Iran’s Capital Is Moving. The Reason Is an Ecological Catastrophe

By Humberto Basilio edited by Claire Cameron

Tehran can no longer remain the capital of Iran amid a deepening ecological crisis and acute water shortage.

The situation in Tehran is the result of “a perfect storm of climate change and corruption,” says Michael Rubin, a political analyst at the American Enterprise Institute.

“We no longer have a choice,” Iranian President Masoud Pezeshkian reportedly told officials on Friday.

---

On supporting science journalism

If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

---

Instead, Iranian officials are considering moving the capital to the country’s southern coast. But experts say the proposal does not change the reality for the nearly ten million people who live in Tehran, who are now suffering the consequences of a decades-long decline in water supply.

Since at least 2008, scientists have warned that unchecked groundwater pumping for the city and for agriculture was rapidly draining its aquifers. The overuse did not just deplete underground reserves—it destroyed them, as the land compressed and sank irreversibly. One recent study found that Iran’s central plateau, where most of the country’s aquifers are located, is sinking by more than 35 centimeters each year. As a result, the aquifers lose about 1.7 billion cubic meters of water annually as the ground is permanently crushed, leaving no space for underground water storage to recover, says Darío Solano, a geoscientist at the National Autonomous University of Mexico.

“We saw this coming,” says Solano.

Other major cities like Cape Town, Mexico City, Jakarta and parts of California are also facing day zero scenarios as they sink and run out of water.

This is not the first time Iran’s capital has moved. Over the centuries, it has shifted many times, from Isfahan to Tabriz to Shiraz. Some of these former capitals still thrive while others exist only as ruins, says Rubin. But this marks the first time the Iranian government has moved the capital because of an ecological catastrophe.

Yet, Rubin says, “it would be a mistake to look at this only through the lens of climate change.” Water, land and wastewater mismanagement and corruption have made the crisis worse, he says. If the capital moves to the remote Makran coast in the south, it could cost more than $100 billion dollars. The region is known for its harsh climate and difficult terrain, and some experts have doubts about its viability as a national center. Relocating a capital is often driven more by politics than by environmental concerns, says Linda Shi, a social scientist and urban planner at Cornell University. “Climate change is not the thing that is causing it, but it is a convenient factor to blame in order to avoid taking responsibility” for poor political decisions, she says.

A new generation of stratospheric balloons and high-altitude uncrewed aerial vehicles (UAVs) could soon connect the world’s unconnected with high-speed internet at a fraction of the prices commanded by operators of satellite megaconstellations such as SpaceX’s Starlink.

High-altitude platform stations, or HAPS, have been around for a while, but the technology hasn’t fully taken off yet. Google spent 10 years trying to develop balloons that would hover in the stratosphere above remote rural areas and beam internet to residents but abandoned that project, called Loon, in 2021, concluding that it couldn’t be made sustainable.

Richard Deakin, the CEO of World Mobile Stratospheric, said HAPS have failed to make it so far because they couldn’t support power-hungry antennas needed to beam down high-bandwidth internet across vast swaths of land. Previously-tested high-altitude balloons and airships have relied on photovoltaics to generate power, which only provide "a couple of hundred watts," according to Deakin.

He said that his company’s HAPS, an autonomous plane called the Stratomast, will be powered by liquid hydrogen, allowing it not only to hover for six days at an altitude of 60,000 feet (18 km), but also generate enough electricity to support a 10-by-10-foot (3 by 3 meters) phased-array antenna that could connect 500,000 users on Earth at the same time. After six days, a new aircraft would arrive to take over the service while the first returns to the base for refuelling.

Deakin says users will get 200 megabits per second (Mbps) of connectivity directly into their smartphones from Stratomast. That would be a vast improvement over Starlink‘s current direct-to-device offering of 17 Mbps, which is currently only capable of supporting emergency text messaging. Even AST SpaceMobile, which is building a constellation of giant orbiting antennas to beam internet directly to smartphone devices, can sustain only about 21 Mbps.

"When the Stratomast is flying, all these old satellites are going to be in museums," Deakin said.

The 4-metric-ton (4.4 tons) Stratomast aircraft, made of lightweight carbon fiber, has a 184-foot (56 m) wingspan, equivalent to that of a 120-metric-ton (132 tons) Boeing 787 Dreamliner plane. A single Stratomast will cover an area of 6,000 square miles (15,000 square km). Such a wide reach means that the whole of Scotland could be covered with just nine Stratomast platforms, Deakin said. World Mobile Stratospheric estimates the cost of running such a system would be around just £40 million ($52 million US) per year, allowing the company to provide 200 Mbps of internet connectivity to Scotland’s 5.5 million inhabitants at a cost at about 60 pence per person per month.

