Bitcoin and the Art of Unlearning the Fed

Sentiment in crypto is depressed. Even the Bitcoin permabears seem reluctant to keep kicking it while it is down. That is usually when I start paying attention. When sentiment gets this bad, and when bearish narratives have had eight months to harden into consensus, the next rally rarely begins with everyone suddenly finding a new reason to be bullish. It usually starts quietly. The pressure stops getting worse, the marginal seller runs out of ammunition, bad news loses its impact, and price begins to stabilize before the narrative changes. Like watching a tree grow, the turn happens slowly enough that most people do not notice it until it is already visible.

Since the fourth quarter of 2025, one major macro trend has acted as an extreme headwind for Bitcoin: the rise of Opus 4.5 and the beginning of the agentic AI world. That catalyst created three related pressures. First, it accelerated the collapse in software names as investors began discounting AI disruption. Second, it drove a major jump in capex expectations to support the parabolic increase in inference demand from AI agents. Third, it pulled attention, capital, and market leadership toward the physical infrastructure side of AI, including semiconductors, power, cooling, data centers, electrical equipment, and everything needed to build the agentic economy.

The final pressure was the most important for Bitcoin. Higher AI capex numbers, stronger nominal GDP, and the earnings strength of the AI infrastructure complex changed the market’s view of the Fed. Over the last six months, investors moved from expecting multiple rate cuts by the end of 2026 to debating the possibility of multiple rate hikes. Bitcoin became a casualty of the AI agentic rise because the same force that proved AI was real also pulled capital away from crypto and pushed the market toward a more hawkish Fed interpretation.

That leaves Bitcoin with two potential catalysts now. The first is capital rotation. The AI trade has sucked the oxygen out of many parts of the market. For the last two years, exposure to the right infrastructure names, semiconductor names, power and cooling names, or physical-world beneficiaries was rewarded. Software, smaller growth companies, crypto, and anything not directly attached to the AI infrastructure boom were mostly left behind as capital crowded into the dominant earnings and capex story.

That crowding now matters. AI remains the dominant structural trend, but the easy-money phase has passed. Equity momentum is crashing, the infrastructure trade is choppy, and the mid-cycle slowdown is making the same AI beta harder to harvest. Momentum investors still need beta. They do not want to sit in cash forever. As the AI infrastructure trade becomes more volatile, crowded, and difficult to own, they will start looking for a new expression of the same technological regime shift that has not already been fully harvested.

Bitcoin fits that search. It sits outside the crowded AI equity trade while still connecting to the next phase of the digital economy. It can benefit if capital starts looking for a cleaner expression of AI-driven change. It can benefit if investors want exposure to the agentic economy without owning the same crowded infrastructure names. It can benefit if the market begins to realize that AI agents will change software, labor, corporate margins, and eventually the financial architecture required for autonomous economic activity.

The second catalyst is the rate-expectation reversal. Over the last six months, Bitcoin has traded like an asset tied to the disappearance of rate cuts. As cuts were priced out and hikes entered the conversation, Bitcoin moved lower. That does not reduce Bitcoin to a simple rates trade. It means the marginal buyer stepped away as real-rate pressure rose, the dollar narrative improved, and Fed watchers turned the AI infrastructure boom into an argument for a more restrictive central bank.

This is where the art of unlearning the Fed becomes central to the next Bitcoin narrative. In my recent paper, “The Art of Unlearning the Fed,” I argued that the next phase of Fed watching will require something harder than learning a new framework. It will require unlearning an inherited one. For decades, investors listened to the Fed through the same language: hawkish or dovish, restrictive or accommodative, higher for longer or pivot, dots or no dots. That language worked in a more linear economy where policy, inflation, labor markets, and productivity moved with long and variable lags. AI is making that world less stable, less measurable, and more reflexive.

That matters for Bitcoin because the market is still using the old language to price a new economy. Fed watchers are focused on the demand side of AI. They see data centers, chips, energy, cooling, construction, electrical equipment, and grid investment. They see a private-sector capex boom landing on top of 5–6% fiscal deficits, rising debt, and growing interest expense. Inside the inherited framework, that combination looks inflationary. AI capex plus fiscal deficits equals stronger nominal GDP. Stronger nominal GDP equals a hawkish Fed. A hawkish Fed equals higher real rates, a stronger dollar, and lower Bitcoin.

That has been the dominant equation for the last six months. It is also one of the main reasons Bitcoin has struggled. The market has treated AI primarily as a demand shock, but the agentic phase introduces a second force: supply expansion through digital labor. AI agents, copilots, autonomous workflows, coding agents, compliance agents, customer service agents, research agents, and operational agents are the first visible form of digital labor entering the production function. The infrastructure buildout increases demand today, while agent adoption increases potential output tomorrow.

Kevin Warsh has been unusually clear about this. He has described the current moment as “the most disruptive moment in modern economic history in the U.S. and the world,” and later said, “This is a big paradigm shift both for the conduct of our policy and for our economies.” He is treating AI as a force powerful enough to require a new central banking paradigm, which is why his demand-versus-supply framework matters so much for Bitcoin.

Warsh acknowledged that AI infrastructure spending would affect demand and could add “a few tenths of 1%,” but he immediately separated that from the larger supply-side possibility, saying AI’s ability to increase potential output “could be considerably bigger.” That distinction is the whole Bitcoin setup. The market has priced the AI demand impulse first. The next repricing may come from the AI supply impulse.

