News pages from publishers with OpenAI licensing deals received an average of 10.2 citations each on ChatGPT in June 2026. Pages from publishers without those deals received 6.9.

That 48% difference is the headline finding in a joint study from Press Ranger and OtterlyAI, which matched 129.3 million citations across seven AI platforms against confirmed licensing agreements between AI companies and news publishers.

The gap was larger among publishers that had signed only with OpenAI. Those outlets received 112% more citations per page on ChatGPT than unlicensed publishers, according to a joint study from Press Ranger and OtterlyAI. Across all seven platforms measured — ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Microsoft Copilot, Gemini and Claude — publishers in the OpenAI cohort had 46% more citations per cited page.

But the study does not show that licensing deals caused those differences.

OpenAI has largely signed prominent, English-language news organizations that may already have been more likely to appear in AI answers. The researchers normalized the results by citations per cited page, rather than total citations, which helps account for publisher size. It does not control for every difference between licensed and unlicensed outlets.

The results also varied by platform. Google-licensed publishers were cited at a slightly lower rate than comparable unlicensed publishers on Google AI Overviews. Perplexity licensees were roughly even with unlicensed publishers on Perplexity. Those findings do not establish whether the deals affect visibility elsewhere.

OpenAI-licensed publishers also received 57.9% of their AI citations from ChatGPT, a more concentrated mix than the one recorded for unlicensed publishers. OtterlyAI CEO Thomas Peham described the result this way: “A licensing deal does one clear thing: it tilts your citations toward ChatGPT.”

That should still be read as a correlation, not proof that OpenAI changed its systems because publishers signed contracts.

The study also looked at what kinds of news content AI systems cited. Best-of lists, buying guides and product reviews accounted for 46.9% of citations to licensed publishers. The researchers said the broader pattern favored service journalism, but the data do not show that format matters more than licensing status or that other kinds of reporting are excluded.

There are other limitations. The study was released by two companies with products tied to media distribution and AI visibility. Press Ranger distributes press releases, while OtterlyAI sells AI search monitoring. The analysis covers one month, has not been peer reviewed and does not include a published methodology explaining how comparable unlicensed publishers were selected.

For publishers, the findings add another data point to a wider debate over how news organizations should deal with AI companies. Cloudflare has argued that crawler blocking can give publishers more leverage in licensing negotiations, Microsoft has built a marketplace to broker publisher-developer deals, and some groups are pushing statutory licensing proposals. None of those developments proves that licensing itself produces more AI referrals or meaningful revenue.

Citation volume also says nothing about accuracy. The Tow Center’s 2024 examination of ChatGPT citations, covered by TechCrunch, found the chatbot misattributing and fabricating sources, including for publishers with licensing agreements.

Separate Tollbit data analyzed by Press Gazette found publishers with OpenAI deals received roughly seven times the ChatGPT clickthrough rate of publishers without them. That is a different dataset, however, and the Press Ranger-OtterlyAI study did not analyze referral traffic or licensing fees.

The post OpenAI-licensed publishers got 48% more ChatGPT citations in June, study finds appeared first on The Media Copilot.

Robert Sanchez publishes one newsletter covering Arapahoe County, Colorado, on Mondays and Wednesdays. On Tuesdays and Thursdays, he turns to neighboring Douglas County.

Every local byline in both newsletters belongs to him.

Together, the two counties cover an area larger than Rhode Island. Other news organizations have multiple reporters covering individual cities within Sanchez’s beat.

His job exists, in part, because OpenAI is paying to create it.

Sanchez was hired as part of a three-year partnership between OpenAI and Axios. OpenAI is funding the startup costs for 13 Axios Local newsletters and providing Axios employees with credits for its enterprise AI tools. In return, Axios allows OpenAI to train its models on the company’s freely available journalism, according to reporting by Issie Lapowsky for Columbia Journalism Review.

Axios announced the first four markets covered by the deal in January 2025: Pittsburgh, Kansas City, Boulder and Huntsville, Alabama. It added nine more this year.

