Artificial Intelligence

[phkrause]

🤖 $50M push to regulate AI

Photo illustration: Sarah Grillo/Axios. Stock: Getty Images

Advocates for AI safety and regulation — including AI researchers themselves — are emerging as a force in U.S. elections, spending nearly $50 million this election cycle on candidates who favor regulation, a new analysis shared exclusively with Axios' Maria Curi shows. Why it matters: The people building AI are increasingly spending to push for its regulation — sometimes at odds with the companies they work for. Donors tell Axios they're not primarily motivated by party affiliation or specific policy positions, but by fears that industry money will scare candidates into staying quiet about an unpopular technology. 🔭 Zoom out: Forces favoring AI acceleration — who earlier this year seemed poised to target pro-regulation candidates — haven't spent as much as opponents expected, and safety advocates see an opportunity to fill the void. Keep reading ...

[phkrause]

🤖 Anthropic gives Wall Street an AI bull case

Illustration: Sarah Grillo/Axios. Stock: Getty Images

AI companies for years have promised that smarter models would be worth the staggering cost to build them. Anthropic made that case last night in cash, Axios' Madison Mills writes. The company presented eye-popping financials to investors, saying it was cash-flow positive in the third quarter, with revenue surging nearly 50% to $16.8 billion compared to the April-to-June period. Why it matters: As Wall Street broadly spent the day fearing an AI bubble, Anthropic offered what investors have been desperate to see: evidence that frontier AI can be a sustainable business. 🔬 Zoom in: Consider just one of the stats reported by Anthropic: the company is cash-flow positive even as some of the world's most profitable companies are struggling to keep their AI spending in check. For context, Google's parent, Alphabet, long a money-printing machine, turned free cash flow negative in its second quarter as AI infrastructure investment soared. The big picture: Anthropic's early bets on coding expertise, business customers and spending discipline have paid off. Claude Code, a viral hit that brought users to Anthropic in droves, reached an estimated annualized revenue of $6 billion to $7 billion by late summer, according to Bloomberg Intelligence. Customers have generally been reluctant to switch once they've adopted it, according to PitchBook data. 🥊 Reality check: Anthropic still faces enormous training costs and stock-based compensation expenses, which could weigh on its underlying profitability. And Anthropic faces more than $500 billion in planned infrastructure and computing spending over the next decade.

[phkrause]

⚡ Scoop: "OG" Bill Gates returns for AI fight

Photo illustration: Sarah Grillo/Axios. Photo: Getty Images

Bill Gates is gearing up to dramatically expand his AI safety push, personally leaning on business leaders and considering a major deployment of his resources, sources tell Axios' Bradley Olson. Why it matters: The formidable, sharp-elbowed Bill Gates who built Microsoft is back. He's directing his competitive fire toward what he has come to see as one of the most important causes of his lifetime. "I can reach CEOs on this better than you can," Gates told one person he consulted on AI safety, showing an eagerness for combat that associates say is a return to "OG" form. 🔬 Zoom in: Once a voice for optimism on AI, Gates has grown dismayed by an escalation of model capabilities, alarming security incidents and what he sees as insufficient regulation, sources say. As he prepared to make the case for a far more robust regulatory framework on AI, he consulted extensively with leading AI executives, researchers, and safety advocates, including Anthropic CEO Dario Amodei. Behind the scenes: People Gates has spoken to called his advocacy a "comeback tour," an attempt to rebuild a legacy that has been tarnished by his past association with Jeffrey Epstein. Gates has said he's aware that some people may dismiss his warnings due to the Epstein ties. But he sees the stakes as too high to stay on the sidelines and believes his Rolodex, foundational tech credibility and wealth can do an immense amount of good. 🔮 What's next: Gates' calls for robust regulation will put him on a collision course with the Trump administration, which has favored a lighter regulatory touch. Keep reading.

