Fortune favors the lobbyists

TL;DR: The playbook worked for crypto. Now AI is running it back, with companies spending hundreds of millions to shape public opinion, weaken regulation, and ultimately send one message: The AI industry is here to stay, so you might as well like us.

What happened: A few years ago, crypto ads dominated the Super Bowl, with spots featuring A-list stars like LeBron James and Matt Damon (who implored us to be brave by buying crypto). They aged poorly. Now it’s AI’s turn, and the industry is using a similar spending spree to capture hearts, minds, and politicians. Anthropic just made its Super Bowl debut with a promise that Claude won’t have ads (who knows for how long)—a stark philosophical contrast to OpenAI. Speaking of, OpenAI spent about $14 million on a game night ad last year, and it’s shelling out for another 60-second spot for this Super Bowl.

But AI companies aren’t just dropping millions for TV’s biggest night. According to iSpot, an ad measurement firm, OpenAI, Meta, Google, and Microsoft spent a total of $382.8 million on traditional TV ads promoting AI last year. There’s a broader PR blitz as the industry competes not just to build sprawling data centers, but also for greater AI adoption—or at least acceptance—among consumers. Meta just spent a cool $6 million on TV ads touting job creation as a virtue of AI data centers, which have faced fierce local backlash. Amazon funded similar ads in Virginia, the New York Times reports. Both Microsoft and Google recently announced they’d be willing to pay more for the higher utility bills their data centers cause (how magnanimous).

Just the tip of the iceberg: Shaping public opinion not only helps stem loud public criticism of job displacement related to AI as well as its environmental impacts—it can also determine the outcome of policy pushes and AI infrastructure buildouts. Beyond tear-jerking ads suggesting data centers will save your small town, the AI industry has raised a massive $125 million for the upcoming midterm elections. The primary super PAC, Leading the Future, is essentially the AI industry’s answer to Fairshake PAC; the crypto super PAC and two aligned groups spent roughly $290 million in the 2024 elections—almost half of all corporate contributions that cycle. Leading the Future is backed by some familiar tech names: OpenAI president Greg Brockman and his wife ($25 million), Andreessen Horowitz ($12.5 million from each of its two co-founders), and Palantir co-founder Joe Lonsdale ($250,000), among others.

Leading the Future is co-led by Josh Vlasto, a former Fairshake adviser, and so far the spending has been bipartisan. Its Democratic arm has opposed New York Assemblyman Alex Bores, sponsor of an AI safety bill, in his New York congressional bid, while its Republican arm has backed Chris Gober, a Trump-aligned, pro-innovation conservative running in a Texas congressional primary.

What this money could buy: A lot less regulation. There were well over 700 AI bills floating through state legislatures as of last spring. Around that time, OpenAI called for a single, nationwide set of rules for the most powerful models—ones that would rely heavily on voluntary reporting rather than strict enforcement and also override tougher state laws. President Donald Trump has largely been on the same page, signing an executive order that tries to preempt states from setting their own AI standards. (Whether it’s enforceable is another question.)

One category of industry-friendly bills: “right-to-compute” laws that treat access to computing power like a civil liberty, with some politicians comparing it to the First Amendment. These bills raise the legal bar so high that governments will have to clear strict constitutional hurdles before they can limit how AI systems are built or used.

The takeaway: AI companies tell us they’re building the future. But all this spending shows that their vision of the future isn’t just about improving models and constructing physical infrastructure—it includes using reputational gloss and political influence to scale fast and ask questions later. —WK