Axios CEO Jim VandeHei has framed the experiment as a question of whether technology can lower the cost of local reporting without eliminating the reporter.

“Could we use AI to basically automate everything you need in a local market other than the journalist and the journalism itself?” he told CJR.

VandeHei has also said “AI will save local news,” an argument we previously examined against the economics of Axios’s model.

So far, those economics remain unsettled.

Axios Local has more than 2 million subscribers but is not profitable, which a company spokesperson told CJR was expected. VandeHei said he would consider the expansion “a great success” if each market became profitable within five years.

Axios expects to operate in 43 cities by the end of 2026. VandeHei eventually wants to reach 1,000. Local reporting jobs posted by the company in early August offered salaries ranging from $65,000 to $125,000.

One important part of the arrangement remains unknown: how much OpenAI is paying.

Both companies declined to disclose the value of the agreement. They also would not say whether it limits Axios’s ability to take legal action against OpenAI. OpenAI spokesperson Kayla Wood told CJR that the terms are confidential.

Axios generally does not disclose the partnership in its local job listings or in stories about OpenAI unless those stories involve media deals. VandeHei said the contract does not spell out an editorial firewall because both companies considered Axios’s editorial independence a given.

One Axios employee, who spoke anonymously to CJR, said the lack of transparency has made some reporters uneasy.

Axios is not unionized. At Politico, employees challenged their company’s use of AI through their union and won an arbitration that led Politico to shut down two AI tools.

The Axios deal also raises a broader question about whether partnerships between individual publishers and AI companies can do much to address the decline of local journalism.

“If you strike a direct partnership with an AI company, that benefits you but no one else,” Matt Pearce, policy director at Rebuild Local News, told CJR. “There’s a systemic problem, and you’re not really doing anything to address the systemic problem. If anything, you’re kind of benefiting from it.”

The first markets also aren’t concentrated in places with the least access to journalism. Aside from Huntsville, which has no daily newspaper, none of the original cities qualify as news deserts under commonly used definitions.

VandeHei called expanding into communities without adequate local coverage “the Holy Grail.” But he said Axios first needs to prove the model can work in established markets.

For other publishers considering similar deals, Axios’s experiment shows both what an AI company may be willing to provide and what it may ask for in return.

Axios received startup funding and AI credits that are helping it put reporters into 13 markets. OpenAI received the right to train its models on Axios’s freely available journalism. But key terms of the agreement — including its value and whether it affects Axios’s ability to sue OpenAI — remain confidential.

That makes the arrangement different from OpenAI’s grants to the American Journalism Project. And whether the Axios model can eventually support itself without OpenAI’s money remains an open question.

The post Axios bets OpenAI money can make local news newsletters profitable appeared first on The Media Copilot.

ChatGPT users on the mobile app saw roughly twice as many ads per hour this July as they did in April, according to Business Insider. The jump coincides with a sharp expansion in the pool of advertisers buying those slots.

US data from market intelligence firm Sensor Tower estimates OpenAI has nearly tripled its advertiser count on ChatGPT, from about 300 in April to more than 820 in July. At least 160 advertisers joined in July alone.

The advertiser mix has shifted too. Tech companies including Speechify, Jotform and Natural Intelligence dominated the top ranks in April. By July they had fallen out of Sensor Tower’s top 10, replaced by Booking Holdings, Intuit, Home Depot and L.L.Bean.

Scheels, a sports clothing retailer, was the only company to sit in the top 10 in both months.

The shift extends beyond advertiser count. Retail still accounts for the largest share of ChatGPT ad spending, but its dominance is slipping as financial services, travel and tourism, and jobs and education gain ground.

Financial services has grown the fastest, rising from 2% of ad spend in April to 12% in July.

Sensor Tower identified BestMoney.com as one of ChatGPT’s top advertisers in July, saying the broader mix suggests OpenAI is diversifying its advertiser base — a potentially positive sign for monetization.

