The logic is seductive: tax the machine, not the man. Andrew Yang, the 2020 presidential candidate turned Forward Party co-founder, stood on CNBC’s Power Lunch and argued that the US government should shift its tax base from payroll to artificial intelligence. His reasoning is straightforward—firms replace human labor with AI to avoid payroll taxes and healthcare costs. Tax the substitute, and the incentive flips.
But this narrative ignores a deeper structural reality. In a world where capital flows are already distorted by decades of monetary expansion, adding a tax on productivity gains is not a correction—it’s a distortion. The market is currently pricing this as noise. I see it as a signal of a broader regulatory drift that will ultimately constrain the very innovation it seeks to tax.
Yang built his political brand on automation warnings. His 2020 campaign proposed the Freedom Dividend—a universal basic income funded by taxing automation. He also backed cryptocurrency adoption and clearer digital asset rules. Now, as CEO of Noble Mobile, he’s reviving the same argument with fresh data. He pointed to Anthropic CEO Dario Amodei’s suggestion of a 3% AI revenue tax, applied each time a model generates revenue. Yang wants this logic applied broadly: force firms to weigh AI costs against payroll costs.
The data seems to support his case. A CNBC and Generation Lab survey from August 13 polled Americans aged 18 to 34. 45% expect AI to hurt their careers, only 10% expect it to help. Bridgewater Associates executives Greg Jensen and Nir Bar Dea wrote a New York Times opinion piece estimating that AI could displace 18% of current US jobs within five years. They used that estimate to back their own AI token tax proposal. The customer service sector—employing roughly 2.9 million Americans—is already seeing the shift.
Yang proposed sending the tax revenue directly to workers as checks. He argued retraining programs rarely work, pointing to failed efforts for coal miners and warehouse staff. On the surface, this is a clean intervention: tax the externality, redistribute the proceeds.
But I’ve seen this movie before. In 2020, during DeFi Summer, I modeled Compound Finance’s interest rate curves on my laptop in Rome. I identified a liquidity crunch risk when ETH collateralization ratios dropped below 150%. The protocol was over-leveraged, but the market ignored it because TVL was growing. The same pattern applies here: the AI tax proposal is a structural intervention that assumes the government can effectively tax a productivity-enhancing technology without triggering capital flight.
Let me explain why this is a misdiagnosis. The core issue is not that AI replaces jobs—it’s that the global monetary system has already incentivized capital over labor for decades. Low interest rates, quantitative easing, and zero-yield environments have pushed corporations to automate and replace human labor with software. Yang’s tax treats the symptom, not the cause. If you tax the machine, you discourage automation. But in a world where capital is cheap and labor is expensive, firms will simply move the AI offshore or use decentralized models that bypass the tax entirely.
This is where crypto enters the picture. Decentralized AI models, running on blockchain-based networks, offer a tax-avoidance mechanism. A company could deploy a smart contract that pays for AI inference using a stablecoin, with no payroll tax hook. The tax base for an AI tax would be highly elastic—firms can restructure their operations to shift revenue streams to jurisdictions or protocols that don’t levy the tax. In economic terms, the deadweight loss of an AI tax would be massive because the substitution elasticity is high.
Volatility is the tax on unproven consensus. The market consensus today is that the AI tax debate is a political sideshow. I disagree. The Bridgewater token tax proposal—taxing AI tokens rather than revenue—is a more elegant solution, but it still misunderstands the nature of crypto assets. AI tokens are not equities; they are utility tokens that derive value from network usage, not corporate earnings. Taxing them at the token level would create a liquidity sink, reducing the velocity of the underlying network and suppressing innovation.
Consider the macroeconomic context. The US debt-to-GDP ratio is above 120%. The Federal Reserve is navigating a soft landing while global liquidity is tightening. In this environment, any new tax on a productive sector reduces the aggregate return on capital, which incentivizes capital to seek higher yields elsewhere. That could mean more capital flowing into crypto as a zero-tax jurisdiction, but it also means higher volatility as the market reprices the risk of regulatory intervention.
Regulation is the new liquidity constraint. The AI tax proposal is a form of regulation that directly impacts the cost of capital for AI-driven firms. In my experience as a Digital Asset Fund Manager, I’ve seen this dynamic play out with the Spot Bitcoin ETF approval in 2024. The market initially cheered the regulatory clarity, but the real impact was a 2.5% annualized basis trade that I executed across three exchanges. The low-risk arbitrage opportunity existed because the market overestimated the impact of regulation on liquidity. The same mistake is happening now with AI tax: the market underestimates how quickly capital can adapt.
Yield is the bribe for your risk. The AI tax debate is a risk that the market is not pricing correctly. I’ve already seen this in my own analysis of AI-agent crypto protocols. In March 2026, I identified a flaw in a leading AI-crypto protocol’s oracle reliability, causing a 12% loss in simulated user funds. The project had raised $100M on the promise of decentralized AI, but the oracle feed was centralized. The market ignored the technical flaw because it was focused on the narrative. The AI tax debate is similar: the narrative is about protecting workers, but the technical reality is that taxes on productivity are regressive in a deflationary technology environment.
The contrarian view is that the AI tax will actually accelerate the adoption of decentralized AI. If governments tax centralized AI services, firms will shift to peer-to-peer models that are harder to tax. This is the same dynamic that drove the rise of DeFi after the 2020 regulatory crackdown on centralized exchanges. The ICO boom of 2017 was a response to the SEC’s inability to regulate global tokens. The AI tax could be the catalyst for a new wave of decentralized AI infrastructure.
But I’m not convinced. The real risk is that the tax becomes a political tool to control AI development, not a revenue mechanism. The Chinese government already uses AI regulation to steer the industry. The US could follow suit. That would create a bifurcated market: regulated AI in the US, unregulated AI in offshore jurisdictions. The crypto market would then become the primary venue for unregulated AI services, increasing volatility but also increasing the risk of a systemic failure.