"That’s enough for TV, computer broadband, the whole thing…,"Gregory Gottlieb, the head of aerial platforms at World Mobile, told Space.com.

In comparison, the cheapest Starlink subscription, only covering areas with low demand, currently costs $40 a month. And price is only one of the drawbacks of LEO satellite internet. To be able to connect to Starlink satellites, users need dedicated terminals. Although Starlink’s downlink speeds reach up to 250 Mbps, the bandwidth gets diluted as the number of users grows. For example, troops on the frontline in eastern Ukraine complain that Starlink bandwidth limits the use of ground robots, as most terminals there only get about 10 Mbps.

"There really isn’t any satellite constellation that can serve more than one person per square kilometer [0.4 square miles]," Mikkel Frandsen, the founder and CEO of another HAPS developer, New Mexico-based Sceye, told Space.com. "That’s kind of the upper end."

Sceye, founded in 2014, has developed an airship-like HAPS powered by solar energy, which has already completed several successful test flights. In August last year, the Sceye airship became the first stratospheric platform that successfully survived a night in the stratosphere without sinking after sunset and remained in a required position above a fixed spot on Earth. The problem of drift and difficulties with station-keeping were among the issues that led to the demise of the Google Loon project, according to Frandsen.

In June, Sceye received a "strategic investment" from Japanese telecommunications operator SoftBank, which hopes the technology will allow it to provide next-generation connectivity to users even in the most underserved areas. Sceye also recently won a contract from NASA to host Earth-observation payloads.

Frandsen said that Sceye doesn’t want to compete with satellite internet providers but thinks that megaconstellations, even when fully deployed, will not be able to satisfy the world’s need for connectivity.

"All satellite constellations, when they’re combined, will not do anything other than [make] a little dent in the global demand for connectivity," he said. "They’re going to do fine business at the prices they’re charging, but they’re not going to serve billions of people. Space isn’t all that scalable. They’re going to serve millions of people."

LEO satellite megaconstellations such as SpaceX‘s Starlink orbit a few hundred kilometers above Earth’s surface. In the past five years, they have replaced distant geostationary satellites, which orbit 22,236 miles (35,786 km) above Earth, as the dominant technology for delivering internet connectivity from space. But the growth in the number of satellites concerns space sustainability experts. The more objects hurtling around the planet, the higher the risk of collisions that could pollute near-Earth space with thousands of dangerous fragments. Moreover, atmospheric physicists worry about the growing amount of metal being burned up in the atmosphere during satellite reentries.

"HAPS are a really interesting domain because I think, in many ways, they cover the best of terrestrial and the best of satellite — high-altitude systems without suffering from the drawbacks," said Deakin.

Gottlieb said that HAPS could provide a convenient, flexible and easily replaceable alternative to satellite internet at times of growing geopolitical tensions.

"There is a view that, within 24 hours of any major conflict, low Earth orbit would be unusable for military purposes," Gottlieb said. "We can deploy aircraft at very short notice. We can be agile in terms of spectrum that we’re using, and all sorts of different boxes that can go onto our platforms."

Deakin’s team has been developing Stratomast since 2019 as part of a collaboration with German telecommunications provider Deutsche Telecom. During tests in Germany and Saudi Arabia, they demonstrated the workings of their novel antenna technology.

Earlier this year, the company was acquired by the U.S.-based telecom provider World Mobile. The company recently partnered with Indonesian telecommunications provider Protelindo to get Stratomast off the ground. The partnership plans to conduct flight tests with their antenna at a lower altitude next summer and hopes to begin stratospheric test flights in 2027.

Sceye, in the meantime, is working on increasing the endurance of its airship and hopes to commence commercial service in 2027.

Are you curious about the recently released emails of Jeffrey Epstein but don’t feel like clicking through thousands of cumbersome files? There’s now a way to search Epstein’s inbox in a much easier way. And it’s both interesting and creepy at the same time.

It’s called Jmail.world and it’s a project by Riley Walz and Luke Igel. They’ve built a tool that might best be described as Jeffrey Epstein Inbox Simulator ’25. At least that’s how The Atlantic’s Charlie Warzel puts it. And it feels like an apt description.

Basically it looks like Gmail and it has many of the same features. And while it’s a clever way to present the tens of thousands of documents released by the House Committee on Oversight and Government Reform, it’s a very odd feeling to essentially be placed in the late sex trafficker’s shoes.