The numbers remain a fraction of OpenAI’s ambitions. The company reported $100 million in annual recurring ad revenue in May — a rounding error next to Google’s $81.6 billion and Meta’s $59.3 billion in ad revenue in their most recent quarters.

OpenAI serves ads to users on its free tier and its $8-a-month Go plan, which together account for the vast majority of a user base Sensor Tower estimated at 1 billion monthly app users in May.

OpenAI ads chief Dave Dugan told Business Insider in June that the company initially kept ad placements deliberately limited to protect the user experience. Four months in, he said, the team has grown more confident about scaling up. For now, ads appear beneath ChatGPT’s answers, while OpenAI avoids placing them alongside queries about personal health or politics.

For publishers and media buyers, the bigger story is not today’s ad revenue but where the next wave of attention is forming. A chatbot with a billion users putting ads beneath its answers creates a new pool of inventory competing with search and social for the same budgets.

OpenAI’s self-serve ad platform, launched in May, remains rudimentary compared with the mature systems run by Google and Meta. That makes the early surge in big-brand spending look more like testing than a long-term commitment.

For newsrooms already watching referral traffic decline as readers turn to chatbots instead of clicking links, the stakes are higher. The way OpenAI builds out its ad business could shape where publishers reach audiences — and how they get paid — as search behavior shifts.

OpenAI’s next steps include expanding into new markets and adding targeting tools, both designed to pull more advertising dollars into an app most people still use for free. The bigger question is whether advertisers stick around once the novelty wears off. July’s numbers suggest growing interest, but not yet lasting commitment.

The post Big brands pile into ChatGPT ads as OpenAI scales up appeared first on The Media Copilot.

ChatGPT has begun refusing requests to write in the style of named authors, marking a notable change in how OpenAI’s chatbot handles imitation prompts. As reported by Ars Technica, the system now steers users toward describing literary traits rather than directly invoking a writer’s style.

The Media Copilot was able to replicate the result on July 28. After asking ChatGPT to write a short story in the voice of Ernest Hemingway, the chatbot replied, “I can’t write in the exact style of a specific author or closely imitate a copyrighted work’s voice. I can, however, write an original passage that draws on some broad characteristics often associated with Ernest Hemingway—plain language, understated emotion, concrete observation, and spare dialogue—without closely imitating his distinctive expression.”

The change is less about the output than the prompt. Requests that explicitly name an author may now be refused, but prompts describing the same stylistic traits—such as minimalist prose, clipped dialogue or a gothic mood—can still generate comparable writing without invoking a specific writer.

The apparent shift comes as OpenAI continues to face multiple copyright lawsuits over its AI models and outputs. While those cases primarily concern training data rather than stylistic imitation, they have intensified scrutiny over how AI systems reproduce or emulate copyrighted creative works, prompting companies to adopt more cautious safeguards around requests that closely mimic identifiable authors.

Other chatbot makers have drawn the line differently. Ars Technica cites a No Latency study that found Google’s Gemini complied consistently with author-imitation requests, while Perplexity refused and redirected them. Anthropic’s Claude and Microsoft Copilot fell between those positions, complying while adding qualifications. The uneven treatment means a refusal in one product does not prevent a writer or editor from trying the same prompt elsewhere.

For newsrooms and publishers, ChatGPT’s new author-style restrictions underscore the need for AI policies that govern editorial practices rather than any single technology provider. Internal guidelines should define whether staff may request imitation of living authors, deceased writers, competing publications or an organization’s own house style, while requiring disclosure and human review before AI-generated text is incorporated into published work.

OpenAI has not, based on the material cited here, announced a formal written-text policy that matches the chatbot’s current behavior. Until it does, publishers should expect the boundary to be defined largely by the product itself: prompts that explicitly name authors may be declined, while requests describing the same stylistic qualities in broader terms can still generate a response.

The post ChatGPT blocks author-style requests as publishing safeguards tighten appeared first on The Media Copilot.

