---
title: "Bridgewater CIO Interview: AI and Modern Mercantilism Ignite Global 'Resource Grab,' Era of Compute Rationing Is Coming"
type: "News"
locale: "en"
url: "https://longbridge.com/en/news/286860545.md"
description: "Bridgewater Associates' Chief Investment Officer warns that modern mercantilism and AI are driving the largest capital expenditure wave since World War II. The software industry has reached its 'Barnes & Noble moment,' and severe compute shortages are spawning a 'rationing system'—those who fail to secure compute will be completely eliminated in a harsh winner-take-all game"
datetime: "2026-05-19T05:56:32.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/286860545.md)
  - [en](https://longbridge.com/en/news/286860545.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/286860545.md)
---

# Bridgewater CIO Interview: AI and Modern Mercantilism Ignite Global 'Resource Grab,' Era of Compute Rationing Is Coming

The global macro environment is undergoing a dramatic reshaping driven by two powerful forces.

In a recent interview, Greg Jensen, Co-Chief Investment Officer of Bridgewater Associates, pointed out that **"modern mercantilism" and "artificial intelligence (AI)" are deeply intertwined, jointly accelerating a global "resource grab."**

Bridgewater believes that AI is experiencing the largest capital expenditure expansion since World War II. In this unprecedented technological disruption, the software industry has officially entered its "Barnes & Noble moment," while severe supply shortages of computing power are giving rise to a "compute rationing system." This will further intensify the "winner-take-all" monopoly dynamics in the global business world.

## **Key Takeaways**

> **Intertwined Dual Forces:** The geopolitical "choke point" games of modern mercantilism are deeply combined with AI's extreme hunger for resources, weakening global productivity and pushing up inflation.
> 
> **Largest Capex Surge Since WWII:** The macro baseline expectation (modal outcome) remains high profits, high growth, high inflation, and rising interest rates. However, the unpredictable "transactional statecraft" of the Trump administration has significantly increased the risk of escalating conflicts.
> 
> **Software Industry's "Barnes & Noble Moment":** The market is repricing for "AI eating the world." Valuation premiums in the software industry are being rewritten, but this is just the beginning of disruptions across many industries.
> 
> **Compute Rationing Intensifies "Winner-Take-All":** The growth rate of compute demand has surpassed the massive growth rate of supply. By the end of 2026, the supply-demand gap will be unimaginably large, and companies unable to access cutting-edge compute resources will face structural disadvantages.

## "Choke Point" Confrontation Under Modern Mercantilism and Shifting Alliances

In the dialogue, Bridgewater researcher Jim Haskel first raised the question of how to assess the two major forces of current geopolitics and technological disruption.

Greg Jensen pointed out that as early as November 2024, Bridgewater had stated in its Daily Observations report that "We're all mercantilists now." **Greg Jensen stated,** "In this new landscape, trade deficits are viewed as pure transfers of wealth, rather than part of a cooperative trading system.

Entering 2026, from resource grabs in Greenland to games over commodity channels in Venezuela, and the evolving situation in Iran and the Strait of Hormuz, **the trend of "weaponizing key choke points" has significantly intensified.**

While this series of "chaotic and unpredictable" policies may serve as negotiating leverage for the US in the short term, they also bring huge risks of miscalculation and conflict escalation. Bridgewater warns that as large amounts of capital concentrate in the US, overseas investors in Europe and Canada have begun to question: **Is it still wise to bet the vast majority of asset portfolios on US equity assets?**

## Largest Capex Expansion Since WWII: How to Build a Portfolio to Handle the "Modal Outcome"?

The superposition of geopolitical conflicts and the AI boom has pushed the world into the largest **capital expenditure (Capex) expansion period** since World War II.

From a systematic investment process perspective, Bridgewater has a clear depiction of the macro chain reactions of this capital expenditure boom (i.e., the baseline "modal outcome"):

> **Short-term High Profits and Growth:** Companies focused on capital expenditure see their outflows immediately converted into profits for other supply chain companies in the short term, before becoming costs for other enterprises.
> 
> **Pushing Up Inflation and Interest Rates:** Due to the extremely high capital demand for building infrastructure across society, this typically drives up inflation and creates sustained upward pressure on interest rates.
> 
> **Currency Appreciation for Leaders:** Countries dominating in capital expenditure and technological explosions often see their currencies appreciate.

However, Greg Jensen emphasized that while this is the most likely base case, **its probability may have dropped below 50%**. This is because fiscal policy limits are being repeatedly tested, traditional institutions are being weakened, and there are significant geopolitical security risks.

