Science – Page 7

Researchers find what makes AI chatbots politically persuasive

AI, Behavioral science, chat gpt, Computer science, grok, Human behavior, large language models, Science / Beth Washington / December 6, 2025

A massive study of political persuasion shows AIs have, at best, a weak effect.

Roughly two years ago, Sam Altman tweeted that AI systems would be capable of superhuman persuasion well before achieving general intelligence—a prediction that raised concerns about the influence AI could have over democratic elections.

To see if conversational large language models can really sway political views of the public, scientists at the UK AI Security Institute, MIT, Stanford, Carnegie Mellon, and many other institutions performed by far the largest study on AI persuasiveness to date, involving nearly 80,000 participants in the UK. It turned out political AI chatbots fell far short of superhuman persuasiveness, but the study raises some more nuanced issues about our interactions with AI.

AI dystopias

The public debate about the impact AI has on politics has largely revolved around notions drawn from dystopian sci-fi. Large language models have access to essentially every fact and story ever published about any issue or candidate. They have processed information from books on psychology, negotiations, and human manipulation. They can rely on absurdly high computing power in huge data centers worldwide. On top of that, they can often access tons of personal information about individual users thanks to hundreds upon hundreds of online interactions at their disposal.

Talking to a powerful AI system is basically interacting with an intelligence that knows everything about everything, as well as almost everything about you. When viewed this way, LLMs can indeed appear kind of scary. The goal of this new gargantuan AI persuasiveness study was to break such scary visions down into their constituent pieces and see if they actually hold water.

The team examined 19 LLMs, including the most powerful ones like three different versions of ChatGPT and xAI’s Grok-3 beta, along with a range of smaller, open source models. The AIs were asked to advocate for or against specific stances on 707 political issues selected by the team. The advocacy was done by engaging in short conversations with paid participants enlisted through a crowdsourcing platform. Each participant had to rate their agreement with a specific stance on an assigned political issue on a scale from 1 to 100 both before and after talking to the AI.

Scientists measured persuasiveness as the difference between the before and after agreement ratings. A control group had conversations on the same issue with the same AI models—but those models were not asked to persuade them.

“We didn’t just want to test how persuasive the AI was—we also wanted to see what makes it persuasive,” says Chris Summerfield, a research director at the UK AI Security Institute and co-author of the study. As the researchers tested various persuasion strategies, the idea of AIs having “superhuman persuasion” skills crumbled.

Persuasion levers

The first pillar to crack was the notion that persuasiveness should increase with the scale of the model. It turned out that huge AI systems like ChatGPT or Grok-3 beta do have an edge over small-scale models, but that edge is relatively tiny. The factor that proved more important than scale was the kind of post-training AI models received. It was more effective to have the models learn from a limited database of successful persuasion dialogues and have them mimic the patterns extracted from them. This worked far better than adding billions of parameters and sheer computing power.

This approach could be combined with reward modeling, where a separate AI scored candidate replies for their persuasiveness and selected the top-scoring one to give to the user. When the two were used together, the gap between large-scale and small-scale models was essentially closed. “With persuasion post-training like this we matched the Chat GPT-4o persuasion performance with a model we trained on a laptop,” says Kobi Hackenburg, a researcher at the UK AI Security Institute and co-author of the study.

The next dystopian idea to fall was the power of using personal data. To this end, the team compared the persuasion scores achieved when models were given information about the participants’ political views beforehand and when they lacked this data. Going one step further, scientists also tested whether persuasiveness increased when the AI knew the participants’ gender, age, political ideology, or party affiliation. Just like with model scale, the effects of personalized messaging created based on such data were measurable but very small.

Finally, the last idea that didn’t hold up was AI’s potential mastery of using advanced psychological manipulation tactics. Scientists explicitly prompted the AIs to use techniques like moral reframing, where you present your arguments using the audience’s own moral values. They also tried deep canvassing, where you hold extended empathetic conversations with people to nudge them to reflect on and eventually shift their views.

The resulting persuasiveness was compared with that achieved when the same models were prompted to use facts and evidence to back their claims or just to be as persuasive as they could without specifying any persuasion methods to use. I turned out using lots of facts and evidence was the clear winner, and came in just slightly ahead of the baseline approach where persuasion strategy was not specified. Using all sorts of psychological trickery actually made the performance significantly worse.

Overall, AI models changed the participants’ agreement ratings by 9.4 percent on average compared to the control group. The best performing mainstream AI model was Chat GPT 4o, which scored nearly 12 percent followed by GPT 4.5 with 10.51 percent, and Grok-3 with 9.05 percent. For context, static political ads like written manifestos had a persuasion effect of roughly 6.1 percent. The conversational AIs were roughly 40–50 percent more convincing than these ads, but that’s hardly “superhuman.”

While the study managed to undercut some of the common dystopian AI concerns, it highlighted a few new issues.

Convincing inaccuracies

While the winning “facts and evidence” strategy looked good at first, the AIs had some issues with implementing it. When the team noticed that increasing the information density of dialogues made the AIs more persuasive, they started prompting the models to increase it further. They noticed that, as the AIs used more factual statements, they also became less accurate—they basically started misrepresenting things or making stuff up more often.

Hackenburg and his colleagues note that we can’t say if the effect we see here is causation or correlation—whether the AIs are becoming more convincing because they misrepresent the facts or whether spitting out inaccurate statements is a byproduct of asking them to make more factual statements.

The finding that the computing power needed to make an AI model politically persuasive is relatively low is also a mixed bag. It pushes back against the vision that only a handful of powerful actors will have access to a persuasive AI that can potentially sway public opinion in their favor. At the same time, the realization that everybody can run an AI like that on a laptop creates its own concerns. “Persuasion is a route to power and influence—it’s what we do when we want to win elections or broke a multi-million-dollar deal,” Summerfield says. “But many forms of misuse of AI might involve persuasion. Think about fraud or scams, radicalization, or grooming. All these involve persuasion.”

But perhaps the most important question mark in the study is the motivation behind the rather high participant engagement, which was needed for the high persuasion scores. After all, even the most persuasive AI can’t move you when you just close the chat window.

People in Hackenburg’s experiments were told that they would be talking to the AI and that the AI would try to persuade them. To get paid, a participant only had to go through two turns of dialogue (they were limited to no more than 10). The average conversation length was seven turns, which seemed a bit surprising given how far beyond the minimum requirement most people went. Most people just roll their eyes and disconnect when they realize they are talking with a chatbot.

Would Hackenburg’s study participants remain so eager to engage in political disputes with random chatbots on the Internet in their free time if there was no money on the table? “It’s unclear how our results would generalize to a real-world context,” Hackenburg says.

