likely on the long end. It certainly has
[25:44] probably a steeper yield curve that
[25:46] allows the banks to lend. That allows
[25:49] the financing of the massive capex boom.
[25:53] And it also allows for the Treasury to
[25:55] roll its debts. So it's all connected
[25:58] and it's all part of the financing
[26:00] regime.
[26:02] And again, as we've said before, is this
[26:04] payment of interest payments
[26:07] it leads by 3 years and we have to keep
[26:09] going. We have to keep going. So we I
[26:12] don't see any path right now of which it
[26:15] slows down.
[26:17] And this is the reasons I'm starting to
[26:18] believe it's a super cycle.
[26:21] So again
[26:23] the federal rollover
[26:26] everything is committed in advance
[26:28] to the amount of liquidity that needs to
[26:30] come into the system.
[26:32] The hyperscaler buildout is committed in
[26:34] advance. They've already started. What's
[26:36] interesting
[26:38] is the hyperscalers can't break ground
[26:41] fast enough on the data centers or get
[26:43] permitting fast enough. So there are
[26:46] bottlenecks that happen within this
[26:48] cycle. Getting linked up to the power
[26:50] grid, all of that stuff. There will be
[26:52] definitely rotations.
[26:55] And you'll have rotations because things
[26:56] like if they haven't built the data
[26:58] centers fast enough, then they'll stop
[27:01] ordering chips for a bit, and then that
[27:03] will slow the expenditure so they can
[27:06] then fund more of the CapEx and then
[27:09] order the chips once the data centers
[27:10] are closer. Cuz if not, you've got chips
[27:13] too far in advance. So, there's going to
[27:15] be a balance, and I think, you know, at
[27:16] some point the semis will correct and
[27:18] will do what we did in the 2000 in the
[27:20] in the late '90s into 2000 where we had
[27:24] many massive rotations, and the semis
[27:27] were just part of it, even though they
[27:28] were very large part of the internet
[27:30] boom.
[27:32] So, everybody is coalescing around the
[27:34] same thing, this whole compute cycle,
[27:37] and the liquidity is there all of it.
[27:40] Also, let's go forwards to revenues.
[27:44] Because people are worried about CapEx
[27:46] not producing revenues, but it looks
[27:48] pretty good. And here's the cloud
[27:50] revenues from the big cloud providers,
[27:53] and here's their CapEx. So, CapEx is
[27:55] going bananas, you know,
[27:57] growing massively to what, 80% a year,
[28:01] and then
[28:02] what we're seeing is revenues are
[28:04] growing with them.
[28:05] With a lag, 18 months, and that makes
[28:07] sense.
[28:09] So, the likelihood is that CapEx put in
[28:11] the ground now will have returns later.
[28:14] And I know some people don't want to
[28:16] believe that's the case. They want to
[28:17] believe it's malinvestment. Of course,
[28:18] there's going to be some malinvestment
[28:19] somewhere, but at generalistic level,
[28:22] that's unlikely to happen. And what this
[28:24] is producing is something truly
[28:26] extraordinary. It's never happened
[28:27] before.
[28:28] Look at this chart of Anthropic's run
[28:30] rate of revenue.
[28:32] It is the fastest ramp of revenue
[28:35] ever recorded. It's a double
[28:37] exponential. It actually reads law in
[28:39] revenue.
[28:41] So, that tells you there is demand
[28:44] for intelligence.
[28:47] And the intelligence
[28:49] is going to compound into profits.
[28:51] And that's how I see it.
[28:53] We're also seeing this extraordinary
[28:56] Jevons paradox playing out
[28:58] where token volumes
[29:00] are
[29:01] going up massively. So, I everybody
[29:04] starting to use tokens. We've barely
[29:06] started the process.
[29:07] We've barely started the process of
[29:10] AI use across every single substrate,
[29:13] you know, from your fridge to your
[29:14] computer to your watch to everything.
[29:16] It's all coming. So, token volume keeps
[29:19] exploding as
[29:21] token costs keep imploding.