[phkrause]

Internet loses human touch

Illustration: Brendan Lynch/Axios. Stock: Getty Images The internet is undergoing a foundational shift as websites evolve from human destinations to information hubs for AI agents, Axios' Sara Fischer and Kerry Flynn write in an Axios Media Trends Executive briefing. Why it matters: Human discretion, a critical safeguard for establishing trust online, is rapidly disappearing. The big picture: Bad actors are using AI to create new information sources designed to manipulate LLMs. 🔮 State of play: Premium publishers and brands are looking to get ahead of this problem in the short term by becoming reliable information warehouses for bots while serving premium experiences for humans. At TIME, a news publisher that has leaned heavily into agentic optimization, human visitors encounter articles with imagery and storytelling while AI agents receive structured information, COO Mark Howard told Axios. TIME's Agent Ads product extends that approach to advertising, providing machine-readable information about sponsors, including FAQs. Reality check: Smaller companies may struggle to handle an influx of bot traffic or build the infrastructure needed to serve both humans and machines. The bottom line: Few consumer websites will survive as direct destinations in the agentic era. Without a trust layer to vet information that machines retrieve and analyze, the internet risks becoming a breeding ground for content that's useless to humans.

[phkrause]

AI governing the AI world

Illustration: Sarah Grillo/Axios. Stock: Getty Images

Slowly but surely, the leading AI companies are creating a national self-governing structure to oversee their technology and its societal impact, Jim VandeHei and Mike Allen write in a "Behind the Curtain" column. With the federal government largely hands-off, Anthropic, OpenAI and others are self-regulating their technology and now setting up their own plans to redistribute their wealth. Why it matters: Not since the age of the robber barons have America's industrial titans wielded such sweeping power to shape society — and decide how their private fortunes should serve the public good. In the modern era, Washington has largely exercised those powers through regulation, taxation and public spending. 🖼️ The big picture: President Trump has no interest in new regulations. Instead, he has blessed a basic framework for the AI companies to regulate themselves — and potentially each other — outside of formal government monitoring. Anthropic, OpenAI, Google, Elon Musk's XSI, Meta and Nvidia are working on different versions of this. To be clear, this is all voluntary and rests on the public trusting companies the size of industrialized nations to responsibly police themselves. These companies are in a frantic race to create human-surpassing superintelligence. All are pressing ahead, even as their leaders — particularly at Anthropic — warn of potentially catastrophic risks and profound economic disruption. At the same time, some AI leaders are promising to direct a portion of their vast wealth to benefit those outside their companies. Zoom in: At least three competing visions are emerging for sharing the spoils of the AI boom, each with a different answer to who should control the wealth. Anthropic: The company. On the eve of its blockbuster IPO, Anthropic pledged 5% of its shares to a new public benefit fund that could grow into one of the largest philanthropic institutions in history. At a hypothetical $2 trillion valuation, its stake would be worth $100 billion — more than the Gates Foundation's current endowment. Including earlier founder and employee pledges, roughly 20% of Anthropic's shares are earmarked for the public good. Musk: The people. Musk responded to Anthropic's announcement by urging the company to give shares "directly to the public." Yesterday, Musk pledged to give every American child a share of SpaceX next year — a staggering $10.5 billion giveaway at current prices. Musk called it the "first of many," though he offered no details on how the unprecedented distribution would work. OpenAI: The government. CEO Sam Altman has floated giving the U.S. government a 5% stake in OpenAI through a sovereign wealth fund — potentially handing Washington a multibillion-dollar interest in one of the world's most powerful AI companies. Altman envisions an Alaska-style fund that could distribute the industry's investment returns directly to Americans, though the plan remains preliminary and would likely require congressional approval. 🥊 Reality check: The numbers are staggering. So are the assumptions about future wealth, billionaire follow-through and the public's willingness to trust its benefactors. Bill Gates and Warren Buffett launched the Giving Pledge in 2010, a voluntary promise by billionaires — including Musk — to give away most of their fortunes. Gates built a philanthropic empire, only to become a lightning rod for conspiracy theories. Anthropic's ambitions depend on extraordinary growth projections — its own employees envision 17% GDP growth in 2030 — while Musk has a history of making spectacular promises that fail to materialize. Anthropic's roots in effective altruism, a philosophy deeply distrusted by the Trump administration and its allies, also raise a more fundamental question: Who gets to define the public good? Between the lines: No industry requirements or consensus exist on how best to redistribute their expected wealth. But Trump Accounts — which allow families, businesses and individuals to invest in tax-advantaged funds for children — are emerging as a favorite vehicle for free-market AI leaders. Billionaire investor Brad Gerstner, a driving force behind the program, is championing a vision of broad-based ownership in which ordinary Americans share directly in the economic upside of the AI revolution. SpaceX President Gwynne Shotwell and her husband have already pledged roughly $320 million in company shares to 2 million children through Trump Accounts — a potential blueprint for distributing an even larger slice of Musk's fortune. 👀 What to watch: The Trump administration hasn't fully relinquished its regulatory authority, potentially complicating the emerging experiment in AI self-governance. After Anthropic disclosed that its AI had submitted a false homicide tip to Philadelphia police and improperly interacted with government websites, the White House declared that AI companies must report security incidents and remedy any harm. "This notification and remediation process is not optional," Trump's Super Intelligence Force warned — a striking assertion of federal authority just weeks after the president blessed a voluntary system of industry oversight.