Science, 2025. DOI: 10.1126/science.aea3884

Jacek Krywko is a freelance science and technology writer who covers space exploration, artificial intelligence research, computer science, and all sorts of engineering wizardry.

Researchers find what makes AI chatbots politically persuasive Read More »

Rare set of varied factors triggered Black Death

black death, epidemiology, medieval history, Science, volcanology / Kelly Newman / December 5, 2025

The culprit is a bacterium called Yersinia pestis, and it’s well known that it spreads among mammalian hosts via fleas, although it only rarely spills over to domestic animals and humans. The Black Death can be traced to a genetically distinct strain of Y. pestis that originated in the Tien Shan mountains west of what is now Kyrgyzstan, spreading along trade routes to Europe in the 1340s. However, according to the authors of this latest paper, there has been little attention focused on several likely contributing factors: climate, ecology, socioeconomic pressures, and the like.

The testimony of the tree rings

Taking tree samples from the Pyrenees. Credit: Ulf Büntgen

“This is something I’ve wanted to understand for a long time,” said co-author Ulf Büntgen of the University of Cambridge. “What were the drivers of the onset and transmission of the Black Death, and how unusual were they? Why did it happen at this exact time and place in European history? It’s such an interesting question, but it’s one no one can answer alone.”

Büntgen et al. collected core and disc samples from both living and relict trees at eight European sites to reconstruct summer temperatures for that time period. They then compared that data with estimates of sulphur injections into the atmosphere from volcanic eruptions, based on geochemical analyses of ice core samples collected from Antarctica and Greenland.

They studied a wide range of written sources across Eurasia—chronicles, treatises, historiography, and even a bit of poetry—looking for mention of atmospheric and optical phenomena linked to volcanic dust veils between 1345 and 1350 CE. They also looked for mentions of extreme weather events, economic conditions, and reports of dearth or famine across Eurasia during that time period. Information about the trans-Mediterranean grain trade was gleaned from administrative records and letters.

Rare set of varied factors triggered Black Death Read More »

New report warns of critical climate risks in Arab region

arabian peninsula, climate change, extreme heat, heat index, Science, syndication / Kelly Newman / December 5, 2025

The new WMO report shows that the foundations of daily life across the Arab region, including farms, reservoirs, and aquifers that feed and sustain millions, are being pushed to the brink by human-caused warming.

Across northwestern Africa’s sun-blasted rim, the Maghreb, six years of drought have slashed wheat yields, forcing countries such as Morocco, Algeria, and Tunisia to import more grain, even as global prices rise.

In parts of Morocco, reservoirs have fallen to record low levels. The government has enacted water restrictions in major cities, including limits on household use, and curtailed irrigation for farmers. Water systems in Lebanon have already crumbled under alternating floods and droughts, and in Iraq and Syria, small farmers are abandoning their land as rivers shrink and seasonal rains become unreliable.

The WMO report ranked 2024 as the hottest year ever measured in the Arab world. Summer heatwaves spread and persisted across Syria, Iraq, Jordan, and Egypt. Parts of Iraq recorded six to 12 days with highs above 50° Celsius (122° Fahrenheit), conditions that are life-threatening even for healthy adults. Across the region, the report noted an increase in the number of heat-wave days in recent decades while humidity has declined. The dangerous combination speeds soil drying and crop damage.

By contrast, other parts of the region—the United Arab Emirates, Oman, and southern Saudi Arabia—were swamped by destructive record rains and flooding during 2024. The extremes will test the limits of adaptation, said Rola Dashti, executive secretary of the Economic and Social Commission for Western Asia, who often works with the WMO to analyze climate impacts.

Climate extremes in 2024 killed at least 300 people in the region. The impacts are hitting countries already struggling with internal conflicts, and where the damage is under-insured and under-reported. In Sudan alone, flooding damaged more than 40 percent of the country’s farmland.

But with 15 of the world’s most arid countries in the region, water scarcity is the top issue. Governments are investing in desalination, wastewater recycling, and other measures to bolster water security, but the adaptation gap between risks and readiness is still widening.

The worst is ahead, Dashti said in a WMO statement, with climate models showing a “potential rise in average temperatures of up to 5° Celsius (9° Fahrenheit) by the end of the century under high-emission scenarios.” The new report is important, she said, because it “empowers the region to prepare for tomorrow’s climate realities.”

This article originally appeared on Inside Climate News, a nonprofit, non-partisan news organization that covers climate, energy, and the environment. Sign up for their newsletter here.

New report warns of critical climate risks in Arab region Read More »

Rocket Report: Blunder at Baikonur; do launchers really need rocket engines?

isar aerospace, launch, openai, rocket report, Science, South Korea, Space, spacex, starship, stoke space, vega c / Kelly Newman / December 5, 2025

The Department of the Air Force approves a new home in Florida for SpaceX’s Starship.

South Korea’s Nuri 1 rocket is lifted vertical on its launch pad in this multi-exposure photo. Credit: Korea Aerospace Research Institute

Welcome to Edition 8.21 of the Rocket Report! We’re back after the Thanksgiving holiday with more launch news. Most of the big stories over the last couple of weeks came from abroad. Russian rockets and launch pads didn’t fare so well. China’s launch industry celebrated several key missions. SpaceX was busy, too, with seven launches over the last two weeks, six of them carrying more Starlink Internet satellites into orbit. We expect between 15 and 20 more orbital launch attempts worldwide before the end of the year.

As always, we welcome reader submissions. If you don’t want to miss an issue, please subscribe using the box below (the form will not appear on AMP-enabled versions of the site). Each report will include information on small-, medium-, and heavy-lift rockets, as well as a quick look ahead at the next three launches on the calendar.

Another Sarmat failure. A Russian intercontinental ballistic missile (ICBM) fired from an underground silo on the country’s southern steppe on November 28 on a scheduled test to deliver a dummy warhead to a remote impact zone nearly 4,000 miles away. The missile didn’t even make it 4,000 feet, Ars reports. Russia’s military has been silent on the accident, but the missile’s crash was seen and heard for miles around the Dombarovsky air base in Orenburg Oblast near the Russian-Kazakh border. A video posted by the Russian blog site MilitaryRussia.ru on Telegram and widely shared on other social media platforms showed the missile veering off course immediately after launch before cartwheeling upside down, losing power, and then crashing a short distance from the launch site.