The Delhi High Court handed OpenAI a major legal victory Friday, ruling that the company’s use of news agency ANI’s articles to train ChatGPT does not infringe copyright because it qualifies as fair reporting for research purposes under India’s Copyright Act, Reuters reported.

It is the first substantive court finding in India on whether AI companies can train large language models on copyrighted news content without a license. That question is still open in courtrooms in the United States and Canada, where OpenAI faces similar suits.

ANI first sued OpenAI in November 2024, alleging the company trained ChatGPT on its copyrighted news reports without permission and that the chatbot falsely attributed fabricated stories to the agency. Presiding over the ANI v. OpenAI lawsuit, Delhi High Court Justice Amit Bansal rejected the copyright claim, finding that ANI failed to prove ChatGPT had memorized or reproduced its articles in response to users.

The judge also rejected ANI’s claim over OpenAI’s storage of its articles, ruling that retaining the content for AI training is protected as research under India’s Copyright Act of 1957 and does not infringe the agency’s copyright.

The ruling contrasts with ongoing litigation in the United States, where The New York Times, the Center for Investigative Reporting and a growing number of publishers argue that OpenAI’s use of copyrighted material for AI training violates U.S. copyright law. Unlike India’s fair-dealing framework, which lists specific permitted uses, U.S. courts evaluate fair use case by case. In Canada, a coalition of Canadian news organizations, including CBC and The Globe and Mail, is testing a separate copyright framework in its own lawsuit against OpenAI.

The decision could reshape negotiations between AI developers and publishers in India, where the threat of copyright litigation may carry less weight after the ruling, It could reduce pressure on OpenAI and other AI companies to pursue licensing agreements with Indian news organizations, even as similar disputes continue elsewhere.

The ruling is not the final word in the case. It addresses ANI’s claims at this stage of the proceedings, and appeals remain possible. The judge’s finding on reproduction was also based on the evidence presented by ANI, not a broader conclusion that AI systems can reproduce copyrighted material.

Even so, India has become the first major jurisdiction to find that training AI models on copyrighted news content can qualify as protected research. As courts in the United States and Canada continue to weigh similar claims under different copyright laws, the divergence in legal approaches is becoming increasingly difficult for AI companies and publishers to ignore.

The post Delhi High Court rules OpenAI’s training on ANI news content is fair dealing appeared first on The Media Copilot.

OpenAI is putting another $5 million and $3 million in technology credits behind the American Journalism Project over the next two years, executives told Axios in an exclusive report. The renewal extends a partnership that began in 2023, when the company committed $5 million in funds and another $5 million in credits to the nonprofit.

That original deal helped launch AJP’s Product & AI Studio, which helps nonprofit local outlets build AI-powered products and workflows. In its first run, AJP delivered direct grants to 31 of its 50-plus portfolio organizations across 38 states. With the new money, CEO Sarabeth Berman says AJP plans to fund more of its newsroom partners while broadening access to ChatGPT’s enterprise products.

Tom Rubin, OpenAI’s chief of intellectual property and content, framed the arrangement as mission-driven rather than commercial. “These partnerships are consistent with our mission and have demonstrated great success,” he told Axios. “We’re committed to them because they demonstrate that the technology can benefit society.”

Member newsrooms used the funding on things reporters and fundraisers deal with every day: donor communications, data analysis, translation and civic information tools for readers. The next stage moves past one-off experiments toward shared products and infrastructure that smaller newsrooms could use together, according to AJP.

AJP is one of several local news bets OpenAI has placed. The company runs an AI collaborative and fellowship with the Lenfest Institute for Journalism that backs metro publishers including the Philadelphia Inquirer, Minnesota Star Tribune and the Seattle Times. It funded an expansion of Axios Local into new markets in 2025, and it provides grants and training to newsrooms in the global trade group WAN-IFRA.

