Having demonstrated that it has the operational capability to transport humans safely to the moon and back, the United States is moving on to its next major aim: It wants nuclear reactors in orbit and on the lunar surface by 2030. For such a feat, the National Aeronautics and Space Administration will have to work in conjunction with the Department of Defense and the Department of Energy.

In a post on X, the White House Office of Science and Technology Policy (OSTP) unveiled a document with new guidelines for federal agencies to establish the space nuclear technology road map for the coming years. This, they say, will ensure “US space superiority.”

At present, space instruments use solar power to operate. However, this is considered impractical for more complex purposes. Although technically there is always sunlight, the power is intermittent and almost always requires bulky batteries to store it.

Reactors produce fairly continuous energy for years through nuclear fission. They can also be used for so-called nuclear electric propulsion. Continuous output makes them the most viable option for lunar base subsistence, but they can also allow spacecraft to undertake long or complex missions without worrying about depleting a limited supply of chemical fuel.

Nuclear technology, in short, makes it possible to go farther, with more payload, for longer, and with fewer constraints.

According to the memorandum, the US goal is to put a medium-power reactor in orbit by 2028, with a variant designed for nuclear electric propulsion, and a first functional large reactor on the surface of the moon by 2030. To achieve this, both NASA and the Pentagon will develop energy technologies in parallel, using the current strategy of competition among contractors.

The reactors will have to be modular and scalable, and will have to include applications for both future life on the moon and space propulsion. For its part, the DOE will have to ensure that these projects have the fuel, infrastructure, and safety features necessary to achieve their objectives. In addition, the agency will evaluate whether the industry has the capacity to produce up to four reactors in five years.

The plan contemplates technologies that produce at least 20 kilowatts of electricity (kWe) for three years in orbit and at least five years on the lunar surface. In the meantime, they should have a design capable of raising power to 100 kWe. The first designs should arrive within a year.

Finally, the order tasks the OSTP with creating a road map for the initiative, noting obstacles and recommendations for addressing them.

“Nuclear power in space will give us the sustained electricity, heating, and propulsion essential to a permanent presence on the moon, Mars, and beyond,” OSTP posted. For his part, NASA administrator Jared Isaacman posted, “The time has come for America to get underway on nuclear power in space.” The message was followed by an emoji of a US flag.

The plan provides a common framework for each agency to work within. In the background, the race for space infrastructure is evidence of technological competition with China, which is also seeking advanced energy capabilities for the moon.

This story originally appeared in WIRED en Español and has been translated from Spanish.

Nvidia is the undisputed king of AI chips. But thanks to the AI it helped build, the champ could soon face growing competition.

Modern AI runs on Nvidia designs, a dynamic that has propelled the company to a market cap of well over $4 trillion. Each new generation of Nvidia chip allows companies to train more powerful AI models using hundreds or thousands of processors networked together inside vast data centers. One reason for Nvidia’s success is that it provides software to help program each new generation of chip. That may soon not be such a differentiated skill.

A startup called Wafer is training AI models to do one of the most difficult and important jobs in AI—optimizing code so that it runs as efficiently as possible on a particular silicon chip.

Emilio Andere, cofounder and CEO of Wafer, says the company performs reinforcement learning on open source models to teach them to write kernel code, or software that interacts directly with hardware in an operating system. Andere says Wafer also adds “agentic harnesses” to existing coding models like Anthropic’s Claude and OpenAI’s GPT to soup up their ability to write code that runs directly on chips.

Many prominent tech companies now have their own chips. Apple and others have for years used custom silicon to improve the performance and the efficiency of software running on laptops, tablets, and smartphones. At the other end of the scale, companies like Google and Amazon mint their own silicon to improve the performance of their cloud-computing platforms. Meta recently said it would deploy 1 gigawatt of compute capacity with a new chip developed with Broadcom. Deploying custom silicon also involves writing a lot of code so that it runs smoothly and efficiently on the new processor.

Wafer is working with companies including AMD and Amazon to help optimize software to run efficiently on their hardware. The startup has so far raised $4 million in seed funding from Google’s Jeff Dean, Wojciech Zaremba of OpenAI, and others.

Andere believes that his company’s AI-led approach has the potential to challenge Nvidia’s dominance. A number of high-end chips now offer similar raw floating point performance—a key industry benchmark of a chip’s ability to perform simple calculations—to Nvidia’s best silicon.

“The best AMD hardware, the best [Amazon] Trainium hardware, the best [Google] TPUs, give you the same theoretical flops to Nvidia GPUs,” Andere told me recently. “We want to maximize intelligence per watt.”

Performance engineers with the skill needed to optimize code to run reliably and efficiently on these chips are expensive and in high demand, Andere says, while Nvidia’s software ecosystem makes it easier to write and maintain code for its chips. That makes it hard for even the biggest tech companies to go it alone.

When Anthropic partnered with Amazon to build its AI models on Trainium, for instance, it had to rewrite its model’s code from scratch to make it run as efficiently as possible on the hardware, Andere says.

Of course, Anthropic’s Claude is now one of many AI models that are now superhuman at writing code. So Andere reckons it may not be long before AI starts consuming Nvidia software advantage.

“The moat lives in the programmability of the chip,” Andere says in reference to the libraries and software tools that make it easier to optimize code for Nvidia hardware. “I think it’s time to start rethinking whether that’s actually a strong moat.”

Besides making it easier to optimize code for different silicon, AI may soon make it easier to design chips themselves. Ricursive Intelligence, a startup founded by two ex-Google engineers, Azalia Mirhoseini and Anna Goldie, is developing new ways to design computer chips with artificial intelligence. If its technology takes off, a lot more companies could branch into chip design, creating custom silicon that runs their software more efficiently.

A new generation of experimental, frontier AI models are rapidly developing the ability to discover and exploit software vulnerabilities and business leaders need to start to pay attention, the UK government has warned.

In an open letter to Britain’s business leaders published on 15 April, business secretary Liz Kendall said the threats organisations face in cyber space are changing and their responses need to change, too.

“For years, the most serious cyber attacks have relied on a small number of highly skilled criminals. That is now shifting,” she said. “AI models are becoming capable of doing work that previously required rare expertise: finding weaknesses in software, writing the code to exploit them, and doing so at a speed and scale that would have been impossible even a year ago.”

Following the recent debut of Anthropic’s frontier model, Mythos, and its accompanying Project Glasswing – which is intended to give some of the world’s largest technology companies a head start on addressing the vulnerabilities it can supposedly uncover – Kendall revealed that the UK’s AI Security Institute (AISI) operated by the Department for Science, Innovation and Technology (DSIT) has been testing out its capabilities.

She said AISI had found Mythos to be “substantially more capable at cyber offence than any model we have previously assessed.”

According to the AISI, frontier model capabilities are doubling every four months, down from eight months in the recent past.

“This finding is significant both for what it means today, but also because it highlights the speed at which AI capabilities are increasing and the threats they potentially pose,” said Kendall

“OpenAI also announced scaling up their Trusted Access for Cyber programme last night, showing that AI’s accelerating impact on cyber is not isolated to a single company, and we expect more to follow.

“The trajectory is clear and therefore it is vital that we are prepared for frontier AI model capabilities to rapidly increase over the next year, and plan accordingly for that outcome,” she said.