An unenviable track record … Analysts say the circumstances of the launch suggest it was likely a test of Russia’s RS-28 Sarmat missile, a weapon designed to reach targets more than 11,000 miles (18,000 kilometers) away, making it the world’s longest-range missile. The Sarmat missile is Russia’s next-generation heavy-duty ICBM, capable of carrying a payload of up to 10 large nuclear warheads, a combination of warheads and countermeasures, or hypersonic boost-glide vehicles, according to the Center for Strategic and International Studies. Simply put, the Sarmat is a doomsday weapon designed for use in an all-out nuclear war between Russia and the United States. The missile’s first full-scale test flight in 2022 apparently went well, but the program has suffered a string of consecutive failures since then, most notably a catastrophic explosion last year that destroyed the Sarmat missile’s underground silo in northern Russia.

The easiest way to keep up with Eric Berger’s and Stephen Clark’s reporting on all things space is to sign up for our newsletter. We’ll collect their stories and deliver them straight to your inbox.

Sign Me Up!

ESA fills its coffers for launcher challenge. The European Space Agency’s (ESA) European Launcher Challenge received a significant financial commitment from its member states during the agency’s Ministerial Council meeting last week, European Spaceflight reports. The challenge is designed to support emerging European rocket companies while giving ESA and other European satellite operators more options to compete with the continent’s sole operational launch provider, Arianespace. Through the program, ESA will purchase launch services and co-fund capacity upgrades with the winners. ESA member states committed 902 million euros, or $1.05 billion, to the program at the recent Ministerial Council meeting.

Preselecting the competitors … In July, ESA selected two German companies—Isar Aerospace and Rocket Factory Augsburg—along with Spain’s PLD Space, France’s MaiaSpace, and the UK’s Orbex to proceed with the initiative’s next phase. ESA then negotiated with the governments of each company’s home country to raise money to support the effort. Germany, with two companies on the shortlist, is unsurprisingly a large contributor to the program, committing more than 40 percent of the total budget. France contributed nearly 20 percent, Spain funded nearly 19 percent, and the UK committed nearly 16 percent. Norway paid for 3 percent of the launcher challenge’s budget. Denmark, Portugal, Switzerland, and the Czech Republic contributed smaller amounts.

Europe at the service of South Korea. South Korea’s latest Earth observation satellite was delivered into a Sun-synchronous orbit Monday afternoon following a launch onboard a Vega C rocket by Arianespace, Spaceflight Now reports. The Korea Multi-Purpose Satellite-7 (Kompsat-7) mission launched from Europe’s spaceport in French Guiana. About 44 minutes after liftoff, the Kompsat-7 satellite was deployed into SSO at an altitude of 358 miles (576 kilometers). “By launching the Kompsat-7 satellite, set to significantly enhance South Korea’s Earth observation capabilities, Arianespace is proud to support an ambitious national space program,” said David Cavaillolès, CEO of Arianespace, in a statement.

Something of a rarity … The launch of Kompsat-7 is something of a rarity for Arianespace, which has dominated the international commercial launch market. It’s the first time in more than two years that a satellite for a customer outside Europe has been launched by Arianespace. The backlog for the light-class Vega C rocket is almost exclusively filled with payloads for the European Space Agency, the European Commission, or national governments in Europe. Arianespace’s larger Ariane 6 rocket has 18 launches reserved for the US-based Amazon Leo broadband network. (submitted by EllPeaTea)

South Korea’s homemade rocket flies again. South Korea’s homegrown space rocket Nuri took off from Naro Space Center on November 27 with the CAS500-3 technology demonstration and Earth observation satellite, along with 12 smaller CubeSat rideshare payloads, Yonhap News Agency reports. The 200-ton Nuri rocket debuted in 2021, when it failed to reach orbit on a test flight. Since then, the rocket has successfully reached orbit three times. This mission marked the first time for Hanwha Aerospace to oversee the entire assembly process as part of the government’s long-term plan to hand over space technologies to the private sector. The fifth and sixth launches of the Nuri rocket are planned in 2026 and 2027.

Powered by jet fuel … The Nuri rocket has three stages, each with engines burning Jet A-1 fuel and liquid oxygen. The fuel choice is unusual for rockets, with highly refined RP-1 kerosene or methane being more popular among hydrocarbon fuels. The engines are manufactured by Hanwha Aerospace. The fully assembled rocket stands about 155 feet (47.2 meters) tall and can deliver up to 3,300 pounds (1.5 metric tons) of payload into a polar Sun-synchronous orbit.

Hyundai eyes rocket engine. Meanwhile, South Korea’s space sector is looking to the future. Another company best known for making cars has started a venture in the rocket business. Hyundai Rotem, a member of Hyundai Motor Group, announced a joint program with Korean Air’s Aerospace Division (KAL-ASD) to develop a 35-ton-class reusable methane rocket engine for future launch vehicles. The effort is funded with KRW49 billion ($33 million) from the Korea Research Institute for Defense Technology Planning and Advancement (KRIT).

By the end of the decade … The government-backed program aims to develop the engine by the end of 2030. Hyundai Rotem will lead the engine’s planning and design, while Korean Air, the nation’s largest air carrier, will lead development of the engine’s turbopump. “Hyundai Rotem began developing methane engines in 1994 and has steadily advanced its methane engine technology, achieving Korea’s first successful combustion test in 2006,” Hyundai Rotem said in a statement. “Furthermore, this project is expected to secure the technological foundation for the commercialization of methane engines for reusable space launch vehicles and lay the groundwork for targeting the global space launch vehicle market.”

But who needs rocket engines? Moonshot Space, based in Israel, announced Monday that it has secured $12 million in funding to continue the development of a launch system—powered not by chemical propulsion, but electromagnetism, Payload reports. Moonshot plans to sell other aerospace and defense companies the tech as a hypersonic test platform, while at the same time building to eventually offer orbital launch services. Instead of conventional rocket engines, the system would use a series of electromagnetic coils to power a hardened capsule to hypersonic velocities. The architecture has a downside: extremely high accelerations that could damage or destroy normal satellites. Instead, Moonshot wants to use the technology to send raw materials to orbit, lowering the input costs of the budding in-space servicing, refueling, and manufacturing industries, according to Payload.

Out of the shadows … Moonshot Space emerged from stealth mode with this week’s fundraising announcement. The company’s near-term focus is on building a scaled-down electromagnetic accelerator capable of reaching Mach 6. A larger system would be required to reach orbital velocity. The company’s CEO is the former director-general of Israel’s Ministry of Science, while its chief engineer was the former chief systems engineer for David’s Sling, a critical part of Israel’s missile defense system. (submitted by EllPeaTea)

A blunder at Baikonur. A Soyuz rocket launched on November 27 carrying Roscosmos cosmonauts Sergei Kud-Sverchkov and Sergei Mikayev, as well as NASA astronaut Christopher Williams, for an eight-month mission to the International Space Station. The trio of astronauts arrived at the orbiting laboratory without incident. However, on the ground, there was a serious problem during the launch with the ground systems that support processing of the vehicle before liftoff at Site 31, located at the Baikonur Cosmodrome in Kazakhstan, Ars reports. Roscosmos downplayed the incident, saying only, in passive voice, that “damage to several launch pad components was identified” following the launch.