[phkrause]

AI is hitting an energy wall

Illustration: Lindsey Bailey/Axios

Silicon Valley moves fast and breaks things, but the U.S. electrical grid moves slowly and requires permits, Axios AI+ co-author Madison Mills writes. The explosive growth of AI is colliding with the glacial pace of building out new power infrastructure, prompting a new warning from Goldman Sachs that energy bottlenecks are set to stall the AI race. Why it matters: Amid a wave of anxiety from Wall Street about the debt-fueled AI buildout, Goldman points to power as the biggest risk to trillions in spending in a new report shared exclusively with Axios. 🔬 Zoom in: The U.S. has roughly 2,600 gigawatts of proposed power generation and storage projects awaiting grid connections, more than twice the grid's existing installed capacity, though many projects may never materialize. The median time from an application for a grid connection to commercial operation is roughly five years. 💰 Follow the money: The massive AI expansion is being financed with debt, raising the stakes if construction timelines slip. Goldman projects AI tech giants will borrow $420 billion globally in 2027, up more than 60% from its 2026 estimate. Keep reading.

[phkrause]

Battle of the bots

Illustration: Aïda Amer/Axios. Stock: Getty Images Axios' Ina Fried recently tested two of the newest personal AI agents, OpenAI's Dots and Meta's Muse. Ina writes: I asked each assistant to plan a trip to Atlanta to see my beloved WNBA Golden State Valkyries, organize a brunch watch party and create souvenir programs. ☀️ Where they shined: When I asked both agents to find flights and tickets for Game 3 in Atlanta, each caught that the game had been moved up a day and that tickets weren't yet on sale. Both found good flight options. 🧐 Where they were creative, but imperfect: For my brunch, Muse designed a colorful 10-page recipe booklet featuring ube waffles and lavender-honey butter. Dots produced a more practical eight-page packet with detailed cooking instructions and attention to dietary preferences. Both designed invitations that needed aesthetic and language edits. 👎 Where they struggled: I asked the agents to help me buy Valkyries trading cards. Both helped me compare options and prices. I chose Target, where I had a 5% discount and free shipping. Muse tried to complete the purchase, but got stuck at Target's sign-in page. Dots directed me to the Target app, where my account and payment information were already saved. 💵 The bottom line: Both agents could have done more if I gave them greater access to my accounts and wallet — but I don't trust them enough for that yet. Go deeper.

[phkrause]

Ahead of massive IPOs, AI companies are sketching plans to share their wealth

As they steam toward massive public offerings on Wall Street, the companies at the forefront of artificial intelligence have begun sketching plans to share their wealth with the public — a commitment they say is imperative to address disruptions the technology could bring to the workforce and people’s livelihoods. Read More.

