Startup Loft Orbital Is Launching AI-Powered Satellites This Fall

Building space-based AI. Designing new molecules with quantum computers. Why pink noise is bad for your sleep. All that and more in this week’s edition of The Prototype. To get it in your inbox, sign up here.Loft's Mission Operations in Toulouse, FranceLoft OrbitalSpace is the latest place for AI, with major tech companies like SpaceX, Nvidia, Google, Blue Origin and more announcing plans for space-based data centers. But as I discussed a few weeks ago, orbital data centers face major economic and engineering hurdles thanks to the laws of physics. San Francisco-based Loft Orbital, which builds standardized satellite platforms for a variety of customers, is offering up a different vision of AI in space, at least in the near term. Later this year, it plans to launch a 10-satellite constellation focused on doing AI in space for real-world applications like environmental monitoring or law enforcement. It won’t be easy, Paul Lasserre, who heads the project, told me. Doing AI in space with current technology comes with significant constraints. One is power–the satellites are only going to have about 500 watts available to them. That’s less power than a high-end gaming computer. Plus all the chips and hardware need to be built to withstand radiation in space. On top of all that, there are severe limitations in bandwidth that slow the speed of data uploading and downloading. The upshot of these constraints is that Loft is not going to be using large models like Claude or ChatGPT in space. Instead, it will be working with lightweight AI models that can run on the satellites themselves. These models will take information from cameras and other sensors to alert people on the ground to actionable information, such as possible wildfires or piracy. That could mean immediate action by law enforcement, or they might instead opt to prioritize downloading satellite data about that area to crunch at Earth-bound data centers. Lasserre acknowledges that because these models are smaller, they could be prone to false positives. But he likens the approach to early AI systems in medical imaging, which were used to prioritize patients for doctors to examine. Even when they weren’t always right about the risk, they made it more likely that doctors were seeing patients with more urgent needs. In addition to providing these services to customers, Loft is also planning an AI application marketplace on its satellites. This would enable government and commercial partners to develop and deploy their own algorithms on Loft’s systems. Lasserre also said the company is working on developing better visual AI, which could combine data from different kinds of sensors to better understand what’s happening on the Earth’s surface. “This is about turning satellites into intelligent, autonomous systems,” he said. “Instead of just being cameras in orbit.”Discovery of the Week: Nanoengineering New MoleculesIBM Research and the University of ManchesterA team of international researchers from universities and IBM Research nanoengineered a new kind of molecule whose electrons have a structure never before seen in nature. The results were published this week in Science. The molecule is made from carbon and chlorine, and was built by precisely manipulating another chemical atom-by-atom in a vacuum. The result was that the molecule’s electrons have a structure containing 90-degree “twists”, giving it a topology called a half-Möbius. This finding is “significant,” Igor Alabugin, a professor at Florida State University not involved in the study, told me, because most molecules have one of two different electronic structures. He went on to say that new tech advances “allowed the authors to realize a molecular architecture that would have seemed nearly impossible not long ago.”After building the new molecule, the researchers used an IBM Heron quantum processor to interpret their data, simplifying the numbers to the point where a conventional supercomputer could finish crunching them. This was crucial to the process, Alessandro Curioni, director of IBM’s lab in Zurich, told me. In the future, the team plans to try to use quantum computers to design the molecules first before engineering them.The molecule itself is also interesting, Curioni said, and the team plans to do more work with it and other molecules like it. That’s because the electron structure can be switched to be “left-handed” or “right-handed”. That switching could have applications for computing or sensing. Additionally, the twists in the electron structure could be used to generate large-scale magnetic fields on a microscopic level. Coupled with quantum computing, this could open up a number of interesting future applications. “I strongly believe that we are entering a new world in which quantum computing is going to become a fundamental instrument,” he said, enabling chemists to precisely engineer, on the nanoscale, molecules with specific properties.

Making The Web Quantum-ProofGoogle has announced plans to update Chrome so that HTTPS certificates will be secured against quantum computer attacks. Conventional encryption is vulnerable to quantum computers once they reach a certain level of scale. That was once thought to be decades away, but recent research suggests that much smaller quantum computers might be able to crack it, speeding up the timeline.For Google, the challenge is that HTTPS certificates for browsing need to be bigger to secure them against quantum computers. But the bigger they are, the slower the browser will be. That risks people disabling encryption so they can surf the web faster. To avoid this, Google is turning to a mathematical innovation called Merkle trees, which can theoretically provide better security in a size not much bigger than current certificates. Google says that it’s already experimenting with the quantum-proof certificates on the web today, and intends to gradually deploy it in collaboration with Cloudflare with a goal of full-rollout by the third quarter of 2027. If all works well, that should give website owners time to secure their traffic against large-scale quantum computers years before they enter the market. Plus Google and Cloudflare’s market power could make this a de facto standard for the web, even if you’re not using Chrome.

