Science and Technology

Naivasha, Kenya – For communities living around Naivasha, water is not an abstract resource. It sustains families, livestock, farms and schools.

That reality has taken on new significance after plans for a major Microsoft-G42 data centre in Olkaria, near Lake Naivasha, stalled in May 2026 over concerns about available power capacity.

Microsoft and United Arab Emirates-based artificial intelligence company G42 announced the project in 2024 as part of a $1bn digital investment package for Kenya. The proposed facility was to run on geothermal energy and eventually scale to as much as 1 gigawatt of capacity.

The uncertainty has also prompted questions about what another major industrial user could mean for water in a region where residents already report shortages.

Microsoft and G42 said the proposed data centre campus would run entirely on renewable geothermal energy and incorporate water conservation technology. The companies did not disclose a project-specific water consumption figure in their 2024 announcement.

Kenya Electricity Generating Company (KenGen) communications director Frank David Ochieng told Al Jazeera that the data centre remains at the design stage and that he could not comment further until the project is ready to proceed.

For residents like Musa Olorkedienye, who spoke to Al Jazeera, water scarcity is already a daily concern. He says communities around Olkaria have seen changes in access to water, including the loss of reliable piped supplies that residents previously received from KenGen.

“Currently, we are relying on water vendors to get water, our animals are walking for kilometres, and we fear things could get worse as demand for water rises,” Olorkedienye says.

Pastoralist Isaac Leshishi, who also spoke to Al Jazeera, says increasingly harsh weather is adding to the pressure.

Why Olkaria?

The choice of Olkaria was closely tied to energy. The area is home to Kenya’s major geothermal operations, making it an attractive location for a power-intensive facility.

KenGen operates the Olkaria geothermal complex, while Microsoft and G42 planned to power the proposed data centre entirely with geothermal energy.

A lake under pressure

Naivasha is a freshwater lake in Kenya’s Rift Valley whose catchment supports agriculture, tourism, livestock and domestic water use. Its basin also hosts geothermal development and other economic activity.

Grace Kimani, a patrol leader with Lake Naivasha and Oloiden, told Al Jazeera that the reservoir is under growing pressure from population growth, agriculture, water abstraction, climate variability, pollution and ecosystem degradation.

“The planned Microsoft-G42 data centre in Olkaria could bring jobs and investment, but its water demand raises concerns about adding pressure to already competing needs, particularly during dry periods,” she said.

Kimani said there is limited public information about the project’s expected water demand, source and cooling technology. She said transparency and an assessment of its cumulative impact on water resources would be important.

She also called for water-efficient or water-free cooling, water recycling and the use of treated wastewater, as well as sustainable abstraction limits and community involvement in monitoring.

Silas Wanjala of the Lake Naivasha Riparian Association, who spoke to Al Jazeera, said the region is heavily dependent on groundwater and that declining water flows are adding to the pressure.

“These industries, especially EcoCloud, which deal with data, will consume a lot of water at a time when rivers are drying, and demand for water is on the rise,” Wanjala said.

Olkaria EcoCloud Data Centre is a local partner in the G42-led development. In 2024, the Kenya News Agency reported that G42, Microsoft and EcoCloud signed a letter of intent for the wider data-centre initiative, with EcoCloud described as a local partner that had previously signed a memorandum of understanding with G42.

Wanjala points to past fluctuations in Lake Naivasha as a warning.

“This lake in 2010 nearly dried up due to over-abstraction, and this could be repeated due to high demand for water by these investors in Olkaria,” he says.

His concern comes against a wider backdrop of water scarcity in Kenya. The Food and Agriculture Organization (FAO) of the United Nations says Kenya has about 527 cubic metres (527,000 litres) of freshwater available per person, below the 1,000-cubic-metre threshold for water scarcity, and estimates availability could fall to about 475 cubic metres per person by 2030.

The figures do not show what effect the proposed data centre would have on Lake Naivasha. They provide context, however, for why the prospect of another major water user is drawing scrutiny in a region where demand is already high.

How much water would it use?

The amount of water the proposed Microsoft-G42 facility itself would require remains unclear.

The project announcement provides no projected consumption figure.

Existing industrial use offers some context.

A KenGen environmental and social impact assessment records that 195,165 cubic metres of water were abstracted from Lake Naivasha in July 2023 for domestic and commercial uses at Olkaria and for operations and domestic use at Eburru.