Repairs needed … However, video imagery of the launch site after liftoff showed substantial damage, with a large service platform appearing to have fallen into the flame trench below the launch table. According to one source, this is a platform located beneath the rocket, where workers can access the vehicle before liftoff. It has a mass of about 20 metric tons and was apparently not secured prior to launch, and the thrust of the vehicle ejected it into the flame trench. “There is significant damage to the pad,” said this source. The damage could throw a wrench into Russia’s ability to launch crews and cargo to the International Space Station. This Soyuz launch pad at Baikonur is the only one outfitted to support such missions.

China’s LandSpace almost landed a rocket. China’s first attempt to land an orbital-class rocket may have ended in a fiery crash, but the company responsible for the mission had a lot to celebrate with the first flight of its new methane-fueled launcher, Ars reports. LandSpace, a decade-old company based in Beijing, launched its new Zhuque-3 rocket for the first time Tuesday (US time) at the Jiuquan launch site in northwestern China. The upper stage of the medium-lift rocket successfully reached orbit. This alone is a remarkable achievement for a new rocket. But LandSpace had other goals for this launch. The Zhuque-3, or ZQ-3, booster stage is architected for recovery and reuse, the first rocket in China with such a design. The booster survived reentry and was seconds away from a pinpoint landing when something went wrong during its landing burn, resulting in a high-speed crash at the landing zone in the Gobi Desert.

Let the games begin … LandSpace got closer to landing an orbital-class booster than any other company on their first try. While LandSpace prepares for a second launch, several more Chinese companies are close to debuting their own reusable rockets. The next of these new rockets, the Long March 12A, is awaiting its first liftoff later this month from another launch pad at the Jiuquan spaceport. The Long March 12A comes from one of China’s established rocket developers, the Shanghai Academy of Spaceflight Technology (SAST), part of the country’s state-owned aerospace enterprise.

China launches a lifeboat. An unpiloted Chinese spacecraft launched on November 24 (US time) and linked with the country’s Tiangong space station a few hours later, providing a lifeboat for three astronauts stuck in orbit without a safe ride home, Ars reports. A Long March 2F rocket lifted off with the Shenzhou 22 spacecraft, carrying cargo instead of a crew. The spacecraft docked with the Tiangong station nearly 250 miles (400 kilometers) above the Earth about three-and-a-half hours later. Shenzhou 22 will provide a ride home next year for three Chinese astronauts. Engineers deemed their primary lifeboat unsafe after finding a cracked window, likely from an impact with a tiny piece of space junk.

In record time … Chinese engineers worked fast to move up the launch of the Shenzhou 22, originally set to fly next year. The launch occurred just 16 days after officials decided they needed to send another spacecraft to the Tiangong station. Shenzhou 22 and its rocket were already in standby at the launch site, but teams had to fuel the spacecraft and complete assembly of the rocket, then roll the vehicle to the launch pad for final countdown preps. The rapid turnaround offers a “successful example for efficient emergency response in the international space industry,” the China Manned Space Agency said. “It vividly embodies the spirit of manned spaceflight: exceptionally hardworking, exceptionally capable, exceptionally resilient, and exceptionally dedicated.”

Another big name flirts with the launch industry. OpenAI chief executive Sam Altman has explored putting together funds to either acquire or partner with a rocket company, a move that would position him to compete with Elon Musk’s SpaceX, the Wall Street Journal reports. Altman reached out to at least one rocket maker, Stoke Space, in the summer, and the discussions picked up in the fall, according to people familiar with the talks. Among the proposals was for OpenAI to make a multibillion-dollar series of equity investments in the company and end up with a controlling stake. The talks are no longer active, people close to OpenAI told the Journal.

Here’s the reason … Altman has been interested in building data centers in space for some time, the Journal reports, suggesting that the insatiable demand for computing resources to power artificial-intelligence systems eventually could require so much power that the environmental consequences would make space a better option. Orbital data centers would allow companies to harness the power of the Sun to operate them. Alphabet’s Google is pursuing a similar concept in partnership with satellite operator Planet Labs. Jeff Bezos and Musk himself have also expressed interest in the idea. Outside of SpaceX and Blue Origin, Stoke Space seems to be a natural partner for such a project because it is one of the few companies developing a fully reusable rocket.

SpaceX gets green light for new Florida launch pad. SpaceX has the OK to build out what will be the primary launch hub on the Space Coast for its Starship and Super Heavy rocket, the most powerful launch vehicle in history, the Orlando Sentinel reports. The Department of the Air Force announced Monday it had approved SpaceX to move forward with the construction of a pair of launch pads at Cape Canaveral Space Force Station’s Space Launch Complex 37 (SLC-37). A “record of decision” on the Environmental Impact Statement required under the National Environmental Policy Act for the proposed Canaveral site was posted to the Air Force’s website, marking the conclusion of what has been a nearly two-year approval process.

Get those Starships ready … SpaceX plans to build two launch towers at SLC-37 to augment the single tower under construction at NASA’s Kennedy Space Center, just a few miles to the north. The three pads combined could support up to 120 launches per year. The Air Force’s final approval was expected after it released a draft Environmental Impact Statement earlier this year, suggesting the Starship pads at SLC-37 would have no significant negative impacts on local environmental, historical, social, and cultural interests. The Air Force also found SpaceX’s plans at SLC-37, formerly leased by United Launch Alliance, will have no significant impact on the company’s competitors in the launch industry. SpaceX also has two launch towers at its Starbase facility in South Texas.

Next three launches

Dec. 5: Kuaizhou 1A | Unknown Payload | Jiuquan Satellite Launch Center, China | 09: 00 UTC

Dec. 6: Hyperbola 1 | Unknown Payload | Jiuquan Satellite Launch Center, China | 04: 00 UTC

Dec. 6: Long March 8A | Unknown Payload | Wenchang Space Launch Site, China | 07: 50 UTC

Stephen Clark is a space reporter at Ars Technica, covering private space companies and the world’s space agencies. Stephen writes about the nexus of technology, science, policy, and business on and off the planet.

Rocket Report: Blunder at Baikonur; do launchers really need rocket engines? Read More »

Welcome to “necroprinting”—3D printer nozzle made from mosquito’s proboscis

3D printing, Biology, Biomechanics, materials science, mosquitos, printing, Science / Kelly Newman / December 4, 2025

“To integrate the proboscis, we first removed it from an already euthanized mosquito under a microscope,” Cao explains. Then the proboscis/nozzle was aligned with the outlet of the plastic tip. Finally, the proboscis and the tip were bonded with UV-curable resin.

