Fuller’s energy-intelligence fusion, McLuhan’s tool-building refrain, and in search of the Holy Grail

“Wealth is the product of energy times intelligence: energy turned into artifacts that advantage human life.” – attributed to Buckminster Fuller

Download this PDF here.

Wealth is the product of energy times intelligence: energy channeled into artifacts that enhance human life. American visionary Buckminster Fuller’s idea hints at a fire that powers value itself. As Canadian media theorist Marshall McLuhan warned, the tools we build can also build us. Our tools don’t merely enable prosperity; they reshape society. With artificial intelligence (AI) dominance now tied to national security and economic growth, policy moves—like U.S. President Donald Trump’s supply-side measures and 100 per cent expensing of capital expenditures (CapEx) through 2031—set the stage for a CapEx supercycle. Markets don’t peak simply because valuations are high; they peak when liquidity dries up. Investors should expect persistent, mispriced “bubble” calls from consensus for the rest of the decade.

We can learn from history. The Gilded Age of the 1890s offers a warning: the railroad bubble burst in 1893, triggering a financial crisis. Will there be a moment when the Holy Grail at the end of the quest, Artificial General Intelligence (AGI), is deemed unachievable, causing hyperscalers to be repriced? In my view, the reckoning won’t arrive for at least another five years. If it does at all.

We stand at a historical inflection point where energy and intelligence fuse. Fuller’s axiom feels prophetic: energy and AI are no longer separate inputs; they form a single dynamic that creates new foundations for value. The era of AI, blockchain, and digital infrastructure isn’t a sidestep from history; it’s a leap forward. As previous revolutions exhaust their engines, the next one begins. Each new revolution builds fresh infrastructure, steering productivity and innovation in a new direction. We are only now allocating capital to construct this new core infrastructure.

This isn’t speculative capital chasing glitter. It’s a global refactoring of the economy. At the nexus of abundant, cheaper energy and self-improving digital intelligence, the true engines of prosperity emerge including data centres the size of stadiums, ultra-fast algorithmic systems, and cross-border blockchains that store trust beyond borders. These platforms aren’t abstract trends; they’re the scaffolding of the coming decades. The infrastructure for the AI-driven digital economy is being built. Yes, we’re in the early innings. No, we’re not yet in a bubble.

We’re entering a CapEx supercycle, where policy and power collide with incentives that tilt investment toward physical assets and digital infrastructure, and affordable energy that underpins AI dominance. The road ahead requires redefining wealth—not as a simple ledger but as a dynamic state where energy efficiency, intelligence amplification, and global networks co-create value at unprecedented scales.

Why “AI bubble” talk misses the real problem

Pundits have pivoted from inflation fears to AI bubble fears. No one should be surprised.

Some observers claim the AI hype is outpacing reality. The opposite is closer to the truth: there isn’t enough computer power to meet escalating AI usage. We’re in an AI arms race. While AI investment as a share of U.S. gross domestic product (GDP) isn’t extreme, hyperscalers allocate nearly 20 per cent of revenue to CapEx, well above historical norms. But with 100 per cent expensing of CapEx, what can historical norms tell us?

Bubbles don’t pop because of valuation; they pop when liquidity dries up. In a global reflation scenario where debt is inflated away, investors should ignore near-term bubble chatter. Valuations will matter eventually, just not yet.

To resolve today’s fiscal crisis, we must grow our way out—as was accomplished after the Second World War— by prioritizing broad-based supply-side expansion and non-inflationary growth. U.S. Secretary of the Treasury Scott Bessent has emphasized leveraging the full balance sheet of the nation to achieve this, advocating for running the economy hot with a re-rating in gold, integrating Bitcoin, purchasing equity in companies, and implementing policies that catalyze productive investment and real output.

Think of AI computers as highways and tokens as cars. In the past year, AI traffic surged roughly 100 times. Even with smarter signals and more lanes, demand outpaces supply. Companies effectively borrowed computing capacity from other operations to handle the surge. The old telecom-era view of exponential improvement is over. Today, only near-term upgrades from firms like Nvidia and AMD make highways faster, and even those improvements can’t keep up with the flood of new cars.

By 2030 we may have ten times more computing “power plants” but if demand stays the same, we’ll still be behind. Every new lane is instantly utilized. Those who claim excess AI capacity aren’t seeing the core issue: this isn’t a supply-dominant bubble—it’s a genuine shortage of computing power for the current surge in use. AI workloads exploded dramatically last year, akin to a city’s traffic swelling so fast that even better roads and signals can’t add enough lanes. The era of computers doubling in speed every few years is behind us; substantial, targeted improvements are now required simply to keep up.

In short, skeptics debate bubbles but the underlying reality is a pronounced underground shortage. AI developers are sprinting to close the gap, but for now, supply lags demand. There is no bubble, there’s a race to build enough “roads” before the next AI traffic jam.