[phkrause]

Nvidia's global power play

Illustration: Aïda Amer/Axios. Stock: Getty Images

Nvidia CEO Jensen Huang is crisscrossing Europe with a multibillion-dollar checkbook and a grand ambition: Make his company the indispensable backbone of the global AI economy, Axios' Zachary Basu writes. Why it matters: As fears of an AI bubble rattle Wall Street, Huang is working to spread the boom beyond the U.S. and China — courting governments, bankrolling startups and cultivating massive new markets for Nvidia's chips. 🖼️ The big picture: The godfather of AI is trying to single-handedly build a new market for Nvidia chips, betting that governments and institutions around the world will want their own, highly customized data centers running open models. If it works, it could shore up demand for Nvidia chips just as another big customer group — neoclouds — appears financially shaky. Part showman, part salesman, Huang has spent the past week preaching the gospel of AI across Europe ahead of a keynote this morning at Nvidia's flagship conference in Berlin. In France, Huang announced investments in four startups and joked, whisky glass in hand, that he's "the biggest VC in Europe." President Emmanuel Macron welcomed him to Paris, while France's economy minister compared the AI revolution to the Gold Rush. In Britain, Huang channeled Oprah Winfrey while announcing stakes in five startups. "I am giving you money! And I am giving you money! And I am giving you money!" he shouted from the stage of a converted London church. In Sweden, Huang unveiled a supercomputer backed by the country's industrial elite, likening Nvidia's technology to a "time machine" capable of compressing years of scientific research into days. In Germany, Huang challenged Chancellor Friedrich Merz to loosen Europe's regulatory instincts. "There's nothing to regulate if the industry doesn't thrive," he warned Monday. "Give Germany a chance. Give this industry a chance." 🔬 Zoom in: At the heart of Huang's global strategy is "sovereign AI" — the idea that every nation should build and control its own AI infrastructure, data and models, rather than depend on U.S. or Chinese tech giants. For Huang, the prize is a vast new class of customers, including entire countries that see AI infrastructure as an economic and national security imperative. 👀 What to watch: Huang has emerged as the AI industry's chief cheerleader during periods of market turmoil, countering predictions of a bubble with assurances that demand is still accelerating. His Berlin keynote offers another chance to calm investors by arguing that the next wave of AI spending will come from governments treating the technology as a national necessity, not a speculative bet.

[phkrause]

🤖 Execs trust AI over themselves

Illustration: Sarah Grillo/Axios. Stock: Getty Images

Executives are twice as likely as rank-and-file workers to trust AI over their own judgment, Axios' Emily Peck writes from a new survey by Charter, a workplace research company. 32.8% of execs said that they trust AI more. That number was just 16.3% for individual contributors.

[phkrause]

AI gets better at hacking vital systems

Illustration: Sarah Grillo/Axios It's only a matter of time before freely available AI models will make it easier to hack critical services like power grids and water systems, Axios' Sam Sabin reports from new research. Steve Swick, former top security executive at American Electric Power, told reporters during a briefing: "It's a clock that is moving faster than we've ever seen." 💧 Researchers at cybersecurity firm Frenos tested 10 AI models' ability to plan attacks against industrial systems, such as software that monitors water treatment processes. The models were fed information about potential hacking targets, and were asked to piece together the steps needed to achieve specific objectives. Three out of the 10 models successfully constructed attack plans in about 90% of attempts. Frenos did not disclose the names of the AI models tested. 👾 Frenos co-founder and CTO Harry Thomas says that once attackers gain access to an industrial network, AI could help them quietly map out systems and identify weaknesses without the same time constraints facing human hackers. "AI doesn't have the problem with patience. It doesn't need to eat. It doesn't need to sleep, and it doesn't need to use the restroom." Go deeper ...

[phkrause]

💰 Scoop: AI agents slip into bank accounts

Illustration: Sarah Grillo/Axios. Stock: Getty Images

Major U.S. banks are detecting unauthorized third-party AI agents operating inside customer accounts — with some agents attempting to take actions that appear to be independent of user instructions, sources tell Axios' Madison Mills. Why it matters: AI agents are quickly becoming a nightmare for banks, slipping into their systems faster than they can respond. Because it's hard to distinguish a human from a bot, banks are blocking the technology and rapidly rewriting their terms of service. Morgan Stanley, Bank of America, JPMorgan and Wells Fargo are among the banks racing to contain agent risks. Some are aware of unauthorized access, people familiar with the issue said. 🔭 Zoom in: These are generally situations when an agent or AI assistant accesses a human's account with the user's permission and login credentials. That violates guidelines of most banks, which require humans to log in to their own accounts. Banking execs fear what agents with login access might do, especially if a user has granted purchase authority or the ability to shift funds. Some banks have tracked activity when agents — programmed to fulfill requests or complete tasks on a user's behalf — contact call centers or appear to take actions independent of user instructions, the people said. Keep reading.