[29:24] And so, what we've got is Jevons
[29:25] paradox. The lower the cost goes, the
[29:27] further up it goes. And when you add in
[29:30] the open source models and the
[29:31] competition, they force the price down.
[29:35] So, what you've got is a massive scaling
[29:38] of
[29:39] intelligence across every level. And
[29:41] we've really only a couple of years into
[29:43] this. So, it's going to go wild.
[29:46] And that's because this intelligence of
[29:48] unit per energy
[29:50] ends up becoming Reed's law. It's the
[29:52] first time Reed's law, which is
[29:54] Metcalfe's law squared, has ever been
[29:58] observable. It's not observable in
[29:59] biology. It's not observable anywhere.
[30:02] But now we're seeing it. We're seeing
[30:04] that
[30:06] the amount of intelligence that comes
[30:08] out of a silicon chip
[30:10] is now
[30:11] growing at such a rate, it's doubling
[30:14] every 6 and 1/2 months.
[30:16] I wanted to try and figure out
[30:19] how to measure intelligence per unit of
[30:21] energy.
[30:22] And through and to see if that if it
[30:25] does observe Moore's law. So, I did a
[30:27] lot of work on this and I got to
[30:31] using compute capability, cost, energy,
[30:34] and adoption, we put it all together
[30:36] into one index to show it.
[30:39] And this is the exponential age index.
[30:43] And what we found is intelligence was
[30:46] scaling at Moore's law.
[30:48] The famous Moore's law, which is the
[30:49] blue line.
[30:51] And somewhere around 2020
[30:54] as intelligence started to scale,
[30:57] we basically saw something different.
[31:00] And this has been this bending up. Now,
[31:02] remember, this is a log scale. So, we've
[31:05] got a double exponential here.
[31:09] And that double exponential is only
[31:11] going to keep going. As we get
[31:14] self-recursive learning, we get new
[31:16] improvements in in um
[31:18] in chips, and we get more efficiency in
[31:22] energy.
[31:23] Those things start driving this, and
[31:26] they keep going. As the AI builds AI, it
[31:29] just keeps going. So, we've got an
[31:31] extraordinary change coming here in what
[31:34] in how the world works.
[31:36] Here's the component parts of it. So,
[31:38] you've got the the headline index
[31:39] itself, the economic cost of
[31:41] intelligence, um the per dollar um um
[31:45] capital per dollar of hardware, the
[31:47] physical the per unit of energy. So, you
[31:50] can see the energy curve, which is the
[31:51] top one, is really amazing. So, we're
[31:54] getting a lot of intelligence out per
[31:56] unit of energy because the cost of
[31:58] energy is collapsing according to
[32:00] Wright's law. So, we've got Wright's law
[32:02] in play with Metcalfe's law, and that is
[32:04] what's creating the double exponential.
[32:06] As the cost is collapsing, remember we
[32:08] had this chart,
[32:10] the cost of the tokens are collapsing,
[32:11] the volume of the tokens are collapsing,
[32:13] the cost of electricity is collapsing,
[32:15] energy generation is collapsing, uh the
[32:17] efficiency of producing tokens is
[32:18] collapsing, and the whole thing just
[32:20] keeps going exponentially. It also plays
[32:22] into markets, obviously. So, this
[32:24] exponential age index, I think it
[32:27] probably lags the socks, the semis.
[32:30] And if you think about it, makes sense.
[32:32] Semis are the pointy end of the stick,
[32:33] the first thing that drives this whole
[32:36] equation. Somebody has to order some new
[32:39] silicon because out of the silicon you
[32:41] create intelligence.
[32:43] So, there is a lag here
[32:45] and it's a lag of about a year.
[32:47] And then after a year, what happens is
[32:50] intelligence starts to pick up as those
[32:51] chips are being used. And we're going to
[32:53] see a massive rise in this. So, this is
[32:55] the year-on-year
[32:57] uh rate of change about 200%
[32:59] uh in intelligence right now
[33:01] um and it looks like we're going to see
[33:03] a sharp acceleration in the future as
[33:05] all this new compute comes on. Whether
[33:07] it's this year or next year or the year
[33:09] after, it'll keep going. So, we've got
[33:12] an acceleration point that looks like
[33:13] it's on its way.