Forget the bubble nonsense. AI is nowhere near overhyped. At 1.5 per cent of GDP, it’s only getting started. Railroad CapEx in the 1800s peaked at six per cent of GDP and reshaped the world but AI is scaling with an even loftier ambition.

The Holy Grail: The AGI energy arms race

Artificial General Intelligence (AGI)—the Holy Grail of AI—rests on a simple, brutal premise: an objective function that defines what a system should achieve. As ambitions scale, so does the energy bill. Today’s state of the art models gulp electricity on a scale comparable to entire cities, driven by a global web of data centres that swells with every training run, every inference, every new interaction.

AGI is not just smarter software; it’s a shift toward autonomous minds that set their own objectives. Where today’s AI follows human specified goals, AGI would learn, imagine new aims, adapt across domains, and tackle open-ended problems. The winner of this race won’t just win a market niche; it will gain leverage across industry, society, and geopolitics.

Energy is the common thread binding ambition to outcome. Cheap, secure energy becomes the strategic input that powers a generational CapEx boom. Canada, traditionally seen as a resource nation, could pivot toward becoming both a resource powerhouse and an AI powerhouse. But to get there, we must shed the biases and constraints of past generations and think at planetary scale.

AGI-grade capability will not merely increase, it will explode energy needs. Power becomes not a line item in a budget but the central driver of global capital expenditure. Trillions will flow into grid expansions, chip and storage manufacturing, hyperscale data centres, and secure digital infrastructure. For many regions, winning this race will be tantamount to national security and economic sovereignty—on par with the strategic quests that once defined energy and industrial dominance.

In the McLuhan sense, the “architecture of objective setting” is the medium through which intelligence shapes its own pursuits—and it will be as consequential as the outcomes we chase. We’re not far from the moment McLuhan warned about: Every tool we build has the potential to build us. In the coming five to seven years, a CapEx supercycle will unfold at the intersection of energy, intelligence, and infrastructure. The tools will metastasize into shapers of society, markets, and ambition itself.

The stakes extend well beyond technology and commerce. They threaten to redraw the global economic order and questions of sovereignty, depending on who builds, who powers, and who ultimately controls self-motivated, self-directed intelligence.

Today’s trajectory is already a story of relentless amplification: the tools we invent redefine what’s possible, which in turn amplifies the value, and the cost, of every dollar and watt spent. America’s data centres are primed to double their power usage in the next five years, while the grid must add roughly 95–100 gigawatts of new generation capacity. Grid constraints have become strategic bottlenecks: utilities plan years ahead, and regions with cheap, abundant power such as Texas and Quebec are becoming magnet zones for capital and invention. China is forging ahead, full steam. This isn’t a speculative bubble; it’s hard deployment, as tangible and unstoppable as the construction of railways and steel mills once were.

As I said, this is an arms race for global AI dominance. To fuel this race, we need gigawatts of capacity. This is an arms race that lasts until at least 2031. We are just getting started.

Not your 1990s bubble

The dot-com era left scars and lessons. It was a time of hollow hope, where hype outpaced product and profit. The collapse was brutal: most dot-com firms vanished, networks of fibre lay idle, and capital left behind “dark fibre” as a memorial to overreach. And yet the past isn’t a perfect mirror for today. When today’s giants, Google, Amazon, or Meta stumble, they absorb it within business as usual. Their platforms are indispensable, revenues diversified, cash flows robust. Failures don’t spell extinction; they catalyze reinvestment and expansion. We’re not chasing empty optimism; we’re fortifying digital ecosystems that anchor commerce, logistics, research, and governance. The risk isn’t a wipeout of hollow models but an accelerating transformation where capital, intelligence, and energy converge to shape us in the image McLuhan warned about.

Bubbles aren’t rare blips; they’re the norm. Every leap in computation draws in capital ahead of profits. With technology revolutions, railroads, canals, internet, blockchain, and AI, the cycle repeats: overinvestment, exuberance, and then the next foundational platform.

Serial bubbles are a feature, not a bug. Speculative capital funds the pipes, servers, and algorithms that will power real value when deployment catches up with hype. The dot-com bust seeded cloud platforms. The blockchain bust built enduring infrastructure. Today’s AI bubble is refining the backbone of computation, data, and intelligent automation that will outlive the hype.

Beyond bytes and code, power decides outcomes. Cheap, reliable energy has long dictated industrial dominance, and the digital era is no exception. AI and blockchain hunger for watts, and only those with abundant, low-cost energy will stay ahead. Investors should target utilities, renewables, and regional champions capable of delivering scale and resilience. The economics of platform growth now hinge on watts and efficiency, and who commands them will shape innovation and reap outsized profits. Regulatory policy has moved from backdrop to engine. Incentives for full expensing of capital equipment, reshoring of manufacturing and energy supply chains, and strategic protections compress investment cycles and demand speed and scale. National security is no longer peripheral; it’s central to how and where capital flows, defining economic sovereignty.

