Jordi Visser argues that AI has already moved from the chatbot era into an agentic era that is massively increasing compute demand, reshaping labor, and changing how markets should value software, hardware, crypto, and commodities.
Watch on YouTube ›Get the market thesis, key claims, assets, contradictions, and follow-up questions from any financial video — then unlock a version personalized to your portfolio, watchlist, and favorite speakers.
This interview centers on Jordi Visser’s framework that AI is no longer just a productivity tool but a structural regime shift. He says the transition from chatbot to agentic AI in late November/December was the key inflection point, because agents can act continuously across workflows and require vastly more compute. In his view, that compute demand is non-linear, underappreciated, and already showing up in prices for DRAM, semiconductors, and industrial inputs. He argues this is not simply an Nvidia story: the AI boom extends to a broader set of hardware, memory, and commodity beneficiaries. Visser is bearish on the traditional software model and skeptical of using old discounted cash flow frameworks for many AI-exposed companies. …
Near term, the setup is for inflation prints, AI-capex headlines, and recession chatter to drive volatility and rotation. Tactical opportunity seems to favor AI-linked hardware, memory, silver, and Bitcoin over software and broad-duration equity exposure.
Over the next few months, the base case is continued repricing of AI winners versus losers as agentic adoption and compute shortages become more visible. Confirmation would come from persistent capex, rising inflation, and further multiple compression in software; the main invalidation would be a slowdown in AI spend or a clear drop-off in demand.
Structurally, Visser is arguing that AI changes the economic production function, making physical compute, power, memory, and strategic hardware more important than legacy software narratives. If that regime holds, capital allocation, labor organization, and asset valuation norms all shift toward scarce, AI-enabling inputs rather than cash-flow-heavy software franchises.
The end of November into December marked the transition from the chatbot era to the agentic era.
He treats that period as the key inflection point for AI adoption and compute demand.
Agentic AI requires roughly a thousand times more compute than the prior chatbot phase.
He uses this to justify a major new wave of infrastructure demand.
AI will not necessarily drive a major recession or a materially higher unemployment rate.
He argues job losses are already happening in a different form, and the labor market is constrained by demographics and immigration.
Can you share a bit about your background and journey to where you are today?
Jordi Visser started at Morgan Stanley in the early 90s in equity derivatives, opened an office in Brazil, ran the S&P options book, then set up his own macro fund and was CIO at a multi-strategy fund. Now he spends all his time with LLMs, teaching people about AI and studying its impact on everything.
What's your broad framework for navigating macro in the age of AI — what's hyperbole, what's real, what's understated?
Jordi says last year people doubted AI and called it a bubble without using it, which signaled an arbitrage opportunity. He argues AI is not a bubble — Corning and Nvidia gains are justified by trillions of dollars flowing through the economy. Oil prices don't matter to the global economy the way they used to, and rising gas prices won't stop AI spend. He notes the S&P 500 rose 15% while household net worth rose $15 trillion.
How are you thinking about this productivity boom — is it sufficiently discounted or understood, and what does it actually look like?
Jordi says productivity is not good for enterprises in his opinion. The reason he's there is because he's 100% a believer Bitcoin is the endgame for growth assets. Enterprises are seeing share prices fall — the productivity boom benefits people with no employees (like himself: $12k/year for five LLMs plus hardware). The labor arbitrage for enterprises is negative. Entrepreneurs with good ideas can make money, but enterprises will have a hard time adopting because they need to get rid of people.
Unlock the full claims, asset map, scores, related transcripts, follow-up questions, and AI chat — shaped around your portfolio, watchlist, favorite speakers, and risks.