[33:15] And that would tell you that all of the
[33:17] new data centers, all of the compute
[33:20] massively increases the rate of
[33:21] intelligence coming out. And then the
[33:23] energy [snorts] efficiency of it uh
[33:25] overall
[33:27] adds to the whole equation.
[33:31] But this whole kind of idea that nobody
[33:34] can win, nobody can lose
[33:36] is actually so big
[33:40] that is the focus of the great race.
[33:43] This is the race between
[33:45] the largest, most powerful nations on
[33:48] Earth and they can't lose it either.
[33:51] You see, China and the US are now locked
[33:53] into a race where nobody can stop
[33:55] spending and nobody can stop
[33:57] accelerating.
[33:58] If the rate of intelligence is doubling
[34:00] every 6 months-ish,
[34:03] then if one country has a recession
[34:07] and can't spend on intelligence
[34:10] and that lasts 18 months, they're going
[34:13] to be more than 3 years behind in a
[34:16] race.
[34:17] And so, that means the leader becomes so
[34:20] powerful
[34:21] that as AGI moves towards ASI,
[34:25] one system rules all.
[34:28] Nobody wants one system, whether it's
[34:29] Chinese or United States, to dominate
[34:33] everybody else's culture, economics, and
[34:36] everything. Diversity is key in all of
[34:38] this. You need the tension of the two.
[34:41] But you see neither can slow down here.
[34:44] And they both know it. It's one of the
[34:46] reasons why China and the US really
[34:49] wanted the Iran situation resolved. They
[34:52] want a lot of these friction points gone
[34:54] because they don't want to worry about
[34:55] the energy cuz the energy is just the
[34:57] input. The output is the intelligence.
[35:00] And that is the race itself. That is why
[35:02] Taiwan matters. That's why the US has
[35:04] built TMC plants
[35:06] in Arizona
[35:08] as we split apart our claims over these
[35:11] things. That's where the rare earth
[35:13] metals agreements have to come from. So
[35:15] everybody is in this race and this race
[35:17] cannot stop.
[35:18] And things like wars,
[35:21] they become
[35:23] too capital intensive when you need all
[35:25] of the capital for this.
[35:27] It's a fascinating time. China is
[35:30] winning its part of the race by
[35:33] massively lowering the cost of of energy
[35:36] creation via its enormous solar ramping
[35:39] and other and other means. It's also
[35:41] producing open-source models that are
[35:44] highly efficient. Because of this, the
[35:46] US is in front of intelligence itself.
[35:49] Frontier intelligence comes out of the
[35:51] US.
[35:52] So that is the race that's going on.
[35:55] And the race just keeps going.
[35:58] And this is the chart that I think will
[35:59] play out like Moore's law did and like
[36:02] Kurzweil's chart that intelligence my
[36:05] exponential age index will just keep
[36:07] following the trend all the way out. Now
[36:10] because it's a double exponential,
[36:12] on the on the uh chart now, it's about
[36:14] at half a million. And that entire thing
[36:17] by 2040 gets to I don't know, 75
[36:21] trillion?
[36:22] So the acceleration in intelligence is
[36:24] beyond anything we understand. And
[36:27] around here is where Ray Kurzweil says
[36:29] we have the singularity, where we merge
[36:31] with the machines. This is the time of
[36:33] ASI, and it's not far away, and you can
[36:35] see the path there.
[36:38] But let's look about the US-China race.
[36:40] The US-China race
[36:42] um it's kind of neck and neck. You can
[36:44] see back in the early early '20, '20,
[36:47] '21, '22, '23,
[36:49] it was the US was well in the lead,
[36:51] China was well behind.