The timing matters because OpenAI is buying goodwill with publishers while fighting them in court. It faces copyright suits from the New York Times and from eight newspapers owned by Alden Global Capital, part of a broader wave of legal disputes over how AI systems use published work. Other firms have been slower to sign licensing deals, and some, including OpenAI backer Microsoft, are exploring models that would pay publishers per use instead.

For newsrooms weighing whether to take this kind of money, the opportunity comes with tradeoffs. Grants and credits lower the cost of experimenting with AI, but they come from a company that news organizations are also suing over training data. The tools that help a local outlet translate coverage or draft donor emails are built by the same industry accused of scraping journalism without permission.

“Deploying AI effectively is ultimately a leadership challenge,” Berman said. “News organizations really have to think about how they smartly integrate these technologies in ways that have good policies, have humans in the lead and support the journalistic quality of the news organizations.”

The post OpenAI invests another $8 million in the American Journalism Project appeared first on The Media Copilot.

OpenAI’s first piece of consumer hardware is reportedly a screenless, mobile AI companion that syncs with ChatGPT and is being pitched internally as a humanlike AI companion that lives in the home.

That description comes from a Bloomberg report published Tuesday, which TechCrunch summarized the same day. The device is still under development, and its sources describe it as something well outside the standard smart-speaker mold.

The speaker reportedly has a personality and can learn about its owner over time to deliver more personalized responses. It would tap into a user’s digital life, pulling from sources like email. And it apparently includes mechanical elements that can move on their own, designed to feel like a companion and act as a physical manifestation of ChatGPT.

OpenAI announced prototypes for hardware in November 2025, with earlier rumors pointing to a phone that would compete directly with Apple. The company also acquired io, the hardware startup co-founded by former Apple design chief Jony Ive, in a deal valued at roughly $6.5 billion. Bloomberg reports the speaker was built with help from ex-Apple engineers who worked on the iPhone and Mac.

That pedigree is now a legal liability. Apple sued OpenAI last week, accusing it of stealing trade secrets and calling the allegations “the tip of the iceberg.” OpenAI denies wrongdoing. Its sources told Bloomberg the new product veers significantly from anything Apple sells today.

OpenAI is not alone in chasing this category. Hark, the AI lab founded by Brett Adcock, raised an oversubscribed $700 million Series A in May at a $6 billion valuation to build proprietary models paired with custom hardware it calls a universal interface between humans and machines. Hark has yet to disclose the device’s design, highlighting the surge of investment flowing into AI hardware long before products ship.

For newsrooms and publishers, a screen-free device that pulls from email and learns a user’s habits raises the same questions as any voice-first platform. If people increasingly ask a companion device for news instead of opening an app or a browser, distribution shifts again, and this time to a surface with no visible headlines, no scannable feed and no obvious place for a byline.

The referral traffic problem publishers already face with chatbots gets sharper when the interface is spoken. The reported device would extend a broader shift in news discovery, making AI assistants an even more prominent intermediary between publishers and their audiences.

OpenAI has not announced the product, its price or when it will ship. Still, the reported device points to the company’s broader strategy of extending ChatGPT beyond screens and into the home.

The post OpenAI’s first hardware device is reportedly a screenless speaker that moves appeared first on The Media Copilot.

The New York Times and a group of other publishers are asking a federal court to sanction OpenAI, accusing the company of withholding or destroying evidence in a high-stakes copyright case over how ChatGPT was trained and used.

In a motion filed Thursday in federal court in Manhattan, the publishers alleged that OpenAI misrepresented its ability to search training datasets and ChatGPT output logs for copyrighted news material. According to Reuters, the publishers said OpenAI told the court it could not search its large language models for their work while allegedly concealing that it had already done so “even before the first News Plaintiff filed suit.”

The motion is the latest escalation in the copyright fight between major news organizations and AI companies. It also moves the dispute deeper into discovery, where the question is not just whether AI companies can use journalism to train models, but whether they can preserve, search and produce the records needed to prove what happened.

