The future of AI is dividing the Democratic Party, as 2028 hopefuls and party leaders stake out clashing positions in what's already shaping up as a major policy battle in the primary.

Why it matters: If Democrats win back the White House in 2028, where they land on AI will shape how the country approaches the new technology — with big consequences for the economy and workers.

The big picture: Two main arguments are now playing out within the Democratic Party:

1. Democrats should embrace AI to beat China and capture the jobs that come with the many data centers AI companies are building. (The Trump administration has a similar argument, though most Democrats say the White House has given AI companies too much latitude.)
2. Democrats should slow down and push for more regulation of the AI industry, given its potential power to displace millions of workers and the volume of natural resources being sucked up by new data centers to power the technology.

Driving the news: Swing state governors such as Pennsylvania's Josh Shapiro and Michigan's Gretchen Whitmer — both potential 2028 presidential contenders — have welcomed AI companies.

• "The future of AI will run through Pennsylvania — and together, we will defeat China in the battle for AI supremacy," Shapiro declared in September. He's touted AI investments from companies such as Amazon in his state.
• In October, Whitmer boasted about a multibillion-dollar OpenAI Stargate project: "Today, we won the largest economic project in Michigan history," she said.
• She said the project is expected to "create 2,500 good-paying union construction jobs, more than 450 permanent high-skill, high-paying jobs on site, and 1,500 more in the community."

Some labor unions, especially those in the building trades, have also partnered with the AI industry and applauded the potential jobs it could bring to their members.

• Brent Booker, the president of the Laborers' International Union of North America (LIUNA), told Axios that "data centers are coming. The country should plan responsibly for them — that means that they are constructed with union labor and developed in the context of expanding energy supply for the communities impacted."

The other side: Some progressive Democrats eyeing the 2028 race, such as New York Rep. Alexandria Ocasio-Cortez and California Rep. Ro Khanna, have been much more critical of AI companies.

• They've called for significant government regulations and new policies to protect workers who may be displaced by the technology.
• Vermont Sen. Bernie Sanders called for a "moratorium on the construction of data centers that are powering the unregulated sprint to develop and deploy AI," as he posted on X.
• Asked whether Ocasio-Cortez supports a moratorium, her spokesperson declined to say but pointed to her remarks last week at a congressional hearing where she argued that "there are real public health consequences for the continued failure of regulating or there being congressional action in the AI space."
• She added: "AI chatbots are causing children to take their own lives, as AI data centers are skyrocketing electricity costs and polluting local communities."

While some unions have warmed up to AI, others are fighting it because they believe it will replace their members' jobs.

• The Teamsters have called for all self-driving trucks to have a human operator in the vehicle — setting up a clash with AI companies.
• Some Democrats, including Pennsylvania Sen. John Fetterman, have backed the union's demands, while others have been quiet on the issue.
• Asked whether Shapiro agrees with Fetterman, the governor's spokesperson declined to say.

Declines in tech stocks? Healthy movement. Local officials stopping data centers? Prevents overbuild. Valuations high? Well, they deserve to be.

Why it matters: Every risk for the AI trade is framed as a positive by Wall Street bulls who are adamant we are in the early stages of the AI revolution.

Between the lines: Let's run through the threats to the AI trade and then the bull case Wall Street is attaching to each of those.

1. Stock valuations are too high.

• Wrong! "Nvidia is being valued like a mediocre growth stock. It's obviously not," Vivek Arya, a senior analyst at Bank of America, mentioned during a 2026 semiconductor outlook call.
• When you consider the 50% sales growth expected for Nvidia, it is actually undervalued, several strategists who spoke with Axios said.
• Nvidia trades at 23 times earnings, the second-lowest ratio among the Magnificent 7. Cisco traded at 140 times ahead of the dot-com bubble.
• Tech stocks are also not ferociously rallying the way they once did, which suggests investors are choosier, which is "healthy," Gil Luria, senior analyst at D.A. Davidson, told Axios.

2. Demand for AI will not materialize.

• Nope. OpenAI's weekly ChatGPT users just hit a record of 800 million.

3. Data centers are getting overbuilt.

• Great news! Local officials and state lawmakers are going to put a cap on data center growth as they start to push back on that infrastructure from both sides of the aisle.
• "That's a very healthy natural governor in this market," Jeffrey Favuzza, a tech strategist at Jefferies, told Axios.

4. Too much money is being spent.

• Sure, these top tech firms are on average spending about two-thirds of their cash flow on AI infrastructure. Traditionally they spend very little.
• But "that cash was just collecting dust" before the rollout of AI, Arya at Bank of America said. In that view, AI spending is a way to create future growth and sales opportunities necessary to survive the AI revolution.
• More capex is bullish. Arya expects at least $1.2 trillion in capex by 2040.

5. Is AI ever going to make money?

• That comes down to monetization, likely in the form of advertising and probably coming first from OpenAI, which is widely believed to come early next year.
• New product releases are also expected to deliver more opportunities to monetize.
• Favuzza is also watching how OpenAI works with enterprises: The company recently hired former Slack CEO Denise Dresser as its chief revenue officer, indicating a push to sell to more businesses.

Reality check: Reframing negative signals as positive is "classic financial sentiment," said Paul Kedrosky, a venture capitalist who believes we're already seeing signs of the AI bubble bursting.

The bottom line: Throw out your business school investing textbooks. The rules of markets are changing in the face of the AI revolution, strategists argue.

• Google parent Alphabet will acquire data center company Intersect for $4.75 billion in cash as AI firms continue their aggressive hunt for power to serve energy-hungry data centers. (Axios)
• AI is already reshaping childhood and the safety guardrails aren't there. (Axios)
• Flock's AI-powered surveillance cameras are livestreamed on the internet and accessible without a password or login, according to an investigation by 404 Media.

Thank you for reading (all the way to the end!), responding and attending our events this year.

• Subscribe to Mike Allen's AM newsletter to read Ina's predictions for AI in 2026 while we're on vacation.