The necroprinter achieved a resolution ranging from 18 to 22 microns, which was two times smaller than the printers using the smallest commercially available metal dispensing tips. The first print tests included honeycomb structures measuring 600 microns, a microscale maple leaf, and scaffolds for cells.

But there were still areas in which human-made technology managed to beat Mother Nature.

Glass and pressure

The first issue with mosquito nozzles was their relatively low resistance to internal pressure. “It was impressive but still too low to accommodate some high viscosity inks,” Cao said.

These inks, which look more like a paste than a typical fluid, hold shape better, which translates into more geometrically accurate models that do not slump or spread under their own weight. This was a problem that Cao’s test prints experienced to an extent.

But this wasn’t the only area where human-made technology managed to beat nature. While mosquito nozzles could outperform plastic or metal alternatives in precision, they could not outperform glass dispensing tips, which can print lines below one micron across and withstand significantly higher pressures.

The researchers already have some ideas about how to bridge at least a part of this gap, though. “One possible solution is to use mosquito proboscis as the core and coat it with ceramic layers to provide much higher strength,” Cao said. And if the pressure problem is solved, the 18–22 microns resolution should be good enough for plenty of things.

Cao thinks that in the future, printers like this could be used to print scaffolds for living cells or microscopic electronic components. The idea is to replace expensive, traditional 3D printing nozzles with more affordable organic counterparts. The key advantages of mosquito nozzles, he says, are low cost and ubiquity.

Mosquitoes live almost everywhere on Earth and are easy to rear. The team estimates that organic 3D printing nozzles made from mosquito proboscises should cost around 80 cents; the glass and metal alternatives, the researchers state in the paper, cost between 32 and 100 times more.

“We already started doing more research on mosquitoes themselves and hope to develop more engineering solutions, not only to leverage their deceased bodies but also to solve practical problems they cause,” Cao said.

Science Advances, 2025. DOI: 10.1126/sciadv.adw9953

Welcome to “necroprinting”—3D printer nozzle made from mosquito’s proboscis Read More »

A fentanyl vaccine is about to get its first major test

fentanyl, health, opioids, Science, syndication, vaccines / Mike M. / December 4, 2025

Vaccine trial in the Netherlands hopes to protect against fentanyl-related overdose and death.

Just a tiny amount of fentanyl, the equivalent of a few grains of sand, is enough to stop a person’s breathing. The synthetic opioid is tasteless, odorless, and invisible when mixed with other substances, and drug users are often unaware of its presence.

It’s why biotech entrepreneur Collin Gage is aiming to protect people against the drug’s lethal effects. In 2023, he became the cofounder and CEO of ARMR Sciences to develop a vaccine against fentanyl. Now, the company is launching a trial to test its vaccine in people for the first time. The goal: prevent deaths from overdose.

“It became very apparent to me that as I assessed the treatment landscape, everything that exists is reactionary,” Gage says. “I thought, why are we not preventing this?”

Fifty times more potent than heroin and 100 times more potent than morphine, fentanyl was first approved by the Food and Drug Administration in 1968 as an intravenous pain reliever and anesthetic. Its potential for abuse was recognized even then, and clinicians could get it only in combination with the sedative droperidol in a ratio of 50:1 droperidol to fentanyl.

Cheap to make and incredibly addictive, fentanyl is now found in street drugs and counterfeit pills, because it boosts their potency and cuts costs. The drug is the biggest driver of overdose deaths in the United States and the leading cause of death for Americans aged 18 to 45.

Naloxone, known by the brand name Narcan, can rapidly reverse overdoses caused by fentanyl and other opioids. Widespread distribution of the medication contributed to a 24 percent decline in US drug overdose deaths in 2024. It works by attaching to opioid receptors throughout the body and displacing the opioid molecules that are attached there.

But a vaccine like the one ARMR Sciences is developing would be given before a person even encounters the drug. Gage likens it to a bulletproof vest or a suit of armor—hence the company’s name. (It was previously registered as Ovax but switched names in January.) “This is something that could completely change the paradigm of how we deal with overdose, because it doesn’t require someone to be carrying the treatment on them,” Gage says.

Opioid vaccines were initially proposed in the 1970s, but after early attempts at heroin vaccines failed, much of the research was abandoned. The modern opioid epidemic has led to a resurgence of interest, with backing from the US government.

ARMR’s experimental vaccine is designed to neutralize fentanyl in the bloodstream before it reaches the brain. Keeping fentanyl out of the brain would prevent the respiratory failure that comes with overdose, which causes death, as well as the euphoric high people get while taking fentanyl.

The basic idea behind ARMR’s shot is the same as any other vaccine. It trains the body’s immune system to make antibodies that recognize a foreign invader. But since fentanyl is much smaller than the pathogens our current vaccines target, it doesn’t trigger a natural antibody response on its own. To stimulate antibody production, ARMR has paired a fentanyl-like molecule with a “carrier” protein—a deactivated diphtheria toxin that’s already used in several approved medical products.

If a vaccinated person encounters fentanyl, antibodies in the blood would then bind to the drug and prevent it from traveling to the brain. Normally, fentanyl molecules can pass through the blood-brain barrier with ease, in part because of their small size. But fentanyl molecules with antibodies attached would be too big to get through. The result? No high and no overdose. The antibody-bound fentanyl molecules would eventually be passed in the urine.

The vaccine is based on work from the University of Houston, with collaborators at Tulane University designing an adjuvant derived from E.coli bacteria to boost the immune response to the vaccine. In rats, the shot blocked 92 to 98 percent of fentanyl from entering the brain and prevented the behavioral effects of the drug. The effects lasted for at least 20 weeks in the rats, which Gage thinks could translate to a year of protection in people.

“The big breakthrough in the past five or six years is the advancement of the adjuvant technology that we’re able to utilize now, which causes an extremely robust immune system response,” he says.

ARMR’s Phase 1/2 trial, which is slated to begin in early 2026, will enroll around 40 healthy adults at the Centre for Human Drug Research in the Netherlands. The first part of the trial will evaluate the vaccine’s safety and determine the best dosage. Volunteers will receive a series of two shots in varying doses, and researchers will measure their blood antibody levels. In the second part of the trial, a small group of participants will receive a medical dose of fentanyl so that investigators can study how well the vaccine blocks its effects. Gage says ARMR chose the Dutch site because of its experience conducting studies on naloxone and nalmefene, another medication that reverses opioid overdose.

The company is testing an injectable vaccine in this study but is also looking into an oral formulation, akin to a Listerine strip, for future trials.