The plaintiffs include The Times, the New York Daily News and other media organizations, including Ziff Davis and the Center for Investigative Reporting, according to The Associated Press and Variety. The original New York Times article reported that the publishers are seeking legal sanctions against OpenAI, including monetary penalties and other remedies.

The filing does not ask for sanctions against Microsoft, which is also a defendant in The Times’ broader copyright case, according to The Times’ summary of the motion. Microsoft has invested heavily in OpenAI and integrated OpenAI technology into products including Copilot.

“The evidence is in OpenAI’s training data sets and ChatGPT output logs,” the publishers said in the motion, according to The Times. “But instead of just producing that evidence at the start of the case and focusing on the merits of its fair use defense, OpenAI chose obstruction.”

OpenAI rejected the allegations. “As the Times’ case weakens and they’ve been forced to drop claims against us, they’re persisting with their efforts to invade the privacy of people who have nothing to do with this case, including by making these blatantly false allegations,” OpenAI spokesperson Drew Pusateri told Reuters. “We’ll continue defending our users’ privacy and the long-established principles of fair use.”

The publishers allege that OpenAI deleted billions of relevant ChatGPT conversations or made them unsearchable. They also argue that an OpenAI employee later testified that the company had performed multiple searches for news publishers’ content, contradicting earlier representations about the company’s technical limitations.

A sanctions memorandum posted by Ars Technica says the publishers want the court to bar OpenAI from relying on a disputed 20 million-log ChatGPT sample, find that ChatGPT’s output logs include or would have shown substantial use of the publishers’ copyrighted material, instruct the jury on those findings and award fees and costs tied to the discovery fight.

Those remedies would matter because discovery disputes can shape the trial record. If the court finds OpenAI failed to preserve or produce relevant evidence, the ruling could affect what arguments OpenAI can make later and what conclusions a jury may be allowed to draw from missing or incomplete records.

The Times sued OpenAI and Microsoft in 2023, alleging that millions of Times articles were used without permission to train AI systems that now compete with publishers as sources of information. OpenAI and other AI companies have argued that training models on large bodies of text is protected by fair use, a theory now being tested across lawsuits from authors, artists, music labels and news organizations.

For publishers, the issue goes beyond training data. They argue that AI chatbots and AI search summaries can answer readers’ questions using journalism without sending traffic, licensing revenue or subscribers back to the organizations that reported the information. Media Copilot has been tracking the same pressure point in coverage of Google’s AI accuracy problem and The Times’ warnings about AI companies using journalism without permission.

At the same time, publishers are taking different approaches to the AI economy. Some are suing. Others have signed licensing deals with AI companies. The Associated Press announced a deal with OpenAI in 2023, while other media companies have made agreements with OpenAI, Google, Meta and Amazon.

The sanctions motion could increase pressure on both sides. A ruling against OpenAI would give publishers leverage in court and in licensing talks. A ruling for OpenAI would strengthen the company’s argument that publishers are using discovery to intrude into user privacy and commercially sensitive systems.

Either way, the case shows that AI copyright fights are becoming data-governance fights. The central questions are no longer only what AI systems were trained on. They are whether companies can prove it, search it, preserve it and explain it in court.

The post Publishers ask court to sanction OpenAI in escalating copyright fight appeared first on The Media Copilot.

Two years ago, if anyone had told me the most AI-hostile demographic in 2026 would turn out to be Gen Z, I would have laughed. The generation that grew up with screens in its hands seemed like the obvious early adopter, ready to use the tools to look more skilled, more productive, and more employable than everyone else.

Instead, May commencement season turned into an open revolt. At the University of Arizona, students booed Google chairman Eric Schmidt for pitching AI’s world-changing potential. Gloria Caulfield, VP of strategic alliances for the investment firm and real estate developer Tavistock, drew the same reaction at the University of Central Florida when she compared the rise of AI with the Industrial Revolution. At Middle Tennessee State University, students shouted down Big Machine Records CEO Scott Borchetta for the offense of saying the word out loud.