The Hot Take: Robotics Is OverhypedEach week, I ask investors for their take on tech trends within their industries. Today I’m featuring thoughts from Pegah Ebrahimi, cofounder and managing partner of FPV Ventures, which invests in “mission-driven founders” from a wide swathe of industries. Portfolio companies include logistics startup Flexport, graphic design company Canva and gene-editing biotech Strand Therapeutics.Pegah Ebrahimi© Rafaella Pedroso 2025. All rights reserved.What is being overhyped right now?Robotics. People are underestimating the timeline and overestimating the number of winners. The last time we saw something with similar potential was with autonomous vehicles. They attracted massive amounts of capital but took years of development before we got to where we are today. And now you can count the number of players still in the race on one hand. Robotics will likely follow a similar pattern because they’re costly to build and to maintain. Not to mention, China has a big head start in making machines at lower costs, so the margins might not pan out for U.S. hardware companies. The smarter bet is on the underlying models, not the hardware. It will be big, just not as soon (or as much) as the hype suggests. What should more people be talking about today?Security and privacy in the age of AI. Think about it: When your bank deploys an AI agent to manage transactions, the attack surface isn’t your password anymore, it’s the agent itself. Most people have no idea what they’re exposing because of everyday interactions and conversations they’re having with chatbots. The concept of hacking is about to evolve into something way more complex and the privacy and security space is seriously underinvested given how high the stakes are. What are we all going to be talking about in five years?Healthcare (biology and chemistry). It’s slow-moving by nature, with long time horizons and heavy regulation, but AI in hard sciences is where the real opportunity is. Healthcare is one of the top spending categories globally and costs continue to climb. In five years, we’ll stop talking about software disruption and start talking about the very real problems we can actually solve. Not how can I do my CRM faster, but how can I discover new sources of energy or solve climate change. Plus, five years from now quantum computing will be within reach, making small scale gen AI seem small compared to the things we can potentially solve down the line.

On My RadarAnthropic vs. the Pentagon: As they threatened when I talked about this last week, the Department of Defense declared Anthropic a supply chain risk on Thursday “effective immediately.” This is the first time such a declaration has been made about an American company, and it effectively prevents Anthropic’s AI products from being used by the military or any of its contractors for military purposes (even though Claude was reportedly used during some strikes against Iran in the past few days.) Anthropic CEO Dario Amodei said that his company plans to challenge the declaration in court. Google’s Trillion Dollar AI Bet: My colleague Richard Nieva reported earlier this week that Google is planning to spend nearly $1.9 trillion on AI infrastructure over the next ten years. That’s a major bet that the need for AI data centers will keep growing steadily, which may well be true: A recent MIT CSAIL paper suggests that about 80-90% of performance differences for large models comes from higher training compute. But the real question is, is there enough economic demand to pay for all that computation?Pro Science Tip: Pink Noise Could Hurt Your SleepYou’ll sometimes see pink noise–a kind of noise that includes a wide variety of frequencies and to me sounds like TV–static, promoted as a way to help you sleep. (In fact, the Google AI overview I did after a search for “pink noise” promoted this very idea.) A new study, however, finds that just ain’t so. Researchers from Penn Medicine found that listening to pink noise while you sleep reduces both REM sleep (when you’re dreaming) and deep sleep (when your brain consolidates memories and performs other restorative functions). Meaning overall it’s probably giving you a worse night’s sleep. What’s Entertaining Me This WeekThis week I’ve been listening to the album Now You’re A Circle by Boys Go To Jupiter. It’s a fast follow-up to their 2025 album Meet Me After Practice, which was one of my favorite albums last year. This one is another strong entry from the indie pop trio, starting strong with the first track and never letting up on the band’s unique combination of theatricality and soul.

More From Forbes ForbesNo AI Plan, No Loan.

Small Business Lenders Force Main Street To Face The FutureBy Brandon KochkodinForbesCursor Goes To War For AI Coding DominanceBy Anna TongForbesThis New Company Will Build Ukraine’s Deadly Drone Boats For Western MilitariesBy Iain Martin