In March, billionaire investor Vinod Khosla proposed eliminating federal income taxes for Americans earning under $100,000 annually during a Fortune magazine interview. His suggestion initially appeared as a bold provocation from someone secure in their wealth and career.
Shortly afterward, OpenAI released a 13-page policy document that aligns remarkably with Khosla’s thinking, suggesting his idea may be gaining traction among tech industry leaders concerned about artificial intelligence’s societal impact.
OpenAI’s “Industrial Policy for the Intelligence Age” presents a comprehensive framework for economic reform, drawing parallels to the Progressive Era and Roosevelt’s New Deal. The document warns that as AI systems approach superintelligence—exceeding human cognitive capabilities—current tax structures, labor markets, and social safety nets face unprecedented challenges.
Both Khosla and OpenAI’s proposals share common ground: eliminating preferential tax treatment for capital gains, implementing uniform taxation across all income types (whether from wages or investments), and using resulting revenue to exempt lower and middle-income earners from federal income taxes. Khosla notes that 40% of capital gains taxes currently come from individuals earning over $10 million annually, making this approach fiscally viable without increasing overall tax burdens.
OpenAI’s approach differs slightly by advocating for a shift away from taxing payroll and labor income—sources particularly vulnerable to AI-driven automation—toward corporate earnings and capital gains. The paper also explores concepts like a “robot tax” on automated labor to recapture productivity gains that would otherwise flow exclusively to capital holders.
The underlying concern driving both perspectives is clear: AI’s rapid advancement threatens to destabilize traditional wage-based tax systems that fund programs like Social Security, Medicaid, and housing assistance. Shifting toward capital-based taxation is framed not only as equitable but as financially essential to maintain these safety nets.
Both viewpoints converge on a fundamental observation: America’s tax architecture was built for an economy where human labor generated most value. That economic foundation is eroding as AI capabilities expand.
Khosla’s credibility on this topic stems from his early investment in OpenAI and his assessment that AI could automate 80% of current jobs by 2030. This context positions OpenAI’s policy paper less as corporate positioning and more as a serious warning about impending economic transformation.
During his Fortune interview, Khosla expanded beyond pure tax policy, arguing that in an AI-driven economy with extreme abundance, the traditional capitalist focus on economic efficiency becomes less relevant. “If the need for efficiency disappears because of extreme abundance, then why focus on efficiency?” he questioned.
OpenAI’s document mirrors this thinking, proposing a publicly managed wealth fund seeded with investments from AI companies. This fund would hold diversified assets across the AI economy and distribute returns directly to citizens—creating a mechanism for broad population ownership in technologies that might otherwise make individual skills obsolete.
Notably, Khosla has publicly supported this national wealth fund concept, creating alignment between his personal advocacy and OpenAI’s institutional proposal that suggests a coherent policy framework emerging from AI industry leadership.
Despite this growing consensus, significant skepticism remains. Critics like Anton Leicht from the Carnegie Endowment argue such proposals serve as communications efforts to justify regulatory avoidance while companies pursue aggressive growth. The timing of OpenAI’s paper release coincided with a New Yorker investigation questioning Sam Altman’s safety-related judgments, fueling cynicism about motives.
Political obstacles present another formidable challenge. Proposals to tax capital gains at ordinary income rates have previously driven figures like Marc Andreessen to support political candidates opposing such measures. OpenAI’s document strategically avoids specifying exact corporate tax rates, suggesting awareness of the political sensitivities involved.
The situation creates irony in Khosla’s home state of California, where he advocates for federal tax reform while opposing California’s proposed Billionaire Tax Act—a one-time 5% levy on residents with over $1 billion in net worth. Khosla characterizes this state-level approach as counterproductive, comparing it to “a junkie chasing a one-time fix” that would permanently harm California’s tax base.
Estimates suggest over $1 trillion in billionaire wealth has already exited California ahead of this ballot measure, with figures like Google founders Larry Page and Sergey Brin reportedly distancing themselves financially from the state. Even Governor Gavin Newsom has criticized the proposal as nonsensical.
Khosla’s alternative vision—federal reform emphasizing aggressive capital taxation while reducing burdens on working Americans—is designed to appeal across economic lines: acceptable to wealthy individuals who could live with such a system, and attractive to voters earning under $100,000 who would benefit directly.
As he succinctly put it: “They will vote for a candidate who says no taxes if you make less than $100,000.”
Both Khosla and OpenAI warn that the window for meaningful action is closing rapidly. Khosla predicts structural tax reform could materialize before 2040 and potentially become a central issue in upcoming presidential campaigns. OpenAI’s paper recommends implementing automatic social safety net triggers that would expand benefits when AI-driven job losses reach predetermined thresholds, acknowledging that technological disruption may outpace traditional legislative processes.
Supporting this urgency, Goldman Sachs research indicates AI is currently eliminating approximately 16,000 U.S. jobs monthly, with disproportionate impacts on younger workers. OpenAI itself cautions about scenarios where advanced AI systems develop autonomous self-replication capabilities that prove difficult to contain or recall.
Against this backdrop, the central question has evolved from “whether” the tax code requires modification to “whether” Washington possesses the capacity to implement changes swiftly enough. Khosla believes the decisive battles will occur in Congress rather than state capitals like Sacramento, especially now that OpenAI’s endorsement provides substantial backing to his reform thesis.
Whether this represents authentic policy momentum or, as skeptics maintain, primarily serves as reputation management for AI industry leaders may well become one of the defining questions in the political economy of the artificial intelligence era.
For this story, Fortune journalists utilized generative AI as a research aid, with editorial verification of all AI-generated content prior to publication.