[phkrause]

AI cracks a 22-year-old security puzzle in two days

For more than two decades, a major security flaw in software powering much of the internet went unfixed because it was too difficult for hackers to take advantage of. Now, researchers have used AI to figure out how to exploit the flaw in just two days, potentially allowing attackers to steal sensitive information or take control of websites. The vulnerability has since been patched, but cybersecurity experts warn the discovery shows how AI could make dangerous cyberattacks easier than ever.

+ Analysis: AI quickly exposes a decades-old internet vulnerability

+ Just wondering: What happens if AI agents start speaking their own language?

[phkrause]

AI money tsunami

Illustration: Maura Losch/Axios

Anthropic filed yesterday for what could be the biggest IPO ever, capping a massive day of AI dealmaking that nearly equals 1% of U.S. GDP, Axios' Dan Primack writes. Why it matters: Few industries have ever become so central to the American economy so quickly — and we're watching Wall Street finance the buildout in real time. Anthropic's long-awaited IPO may seek to raise a record $100 billion. Also yesterday, SpaceX secured $40 billion in debt financing to lease GPUs — the chips that power modern AI — and Meta said it would seek up to $95 billion in financing for its AI buildout. Together, that's roughly $235 billion in a single day — more than the entire U.S. airline industry adds to the economy in a year. 🖼️ The big picture: AI's appetite for capital has outgrown what any one part of the financial system can supply. Anthropic alone has committed roughly $750 billion to computing power in the coming years — about seven times what its record-setting IPO may raise. Even private investors are struggling to keep up: OpenAI's record $120 billion funding round took months to close. Zoom in: Most Americans will be financially exposed to these deals, even if they never buy a share. SpaceX and Meta are already major holdings in the top stock indices that fill pensions, 401(k)s and other retirement plans. Anthropic will soon join them. Between the lines: Anthropic's filing spells out how big the stakes are, devoting more than a quarter of the document to risks. The company warns that AI could pose "catastrophic or existential risks to humanity" and that models may try to manipulate the people using them. 🔭 Zoom out: Each of these three deals is a bet that AI use keeps compounding fast enough to justify the industry's trillions in spending. The flip side: Declining use could crack the foundation of America's entire economy. The bottom line: Everyone is on the AI train, whether we want to be or not.

[phkrause]

🤖 AI's Y2K moment

Illustration: Sarah Grillo/Axios. Stock: Getty Images

The AI-discovered software flaws that sent the cybersecurity world into a panic last week have so far led to surprisingly few successful attacks, Axios' Sam Sabin writes. Why it matters: Hackers have made millions of attempts to exploit bugs that OpenAI agents found in PHP, software used by roughly 70% of websites. The flaws have proven unusually hard to weaponize, and companies' rush to patch them appears to have helped. The past week has unfolded in ways that remind some researchers of Y2K, a feared software crisis in 1999. Warnings of Armageddon helped rally companies to engineer a solution. Keep reading.

[phkrause]

