OpenAI Lehane 2028 Essay: A "Third Path" on AI and a Warning on Data Center Moratoriums

OpenAI is already campaigning for the 2028 presidential election, and the pitch is a mix of reassurance and warning. On November 4, 2026, TechCrunch reported that Chris Lehane, OpenAI's chief global affairs officer, published an essay arguing that 2028 could be a defining moment for AI. He wants candidates to write real policy for Americans' worries about the technology, not just attack it or attack data centers.

This post explains what is reported, what critics say, and what we could not verify. We could not find the essay on openai.com. It appears in The Prompt, OpenAI's global affairs newsletter on Substack, but the details below about its contents come from TechCrunch. Treat the specifics as reported, not as a reading of the original text.

TL;DR: the essay at a glance

What does the essay actually argue?

Lehane's central move is historical. He compares the possible economic effect of AI to Franklin D. Roosevelt's New Deal. TechCrunch notes that the original New Deal built worker protections and limits on corporate power, and that it is unclear how Lehane's vision would hold AI companies accountable. The 2028 essay speaks to people who will run for office.

The "compelling third path" is the key phrase. It positions OpenAI between two camps. One camp wants to pause or ban the technology and its buildout. The other wants to accelerate with no new rules. Lehane asks politicians to offer voters something they can see: a share of the gains.

The five policy priorities

TechCrunch lists the following, drawn from the essay:

• Tax code and ownership. Incentives for workers to learn AI skills, easier access to capital for entrepreneurs, and new approaches to employee ownership and profit sharing.
• Child safety. Age-appropriate protections and safeguards for AI companions used by minors.
• Education. An argument against keeping AI out of classrooms, with a warning about inequality for children in rural communities. TechCrunch notes that the claim that AI improves learning is unproven.
• AI safety. Enforceable safety standards, independent evaluations, public oversight, and the ability to restrict systems or shut them down.
• Data centers. Construction jobs and local tax revenue, plus the warning against a moratorium.

The safety item stands out. Shutdown ability and independent evaluation are demands that the safety community has made for years. Seeing them in a lab's political essay shows how far the debate has moved. TechCrunch also notes that recent studies found OpenAI and other labs have not published comprehensive plans for containing a model that escapes human control.

Why the data center line matters most

The Iraq war comparison is the sharpest sentence in the reported text. Lehane warns that backing a moratorium could jeopardize America's lead over China, and he likens that choice to the 2002 vote that authorized the Iraq war. The implication is that politicians who vote the wrong way now will be judged later.

Plot of land marked with a grid, illustrating a data center siting decision for the OpenAI 2028 election debate

The timing is pointed. On November 3, 2026, Ohio voters passed data center bans or limits in at least 14 municipalities, as we covered in our Ohio results post. TechCrunch adds that opposition to data centers was a major bipartisan concern in several midterm races, and that Democrats flipped control of Congress amid voter frustration over the cost of living.

OpenAI has fought this fight before. In August it sent a letter to Texas Governor Greg Abbott on a possible pause, which we examined in OpenAI's letter to Governor Abbott. Axios also reported that OpenAI may rebrand data centers as "AI compute campuses" after the midterms; see our post on that report. Read together, the 2028 essay is the third public step in one strategy: talk about jobs and tax revenue, change the vocabulary, and make a moratorium look like a risky vote.

What do critics say?

TechCrunch quotes Kate Brennan of the AI Now Institute. She says the rhetoric marks a shift from OpenAI's 2025 agenda, which focused on speed, fewer regulatory barriers and more infrastructure. She calls the essay the work of a company advocating for its own benefit, and says the proposals are largely predicated on AI's upside.

She also calls OpenAI's safety stances "disingenuous," citing its 2024 opposition to California's SB 1047. That bill would have required catastrophic-risk evaluations and shutdown capabilities, which are close to what the new essay says it wants. Brennan suggests other tools: a revised corporate tax rate, tax incentives for hiring people instead of automating jobs, and limits on outsized market power.

TechCrunch gives OpenAI's side too. It notes that OpenAI has more recently backed several state safety laws and said it would be open to independent evaluators embedded in the company. Last year it had argued that states could undermine American innovation. The outlet asked OpenAI for more detail on the proposals and its stances, and the company had not responded.

Claimed versus verified

How to check it yourself: look for the essay on OpenAI's own site or in Lehane's public channels and compare the quotes with TechCrunch's. Do not rely on summaries from aggregators.

How this fits the wider 2028 picture

OpenAI is not alone. Anthropic is hiring a political programs lead to brief 2028 candidates of both parties, as we reported in Anthropic's presidential engagement program. OpenAI itself opened an election war room ahead of the midterms; see what is confirmed about the war room.

A ballot box with a green check card dropping in, representing AI as a 2028 voting issue

On the other side are candidates who want to go further than any lab suggests. Michigan Senate candidate Abdul El-Sayed backs a pause and an FDA-style agency, as detailed in our El-Sayed explainer. Geoffrey Hinton has called for similar oversight in his Chornobyl warning. Inside the White House, a task force has 120 days to define the federal role on AI risk, as covered in the super intelligence task force post.

The pattern is simple. Labs now want to be seen as asking for rules. Critics want to know which rules, who writes them, and who enforces them. The gap between those two positions will define how the issue plays in 2028.

What builders and readers should take from this

You do not need to follow campaigns to feel this. Three practical points:

1. Safety terms are becoming policy terms. Independent evaluation, public oversight and shutdown capability may appear in state bills and procurement rules. Teams that ship agents should be able to explain their own evaluation and kill-switch story.
2. Data center siting is a live political risk. If you depend on a region for capacity or price, watch local ballots and permitting, not only chip supply.
3. Read lab policy essays as positions, not as facts. An essay from a company's policy chief states what the company wants. Check each claim against the law, the filings and independent researchers.

If you run AI agents in production and want to limit what they can do on their own, AgentBeam is the agent security platform from the explainx.ai team that stops AI agents before they take dangerous actions. It is one concrete way to turn "shutdown and oversight" from a slogan into a control.

What is still open

• The essay text. We have not read it, so quotes and priorities need a primary check.
• OpenAI's answers to TechCrunch on specifics: which standards, who evaluates, and what authority would restrict or shut down a system.
• Whether any 2028 candidate adopts the "third path" language.
• Whether the data center fight moves from local ballots to federal policy before 2028.

Related reading

• Ohio voters pass data center bans in 14 municipalities
• OpenAI's letter to Governor Abbott on a data center pause
• OpenAI's "AI compute campuses" rebrand report
• OpenAI's election war room: what is confirmed
• Anthropic's 2028 presidential engagement program
• Abdul El-Sayed's AI pause proposal
• White House super intelligence task force
• Source: TechCrunch on the Lehane essay

Details reflect reporting available on November 4, 2026. Check the original essay and later coverage for changes.