[36:53] And then suddenly China started
[36:54] accelerating as they came into this
[36:57] race. And now the gap has narrowed to
[37:00] 0.2 months using my index of of how to
[37:05] measure this. So, it's a it's a much
[37:06] broader measure than who's got the best
[37:08] model, how fast it is, how cheap or it
[37:10] is, but the overall amount of
[37:12] intelligence per unit of energy being
[37:14] produced by the two largest, most
[37:17] powerful nations in the world is roughly
[37:19] equal. And they've done it in different
[37:21] parts. And this race is going to keep
[37:23] going. Right now, China's accelerating,
[37:25] and the US has to figure out the energy
[37:27] side of the equation to start
[37:28] reaccelerating there because the
[37:30] frontier models get copied and and built
[37:33] into the open-source models very fast.
[37:35] So, anyway, that's the race.
[37:37] And then there's a set of dominoes that
[37:39] play around it, which is where it gets
[37:40] really interesting. So, we know that the
[37:42] semis lead the intelligence index.
[37:46] And that leads capex
[37:49] cuz as intelligence starts scaling, they
[37:50] have more confidence in capex, and we're
[37:53] starting to see capex, and then
[37:54] following it is productivity, which is
[37:57] really what we need to see moving. Now,
[37:59] here we've got total factor
[38:00] productivity, uh which is a much slower
[38:03] um measure, and this is like a 5-year
[38:04] average as well, and it's the domino
[38:07] that's not really started moving. But
[38:09] that's the one I think moves, and that's
[38:11] the one that everybody needs to move
[38:13] because if we increase productivity,
[38:15] we increase GDP growth, we lower
[38:17] inflation, and this whole game continues
[38:19] to hyper-accelerate. And that leads us
[38:21] into the economic singularity. That's
[38:24] when all of this comes together and its
[38:26] outcomes are as expected.
[38:30] So, that's GDP growth
[38:33] now, under the new magic formula,
[38:36] is humans and robots plus AI
[38:39] plus productivity and debt plus energy
[38:41] density and um
[38:43] compute efficiency. Those are the sort
[38:46] of productivity outputs of the
[38:49] exponential age.
[38:52] And
[38:53] this is where things change cuz we can
[38:55] now put infinite workers into the
[38:57] population.
[38:59] The robots and the AI and the AI in
[39:02] robots that people are forgetting about,
[39:04] we're going to put AGI brains in
[39:07] physical robots and they're going to be
[39:09] essentially a new species which people
[39:11] aren't ready for, but that's coming. But
[39:13] even just the rise of the
[39:16] agents
[39:17] um is enough
[39:19] to start generating GDP growth and the
[39:22] robots will do more. Obviously,
[39:24] productivity of an aging population not
[39:26] great, but the productivity of these new
[39:27] populations very good. Energy density,
[39:30] this is the big race where you have to
[39:32] lower the cost of energy. This is the
[39:33] race to solar. This is the race to
[39:35] nuclear um and and this is the race to
[39:39] lower uh fossil fuel costs uh by freeing
[39:42] up all of the trade structures, whether
[39:45] it is uh the friction of Iran or the
[39:47] friction of Venezuela, whatever it may
[39:48] be, everything gets cleared because this
[39:51] is the race and computer efficiency does
[39:53] the rest.
[39:55] And that is how we solve the world's
[39:57] problems.
[39:58] Because what we do is we accelerate GDP
[40:01] growth that's a GDP collapse is like it
[40:02] did in the 1950s.
[40:06] But what does the system look like on
[40:08] the other side?
[40:10] So, once capex starts growing, uh not
[40:12] capex, once productivity starts growing,
[40:16] we've got our answer. So, here is the
[40:18] BLS measure of productivity. Now,
[40:20] Washer said he doesn't like the measures
[40:21] of productivity cuz they're not very
[40:23] good. I agree. And they will find
[40:26] another one, and they will show how
[40:28] compute is
[40:29] driving productivity. But, here we've
[40:31] got the total factor productivity the
[40:33] funnel model, and we've got the BLS
[40:36] model. The BLS is turning up. Let's see
[40:38] it accelerate. If it accelerates through
[40:41] sort of
[40:42] 3 4% then we're into the big thing.
