Andrew Yang went on CNBC’s “Squawk Box” on Thursday and said the quiet part out loud: his company is already swapping out entry-level analysts and engineers for AI, and he thinks the federal tax code should treat that automation like a billable line item. In his telling, the fastest way corporate AI investments “pay for themselves” is by cutting headcount, especially junior hires, so he wants lawmakers to rethink payroll taxes and make automation shoulder more of the cost. The stark pitch has shoved the AI, jobs, and tax debate back into the national spotlight.
On air, Yang did not mince words. “The easiest people to fire are the people you haven’t hired yet,” he said. Asked whether companies are in fact replacing junior analysts and engineers with AI, he answered “100%,” according to a transcript of the interview. As reported by 24/7 Wall St., Yang argued the remedy is to tax AI and robots while cutting the tax load on human workers. His appearance is included in CNBC’s “Squawk Box” recap and its companion podcast, Squawk Pod.
Yang’s core policy pitch is to move the tax burden off hiring people and onto deploying automation. Companies already pay payroll taxes and benefits tied to human workers, he noted, while AI systems do not trigger those obligations. He frames an AI or robot levy as a way to fund retraining while also financing payroll tax cuts that would make hiring humans comparatively cheaper. Fortune has previously covered Yang’s broader argument for “stopping taxing labor” and asking automation to carry more of the fiscal weight.
The math behind the pitch
To back up his case, Yang points to the tidal wave of spending by the big cloud and AI infrastructure players. Analysts and company guidance suggest combined hyperscaler AI outlays could approach roughly $1 trillion, according to reporting by 24/7 Wall St. Microsoft told investors that its AI business has already reached a $37 billion annual run rate and that it recorded about $30.9 billion in capital spending in its most recent quarter, figures the company detailed in its April earnings release from Microsoft. Meanwhile, NVIDIA’s record first-quarter revenues and a roughly $91 billion revenue guide for the current quarter highlight the scale of demand for computing power that Yang warns is ultimately being financed by labor cuts, according to NVIDIA.
Industry reaction and the token-tax idea
Some AI leaders are not instantly dismissing the idea that the sector should chip in more. Anthropic CEO Dario Amodei has repeatedly warned that advanced AI could hollow out entry-level white-collar roles and has urged policymakers to consider new revenue tools to cushion that blow, a line of thinking covered by Axios. Proposals floating around policy circles range from per-usage or per-token taxes on AI activity to broader corporate contributions tied to automation. The arguments now turn on familiar questions: who really ends up paying, how to avoid kneecapping US competitiveness, and whether any of this will arrive before the job losses do.
What lawmakers are watching
On Capitol Hill, lawmakers and staff are already gaming out responses. Ideas on the table include workforce training funds, targeted tax credits, and narrower levies meant to help workers and communities absorb AI-driven disruption. The Washington Post has outlined several of the options under discussion, along with the tradeoffs for workers, employers, and the broader US competitive landscape.
Yang’s CNBC cameo boils that policy sprawl into a blunt question: if automation is taking over entry-level jobs, who pays for the transition. Expect more talk show hits, white papers, and hearings as tech leaders and lawmakers try to turn that question into something that looks like a law.