Investors' current response should not be blind optimism, but rather **must construct a diversified asset allocation portfolio that can profit in the base case (modal outcome) while resisting the aforementioned major geopolitical and disruption risks.**

## Software Industry's "Barnes & Noble Moment": Repricing Disruption Risk

Regarding the current high valuations of US tech stocks, Bridgewater proposes an inspiring theory of "disruption risk premium."

Greg Jensen compares the current explosion of AI applications (represented by profitable breakthroughs in programming and code large models like Claude Code and Codex) to **the "Barnes & Noble moment" of 1998.**

> **Historical Mapping:** In 1998, facing the sweeping Amazon, US physical retail and bookstore giants (such as Barnes & Noble) suddenly realized they must invest heavily to move online at all costs.
> 
> During that period, the market frantically digested the expectation that "the internet was eating the world," causing tech stocks and the Nasdaq index to soar, while traditional Dow components plummeted. Although it later proved to be a huge dot-com bubble, the broader trend ultimately became reality over the following decade.

Today, history is repeating itself. **The market is frantically digesting the logic that "AI is eating the world," and this round of disruption begins with the software industry.**

Because AI fundamentally changes how software is built and the demand structure for software engineers, AI supply chain stocks (such as chips and compute infrastructure) have surged, while traditional software and related sectors most vulnerable to AI substitution have fallen sharply.

Bridgewater believes that when software stocks previously peaked, valuations perfectly priced in free cash flows for the next ten years, assuming 100% of software companies could survive securely in an environment without disruption; today's repricing means the market believes only about 82% of companies can survive the transition period of the next decade. **This is not crazy; it is the essential manifestation of the "disruption risk premium" entering the stock market.** And this is just the first wave; as technology diffuses, more industries will face this revaluation moment.

## Severe Mismatch in Compute Supply and Demand: Era of Rationing Arrives, Intensifying "Winner-Take-All"

Finally, regarding the contradiction between compute supply and demand, Bridgewater provided very clear and shocking industry data and chart analysis.

Greg Jensen pointed out: **"The only thing in the world growing faster than 'compute supply' right now is 'compute demand.'"** Although the current scale of semiconductor and data center construction is astronomical (Bridgewater believes the potential for supply growth is not even fully priced by the market), the actual landed "inference demand" is exploding even more fiercely.

Bridgewater predicts that **by the end of 2026, the compute supply-demand gap will be enormous, equivalent to "directly wiping out all of humanity's AI construction achievements from the past year."**

Since there is simply not enough compute in the world, **"compute rationing" has quietly begun**. It is happening in three ways:

-   **Price Adjustment:** Significantly increasing prices for compute and model calls;
    
-   **Access Restrictions:** Throttling and restricting model access permissions;
    
-   **Internal Halts:** Even within top AI labs, due to insufficient compute, some frontier research projects have to be shut down or paused.
    

**This "compute rationing" will have profound implications for future business competition.**

Bridgewater warns that since the world's top AI models and compute resources cannot be distributed fully and equally, **the "winner-take-all" industry monopoly dynamic is being unprecedentedly intensified.**

In this cruel disruption cycle, head giants that can seize the initiative to establish strategic partnerships and have enough capital to buy and secure compute will gain incomparable "structural generational advantages," thereby completely overwhelming competitors who cannot access frontier compute resources. For investors, understanding and locking in this underlying advantage will be the core key to capturing excess returns in the next stage.

Below is the full transcript of the dialogue, translated with AI assistance:

