More news below. Have a wonderful weekend. \u2014Andrew Nusca<\/p>\n
Want to send thoughts or suggestions to\u00a0Fortune Tech?\u00a0Drop a line\u00a0here<\/a>.<\/p>\n Amodei, Altman, Hassabis call for collaboration on AI rules at the G7 summit Ludovic Marin\/AFP\/Getty Images<\/p>\n At the G7 summit in \u00c9vian-les-Bains, France on Wednesday, Anthropic CEO Dario Amodei warned (he\u2019s always warning, isn\u2019t he?) heads of state to \u201cresist the temptation to splinter\u201d in their approaches to AI regulation.<\/p>\n Amodei told attendees\u2014which included archrival Sam Altman of OpenAI and Google DeepMind chief Demis Hassabis\u2014that democratic countries should work together to prevent threat actors from gaining access to the most advanced tools, according to a Financial Times report<\/a>.<\/p>\n His counsel, of course, comes as Donald Trump\u2019s White House continues its abrupt export controls on Anthropic\u2019s new Fable and Mythos AI models on national security concerns.<\/p>\n Despite their competitive tensions, Altman and Hassabis agreed. Altman said G7 nations should have access to AI-powered cybersecurity tools and Hassabis also called for collaboration. (Both were seated to each side of Trump; Amodei joined Salesforce CEO Marc Benioff next to Macron.)<\/p>\n There\u2019s little doubt that incentives are aligned. All three companies, two of which are headed toward blockbuster IPOs, are more than happy to find stability while selling their wares to the governments of the G7.<\/p>\n Never one to pull a punch, host and French president Emmanuel Macron was critical of Trump\u2019s move against Anthropic. The decision, he said, \u201cclarified the stakes\u201d for having consistent AI regulations because one government could simply choose to \u201cturn off the switch,\u201d harming everyone. \u2014AN<\/p>\n Tokens are getting cheaper, leading companies to spend even more on AI<\/p>\n The ghost of a 19th century English economist may be haunting<\/a> yet another part of the AI boom.\u00a0<\/p>\n In 1865, William Stanley Jevons observed that when the Watt steam engine made coal use more efficient\u2014decreasing the amount required for a task\u2014coal consumption actually skyrocketed.\u00a0<\/p>\n More than 150 years later, one economist is citing this phenomenon, dubbed Jevons paradox<\/a>, to explain why the cost of AI will continue to creep up.<\/p>\n Despite the price of a single token dropping more than 90% since 2023, spending on large language models has doubled since late last year, according to the Silicon Data Token Expenditure Index<\/a>.\u00a0<\/p>\n Apollo chief economist Torsten Slok said it\u2019s yet another example of the Jevons paradox in action. \u201cAs tokens get cheaper, companies don\u2019t spend less but instead run more AI agents, automate more workflows and generate more code, pushing aggregate expenditure higher even as the unit cost of intelligence collapses,\u201d Slok wrote in a recent blog post<\/a>.<\/p>\n The cost of tokens has become a major concern for companies racing to leverage AI. The trend of \u201ctokenmaxxing,\u201d in which employees blitz to increase their AI use, has emerged as companies like Meta and Amazon incentivize the technology\u2019s use.\u00a0<\/p>\n But the tactic is proving unsustainable. Uber president and COO Andrew Macdonald recently said the rideshare company burned through its entire AI budget<\/a> in the first four months of the year, leading it to cap monthly AI spending<\/a> at $1,500 per employee.<\/p>\n In a brief last week, Bain and Co. analysts found that while token costs were halved from December 2024 to December 2025, the tokens consumed grew by 450% in the same period.<\/p>\n Companies feel compelled to upgrade their AI models to take advantage of the upgraded technology, they wrote, rather than stick with their current models and pocket the savings.\u00a0<\/p>\n \u201cThe models get cheaper. The usage gets heavier,\u201d they wrote. \u201cThe bill stays stubbornly high.\u201d \u2014Sasha Rogelberg<\/p>\n Roelof Botha joins SpaceX board<\/p>\n Roelof Botha, longtime investor and former steward at Sequoia Capital, has joined<\/a> the board of SpaceX, a filing revealed.<\/p>\n The news comes less than a week after SpaceX\u2019s historic IPO and about seven months after Botha stepped down as the steward of the legendary venture firm Sequoia Capital.\u00a0<\/p>\n His tenure coincided with a notably difficult time in the VC firm\u2019s history. Botha, as an investor, backed winners like YouTube, Instagram, Block, and MongoDB, among others. But his leadership style was not roundly embraced<\/a>, he missed key opportunities<\/a> in artificial intelligence, and the firm took a reputational hit\u2014prompting an apology to investors\u2014after it lost its entire $150 million investment<\/a> in Sam Bankman-Fried\u2019s FTX.\u00a0<\/p>\n Botha, who was featured on the cover of Fortune magazine<\/a> and spoke at Fortune Brainstorm Tech<\/a> in 2024, has a longstanding relationship with Elon Musk\u2014SpaceX founder, Tesla CEO, and world\u2019s richest man by a mile.\u00a0<\/p>\n The two (both originally from South Africa) first worked together in the dotcom era when Botha served as CFO at PayPal, where Musk was a cofounder. Botha and Musk reportedly bumped heads back then, over culture and software, but have continued to work together through the decades. For example, Sequoia is among SpaceX\u2019s key venture backers.\u00a0<\/p>\n \u201cI\u2019ve known Elon for over 25 years,\u201d Botha told Fortune last year<\/a>, amid fury over Musk\u2019s DOGE. \u201cHe was the first person to offer me a job in America. He believed in me when I was an unknown student at Stanford. I have a lot of appreciation and understanding that he\u2019s not perfect. None of us are. He deeply cares about doing the right thing.\u201d \u2014Allie Garfinkle<\/p>\n More tech<\/p>\n \u2014U.S. mulls how to structure gov\u2019t equity stakes in AI firms<\/a>. Commerce Secretary Lutnick\u2019s solution? A sovereign wealth fund.<\/p>\n \u2014Beware \u201cFortibleed,\u201d<\/a> a new hack exposing Fortinet and FortiGate VPN credentials.<\/p>\n \u2014Epic\u2019s Unreal Engine 6, expected next year,<\/a> will unify UE5 and Unreal Editor for Fortnite as well as integrate with Anthropic\u2019s Claude and Google\u2019s Gemini.<\/p>\n \u2014Snap shares closed down 8%<\/a> after the company\u2019s $2,200 Specs AR glasses fail to impress.<\/p>\n \u2014The $99 Google Home Speaker begins shipping June 29<\/a>, nine months after its debut.<\/p>\n \u2014Jeff Bezos on the intelligence age:<\/a> \u201cAI is going to create a labor shortage.\u201d<\/p>\n![\"France's](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/06\/GettyImages-2281424015.jpg\"\/)
<\/a>France’s President Emmanuel Macron (center) with Salesforce CEO Marc Benioff (left) and Anthropic CEO Dario Amodei (right) at a working lunch meeting at the G7 summit in Evian, France, on June 17, 2026.<\/p>\n