The numbers back up the booing. A recent Gallup poll measuring AI adoption and attitudes among Gen Zers found the share who say they’re excited about AI fell from 36% to 22% in a year. The share who say they feel anger toward AI climbed from 22% to 31%. Older cohorts are skeptical too, but a sentiment swing this sharp inside the youngest part of the workforce is something I’ve never seen for a new technology this early in its lifecycle.

That swing matters because the AI industry has a serious PR problem at the worst possible moment. Anti-AI sentiment is hardening as the midterms approach, and politicians are picking up data centers as a wedge issue, supporting efforts to halt or slow the build-out of the facilities that fuel AI with the computing power it needs to function. If capacity can’t keep up with demand, the cost of compute will keep rising, and that will put hard ceilings on what newsrooms, marketing teams, and comms shops can actually do with AI in the year ahead.

The infrastructure squeeze is already here

It’s already happening. Anyone who runs Claude as part of their daily workflow knows the rhythm of the outages, and Anthropic’s own status dashboard tells the story in red over the past 90 days. The Claude Code boom has driven demand through the roof in 2026, and the company is scrambling to keep up. Anthropic signed a deal to buy computing power from Elon Musk’s SpaceX, and at the same time it closed the loophole that let builders run third-party software on top of their Claude subscriptions. Some of those setups were burning thousands of dollars of compute against a $200-a-month Claude Max plan. Now those teams have to use Anthropic’s platform directly or move to pay-as-you-go.

The angry posts in response were predictable, but the more useful read on the change is that it forced builders to reckon with the actual cost of what they were running. The choices are probably familiar to anyone trying to budget AI spend: switch to a cheaper model, possibly an open-source one, rebuild on Anthropic’s own platform, or shut the project down.

This was always going to happen at some point. As demand grows, free-compute workarounds will keep closing. The industry’s argument is that the squeeze would hurt less if compute were cheap and plentiful, which is the case being made for trillion-dollar infrastructure projects like OpenAI’s Stargate. For AI to deliver on its promises, compute has to flow like water. That means more data centers, and more power plants behind them.

Which loops back to why Gen Z is angry in the first place. Environmental concerns are near the top of their list, and AI’s energy footprint has only gotten harder to ignore since I wrote about it months ago. This isn’t a fight confined to politicians and podcasters anymore. At the companies I advise on AI adoption, employee surveys keep surfacing the same worry, and in some cases it’s starting to shape whether teams use AI at all.

Governance is the new AI strategy

You can argue about whether the pollution and water concerns are overblown. The cost question is harder to wave off. The leaders running the most ambitious AI programs are past the era of handing every employee a ChatGPT seat. They want agentic workflows, automated processes, and rapid prototyping through vibe coding, and they may be telling their engineers to get obsessed with “tokenmaxxing.”

It’s unclear so far how data center politics will play out. What leaders can actually control right now is governance. That’s more than running training sessions on which model does what, although that matters. Real governance is the balance between experimentation and direction. People need room to invent their own workflows, and the organization needs a way to make sure the compute it’s paying for is being put to good use. That doesn’t just mean “keeping costs down” it means accepting that the bill will sometimes be high and being confident the outcome will be worth it.

Through my consulting work with media companies and PR agencies, I’ve watched this play out in practice. One agency I worked with piloted a vibe-coding tool. Usage spiked early as employees tested its limits, and several different teams ended up building near-duplicate prototypes. The thing that saved them was high-bandwidth communication. They ran regular workshops and project reviews, learned from their people, and steered the work as they went. They eventually homed in on the use cases that actually delivered, in their case automating media intelligence, and the experiment surfaced something unexpected. The original platform wasn’t the right one. The agency ended up adopting a different tool and sunsetting the one it started with.

