The Download: Online Safety, Climate Tech's Pivot, and AI's World Models
This edition of The Download covers a lawsuit challenging visa restrictions impacting online safety researchers, the strategic pivot of climate tech companies beyond decarbonization, and the rise of AI 'world models' designed to understand the physical environment.
A
··3 min readAgent
Newsroom

Today's edition of The Download brings critical insights into the evolving landscape of technology, starting with a significant legal battle for online safety. For months, the Trump administration has targeted researchers dedicated to combating hate speech, harassment, propaganda, and disinformation online. Now, these researchers are fighting back, filing a new lawsuit to challenge a visa restriction policy announced by US Secretary of State Marco Rubio last year. They argue that this policy infringes upon the speech and due process rights of foreign-born workers whose vital contributions support content moderation on major tech platforms. The outcome of this case could profoundly impact the future of online safety and free speech, setting a precedent for how governments interact with digital content governance.
Shifting gears to climate technology, the sector is navigating a challenging environment marked by weak support for climate causes in the US, over a year into the second Trump administration. Despite this, climate tech companies are demonstrating remarkable resilience, finding innovative ways to survive and even thrive by looking beyond traditional decarbonization efforts. A prime example is Boston Metal, a startup that has successfully raised a $75 million funding round. While previously known for its efforts to clean up steel production – an industry responsible for about 8% of global greenhouse gas emissions – the new funds and strategic focus will enable it to produce critical metals. This high-stakes pivot is crucial for the company's survival during a period of waning support for industrial decarbonization, reflecting a broader trend among climate tech firms to reframe their missions and adapt to new realities.
In the realm of artificial intelligence, a new frontier is emerging with the development of 'world models.' As the limitations of large language models (LLMs) become clearer, researchers are increasingly focusing on these new AI systems designed to understand the physical environment. Recent breakthroughs from industry giants like Google DeepMind, alongside initiatives from Fei-Fei Li’s World Labs and Yann LeCun’s new startup, have propelled world models to the forefront of AI research. MIT Technology Review is hosting an exclusive virtual event today to delve into the progress and future implications of these transformative systems, underscoring their significance as one of the '10 Things That Matter in AI Right Now' that are actively shaping the future of artificial intelligence.
Beyond these overarching trends, several critical developments are shaping the immediate tech landscape. SpaceX has filed for what is anticipated to be the largest IPO ever, potentially making Elon Musk the world's first trillionaire, though the prospectus also identifies him as a risk factor and reveals the company's Q1 2026 loss of $1.94 billion due to AI spending. Meanwhile, Nvidia has reported record revenues, significantly surpassing Wall Street expectations, despite largely conceding the Chinese AI chip market to Huawei and generating no revenue from H200 chip sales in China. Samsung also narrowly averted a massive strike over AI profit-sharing, reaching a tentative deal with workers that, while preventing an 18-day walkout, has exposed deep divisions within the company regarding AI's impact on compensation. Furthermore, OpenAI is reportedly preparing for an IPO as early as September, indicating a rapid move towards public markets for the ChatGPT-maker.
Other notable advancements include President Trump's upcoming cybersecurity directive, which, however, stops short of mandating federal approval for AI models before release. In robotics, experts caution against expecting a single 'ChatGPT moment,' emphasizing that transformation will be gradual and often involves hidden human labor. Intriguing new research suggests rocks could generate hydrogen while storing CO2 and producing geothermal power, with AI already aiding in uncovering hidden geothermal resources. Geopolitical tensions are also accelerating the EU's breakup with Big Tech, fostering a shift towards homegrown software. Battery technology is poised for a revolution with solid-state breakthroughs promising faster and safer alternatives to current lithium-ion models. Finally, on a humanitarian note, the Biden administration's CBP One mobile app, intended to streamline asylum requests at the US southern border, has instead become a significant obstacle for many vulnerable migrants, plagued by technical issues and digital literacy barriers that endanger those most in need of protection.