> **Host:**
> 
> You know, when we turn to the global outlook, there are two main themes. These two themes need to be looked at separately, but they are also extremely closely interconnected. This is what we call "Modern Mercantilism," and—of course—Artificial Intelligence (AI). So, perhaps take this opportunity to talk about how you assess these two major forces?
> 
> **Bridgewater Associates Co-Chief Investment Officer Greg Jensen:**
> 
> Okay. We want to delve into this new chapter we are currently in, namely modern mercantilism and the technological disruption brought by AI. In both cases, the world is changing. As you said, first, modern mercantilism and AI are intertwined; there are tight bonds between them, and both are simultaneously driving and accelerating "Resource Grabs."
> 
> By resource grabs, I mean: in this new modern mercantilist world, the US is looking for global "Choke points" to acquire resources it may need. Therefore, both sides attempt to consolidate their relative advantages through this, but these measures have negative impacts on global productivity synergy and cooperation. This is no small matter; it is driving markets, pushing up inflation, and so on.
> 
> Meanwhile, AI is certainly experiencing the largest capital expenditure (Capex) expansion since World War II. This brings us into a very tricky world. You get a "Modal outcome" (i.e., the most likely base case). If you think about our systematic investment process, you will know this is not the first time in history that a capital expenditure boom has occurred. The impact of capex booms is usually higher profits, because this is one company's profit, which has not yet become another company's cost; it usually brings higher economic growth and higher inflation to some extent, especially in the short term. And the demand for capital for capital expenditure usually pushes up interest rates. Places leading in capital expenditure usually see their currencies appreciate, etc.
> 
> So, there are some chain reactions you can rely on, which I describe as a "modal outcome"—this still seems to be the most likely scenario now. But the probability of this most likely scenario occurring may not exceed 50%, because there are too many extremely important factors to deal with in this process.
> 
> In this process, we face "chaotic modern mercantilism." This is a choice of President Trump (the Trump administration), which not only views the world through a Mercantilist lens. This creates risk; although it may bring some potential benefits in negotiations, it also brings risks such as miscalculation by the other side. Therefore, in a complex global system, huge changes are happening at an extremely fast pace, carrying the risk of conflict escalation. This is something that must be dealt with.
> 
> The second point is that we are testing the limits of fiscal policy in this way, weakening institutions, while accompanied by security risks and widespread AI disruption. Although there is a positive side to normal technological change, the speed of disruption brought by this transformation is unprecedented and must be deeply considered and addressed.
> 
> So, how do you respond? You must build an asset allocation portfolio that can make money in the expected "modal outcome," but at the same time be prepared to deal with the major risks and major changes in the process.
> 
> **Host:**
> 
> Despite all these risks, these risks are not unspoken. However, trading valuation multiples in the stock market (US stocks) remain at extremely high levels, and while US interest rates have slightly retreated, they remain in a fairly stable range. So, how do you assess the impact of, for example, modern mercantilism on today's investment portfolios?
> 
> **Bridgewater Associates Co-Chief Investment Officer Greg Jensen:**
> 
> Good, great. Let's step back. We have been talking about modern mercantilism for quite some time. In November 2024, I wrote a Daily Observations report, which you may remember, titled "We're all mercantilists now." In the report, I explained what we mean by mercantilism and some of the impacts we expected. So I want you to understand the latest developments and how we view the latest chapter.
> 
> First, when I wrote "We're all mercantilists now," I pointed out that we had entered the stage of modern mercantilism. In this stage, governments play a huge role in determining how markets operate, with the core focus—at least in the view of the Trump administration—on how to increase national wealth and power. Trade deficits are viewed as transfers of wealth, not part of a cooperative trading system. The priority is self-sufficiency and a "Transactional" statecraft approach. The situations we listed at that time are becoming reality one by one.
> 
> Entering 2025, you will see modern mercantilism fully implemented, such as the highest Tariffs since World War II appearing on the governing "Liberation day." You not only see tariffs, but also pressure on allies, that transactional approach—ensuring you pay. If you are to protect someone in Ukraine and elsewhere, ensure you get paid, and so on. All this is playing out. This has led to growing distrust of the US among traditional allies and shifts in alliances.
> 
> And all this is happening in a world where a large amount of capital is concentrated in the US. So people begin to question: If you are Canadian, or European, how wise is it really to allocate the vast majority of your asset portfolio to US assets?
> 
> Of course, entering this year, discussions about Greenland have increased, as well as Venezuela—the US saw China as a huge consumer of commodities in the Americas, and the US quickly changed this. Perhaps encouraged by the success of this move, the US took major action in Iran. Obviously, actions in Iran are much more complex. But now, with the evolution of the situation in the Strait of Hormuz, you see the trend of choke points being effectively weaponized (Weaponization) intensifying, which reduces global productivity and obviously causes various tensions.
> 
> **Host:**
> 
> Another question we frequently receive, and which you recently wrote a Daily Observations piece exploring, is the mismatch between Compute supply and demand. Therefore, I think it would be very valuable to hear you talk about the implications of this imbalance.
> 
> **Bridgewater Associates Co-Chief Investment Officer Greg Jensen:**
> 
> Okay, let me start with the latest developments in AI; there is really too much happening. We previously mentioned the resource grab triggered by AI—AI is frantically grabbing resources and building data centers on a large scale, and there are simply not enough Semiconductors in the world, etc. The "Science bet" we discussed last year is paying off, and I will elaborate on this later.