🤖 AI's trick-or-treat earnings season

Illustration: Brendan Lynch/Axios

The latest wave of hyperscaler earnings is spooking Wall Street. But bulls and bears alike have evidence to validate their wildest dreams — or their worst nightmares, Axios' Madison Mills reports. Why it matters: The market's AI obsession has entered its Jekyll-and-Hyde era, with the latest earnings from tech giants sending mixed signals about the AI boom. Google: 🧟 The scary: Google's AI Overviews are cannibalizing the company's core business — ad revenue from search. 🍬 The good: Google Cloud is closing in on Microsoft's Azure cloud revenue, a stunning development given that Google was an also-ran in the cloud game only a few years ago. Microsoft and Amazon: 🍭 The good: Both companies had explosive cloud growth. 👻 The iffy: Amazon's net income fell despite soaring operating profits, as it didn't repeat the pretax gain from its Anthropic investment that boosted earnings a year earlier. Microsoft's OpenAI investment boosted its net income slightly, which lifted its profit growth. Data: FactSet. Chart: Matt Phillips/Axios By the numbers: Either way, this week's flood of AI news failed to shore up investor confidence, Axios markets correspondent Matt Phillips writes. The large-cap, tech-heavy Nasdaq 100 stumbled to its third straight weekly loss. The bottom line: The AI boom championed by President Trump is losing momentum just days before the midterms, when voters will be deciding on AI-related issues at the ballot box for the first time.

[phkrause]

AI boom meets the ballot box

Photo illustration: Brendan Lynch/Axios. Photo: Brandon Bell/Getty Images

Technology has become a defining issue in the 2026 midterms — and AI data centers are the punching bag. 🥊 Why it matters: It's not just about noise pollution and curb appeal. Data centers have become a lightning rod for growing resentment over who gets rich from the AI boom and who bears the cost, Axios' Maria Curi and Erica Pandey report. 💬 Tatishe Nteta, a political scientist at UMass Amherst, says: "It's a reflection of what's increasingly becoming an easy target — tech giants and AI." He calls the backlash "a turn against the billionaire class from both the left and the right." 🗳️ Driving the news: Opposition to data centers reached a fever pitch in the months leading up to the midterms. Once an obscure infrastructure issue, data centers have become a political liability for candidates across the country. Voters in 10 states will decide on 41 data center ballot measures on Election Day, ranging from new requirements for developers to outright bans on construction. All six data center measures already decided this year — in New Hampshire, Wisconsin, Michigan and California — have gone against development, according to Ballotpedia. Data: Ballotpedia. Map: Erin Davis/Axios 🔎 Zoom in: Data center backlash is reshaping races and forcing politicians to change course. In Texas, incumbent Gov. Greg Abbott (R) reversed course after years of courting the industry — halting new approvals pending an audit of projects — after a backlash in rural communities. Democratic challenger Gina Hinojosa supports a moratorium on new data centers until tougher rules are established. In Utah, an unlikely coalition of state leaders, local residents and environmental groups is opposing a data center project backed by Shark Tank's "Mr. Wonderful," Kevin O'Leary. In Ohio, incumbent Sen. Jon Husted (R) is on the defensive over past support for data centers as he fights to keep his seat in a race that could help decide Senate control. The state has 21 data center-related measures on Tuesday's ballot. The backlash is minting a new wave of candidates. Across the country, local activists galvanized by data center fights in their communities are running for office. ↔️ Between the lines: The opposition to data centers cuts across generations and party lines, though some groups are far more skeptical than others, per exclusive polling from the new Axios-CivicScience AI Sentiment Index. Overall, 43% of Americans oppose construction of data centers in their area, compared with 19% who support them. Another 38% are neutral. The generational divide is stark: Nearly two-thirds of baby boomers and older Americans (65%) oppose data centers, compared with 26% of Gen Z. A majority of independents (54%) and nearly half of Democrats (48%) oppose data centers. Republicans are almost evenly split, with 30% opposed and 29% supportive. 🇺🇸 The Trump factor: President Trump has complicated Republicans' efforts to navigate local backlash over data centers, declaring "we want them" at an Ohio rally and forcing the Husted campaign to distance itself from his remarks. Trump has called data centers a "Golden Goose" for the U.S. economy and praised companies like Meta for investing in local communities. Democrats are capitalizing on voter frustration and making data centers a major campaign issue.