“The problem was that the PDFs that the above pointed to were hard to read, so we felt putting it in a familiar form with shareable links and easy to screenshot views would be helpful for people. We also wanted to focus on just the emails in that drop,” Igel told Gizmodo over email.

Idel is the co-founder of Kino, an AI video assistant that allows users to index massive amounts of video. Idel says he previously worked at SpaceX, NASA, and was an undergrad at MIT. And he built this tool with Riley Walz, who has done a number of clever projects in the past.

“Riley and I knocked this out in a few hours Wednesday night using Cursor,” Igel told Gizmodo. “This is our first real coding project together! Riley has done so much amazing stuff like this in the past as you know.”

Igel tells Gizmodo it’s “pretty much all” of the emails from the latest batch released by Congress. So while it’s an interesting project, don’t necessarily assume it’s comprehensive. But it’s still an interesting way to browse the inbox of a truly heinous man.

These emails aren’t the “Epstein files” that everyone has been clamoring for over the past few years. Those are files still held by the Department of Justice, though Congress has passed a law requiring their release. The question is whether they’ll ever see the light of day. President Trump ordered a new investigation into Epstein’s ties to powerful Democrats so many people worry that DOJ will say they can’t release any files because there’s an ongoing investigation. But time will tell.

We have a whole new generation growing up in the dawn of artificial intelligence. The early signs of its impact are alarming.

A British youth charity called OnSide surveyed 5,035 young people between the ages of 11 to 18 for the “Generation Isolation Report,” its annual study on how the youth spend their free time. The results paint a rather bleak picture.

The survey found that two-in-five teens turn to AI for advice, company or support, with 20% of those that do saying that talking to AI is easier than talking to a real person.

“AI support is instant, but no substitute for the trust, empathy and understanding of a human conversation,” OnSide chief executive Jamie Masraff said in the report.

Over half of the young respondents said that they turned to AI specifically for advice on things like clothes, friendships, mental health or to have AI help them through emotions like sadness and stress. One-in-ten said that they were choosing AI because they just wanted someone to talk to.

The study and its findings show a generation that is lonely and one that has unrestricted access to technology that is addictive in nature. According to the study 76% of young people spend most of their free time on screens, and that 34% report feeling high or very high feelings of loneliness.

AI, which is still in its under-regulated Wild West era, is one such technology, and it’s no surprise that lonely young people turn to it for quick companionship and advice.

“It’s clear that the interlinked issues of loneliness, digital dependence and isolation have become entrenched in young people’s lives, raising deeper questions about what it’s like to grow up this way,” Masraff said.

As AI burrows itself deeper into the everyday lives of teens, alarm bells are sounding. AI chatbots have turned out to be dangerously addictive for some adults, whose brains have reached full-functioning capacity. Now imagine how much worse it could get for kids whose pre-frontal cortices are far from completion.

The American Psychological Association has been pushing the FTC to address the use of AI chatbots as unlicensed therapists. The Association wrote in a blog post from March that chatbots used for mental health advice could endanger users, especially “vulnerable groups [that] include children and teens, who lack the experience to accurately assess risks.”

In some instances, the results have allegedly been fatal. Two separate families have filed complaints with artificial intelligence companies Character.AI and OpenAI, claiming that the companies’ chatbots had influenced and aided their sons’ suicide. In one case, OpenAI’s ChatGPT helped a 16-year-old with planning his suicide and even discouraged him from letting his parents know of his suicidal ideation.

Several AI chatbots are also being investigated over sexualized conversations with children. Meta was lambasted earlier this year after a leaked internal document showed that the tech giant had okayed its AI tools to engage in “sensual” chats with children.

Last month, Congress introduced a bipartisan bill called the GUARD Act, with the aim of forcing AI companies to institute age verification on their sites and block users under 18 years of age.

“AI chatbots pose a serious threat to our kids,” Sen. Josh Hawley, who introduced the bill along with Sen. Richard Blumenthal, told NBC News. “More than seventy percent of American children are now using these AI products.”

But even if that bill becomes law, it’s uncertain how effective it will be at keeping children away from AI chatbots. Age verifications and limits used by social media platforms haven’t been the most effective tools at guarding children from the adverse effects of the internet.

Two of the most anticipated studies this year—large-scale, randomized, controlled, and double-blinded trials of the popular obesity and diabetes drug semaglutide for Alzheimer’s disease—have unfortunately ended in a dud.

