Among privately held firms, there’s no beating OpenAI when it comes to generating revenue from selling artificial intelligence. But startups selling the digital picks and shovels of the AI gold rush, including to OpenAI and other AI developers, are doing OK too.

The latest example is Vercel, a cloud service companies use to develop artificial intelligence and commerce applications and websites. The San Francisco-based firm, which also offers website hosting, crossed $200 million in annual recurring revenue last month, up from $100 million in annual recurring revenue 15 months earlier, according to Jeanne DeWitt Grosser, the company’s chief operating officer.

That implies it is generating $16.7 million in subscription revenue per month, which is more impressive than non-recurring annualized revenue figures other software and AI providers typically tout.

The company declined to share details of its cash flows or profitability. (“All the money we receive from our customers from what we sell, as well as the funding from investors, we use to invest” into the business,” said Vercel CFO Marten Abrahamsen.)

Still, the revenue milestone shows how more venture-backed AI companies are running on Vercel’s software, aiming to improve the speed and performance of their apps. The company’s customers include OpenAI, which uses Vercel to host its website homepage; Decagon, which develops automated customer service agents; AI video providers Runway and Pika, AI notetaking tool Granola, and traditional enterprises such as Stanley Black & Decker and Under Armour, a company spokesperson said.

Revenue at other OpenAI vendors such as Databricks, a database provider; Mercor, a staffing firm; and Turing, which helps refine AI models after they are initially developed, has also surged.

Vercel was founded in 2015 and competes with the likes of Cloudflare and Amazon Web Services, whose stocks have soared during the AI boom of the last two and a half years. (Much of Vercel’s services run on AWS servers.)

The Information included Vercel in its list of 50 promising startups in 2020, when the company was focused on e-commerce firms. Last year, as it gained more traction with AI customers, Vercel raised $250 million at a $3.25 billion post-money valuation.

The 650-person startup last year also appointed Stripe CFO Steffan Tomlinson to join its board of directors, which now has six members.

In 2023, Vercel launched v0, an AI coding and programming assistant that helps customers write code, build websites and offer guidance on how to accomplish technical tasks. Recently, the product added features to help with web development tasks. Launching v0 put Vercel in competition with some of its AI customers, notably OpenAI. That’s become a common dynamic in the AI field.

Grosser said that just as legal AI startups like Harvey aim to beat OpenAI on legal tasks, “we’ll probably be the best at producing the future’s web applications.” (For what it's worth, Harvey and Vercel are both OpenAI customers, too).

While OpenAI’s coding assistant needs to “write all sorts of types of code,” Grosser said, Vercel’s version “wants to be excellent at getting your front-end application onto the web.”

The AI-Powered Future of Influencing

AI could create new ways for fans to interact with content creators, such as letting sports video gamers talk to an AI replica of their favorite players, said Ashwin Desai, who represents athletes for OneTeam Ventures, which helps athletes license their intellectual property.

Using AI, “you could watch an NFL game with a [AI version of the] quarterback to explain the playbook to you live, in real time. You could have a fan in Tokyo watching their favorite hitter from the Rangers, explaining a technique to them,” Desai said at The Information's Future of Influence conference in Los Angeles Tuesday.

But key obstacles remain, said other conference participants.

For one, creators are nervous that if they use AI tools to replicate themselves, the AI companies and social media sites might gain the rights to their content.

To reassure creators that their content won’t get regurgitated by its AI models, Roblox uses software and staff that moderate content and manage intellectual property rights to identify and remove AI-generated content that may violate copyright, said Aliana Miller, head of influencer relations from Roblox. “We’d rather over-moderate than have anything go wrong,” she said.

Roblox has set up a content moderation team to prevent its AI systems from generating copyrighted content, said Aliana Miller, head of influencer marketing at Roblox.

There’s also a concern that crafty users would be able to get AI-generated replicas of influencers to say nasty things, Desai said.

He’s right to feel that way: AI developers have yet to create a model that users cannot hack.