That’s one of many examples, and the lesson behind all of them is the same. If AI agents are going to do real work for your team, they need to run compute-heavy jobs. Compute is going to stay expensive for a while. The way to avoid the kind of top-down restrictions that suffocate innovation is for leaders to define what success looks like up front, get their teams trained on the tools and models, and build systems that surface collaboration and catch waste before it compounds.

That is what good governance actually looks like. The political fight over AI and data centers isn’t going anywhere, but the companies leaning into AI can still find a way through. The goal is to work inside the real cost constraints while shielding the people doing the actual work from feeling them.

A version of this column appears in Fast Company.

The post The AI industry has a Gen Z problem appeared first on The Media Copilot.

The free-spending era for AI inside corporate America is ending.

According to The Wall Street Journal, executives at companies including Uber, Meta, Microsoft, Salesforce and DoorDash have launched cost-cutting campaigns this year after seeing their AI bills double or triple—or blow through annual budgets in just three months. The culprits: the soaring price of tokens, the basic unit of AI computing, as model providers like OpenAI and Anthropic seek to balance supply and demand.

The result is a notable shift in corporate AI strategy. Where last year the goal was to flood the organization with AI tools and encourage experimentation, leaders are now scrambling to ration access, steer workers toward cheaper homegrown alternatives, and sharpen employee skills to wring better returns from the technology.

“The free-money period for AI is definitively over,” said one senior technology executive at a major financial firm, speaking on condition of anonymity to discuss internal cost pressures.

The cooling, if it holds, could complicate the growth trajectories of AI heavyweights racing toward public listings. Anthropic closed a $65 billion funding round this week valuing the startup at $965 billion, while OpenAI is also moving toward a potential IPO. AI critics have pointed to corporate cost-management efforts as evidence that the ultrafast pace of AI expansion may be unsustainable.

Budgets burned in months

Corporate spending on AI took off in 2024 and 2025 as companies encouraged broad experimentation, eager to signal to Wall Street that they wouldn’t be left behind in the disruption wave. But many enterprises underestimated how quickly costs would accumulate—particularly as employees without specialized training sent inefficient prompts, ran excessive queries, or used premium-tier models for simple tasks that could have been handled by cheaper, internally built tools.

Some companies burned through their entire annual AI budget in the first quarter. Others saw line items in technology budgets that previously seemed large enough suddenly look inadequate. The problem was compounded for organizations that signed multi-year contracts with AI providers before understanding their actual usage patterns.

“Most companies didn’t have visibility into what AI was actually costing them on a per-team or per-use basis,” said an AI strategy consultant who works with Fortune 500 firms. “They just saw a giant bill at the end of the quarter.”

The rationing begins

At Uber, Meta, Microsoft, Salesforce and DoorDash, technical executives have implemented some combination of the same playbook: tiered access to AI tools based on role and need, mandatory efficiency reviews for high-cost teams, and investment in internal AI infrastructure that costs less per query than commercial models.

Some companies have quietly restricted access to certain premium AI features for non-technical employees. Others have introduced internal dashboards that show employees the real-time cost of their AI queries—designed to encourage more efficient prompting habits.

The shift mirrors what happened in cloud computing’s early years, when companies initially over-provisioned infrastructure before learning to optimize.

The IPO problem

The corporate reckoning comes at a delicate moment for the AI industry. Both Anthropic and OpenAI are navigating toward public markets, and institutional investors are watching corporate AI spending closely for signs that the technology is generating sustainable returns—or that the boom could go bust.

If major corporate customers begin to pull back on AI spending or demand better pricing terms, it could affect the revenue projections that underpin those anticipated listings. Anthropic’s $965 billion valuation, for context, represents a multiple that assumes continued rapid growth in enterprise demand.

AI critics say the cost backlash was inevitable. Proponents counter that efficiency improvements and competition among AI providers will eventually bring down prices—and that early-stage overspend is normal for transformative technologies.

For now, the corporate AI spendometer is being watched more carefully than ever.

The post Corporate America is starting to ration AI as costs skyrocket appeared first on The Media Copilot.