> 
> In addition, you are welcoming this disruptive moment. The launch of Claude Code has brought us into what I call the "Barnes & Noble moment." We will discuss this. The only point—and this also touches on the question you just asked—is that the only thing in the world growing faster than "compute supply" right now is "compute demand." This is leading to shortages and Rationing, which are major headlines.
> 
> Here, it might be helpful to delve into each aspect. I want to start with the "Science bet." What is happening here is that you now hold two bets. The "Science bet" we talked about earlier is driven by the belief of many people devoting themselves to AI fields: they believe that if they develop this technology, they can achieve major technological breakthroughs, making those who develop this technology extremely powerful in the world. This is the bet driving continuous increases in capital expenditure. Scientific breakthroughs were happening continuously, increasing the Conviction in technological breakthroughs, which further increased R&D, model training, and Capex, etc. You were originally cycling in this closed loop.
> 
> And since December, an extremely significant change has occurred: you have achieved a "Profitable breakthrough." Not only in programming applications, Claude Code, and to some extent Codex, have you seen huge breakthroughs that allow AI labs to create revenue through user adoption. Therefore, this is not just a continuous loop—where scientific breakthroughs keep happening and people are more convinced that following this path will achieve Artificial General Intelligence (AGI)—but at the same time, it has welcomed large-scale market adoption. This is what I mean by "the science bet is paying off." You now have a profitable revenue perspective in the programming field.
> 
> The second point I talked about is the "Barnes & Noble moment." Let me explain what this means. The so-called "Barnes & Noble moment" means the science bet is paying off, and you are now welcoming application adoption, which is creating disruption. The "Barnes & Noble moment" of 1998 refers to the moment when physical enterprises (Bricks and mortar companies) realized they must go online. Everyone must invest funds and move online, because Amazon is already online, and I must go online too.
> 
> In the market at that stage, the market began to digest the expectation that "the Internet was eating the world." During that period, Dot-com stocks and the NASDAQ index soared, while assets in the rest of the world, namely the Dow Jones Industrial Average, fell. In terms of market operations, that was an extremely dangerous period to navigate. Of course, in that case, it ultimately proved to be a bubble. Although if you stretch the timeline to the whole decade, that broad trend ultimately did come true, you experienced that huge bubble period.
> 
> If you look at the evolution of today's version, the market is currently digesting the concept that "AI is eating the world," and it starts with Software. You see AI's ability to change the way software is built; companies need to change how they build software, change their software engineer structures, etc. Therefore, you see AI supply chain stocks soaring, and sectors most vulnerable to AI impact falling.
> 
> This constitutes a very important moment that requires effort to think about, because I believe this is just the first of many disruptions coming as this technology rolls out and has an increasingly large impact. You can see this from the decline in software stock returns. You often hear people say, "Wow, such large volatility in software stocks is a mistake." I just hope people think from the perspective of "disruption" and the "Risk premium" you should price in for disruption. It feels like huge volatility in software stocks, but this is actually just a transition.
> 
> If you look at the past, when software stocks peaked, the market basically priced in the full continued Free cash flow for the next ten years, without considering any disruption factors. So, basically, 100% of major software companies were priced at that time as being able to exist securely 10 years later. And today, if you look at the changes in pricing, after adding a "disruption element," about 82% of companies are expected to survive the transition period of the next decade. This is not crazy; this is the essence of disruption. I think this is the risk premium entering the stock market that you will see in different industries, because major technologies are changing the way these industries operate, and every industry is experiencing this competition.
> 
> Then, let me add one more point. We discussed that because the actual usage demand for AI outside the lab is now very large, i.e., actual "Inference demand," what you see is demand far exceeding supply. We have been using massive amounts of data and indicators to understand all supply in the world; this supply growth is extremely huge. I mean, we still believe that the current existing compute supply has not been fully priced by the market, and the growth rate of demand is even faster than the growth rate of supply.
> 
> This is our total demand and total supply chart. By the end of this year, we believe the gap between supply and demand will become so huge that it is equivalent to "missing out on all AI construction scale from the past year." This is the shortage we will face next year. You have actually already seen it. There is not enough AI in the world, not enough compute, so that you are seeing "Rationing" begin to happen. And this is achieved by raising prices and restricting model access permissions. To some extent, inside the labs, they even have to shut down some of their research projects because there is simply not enough compute outside. We are currently only in the early stages of this phenomenon. As demand grows faster than supply, this gap will become larger and will bring huge impacts.
> 
> Finally, let me conclude this AI chapter with the consequences of AI capabilities being "rationed," because these consequences are crucial. By generating AI faster, we are intensifying the "Winner-take-all" dynamic. Because the best models are not fully distributed, the probability of you ending up in a winner-take-all situation is increasing. This is no small matter. This leads to Frontier AI capabilities creating a structural advantage for those who can access them. Establishing these partnerships to access these resources becomes extremely important. Therefore, this also brings the potential for an accelerated and broader "Disruption loop"—companies that can access the best AI tools, afford compute, and secure compute will completely defeat those that cannot. These are crucial implications for the next stage of AI disruption.