Marco Pravetoni, founder and chief scientific officer of CounterX Therapeutics, has been studying opioid vaccines in his lab at the University of Washington but thinks a shorter-acting monoclonal antibody therapy is more commercially viable right now given the Trump administration’s hostility toward vaccines. The injectable antibody his company is developing is meant to provide monthlong protection against overdose. He says the product is intended for high-risk patients, such as those who are in addiction recovery programs. The Seattle-based company is poised to begin an initial human trial in early 2026.

“We think a month of protection is pretty good in terms of providing a safety net,” Pravetoni says. It would be comparable to Vivitrol, a prescription injectable on the market that’s used to prevent relapse in adults with alcohol or opioid dependence, which lasts for about a month.

One big question facing the development of a fentanyl vaccine or antibody treatment is whether a large enough dose of the drug could skirt by antibodies, making its way to the brain. Sharon Levy, an addiction medicine specialist at Boston Children’s Hospital who has worked on fentanyl vaccines and is one of ARMR’s scientific advisers, says it’s possible. “There’s only going to be so many antibodies,” she says.

In addiction treatment, Levy says there’s always a risk of patients trying to override the effects of a prescribed opioid-blocking medication by taking a high dose of an opioid—which is highly dangerous—but she says this is rare.

Levy and her colleagues have been conducting surveys on the acceptability of a fentanyl vaccine. She thinks a major target group would be teenagers and young adults who may be accidentally exposed to fentanyl when taking street drugs. Individuals with an opioid use disorder who are in active treatment would also be good candidates for vaccination.

“Overall, our experience has been that people would be interested in this,” she says.

Mike Selick, director of capacity building and community mobilization for the National Harm Reduction Coalition, worries that a fentanyl vaccine could block the effects of other opioids, leaving vaccinated individuals with few options for pain medications if they ever needed them.

In animal studies, the University of Houston team found no cross-reactivity with other common opioid-based common pain and addiction treatment medications, such as buprenorphine, methadone, morphine, or oxycodone. But there’s a downside to a lack of cross-reactivity. It means that people could still overdose on other types of opioids—and get high from them.

Gage knows that a fentanyl vaccine isn’t a perfect solution. Even if it works, it won’t end the opioid epidemic or cure opioid addiction. It won’t stop people from seeking out drugs entirely. But it could be another tool for helping to prevent overdose deaths.

“What we’re trying to do is put some innovation and new newfound technology behind this problem,” he says “because I think we’re in desperate need of it.”

This story originally appeared on WIRED.com.

Wired.com is your essential daily guide to what’s next, delivering the most original and complete take you’ll find anywhere on innovation’s impact on technology, science, business and culture.

A fentanyl vaccine is about to get its first major test Read More »

12 former FDA chiefs unite to say agency memo on vaccines is deeply stupid

CDC, clinical trials, COVID-19, fda, health, Science, vaccines, vinay prasad / Mike M. / December 4, 2025

On Friday, Vinay Prasad—the Food and Drug Administration’s chief medical and scientific officer and its top vaccine regulator—emailed a stunning memo to staff that quickly leaked to the press. Without evidence, Prasad claimed COVID-19 vaccines have killed 10 children in the US, and, as such, he announced unilateral, sweeping changes to the way the agency regulates and approves vaccines, including seasonal flu shots.

On Wednesday evening, a dozen former FDA commissioners, who collectively oversaw the agency for more than 35 years, responded to the memo with a scathing rebuke. Uniting to publish their response in the New England Journal of Medicine, the former commissioners said they were “deeply concerned” by Prasad’s memo, which they framed as a “threat” to the FDA’s work and a danger to Americans’ health.

In his memo, Prasad called for abandoning the FDA’s current framework for updating seasonal flu shots and other vaccines, such as those for COVID-19. Those updates currently involve studies that measure well-characterized immune responses (called immunobridging studies). Prasad dismissed this approach as insufficient and, instead, plans to require expensive randomized trials, which can take months to years for each vaccine update.

FDA staff who disagree with the plans can “submit your resignation letters,” Prasad wrote. And airing concerns or criticisms is seen as “unethical” and “illegal.”

Together, the former commissioners called Prasad’s memo the “latest in a series of troubling changes at the FDA,” and the planned policy updates “not … coherent.” Prasad’s arguments against immunobridging, they add, “misrepresent both the science and the regulatory record, especially in the case of vaccines that target well-understood pathogens through an established mechanism of action.”

12 former FDA chiefs unite to say agency memo on vaccines is deeply stupid Read More »

Rare win for renewable energy: Trump admin funds geothermal network expansion

geothermal, geothermal energy, geothermal power, Science, syndication / DJ Henderson / December 3, 2025

Progress on the project is a further indicator that, despite opposition to wind and solar, the Trump administration and Republicans in Congress appear to back geothermal energy.

President Donald Trump issued an executive order on his first day in office declaring an energy emergency that expressed support for a limited mix of energy resources, including fossil fuels, nuclear power, biofuels, hydropower, and geothermal energy.

The One Big Beautiful Bill Act, passed by Republicans and signed by Trump in July, quickly phases out tax credits for wind, solar, and electric vehicles. However, the bill left geothermal heating and cooling tax credits approved under the Inflation Reduction Act of 2022 largely intact.

A reorganization of the US Department of Energy announced last month eliminated the Office of Energy Efficiency and Renewable Energy but kept the office for geothermal energy as part of the newly created Hydrocarbons and Geothermal Energy Office.

“The fact that geothermal is on this administration’s agenda is pretty impactful,” said Nikki Bruno, vice president for thermal solutions and operational services at Eversource Energy. “It means they believe in it. It’s a bipartisan technology.”

Plans for the expansion project call for roughly doubling Framingham’s geothermal network capacity at approximately half the cost of the initial buildout. Part of the estimated cost savings will come from using existing equipment rather than duplicating it.

“You’ve already got all the pumping and control infrastructure installed, so you don’t need to build a new pump house,” said Eric Bosworth, a geothermal expert who runs the consultancy Thermal Energy Insights. Bosworth oversaw the construction of the initial geothermal network in Framingham while working for Eversource.

The network’s efficiency is anticipated to increase as it grows, requiring fewer boreholes to expand. That improvement is due to the different heating and cooling needs of individual buildings, which increasingly balance each other out as the network grows, Magavi said.

The project still awaits approval from state regulators, with Eversource aiming to start construction by the end of 2026, Bruno said.

“What we’re witnessing is the birth of a new utility,” Magavi said. Geothermal networks “can help us address energy security, affordability, and so many other challenges.”

This article originally appeared on Inside Climate News, a nonprofit, non-partisan news organization that covers climate, energy, and the environment. Sign up for their newsletter here.