[40:45] So, we're waiting for the final
[40:47] confirmation that productivity is going
[40:48] to pick up. If productivity doesn't pick
[40:51] up,
[40:52] then this game
[40:53] of intelligence would be a disaster.
[40:58] But, I cannot see a world where it
[41:00] doesn't increase productivity. Much like
[41:02] the internet, and it should be much
[41:03] larger than the internet. The internet
[41:05] did. Productivity was super strong in
[41:07] the late '90s, and it should be even
[41:09] stronger this time around. But, there's
[41:11] a lag. Things take time. We've only just
[41:13] started this game. So, I think we'll see
[41:15] productivity accelerating, and
[41:17] productivity accelerating brings rates
[41:19] lower over time, increases GDP growth,
[41:23] allows for ongoing financial repression
[41:25] where GDP growth grows faster than
[41:27] interest rates. It funds the whole
[41:29] system. We have enough earnings from the
[41:32] corporations to drive this. Now, we know
[41:35] it's not going to be completely
[41:36] unilaterally
[41:38] positive, obviously,
[41:41] because we've got whole frictions and
[41:43] sections of the market where people of
[41:46] the economy where people will not
[41:48] benefit as easily, but others will. But,
[41:50] over time the bet is
[41:53] this is the largest deflationary force
[41:55] of all time. And if it is,
[41:58] and it becomes vastly deflationary, then
[42:00] finally people's real earnings will will
[42:04] be growing. And people will feel richer,
[42:06] and we've not had that for a very long
[42:08] time, and we desperately we it. And this
[42:11] is the only answer I can come up with.
[42:13] Subsidy payments and all of that stuff
[42:15] doesn't work in the same way as
[42:17] generating a massive rise in
[42:19] productivity that collapses inflation to
[42:23] disinflation where we have abundance in
[42:25] goods and where earnings cover all the
[42:28] needs that you have.
[42:29] Sounds optimistic. The path there
[42:32] is not easy.
[42:33] Path there is political. The path there
[42:34] is sectors of the economy slowing down.
[42:36] The path there is whole parts of the
[42:39] economy not wanting to accelerate, other
[42:41] parts wanting to accelerate. The
[42:42] political battle that means what it
[42:44] means for people, how they feel
[42:46] disenfranchised or enfranchised. It's a
[42:49] complicated world. But it should work.
[42:53] So let's go back and talk about this
[42:55] again, the super cycle. I don't think
[42:57] there's a bust. This is a phase change.
[42:59] Debt it now is the new mechanism to fund
[43:02] compute.
[43:03] Debt servicing ability in the private
[43:05] sector is the largest it's ever been.
[43:09] So they will use debt.
[43:11] That is
[43:13] an increase in liquidity. The government
[43:15] continues their
[43:18] liquidity injections.
[43:19] The compute that gets built becomes
[43:22] productivity.
[43:24] And the productivity allows the economy
[43:25] to outrun the debt.
[43:28] Rail, electricity, and fiber all ended
[43:30] the same way.
[43:31] The debt compounded faster than revenue.
[43:34] So we don't want that to happen.
[43:37] But we're going to have to see.
[43:39] Because of the speed of this
[43:42] and the speed of the double exponentials
[43:44] the asset of intelligence and compute
[43:48] should compound faster than the interest
[43:51] payments. That's the big bet here. It's
[43:53] very different than the other boom bust.
[43:57] Now, these were they were real booms and
[43:59] they lasted a long time. There were some
[44:01] busts, too.
[44:03] But that technology didn't go away. Just
[44:05] some of the players did.
[44:07] And we had a phase change.