### Related Stocks

- [INTC.US](https://longbridge.com/en/quote/INTC.US.md)
- [MSFT.US](https://longbridge.com/en/quote/MSFT.US.md)
- [SMCI.US](https://longbridge.com/en/quote/SMCI.US.md)
- [META.US](https://longbridge.com/en/quote/META.US.md)
- [AMD.US](https://longbridge.com/en/quote/AMD.US.md)
- [TSM.US](https://longbridge.com/en/quote/TSM.US.md)
- [QCOM.US](https://longbridge.com/en/quote/QCOM.US.md)
- [MU.US](https://longbridge.com/en/quote/MU.US.md)
- [ASML.US](https://longbridge.com/en/quote/ASML.US.md)
- [AVGO.US](https://longbridge.com/en/quote/AVGO.US.md)
- [ASMG.US](https://longbridge.com/en/quote/ASMG.US.md)
- [SOXX.US](https://longbridge.com/en/quote/SOXX.US.md)
- [FTXL.US](https://longbridge.com/en/quote/FTXL.US.md)
- [VGT.US](https://longbridge.com/en/quote/VGT.US.md)
- [AVGU.US](https://longbridge.com/en/quote/AVGU.US.md)
- [SMCL.US](https://longbridge.com/en/quote/SMCL.US.md)
- [AVGG.US](https://longbridge.com/en/quote/AVGG.US.md)
- [NVDL.US](https://longbridge.com/en/quote/NVDL.US.md)
- [AMDL.US](https://longbridge.com/en/quote/AMDL.US.md)
- [PSI.US](https://longbridge.com/en/quote/PSI.US.md)
- [XDAT.US](https://longbridge.com/en/quote/XDAT.US.md)
- [NVDY.US](https://longbridge.com/en/quote/NVDY.US.md)
- [SOXQ.US](https://longbridge.com/en/quote/SOXQ.US.md)
- [AMUU.US](https://longbridge.com/en/quote/AMUU.US.md)
- [QCML.US](https://longbridge.com/en/quote/QCML.US.md)
- [IGPT.US](https://longbridge.com/en/quote/IGPT.US.md)
- [XSW.US](https://longbridge.com/en/quote/XSW.US.md)
- [RSPT.US](https://longbridge.com/en/quote/RSPT.US.md)
- [SMH.US](https://longbridge.com/en/quote/SMH.US.md)
- [DTCR.US](https://longbridge.com/en/quote/DTCR.US.md)
- [SOXL.US](https://longbridge.com/en/quote/SOXL.US.md)
- [IGV.US](https://longbridge.com/en/quote/IGV.US.md)
- [AVL.US](https://longbridge.com/en/quote/AVL.US.md)
- [XSD.US](https://longbridge.com/en/quote/XSD.US.md)
- [SMCX.US](https://longbridge.com/en/quote/SMCX.US.md)
- [AVGX.US](https://longbridge.com/en/quote/AVGX.US.md)
- [NVDU.US](https://longbridge.com/en/quote/NVDU.US.md)
- [XLK.US](https://longbridge.com/en/quote/XLK.US.md)
- [IXN.US](https://longbridge.com/en/quote/IXN.US.md)
- [AMZN.US](https://longbridge.com/en/quote/AMZN.US.md)
- [.IXIC.US](https://longbridge.com/en/quote/.IXIC.US.md)

## Related News & Research

- [Nectar Social Raises $30M Series A to Build the Agentic Operating System for Modern Marketing](https://longbridge.com/en/news/286426940.md)
- [Ex-Meta Chinese star joins race for self-improving AI with US$4.6b start-up](https://longbridge.com/en/news/286510256.md)
- [AMD set to benefit from agentic AI CPU demand surge](https://longbridge.com/en/news/286655054.md)
- [Mercor’s $2M daily payouts fuel AI labor debate](https://longbridge.com/en/news/286719252.md)
- [Sony’s Surprise TSMC Deal Could Ignite Its Next Big AI Growth Wave](https://longbridge.com/en/news/286294426.md)