The attitude toward government in Silicon Valley was one of taxing us, but not controlling us. This seems to be changing in the AI era. OpenAI now wants more regulation, more taxation and even partial government ownership. Is this CEO Sam Altman’s political test from Satan?
OpenAI recently filed for an initial public offering with the Securities and Exchange Commission, an opportunity for early investors and employees to realize capital gains. But the company reportedly wants to reserve 5% of its shares for the “public” – not as private owners but for the government, perhaps a sovereign wealth fund.
Beijing takes stakes in major tech companies known as “golden shares” that give the Chinese Communist Party special voting power and veto rights over corporate decisions. OpenAI investors may not want more political control, but this will inevitably be part of a Faustian deal with the government. Why would Mr. Altman want to make his company a vassal of the state?
One reason for this could be to gain political advantage against competitors. OpenAI faces competition from companies like Anthropic, Google, Meta and XAI. Unlike its top competitors, OpenAI does not have a large pool of cash or profits on its balance sheet to finance the construction of its data center.
OpenAI plans to spend $600 billion on AI infrastructure by 2030, yet it is generating only about $2 billion in revenue per month. An IPO would raise cash, but not enough to fund its ambitions. Perhaps Mr. Altman hopes that giving the government a stake will reduce his company’s borrowing costs.
Government ownership may also provide regulatory support – such as accelerated approval for models, permits for data centers or federal contracts. A government stake could provide a built-in backstop, keeping OpenAI from becoming too big to fail. The government will not want to suffer a loss on its stake, even if it means keeping it alive as a zombie firm.
OpenAI favored a more pragmatic approach to AI regulation in its early years, but the company has been moving toward greater federal involvement. It recently endorsed a requirement that government developers review new “frontier” AI models before they are deployed. Regulation generally hurts upstarts, so OpenAI can hope to create a regulatory moat.
This spring it released a 13-page white paper, “Industrial Policy for the Intelligence Age,” which calls for expanding government control over AI and the private economy. It supports higher taxes on capital gains, corporate income, and “automated labor”. It also suggested that the government “encourage” a 32-hour working week without any loss in pay.
Some of these bad ideas may come from its Democratic appointees and advisers, including Aaron Chatterjee (Biden economist), LaFonza Butler (former California senator and union organizer), Anna Makanju, (Obama and Biden alumnus), and Ann O’Leary (former chief of staff to California Governor Gavin Newsom).
***
The problem is that any government ownership comes with high business costs. If company decisions become controversial at all they are subject to political interference. The French government has used its stake in automaker Renault to prevent the plant from closing. The Trump team made it a condition of approving Nippon Steel’s acquisition of US Steel because the government had a veto over corporate decisions, which it used to prevent the Illinois plant from closing.
AI firms will not be exempt from this political interference, no matter what promises Mr. Altman gets from the Trump White House, J.D. Vance or Mr. Newsom. Politicians are fickle, and they follow the demands of referendums and interest groups.
The American economy would suffer great losses if its most dynamic new industry became a subsidiary of the government. America has long been prosperous because it has escaped the trap of state capitalism. In this way the US defeated the keiretsu model of Japanese business and state ownership of much of Europe. China’s state-owned industries are the weakest part of its economy.
Mr Trump’s move to buy socialist stakes in US companies could become one of his worst legacies. “It’s almost become a partnership with the American public,” he said recently. “The American people can benefit from the success of AI, and they will like it better.” He does not mean the public, but a partnership with American politicians. Wait until President Newsom or JB Pritzker seizes that precedent.
Americans will benefit from AI through pharmaceutical breakthroughs, increased productivity, which will lead to increased wages, and myriad other ways that are hard to predict. They profited from the Internet without the government taking a stake in Google or, thank God, AOL. OpenAI can succeed or fail on its own.