Rare win for renewable energy: Trump admin funds geothermal network expansion Read More »

A spectacular explosion shows China is close to obtaining reusable rockets

china, Commercial space, jiuquan, LandSpace, launch, reusable rocket, Science, Space, zhuque-3 / Mike M. / December 3, 2025

“China’s first rocket recovery attempt achieved its expected technical objectives.”

Nine TQ-12A engines, burning methane and liquid oxygen, power the first Zhuque-3 rocket off the launch pad. Credit: LandSpace

China’s first attempt to land an orbital-class rocket may have ended in a fiery crash, but the company responsible for the mission had a lot to celebrate with the first flight of its new methane-fueled launcher.

LandSpace, a decade-old company based in Beijing, launched its new Zhuque-3 rocket for the first time at 11 pm EST Tuesday (04:0 UTC Wednesday), or noon local time at the Jiuquan launch site in northwestern China.

Powered by nine methane-fueled engines, the Zhuque-3 (Vermillion Bird-3) rocket climbed away from its launch pad with more than 1.7 million pounds of thrust. The 216-foot-tall (66-meter) launcher headed southeast, soaring through clear skies before releasing its first stage booster about two minutes into the flight.

The rocket’s upper stage fired a single engine to continue accelerating into orbit. LandSpace confirmed the upper stage “achieved the target orbit” and declared success for the rocket’s “orbital launch mission.” This alone is a remarkable accomplishment for a brand new rocket.

Learning on the fly

But LandSpace had other goals for this launch. The Zhuque-3, or ZQ-3, booster stage is architected for recovery and reuse, the first rocket in China with such a design. Made of stainless steel, the first stage arced to the edge of space before gravity pulled it back into the atmosphere. After making it through reentry, the booster was supposed to relight a subset of its engines for a final braking burn before a vertical landing at a prepared location about 240 miles (390 kilometers) downrange from the launch pad.

But something went wrong as the booster approached the landing zone.

“According to telemetry data, an anomaly occurred after the first stage initiated its landing burn, preventing a soft landing on the designated recovery pad,” LandSpace wrote on X. “The stage debris came down near the edge of the recovery pad, and the recovery test was unsuccessful. The specific cause is under further investigation.”

Videos shared on Weibo, a Chinese social media platform, showed the final moments of the booster’s supersonic descent. A fireball enveloped the rocket at the start of the landing burn, and it impacted the recovery pad at high speed. But the rocket appeared to survive the most extreme aerodynamic forces of reentry, and it nearly hit a bullseye at the landing pad, situated in a remote dune field in the Gobi Desert.

“During the first stage recovery system verification test, engines thrust throttling operated normally, attitude control remained stable, and the downrange recovery trajectory was nominal,” LandSpace said, adding that no one was harmed in the accident.

LandSpace’s 216-foot-tall (66-meter) Zhuque-3 rocket lifts off from the Jiuquan Satellite Launch Center in northwestern China. Credit: LandSpace

The crash landing may have been disappointing to LandSpace, but it’s actually an auspicious result for a first attempt. The rocket appears to have made it closer to landing than Blue Origin’s first New Glenn booster earlier this year. Blue Origin made a successful landing on its second attempt last month.

It took SpaceX numerous tries before it landed the first Falcon 9 booster 10 years ago this month, pioneering novel guidance algorithms, supersonic retro-propulsion, and experimentation in how to manage the substantial aero-thermal forces of reentry. For example, SpaceX discovered through flight testing that it needed to add grid fins to the Falcon 9 booster. LandSpace’s booster uses grid fins from the start.

Poised for a breakout

China needs reusable rockets to keep up with the US launch industry, which is dominated by SpaceX, a company that flies more often and hauls heavier cargo to orbit than all Chinese rockets combined. There are at least two Chinese megaconstellations now being deployed in low-Earth orbit, each with architectures requiring thousands of satellites to relay data and Internet signals around the world. Without scaling up satellite production and reusing rockets, China will have difficulty matching the capacities of SpaceX, Blue Origin, and other emerging US launch companies.

Just three months ago, US military officials identified China’s advancements in reusable rocketry as a key to unlocking the country’s ability to potentially threaten US assets in space. “I’m concerned about when the Chinese figure out how to do reusable lift that allows them to put more capability on orbit at a quicker cadence than currently exists,” said Brig. Gen. Brian Sidari, the Space Force’s deputy chief of space operations for intelligence, at a conference in September.

Without reusable rockets, China has turned to a wide variety of expendable boosters this year to launch less than half as often as the United States. China has made 78 orbital launch attempts so far this year, but no single rocket type has flown more than 13 times. In contrast, SpaceX’s Falcon 9 is responsible for 153 of 182 launches by US rockets.

LandSpace’s first landing attempt shows China is positioned to close the gap. The company’s engineers will be smarter about landing rockets on the next try.

What’s more, several more Chinese companies are close to debuting their own reusable rockets. The next of these new rockets, the Long March 12A, is awaiting its first liftoff later this month from another launch pad at the Jiuquan spaceport.

The Long March 12A comes from one of China’s established rocket developers, the Shanghai Academy of Spaceflight Technology (SAST), part of the country’s state-owned aerospace enterprise. The Long March 12A has comparable performance to LandSpace’s Zhuque-3 and will also target a landing of its booster stage downrange on its first flight.

A handful of other rocket developers also claim to be weeks or months away from launching their first reusable boosters. One of them, Space Pioneer, might have been first to flight with its new Tianlong-3 rocket if not for the thorny problem of an accidental launch during a booster test-firing last year. Space Pioneer eventually completed a successful static fire in September of this year, and the company recently released a photo showing its rocket on the launch pad.

The Zhuque-3 rocket begins its first flight. Credit: LandSpace

These new rockets can each lift medium-class payloads into orbit. In its first iteration, the Zhuque-3 rocket is capable of placing a payload of more than 17,600 pounds (8 metric tons) into low-Earth orbit after accounting for the fuel reserves required for booster recovery. This makes Zhuque-3 the largest and most powerful commercial rocket ever launched from China.

LandSpace eventually plans to debut an upgraded Zhuque-3 carrying more propellant and using more powerful engines, raising its payload capacity to more than 40,000 pounds (18.3 metric tons) in reusable mode or a few tons more with an expendable booster.

LandSpace has raised more than $400 million since its founding in 2015, primarily from venture capital firms and government-backed investment funds. LandSpace initially developed its own liquid-fueled engines and a light-class launcher named Zhuque-2, which became the world’s first methane-burning launcher to reach orbit in 2023. LandSpace’s Zhuque-2 has logged four successful missions in six tries.

The larger Zhuque-3 is a “new-generation, low-cost, high-capacity, high-frequency, reusable LOX/methane launch vehicle,” LandSpace says. The company plans to reuse its Zhuque-3 boosters at least 20 times, “enabling efficient multi-satellite deployment for Internet constellations and China’s future space programs.”

