Plus: how a blacklisted Chinese company kept buying Nvidia's best AI chips. Our latest Innovators Under 35 list offers a glimpse. Every year, we recognize 35 people from around the world who are doing groundbreaking scientific work and building clever technical fixes for sticky problems. By finding the top young innovators globally and learning what they're focused on, we aim to give readers a sense of the advances to expect in the years to come. This year's honorees were selected from 550 nominations, with 44 expert judges helping our editors evaluate the finalists. Each works in one of four categories: biotechnology, AI, computing and robotics, and climate and energy--and has already made clear progress toward their goals.

Danijar Hafner has a startup in stealth and a long track record of teaching AI agents about our world. Danijar Hafner's office in San Francisco's SoMa district sits mostly empty. His brand-new startup is still in stealth mode and doesn't even have its name on the door. On the day I visit, there's only one other person there, and little in the way of furniture. But what it lacks in decor, it makes up for in robots. Humanoids of various shapes and sizes hang like marionettes from racks that run down the center of the wide-open space.

Plus: OpenAI agents hijacked a German website before the Hugging Face hack. How much hydrogen awaits us underground? A flurry of exploration efforts is searching for underground stores of hydrogen gas, which could provide a valuable source of zero-carbon fuel. The hunt has spread all over the world and engaged dozens of startups, including the Bill Gates-backed Koloma, which has been poking around the US Midwest to reach ancient oceanic rocks associated with hydrogen production. But the search so far has come up short. No one has yet reported finding a commercially viable reservoir of the gas, and public data on what has been found remains in short supply.

With AI inference now driving enterprise workloads, organizations must rethink infrastructure for speed, efficiency, scalability, and performance per watt to unlock AI's real-world potential. The era of AI inference has arrived. Imagine a healthcare system analyzing millions of data points in real time to accelerate life-saving medical research, or an intelligent assistant instantly resolving thousands of complex customer needs at once. These real-world breakthroughs rely on advanced infrastructure acting as the engine of continuous intelligence, powering real-time services while also supporting an increasingly intelligent edge of IoT and consumer devices. However, in this inference-driven landscape, every delay, bottleneck, or wasted watt directly affects human outcomes and operating costs. This shift changes what infrastructure must deliver.

Plus: OpenAI has launched Astra, its most capable model yet. Battlefields in Ukraine are littered with the remnants of drones. But behind all that wreckage, there's a new gold mine for the defense sector: the data those drones generate. Ukraine has begun making millions of data points gathered during tens of thousands of drone flights available to military contractors and commercial companies. It's a quick way to attract funding and partnerships, but it turns the front line into a model training site, using the chaos of war to create conditions that AI companies struggle to reproduce. As this new industry takes shape, we need a regulatory system that ensures battlefield data isn't treated like ordinary commercial material.

The front line is being turned into a place to train AI models, and right now it's effectively a regulation-free zone. Battlefields in Ukraine are littered with the remnants of drones, which are now firmly established as a critical weapon of modern warfare. The data drones generate will far outlast the wars in which they are used to fight, increasingly becoming part of the AI architecture that shapes even civilian life. For every flight, unmanned systems collect thousands of points of data, from images and video to controller inputs. Together, those records show how a machine and a person responded to constantly shifting circumstances. Ukraine has now begun converting that experience into a resource.

Plus: New York City has banned AI in elementary and middle schools. Digital harms have become the defining fear of American parents. In response, they're increasingly turning to content-monitoring apps that scan their children's texts, photos, emails, and chats, issuing alerts whenever an algorithm flags something it deems dangerous. These tools have had genuine successes, preventing suicide attempts, intercepting predators, and averting school shootings. But they can also cause harm themselves, from false alarms and unnecessary interventions to damaged trust and anxiety. Child-safety researchers say better approaches exist.

As agentic AI moves from pilots to enterprise-wide deployment, orchestration, data, governance, and clear business objectives are becoming critical for scaling, says chief operating officer at NiCE Arun Chandra. As agentic AI moves from experimentation toward enterprise deployment, the challenge is figuring out how agents can work together, connect to the systems and data they need, and operate safely across the workflows that run a business. Although agentic AI has been adopted by some 80% of Fortune 500 companies, progress toward meaningful scale remains uneven, with many organizations still working through isolated pilots. For Arun Chandra, chief operating officer at NiCE, the first step is moving beyond experimentation for its own sake. "Everybody's trying to figure out what can we do with this technology?" he says. But scaling requires a clearer connection to business strategy: Organizations need to define whether they are trying to increase revenue, reduce costs, or pursue another strategic or financial objective.

Plus: OpenAI is restricting its next model after rating it a "critical" cyber risk. Can you fare any better? Puzzles and games have always been central to AI development. The term "machine learning" was popularized in a 1959 article about an algorithm that learned to play checkers. Chess and Go are famous AI test beds too. Judged purely on its puzzling skills, AI is improving a lot--and quickly.

A nonprofit organization called the Fermi Explorer Mission announced today that it intends to launch a spacecraft to our nearest star system by the end of 2029. It's a hugely ambitious mission--if all goes well, the spacecraft could take up to 80,000 years to arrive at Alpha Centauri, which is 4.4 light-years away. And the spacecraft will follow a novel trajectory discovered by an AI system developed by Physical Superintelligence (PSI), an AI physics research lab. PSI is launching today with $58 million in funding led by Breakthrough Energy, a climate-focused investment group founded by Microsoft cofounder Bill Gates. It's not the first time this has been tried.