This transformation isn’t without nerve-wracking moments. AI shifting from headlines to core production provokes awe and anxiety alike. Technology isn’t a neutral instrument; it reorders competition, displaces labour, and reframes value creation. AI’s march is thrilling and unsettling as the prospect of self-improving systems invites profound questions about human limits and agency. AI is more than a tool; it’s a medium that will rewrite expectations, competitive landscapes, and the very process of value creation. The true visionaries grasp this: shaping the direction of development isn’t optional. It’s imperative to steer the trajectory, build safeguards, and design institutions that keep pace with rapid innovation.

Where does wealth anchor in this velocity of change? Gold has long stood as a hedge against instability, a universal store of value in volatile times. It remains relevant, yet Fuller’s warning echoes louder: that real wealth is ideas plus energy. In this transformative era, a new class of assets capitalizes on that fusion, led by Bitcoin, a platform for decentralized value rooted in proof-of-work and, in theory, resistant to inflation and manipulation. The smartest move blends the old with the new. Investors are layering conventional hedges with exposures to AI, computing platforms, energy grids, and digital currencies. Wealth shifts from mere scarcity to scale: the ability to amplify human ingenuity with machine power.

The infrastructure surge carries a crisp reality: deflationary forces ride alongside inflation of ambition. Overcapacity drives falling unit costs, squeezing late movers and rewarding those who optimize scale and efficiency. The marginal cost of computation and storage plunges, sometimes dramatically, yet this isn’t a threat; it accelerates value creation. When energy and intelligence scale together, the resulting value compounds faster than input prices erode. Leading platforms enjoy virtuous cycles: more capacity lowers costs, spurs adoption, and fuels further innovation. We are living in an era of exponential growth in innovation. The real risk is inertia. Wall Street is clutching its pearls but investors should recognize that delaying entry risks missing historic returns, policy incentives, and network effects. The moment to reposition portfolios is now, while critical applications stand at the edge of mass diffusion. Investors need to realize we are in a secular bull market. An S&P 500 target for 2026 of 8000 should be expected.

So what’s the play? This isn’t the bubble stage; it’s the accumulation phase. The window remains open to acquire scalable bets in energy infrastructure, chip foundries, power utilities, and data pipelines before their full value is priced in. Policy, technology imperatives, and the compounding advantage of network effects all favour those who move first and scale early. Market leaders will win by conquering three fronts at once: building platforms, securing cheap power, and catalyzing adoption in lagging sectors. Those who grasp the dynamics know the urgency to shape deployment, not merely ride its wave.

What should investors do?

Capitalize on the AI data centre supercycle. AI’s convergence with digital infrastructure has turned the data centre sector into the power grid of a new age, the nexus of energy and intelligence at scale. Growth hinges as much on watts and concrete as on code. Cheap energy is the indispensable competitive advantage. AI workloads are driving explosive demand for computation, with the sector potentially requiring 95–100 gigawatts of new electricity over the next decade. Some AI campuses may consume one to five gigawatts apiece. The challenges of energy procurement, on-site generation, and grid upgrades are pivotal for operators and investors alike.

If you’re looking for a way to participate, the story begins with energy and minerals, grid and utility capacity, and data centre or hardware exposure. The decades ahead will reward those who secure reliable supply chains for uranium, copper, natural gas, and other critical inputs. They will reward those who back energy generation, transmission, and storage. And they will reward those who remain exposed to the leading chipmakers and hyperscale data centre developers powering AI workloads. The strategy is simple in aim, if complex in execution: build real assets at scale, blend hardware, software, and energy plays, and enter early to ride a multi-year cycle—before margins compress and adoption broadens.

The way forward: seize the buildout

The convergence of AI, digital infrastructure, energy transition, and critical minerals is redrawing the investment map. To win, invest where energy and ideas intersect: core energy, minerals, grid-linked assets, and the hardware that will scale AI. The next era of growth is being written in datasets, grids, and chips—an era that will redefine how wealth is created and distributed. Ignore bubble chatter and embrace the CapEx supercycle. The platform for the next economic model is under construction, and those who recognize and act on this truth will find themselves not just along for the ride, but steering the buildout itself.

Investing is about patience. Hours are spent researching, reading, thinking, and waiting. When a big macro theme unfolds, you implement your plan and then wait. My thesis is that this CapEx supercycle ends early next decade. Ignore the calls for bubbles and recognize the size and magnitude of what stands in front of us. The window is not endless, but it’s wide enough for disciplined, patient capital to redefine the framework of wealth, energy, and intelligence.