[phkrause]

Human genius after AI's new dawn

Illustration: Annelise Capossela/Axios

Bradley writes: OpenAI's massive data dump of more than 300 solved math problems triggered a wave of despair about the end of human supremacy. But one mathematician has found reasons to push back against the doom loop. Why it matters: As AI capabilities continue to advance across disciplines, beating humans at one task after another, every new feat requires important contemplation. What problems remain to be solved? Where can humans still have an impact? What does it all mean? For Vanderbilt University's Dan Margalit, the answer within mathematics is that the machines aren't replacing human intellect. They are deeply reliant on it. He sees AI as a tireless finisher of human-led ideas and imagines a future where mathematicians can thrive. ⚡ Catch up quick: The academic world has been reeling since OpenAI released hundreds of AI-generated math papers, offering solutions to major open challenges just a few weeks after also claiming to solve the legendary Navier-Stokes Millennium Prize Problem. The intrigue escalated last week when OpenAI announced its models had disproved the exponential time hypothesis, upending a foundational assumption in computer science. Rival AI giant Anthropic revealed that its own models independently made the same discovery on the same day, pouring 14 million tokens into the effort, according to The Information. The new findings have triggered more shock and awe in academia. Yet in all that frenzy, here are five ways Margalit, who chairs Vanderbilt's math department, says mathematicians are finding hope in the age of AI. 1. 🏁 AI is a finisher, not an originator. The hope: AI didn't solve these problems in a vacuum. It succeeded where humans had brought ideas close enough to a breakthrough. The context: "The problems solved were the ones humans were closest to solving," says Margalit. The AI failed on thousands of other problems where human literature hadn't already pointed the way. 2. ⚠️ The machines still make mistakes. The hope: Mathematicians and others reviewing the AI solutions have found flaws. The context: As mathematicians read the proofs, a number of the released papers are facing retraction because AI made the same logic errors as humans, Margalit said. 3. 🦾 It's brute force, not brilliant intuition. The hope: AI approaches math the way machines approach chess. The context: OpenAI deployed tens of thousands of agents to tirelessly smash together known theorems in millions of configurations. "That's amazing, but like chess, it's about trying all possibilities to get from where we are to where we want to go," Margalit notes. The AI isn't asking interesting questions. 4. 🧮 AI acts as an ultimate calculator. The hope: Humans used to write entire books of integrals until software came along and automated the process. AI can also be a tool for acceleration. The context: This didn't kill math; it freed humans to explore harder concepts. "When an integral appears in a sophisticated problem, I no longer devote my research to it," Margalit says. 5. 💡 Math is about understanding, not just correct answers. The hope: The ancient Greek word "mathematika" translates to "the study of learning," he says. Margalit credited mathematician Bill Thurston, the author of a landmark 1994 essay that has taken on new life, with this observation: If a computer spits out a list of 10,000 prime numbers, it can't tell us about why they exist or the patterns they form. The context: The universe already knows the answers. Math is the process of the human mind grasping them. "If a computer tells me an answer, that doesn't give me insight or satisfaction," Margalit says. "Mathematics is the human endeavor of understanding."

[phkrause]

🤖 AI loses everywhere

Illustration: Aïda Amer/Axios

Voters showed deep distrust of the AI race, approving dozens of data center moratoriums and electing a raft of candidates who favor slowing a $10 trillion buildout, including in Texas — the epicenter of the boom, Axios' Maria Curi writes. Why it matters: The results were a warning shot to some of the nation's wealthiest and most powerful executives, many of whom are already working on ways to build a more palatable version of the revolution they are dead-set on winning. 🔬 Zoom in: All told, nearly two-thirds of 2028 data center capacity could face delays, although most political analysts don't see Democrats permanently blocking construction. Democratic challenger Gina Hinojosa's surprise victory in the Texas governor's race fundamentally scrambles the math for the AI buildout. She campaigned on keeping a moratorium on data center construction in place until at least mid-2027. The state accounts for 41% of new U.S. capacity expected in 2028, according to the research firm SemiAnalysis. Via X Democrats also won gubernatorial races in Ohio, Georgia and Arizona — the three states expected to have the most new data center capacity after Texas. The winners in Ohio and Georgia favor restricting construction. Out of 41 data center measures on the ballot across 10 states, about two dozen proposals designed to restrict or block development appeared to have passed based on initial returns. 💰 The other side: In California, a measure that would've enacted a 5% wealth tax on individuals and trusts with more than $1 billion in assets appeared headed for defeat after facing heavy opposition from Silicon Valley.