A spectacular explosion shows China is close to obtaining reusable rockets Read More »

3D model shows small clans created Easter Island statues

Archaeology, easter island, Science / DJ Henderson / December 2, 2025

Credit: ArcGIS

Easter Island is famous for its giant monumental statues, called moai, built some 800 years ago. The volcanic rock used for the moai came from a quarry site called Rano Raraku. Archaeologists have created a high-resolution interactive 3D model of the quarry site to learn more about the processes used to create the moai. (You can explore the full interactive model here.) According to a paper published in the journal PLoS ONE, the model shows that there were numerous independent groups, probably family clans, that created the moai, rather than a centralized management system.

“You can see things that you couldn’t actually see on the ground. You can see tops and sides and all kinds of areas that just would never be able to walk to,” said co-author Carl Lipo of Binghamton University. “We can say, ‘Here, go look at it.’ If you want to see the different kinds of carving, fly around and see stuff there. We’re documenting something that really has needed to be documented, but in a way that’s really comprehensive and shareable.”

Lipo is one of the foremost experts on the Easter Island moai. In October, we reported on Lipo’s experimental confirmation—based on 3D modeling of the physics and new field tests to re-create that motion—that Easter Island’s people transported the statues in a vertical position, with workers using ropes to essentially “walk” the moai onto their platforms. To explain the presence of so many moai, the assumption has been that the island was once home to tens of thousands of people.

Lipo’s latest field trials showed that the “walking” method can be accomplished with far fewer workers: 18 people, four on each lateral rope and 10 on a rear rope, to achieve the side-to-side walking motion. They were efficient enough in coordinating their efforts to move the statue forward 100 meters in just 40 minutes. That’s because the method operates on basic pendulum dynamics, which minimizes friction between the base and the ground. It’s also a technique that exploits the gradual build-up of amplitude, suggesting a sophisticated understanding of resonance principles.

3D model shows small clans created Easter Island statues Read More »

“Renewable” no more: Trump admin renames the National Renewable Energy Laboratory

Donald Trump, national renewable energy laboratory, NREL, Science, syndication / DJ Henderson / December 2, 2025

The Trump administration has renamed the National Renewable Energy Laboratory, now calling it the National Laboratory of the Rockies, marking an identity shift for the Colorado institution that has been a global leader in wind, solar and other renewable energy research.

“The new name reflects the Trump administration’s broader vision for the lab’s applied energy research, which historically emphasized alternative and renewable sources of generation, and honors the natural splendor of the lab’s surroundings in Golden, Colorado,” said Jud Virden, laboratory director, in a statement.

He did not specify what this “broader vision” would mean for the lab’s programs or its staff of about 4,000.

The renaming is the latest in a series of actions by the Trump administration to deemphasize or cut the parts of the federal government that support renewable energy, while also expanding federal support for fossil fuels.

Asked for details, the Department of Energy said in an email that the renaming “reflects the Department’s renewed focus on ‘energy addition’ rather than the prioritization of specific energy resources.”

A lab spokesman had no additional information about whether there will be changes to programs or headcount at the lab.

Bill Ritter, a Democrat who was governor of Colorado from 2007 to 2011, said it’s reasonable to assume that the name change signals that the federal government is abandoning the lab’s status as a world leader in energy research.

“It’s an iconic research facility,” he said.

Underscoring this point, he recalled a trip to Israel while he was governor.

“The head of their renewable energy laboratory said, ‘I have nothing to tell you because you come from the place that has the best renewable energy laboratory in the world,’” Ritter said.

“Renewable” no more: Trump admin renames the National Renewable Energy Laboratory Read More »

The missile meant to strike fear in Russia’s enemies fails once again

ICBM, launch, russia, Sarmat, Science, Space, vladimir putin / Kelly Newman / December 2, 2025

Therefore, it’s no wonder Russian officials like to talk up Sarmat’s capabilities. Russian President Vladimir Putin has called Sarmat a “truly unique weapon” that will “provide food for thought for those who, in the heat of frenzied aggressive rhetoric, try to threaten our country.” Dmitry Rogozin, then the head of Russia’s space agency, called the Sarmat missile a “superweapon” after its first test flight in 2022.

So far, what’s unique about the Sarmat missile is its propensity for failure. The missile’s first full-scale test flight in 2022 apparently went well, but the program has suffered a string of consecutive failures since then, most notably a catastrophic explosion last year that destroyed the Sarmat missile’s underground silo in northern Russia.

The Sarmat is supposed to replace Russia’s aging R-36M2 strategic ICBM fleet, which was built in Ukraine. The RS-28, sometimes called the Satan II, is a “product solely of Russian industry cooperation,” according to Russia’s Ministry of Defense.

The video of the missile failure last week lacks the resolution to confirm whether it was a Sarmat missile or the older-model R-36M2, but analysts agree it was most likely a Sarmat. The missile silo used for Friday’s test was recently renovated, perhaps to convert it to support Sarmat tests after the destruction of the new missile’s northern launch site last year.

“Work there began in Spring 2025, after the ice thawed,” wrote Etienne Marcuz, an analyst on strategic armaments at the Foundation for Strategic Research, a French think tank. The “urgent renovation” of the missile silo at Dombarovsky lends support for the hypothesis that last week’s accident involved the Sarmat, and not the R-36M2, which was last tested more than 10 years ago, Marcuz wrote on X.

“If this is indeed another Sarmat failure, it would be highly detrimental to the medium-term future of Russian deterrence,” Marcuz continued. “The aging R-36M2 missiles, which carry a significant portion of Russia’s strategic warheads, are seeing their replacement pushed even further into the future, while their maintenance—previously handled by Ukraine until 2014—remains highly uncertain.”

In this pool photograph distributed by the Russian state media agency Sputnik, Russia’s President Vladimir Putin chairs a Security Council meeting at the Kremlin in Moscow on November 5, 2025. Credit: Gavriil Grigorov/Pool/AFP via Getty Images

Podvig, the UN researcher who also runs the Russian Nuclear Forces blog site, agrees with Marcuz’s conclusions. With the R-36M2 missile soon to retire, “it is extremely unlikely that the Rocket Forces would want to test launch them,” Podvig wrote on his website. “This leaves Sarmat.”

The failure adds fresh uncertainty to the readiness of Russia’s nuclear arsenal. If this were actually a test of one of Russia’s older ICBMs, the result would raise questions about hardware decay and obsolescence. In the more likely case of a Sarmat test flight, it would be the latest in a series of problems that have delayed its entry into service since 2018.

The missile meant to strike fear in Russia’s enemies fails once again Read More »