Novo Nordisk, the makers of semaglutide, announced the disappointing trial results early Monday morning. The drug did not meaningfully slow down the disease’s progression compared to placebo, the trials showed. The company will now shut down the extension of these trials, though some experts and advocacy groups are still hopeful about the potential future of GLP-1 therapy for Alzheimer’s.

“While these results are not what we had hoped for, they will contribute to our understanding of this devastating and fatal disease,” said Joanne Pike, president and chief executive officer of the Alzheimer’s Association, in a statement sent to Gizmodo.

Early but unfulfilled promise

Semaglutide is the active ingredient in the diabetes drug Ozempic and the obesity medication Wegovy; it’s also currently available as an oral diabetes medication under the brand name Rybelsus. It mimics the naturally occurring GLP-1 hormone, which helps regulate our hunger and insulin production, among other things. Though not the first approved GLP-1 drug, semaglutide’s improved duration and potency have greatly changed the field of obesity medicine. The drug is significantly more effective at helping obese people lose weight than diet and exercise alone.

For years, various studies have suggested that GLP-1 medications like semaglutide may also be able to prevent or slow down the progression of dementia. On the basis of this promising research, Novo Nordisk commissioned the evoke and evoke+ phase 3 trials four years ago. These trials collectively involved 3,808 older adults (over age 55) with mild cognitive impairment or early Alzheimer’s disease. The participants were randomized to receive a weekly dose of oral semaglutide or a placebo.

As with past trials, semaglutide was generally safe and tolerable (common side effects tend to be gastrointestinal, such as nausea or vomiting). People on semaglutide did appear to have some noticeable improvements in biomarkers related to Alzheimer’s, according to Novo Nordisk. Ultimately, however, these improvements did not translate to real-world results. There was no significant difference in the progression of people’s dementia between the two groups, the company reported.

“Based on the significant unmet need in Alzheimer’s disease as well as a number of indicative data points, we felt we had a responsibility to explore semaglutide’s potential, despite a low likelihood of success. We are proud to have conducted two well-controlled phase 3 trials in Alzheimer’s disease that meet the highest standards of research and rigorous methodology,” said Martin Holst Lange, chief scientific officer and executive vice president of research and development at Novo Nordisk, in a statement.

The future of GLP-1s for dementia

Alzheimer’s disease is one of the most complex and devastating conditions we can develop. Even today, researchers still aren’t in agreement about the exact causes of Alzheimer’s, nor about the best approach to treating it. And this is far from the first time that a promising drug candidate for Alzheimer’s has failed to cross the finish line in late-stage clinical trials. So in many respects, these results aren’t too much of a surprise.

But there may yet be a silver lining to this flop. Given the biomarker findings, it’s still possible that a GLP-1 medication can slow the progression of Alzheimer’s, just not this particular one or in this particular way. There are newer GLP-1-based drugs, including many in development, that are generally more effective at treating obesity and diabetes than semaglutide. So perhaps that greater effectiveness can apply to Alzheimer’s as well. And there is precedent for this happening with other drug classes. After years of failure (and one controversial approval), there are now several anti-amyloid drugs available that can modestly slow Alzheimer’s progression.

Another consideration is timing. It’s possible that giving anti-Alzheimer’s medications to people at high risk of the disease—but years before they show any symptoms—can significantly slow its emergence, a hypothesis that trials of other drugs are currently testing.

Advocates like the Alzheimer’s Association are not fully abandoning their hopes that GLP-1 therapy for Alzheimer’s can help, at least not yet.

“Though this semaglutide pill did not help against Alzheimer’s, the field will continue to investigate this class of drugs, as they may act differently. And, the Alzheimer’s Association remains a fierce leader for this type of innovative research, and we believe it’s critical to continue investigating diverse approaches to treatment and prevention,” said Maria C. Carrillo, Alzheimer’s Association chief science officer and medical affairs lead, in a statement to Gizmodo.

Novo Nordisk’s announcement comes roughly a week before the company is expected to present the primary results of the two evoke trials at the Clinical Trials in Alzheimer’s Disease (CTAD) meeting in early December. These results may provide a better sense of where the drug came up short.

For now, though, these findings should also illustrate why clinical trial research is so important. Many lab or observational studies will tease the potential of an experimental or repurposed drug, only for later, more definitive trials to come up short. These preliminary studies are key to finding the drugs or interventions that can work, but we should always be mindful that they won’t pan out most of the time.

This is an especially relevant reminder with GLP-1 therapy, which has shown promise for treating conditions beyond obesity, such as alcohol addiction. Other studies have validated the benefits of GLP-1s for some health problems, such as heart disease. But this may not be the first and only flop we see with this drug class.