Finally, using AI tools “needs to be 10 times easier than it currently is” before creators widely adopt them, said Tricia Biggio, co-founder and CEO of Invisible Universe, which uses AI to generate shortform videos. To accomplish that, companies will need to make templates for common applications, such as 30-second TikTok ads, and incorporate the latest AI models into their products so the burden is not on users to keep up with the new AI products released every week, Biggio said. (This version has been updated to clarify how Roblox protects creators’ content.)—Rocket Drew

Policy Watch

Hundreds of U.S. state legislators signed a letter Tuesday opposing the Congressional effort to block lawmaking on AI at the state level for 10 years. The preemption measure, part of the “one big beautiful” budget reconciliation bill passed by the U.S. House last month and now being considered by the Senate, has support from Senator Ted Cruz, among others.

The letter, signed by 130 Republicans, 129 Democrats and 1 independent legislator from across the country, was shared with Congress.

“Legislation that cuts off this democratic dialogue at the state level would freeze policy innovation in developing the best practices for AI governance at a time when experimentation is vital,” they wrote. “Barring all state and local AI laws until Congress acts threatens to setback policymaking and undermine existing enforcement on these issues.”

Their demands follow a similar letter that 40 state attorneys general sent to Congress last month.

The preemption attempt has also come under fire from Anthropic, which has been lobbying Congresspeople to oppose the budget reconciliation bill. That puts Anthropic at odds with OpenAI, which has asked the federal government to preempt state AI laws.

The state preemption is expected to face scrutiny in the U.S. Senate, which prevents unrelated measures from creeping into budget reconciliation bills. And Elon Musk called the bill a “disgusting abomination” on Tuesday. President Trump has asked the Senate to pass the bill by July 4.—Rocket Drew

Here’s what else is going on…

Deals and Debuts

See The Information’s Generative AI Database for an exclusive list of private companies and their investors.

Cybernetix Ventures, a venture capital firm, is raising $100 million for its second fund, which will invest in robotics and physical AI companies.

Aerones, which uses robotics to inspect and maintain wind turbines, raised $62 million in a funding round led by Activate Capital and S2G Investments, with participation from Carbon Equity, Overlap Holdings, Lightrock, Blume Equity, Metaplanet, Change Ventures and Extantia.

Speedata, which develops chips for analyzing large datasets, raised $44 million in a Series B funding round, from investors including Walden Catalyst Ventures, 83North, Koch Disruptive Technologies, Pitango First and Viola Ventures.

Yoshua Bengio, one of the “godfathers of AI,” raised $30 million in donations from Jaan Tallin, Eric Schmidt, Open Philanthropy and the Future of Life Institute for his new nonprofit, LawZero, which will research how to develop AI systems that don’t deceive their users.

Treefera, which uses AI to provide companies with data about their supply chains, raised $30 million in a Series B funding round led by Notion Capital, with participation from Albion VC, Triple Point and Twin Path Ventures.

Ciroos, which develops an AI-powered engineer that can autonomously investigate software bugs, raised $21 million in a funding round led by Energy Impact Parters.

Console, which uses AI to automatically respond to employees’ IT questions, raised $6.2 million in seed funding from Thrive Capital.

Meta Platforms is buying the power output from Constellation Energy’s Clinton, Illinois nuclear power plant under a 20-year deal announced on Tuesday.

The U.S. Commerce Department renamed its AI Safety Institute to the Center for AI Standards and Innovation.

Google’s NotebookLM now allows users to share notebooks publicly, allowing others to ask questions and receive AI-generated answers about the notebooks.

OpenAI’s Codex coding software now has access to the internet.

Recommended Newsletter

the recommended newsletter text for the week.

What We’re Reading

TSMC expects record 2025 profit despite US tariffs, currency volatility

DeepSeek may have used Google’s Gemini to train its latest model

Walmart is supercharging revenue — but with fewer workers

Upcoming Events

Tuesday, June 17 — With AI Agents, Cooperation Could Prevail…Or Not

Don't miss The Information’s Kevin McLaughlin and executives from Google Cloud, Outshift by Cisco and Decagon as they examine whether we're heading toward a unified Internet of Agents or fragmented ecosystems with proprietary boundaries.

More details

Tuesday, September 30 — AI Agenda Live NYC: The Next Wave

Save the date for The Information's next AI summit, AI Agenda Live: The Next Wave. AI Agenda Live will help attendees make sense of what’s coming in the field, from new research breakthroughs to bleeding-edge AI applications.

More details

Thank you for reading the AI Agenda Newsletter! I’d love your feedback, ideas and tips: stephanie@theinformation.com.

If you think someone else might enjoy this newsletter, please pass it forward or they can sign up here.