If the market is wrong about that balance, the cross-asset setup changes. A Fed chair who understands AI may sound tough on inflation credibility while recognizing that productivity is improving. He can defend the inflation target while questioning whether the economy’s speed limit has changed. He can be cautious on rates while understanding that digital labor may weaken wage pressure even as output stays strong. The old labels are too small for the moment because the more important issue is whether AI is rewriting the Fed’s reaction function.

That is why Bitcoin becomes interesting here. The market is crowded into the idea that AI means higher rates. A rally can begin before the Fed turns dovish if investors decide the current hawkish path has been over-discounted. If AI agents begin to mean productivity leverage, disinflationary pressure, labor disruption, and a harder dual-mandate problem, then the rate narrative can shift quickly. Bitcoin only needs rate expectations to stop moving against it before the marginal buyer starts to return.

This is how cross-asset momentum regime shifts usually happen. The market builds a clean consensus, and that consensus becomes a position. In this case, the consensus has been that AI capex creates demand, demand creates inflation, inflation creates hikes, hikes support the dollar, and dollar strength pressures Bitcoin. Then the data and narrative begin to complicate the story. Inflation cools. Wage pressure fails to accelerate. Labor-market internals soften. AI adoption accelerates. Enterprise agents move from pilots to production. Productivity leverage starts showing up in workflows and margins. The simple equation loses its power.

At the same time, the AI equity trade remains in a mid-cycle slowdown. When the AI infrastructure trade was going straight up, it absorbed the speculative oxygen of the market. Now that the trade has become more volatile, crowded, and harder to own, investors will begin looking for a cleaner expression of the next stage. Bitcoin can become that expression because it sits outside the crowded AI equity trade while still benefiting from the same technological regime shift.

The rise of AI agents strengthens the case, even though the commerce side of agents is taking longer than the coding side. Coding agents moved faster because the task environment is more structured, feedback loops are clearer, and the ROI is easier to measure. Commerce agents are harder. Once agents start transacting, consuming compute, managing subscriptions, purchasing services, authorizing workflows, or moving value across digital environments, the stakes become much higher. Mark Zuckerberg was recently reported to have told Meta employees at an internal town hall that AI agents were not progressing as quickly as planned. That detail matters because the next phase is about trust, permissions, reliability, and financial guardrails as much as raw model intelligence.

That delay clarifies the Bitcoin argument. Autonomous software will need a financial architecture around it: permissioning, settlement, authentication, auditability, spending limits, and neutral rails. The argument is not that every agent will settle directly on Bitcoin. The argument is that a world of abundant intelligence, autonomous software, and accelerating digital activity increases the value of neutral, scarce, globally available monetary assets. Bitcoin is a scarce asset sitting outside a system that is becoming more automated, more fiscal, more debt-heavy, and more difficult for central banks to manage.

That is why the deficits matter. Even if AI improves productivity, the U.S. still has large fiscal deficits and a worsening debt path. Interest expense is already a structural constraint. Mandatory spending remains the dominant long-term driver. AI may eventually improve healthcare costs, government efficiency, and productivity, but that is not here yet. The government remains dependent on nominal growth, financial repression, or both. In that environment, a Fed that becomes too hawkish risks colliding with fiscal reality and the labor-market disruption caused by AI.

This is the policy trap. AI capex and deficits make the inherited framework worry about inflation, while AI agents and digital employees create productivity leverage, disinflationary pressure, and a more fragile labor market beneath the surface. Warsh’s point is that central banking must become more forward-looking because the old data may not capture the shift in real time. He has emphasized the need for better data and a more contemporaneous understanding of the economy rather than relying only on government data with mismeasurement problems and surveys that may no longer be relevant.

For Bitcoin, the next rally could be about unlearning the Fed. Investors have been trained to assume that strong nominal activity automatically means a more restrictive central bank. In an AI economy, strong output can coexist with weaker labor bargaining power, lower unit labor costs, and rising potential growth. That is a new monetary-policy puzzle. The market has priced the first half of that puzzle through the AI demand impulse. It has not yet priced the second half through the AI supply impulse.

Bitcoin has been pressured by AI capital crowding and by a hawkish repricing of the Fed. Both pressures may now be close to exhaustion. Equity momentum is crashing, and AI beta is becoming harder to harvest through the same crowded infrastructure names. At the same time, the market may have gone too far in assuming the Fed will respond to AI demand without recognizing AI supply. If Warsh’s framework gains traction, investors may begin to see that AI creates more than demand for data centers. It creates productivity leverage, labor disruption, financial automation, and a need for new monetary guardrails.

Bitcoin is one of the few assets that can benefit from all of those shifts at once. The next rally may begin as a rotation trade, then evolve into something bigger. It can start because momentum hunters need a new beta trade as AI equities become too volatile. It can continue because rate expectations stop moving against it. It can accelerate if investors begin to understand that AI is forcing the Fed, the economy, and the financial system into a new paradigm.

The art of unlearning the Fed is also the art of relearning Bitcoin. The Fed-cut question is giving way to a broader question: whether Bitcoin benefits from a world where the Fed’s inherited framework is no longer enough. In an economy defined by deficits, digital labor, productivity leverage, capital concentration, and autonomous agents, Bitcoin’s role becomes clearer. Bitcoin may still trade with liquidity, but the larger argument is that the next monetary regime will be harder to manage than the last one.