[44:09] So, I think over time
[44:11] this is not going to be a bust. There is
[44:13] a probability there are sectors that may
[44:15] suffer, but overall the game is too big
[44:17] for anybody to let it fail. No
[44:19] nation-state can allow it, and no firm
[44:22] can allow it. There's plenty of room for
[44:24] it to be funded via debt growth. The
[44:26] government is doing its part, and
[44:28] productivity will do the rest. And that
[44:31] is where we're going.
[44:33] And then eventually we get to the point
[44:35] where money stops being able to measure
[44:37] what the new system produces.
[44:40] Because we won't even see this economy.
[44:42] It's an invisible economy of agents and
[44:44] robots that we're not even part of half
[44:46] the time. How do we benefit? Well, we're
[44:48] going to have to figure that out. All I
[44:50] know is we have to lean into it. We have
[44:52] to lean into this because the entire
[44:55] world is organizing itself around
[44:58] intelligence, and everything is
[45:00] downstream of it. All of our investing
[45:03] is this. All of our time. We all need to
[45:06] go up the curve. You can't be left
[45:08] behind in the great race because you
[45:09] will be truly left behind unless you
[45:11] choose to opt out of it, which is okay,
[45:13] too.
[45:14] But everything is happening at the
[45:16] fastest speed. It's hard for us to keep
[45:18] up. It's hard for us to even focus on
[45:21] the amount of changes that we see every
[45:23] day right now.
[45:24] And again, it's only going to get
[45:25] faster.
[45:26] So, the the way to deal with it is be
[45:29] part of it.
[45:30] Be part of it. Invest in it.
[45:33] If I mean, at the dumbest level, if you
[45:36] fear the robots and the AI are going to
[45:38] take over your jobs, you might as well
[45:40] invest in them.
[45:42] Because if not, what are you going to
[45:44] do?
[45:45] Angrily shake your fist at the sky
[45:48] while your job gets replaced and you
[45:50] haven't made a plan?
[45:52] And that's been the plan for me, and
[45:53] I've told about this for a long time, is
[45:56] unfucking your future,
[45:57] you know, and how few you we got four or
[46:00] five years. Now, there's plenty of
[46:02] returns to be had. Yes, crypto has been
[46:04] slower in the last couple of years.
[46:06] Bitcoin still outperformed the Nasdaq
[46:07] since the liquidity low in 2022.
[46:11] And it will have its time in the
[46:13] rotation again.
[46:15] But overall, this is the bet. It's the
[46:17] only bet there is. This is the big one.
[46:19] And we need to pay attention because
[46:21] this super cycle means we won't have a
[46:24] bear market
[46:27] or yeah, a recession or something in
[46:29] 2026 or 27 or 28. It just can't happen.
[46:33] The expenses
[46:34] um of the capex is just too big for that
[46:38] to stop. The amount of liquidity is too
[46:39] big that to happen. Yes, we will have
[46:41] market corrections. Yes, we will see the
[46:44] semis down 50% at some point, maybe
[46:46] trading sideways for a year, but other
[46:48] parts will be hyper accelerating because
[46:51] all parts of the intelligence curve take
[46:53] off one after the other and keep
[46:54] rotating through as intelligence powers
[46:57] through and the energy efficiency powers
[46:59] through the economy. So, anyway, good
[47:02] luck out there.
[47:03] Hopefully, I've given you something to
[47:04] think about. There's a lot more on this
[47:06] kind of stuff that I'll release um so
[47:08] you can understand a bit more of my
[47:09] thinking. Um but hopefully, that's
[47:11] enough to get you going. See you next
[47:12] time.
[47:13] So, you obviously liked this video
[47:15] enough that you've got to the end.
[47:17] That's quite a big task. But listen, do
[47:19] me a favor, hit the like and subscribe
[47:21] button and also check out what video's
[47:23] next cuz I think you'll love it. But if
[47:24] you want even more, and when I'm talking
[47:26] more, I'm talking about member-generated
[47:29] ideas, incredible alpha research,
[47:31] everything there to help you in your
[47:33] journey, just head to
[47:34] realvision.com/join
[47:36] for the best financial intelligence out
[47:38] there and the pure alpha that's within
[47:40] the platform.