AI Is Ushering in an Entirely New Economic Paradigm | Jordi Visser

Episode Summary

Executive Summary: Jordi Visser argues that AI is not a temporary cycle but an infinite, economy-restructuring force that will expand compute demand, compress margins over time, and eventually drive deflation while reshaping labor, policy, and markets. He contrasts old macro indicators with a digital economy increasingly measured by profits, concentration, and state-backed tech buildouts, then extends the thesis to Bitcoin as a scarce digital asset in a fully financialized world.

Main Topics: AI as an infinite demand engine (Priority: 5/5): Visser frames AI as a never-ending project: intelligence is being embedded into every machine, with demand expanding as capability improves. He argues the relevant question is not whether there will be enough revenues, but how rapidly intelligence can be applied across industries and problems. Why traditional macro signals are failing (Priority: 5/5): The conversation challenges legacy recession models, LEI, yield-curve style thinking, and GDP-style measures. Visser says these indicators were built for an industrial economy and miss the digital economy's intangibles, wealth concentration, and software-driven output. Profit margins, not GDP, as the best signal (Priority: 4/5): Visser argues productivity is hard to measure directly, so profit margins are a better proxy for technological gains. He points to widening margin dispersion between low-tech businesses and AI leaders like NVIDIA and Mag7 firms as evidence of a structural shift. AI capex, data centers, and shifting competition (Priority: 5/5): He sees AI buildout as analogous to railroads/electricity but with brains rather than wires. He distinguishes LLMs from future VLMs and VLA systems, arguing compute needs will explode as AI moves from text to vision, action, robotics, medicine, and defense. Deflation, not inflation, as the bigger macro risk (Priority: 5/5): Visser expects AI and related productivity gains to weaken pricing power, reduce labor demand, and push the economy toward disinflation/deflation. He downplays power-cost inflation and argues the state and hyperscalers will absorb infrastructure costs. Bitcoin as the digital-age store of value (Priority: 4/5): Visser portrays Bitcoin as one of the few enduring moats alongside gold and religion, and as the preferred asset for younger generations in a financialized system. He links its rise to debasement, democratization, and the erosion of corporate and monetary moats. State support and the merging of public/private AI (Priority: 4/5): He highlights government policy as a major accelerant, saying AI has become tied to national survival and competition with China. He interprets this as effectively removing the political risk from the AI buildout.

Key Arguments: AI demand is effectively infinite because every additional improvement creates new applications, new markets, and new efficiency gains rather than a fixed endpoint. Legacy macro tools such as LEI, yield-curve recession calls, and GDP miss the shift from physical goods to software/intangibles and therefore overstate recession risk. Profit margins are a superior way to track technological progress because they translate productivity into hard accounting outcomes, unlike GDP or broad productivity estimates. The AI buildout is different from the dot-com era because it is funded by highly profitable incumbents with enormous free cash flow, not weak telecom balance sheets. Large tech firms may increasingly resemble utilities: not because growth is slow, but because they control essential infrastructure for a new intelligence layer. The main macro consequence of AI is likely deflation, as smarter systems reduce costs, increase competition, and eventually eliminate many jobs and businesses. Electricity cost inflation is unlikely to be the binding constraint because hyperscalers and states will find ways to fund or internalize power and infrastructure needs. Bitcoin benefits from a world where money printing, financialization, and the destruction of business moats make scarce digital assets more attractive than traditional stores of value. Younger users will likely migrate toward Bitcoin and digital assets because they are native to a game-like, online financial system rather than older gold-based narratives. Public policy is now aligned with AI acceleration, which lowers the odds of a true shutdown or meaningful restriction on the buildout.

Data Points: Household net worth: $178 trillion - Used to argue that wealth concentration can support spending even when income growth is weak. Income growth ex transfers: 1.4% year over year - Visser cites weak private income growth as evidence of a bifurcated economy. Transfer payments growth: 8.5% to 9% year over year - Presented as a form of bottom-end UBI supporting consumption. Transfer payments size: $5 trillion - Visser says transfer payments have become a major income source in the economy. Private income size: $16–17 trillion - Compared against transfer payments to show scale of state support. S&P 500 corporate profit margins: ~14% - Used as a broad market margin benchmark versus high-tech leaders. NVIDIA profit margins: 70%+ - Illustrates margin superiority in the AI/semiconductor layer. Mag7 profit margins: closer to 70% - Used to show extreme margin concentration among leading tech firms. AI compute needs for VLMs vs LLMs: 50x to 1,000x more compute - Visser argues visual language models require far more compute than text-only systems. Forecast AI/data-center spend: ~$5 trillion over five years - Estimate for global spending, including China, on the AI buildout. Compute share of data-center spend: 30% to 60% - Used to estimate the semiconductor revenue pool within data-center capex. NVIDIA data-center market share: ~90% currently - Presented as NVIDIA’s current moat in AI infrastructure. NVIDIA potential market share: ~50% in the scenario discussed - Conservative future-share assumption if TPU competition increases. Potential NVIDIA revenue in 2030: $750 billion - Derived from the assumed AI capex and market-share scenario. Current NVIDIA revenue estimate: $425 billion - Referenced as the market estimate at the time of the discussion. NVIDIA revenue growth: 67% year over year - Last earnings report cited to support the bullish thesis. NVIDIA quarter-over-quarter growth: over 20% - Used to show momentum continuing to accelerate. Russell 2000 market cap: just over $3 trillion - Compared to NVIDIA to illustrate concentration and scale. NVIDIA market cap: $5 trillion - Used to show how large the AI leader already is relative to small-cap equities. Corporate debt issuance related to AI: ~$100 billion - Described as still small relative to mega-cap free cash flow and balance-sheet strength. COVID stimulus: $9 trillion - Cited as a major fiscal/monetary shock that reinforced financialization.

Pivotal Quotes: "The amount of need for AI is infinite. It will never end." — Jordi Visser: Core thesis that AI is an open-ended demand engine rather than a finite investment cycle. "We are literally plugging intelligence into every machine that has ever been built to make it be more efficient." — Jordi Visser: Explains why AI demand expands across industries and use cases. "The faster we have intelligence grow, the closer we're getting to a deflationary spiral." — Jordi Visser: Summarizes his macro view that AI ultimately reduces prices and increases competition.

Implications: Listeners should expect AI to reshape margins, labor, policy, and capital allocation faster than legacy macro models suggest. For investors, the key is tracking who owns the compute layer, who builds on it, and how deflationary pressure and Bitcoin adoption evolve.