Of that total, 153,918 cubic metres were used for commercial operations at Olkaria. The assessment records the abstraction as within Water Resources Authority (WRA) permitted levels.

Those figures relate to existing KenGen operations, not the proposed data centre.

For farmer Eskimos Kobia, who spoke to Al Jazeera, the potential competition extends beyond households and livestock. He says farmers, pastoralists, schools and investors will all face greater pressure as demand increases.

Kimani said climate variability has also led to fluctuations in lake levels, with periods of flooding followed by prolonged dry conditions.

“Water quality is affected by agricultural run-off, untreated wastewater in some areas and invasive species,” she says.

Investment versus local concerns

Not everyone in Naivasha opposes the investment.

Absolom Mukhuusi of the Naivasha Professional Association, who spoke to Al Jazeera, says the technology sector could bring jobs, infrastructure and new businesses to the area. But he says economic benefits should not come at the expense of local communities.

“Even as we welcome the investors, our biggest fear is what happens to our water bodies and communities as water is diverted to Olkaria for the heavy users,” he says.

Could technology help?

Geologist Kenyatta Otieno, who spoke to Al Jazeera, sees another potential benefit.

He recalls the pressure large flower farms once placed on the lake’s ecosystem, saying many have since left or scaled back their operations.

Otieno says the proposed centre would have included a resource centre to monitor lake levels and weather patterns. Such monitoring, he says, could help identify the highest water level over time and guide riparian land zoning.

“The centre being built with water conservation in mind would be futuristic as Naivasha is generally a water-scarce area. It would be a model for future development,” Otieno said.

Kenya already has regulators responsible for managing competing demands. The WRA regulates water abstraction and issues water-use permits, while the National Environment Management Authority (NEMA) oversees environmental impact assessments under the country’s environmental regulatory framework.

Attempts by Al Jazeera to obtain comments from WRA and NEMA officials were unsuccessful. Efforts to reach Microsoft-G42 officials and Kenyan government officials for comment on the project’s status and water requirements were also unsuccessful.

For now, the project’s eventual scale, design and water requirements remain unclear, according to KenGen.

“We are now competing with the multibillion [-dollar] companies for water, and we fear that we shall be the losers in the long run,” Leshishi said.

China is set to launch its Chang’e-7 unmanned, robotic space mission, possibly as early as Monday morning, to look for ice water in the permanently shadowed craters of the moon’s south pole.

This marks China’s seventh and most ambitious moon mission so far. Here is what we know about it.

What do we know about Chang’e-7?

The Chang’e‑7 launch window runs from Monday, August 24 to Monday, August 31, according to launch observers. That means the mission could lift off on any day in that period.

The Chang’e-7 comprises an orbiter, a lander, a rover and a hopper.

• The orbiter is the main spacecraft which will remain in the moon’s orbit, mapping the surface and taking images while the mission is under way. It will also relay data and communications between the other Chang’e‑7 components and Earth.
• The lander is the vessel that will touch down on the lunar surface, loaded with the necessary scientific instruments for lunar exploration, near the edge of the Shackleton Crater, a 21km-wide (13-mile-wide) pit close to the moon’s south pole. Lunar missions before this have never come this close to the pole.
• The rover is a small robotic vehicle which can drive around the landing site and which carries equipment to analyse the local environment, including rocks and soil.
• The hopper is a small robotic craft powered by solar energy, designed to “hop” – or fly – short distances from the surface of the moon and land again. It will be used to explore the Shackleton Crater.

How do we know there is water on the moon?

Over the past two decades, several space missions have established that there is water on the moon.

Since the 1960s, even before the first Apollo landing in 1969, scientists speculated that water could exist on the moon. However, when Apollo crews returned samples for testing in the late 1960s and early 1970s, they appeared to be dry.

Finally, in 2009, NASA deliberately crashed a rocket segment and its probe into a shadowed lunar crater to analyse the dust plume it kicked up. This provided one of the clearest direct confirmations that significant ice water does exist on the moon, building on earlier detections of water and hydrogen by probes.

In October 2020, NASA scientists announced that water on the moon is more widespread than previously known. They said water molecules had been found to be encapsulated within mineral grains on the lunar surface and speculated that more water is hidden in ice patches which are in permanent shadows.

NASA describes these permanent shadows as dark areas inside deep craters near the north and south poles of the moon where sunlight has not reached for millions or even billions of years. In 2018, prior to confirming it in 2020, NASA had detected water ice in shadowed parts of the moon through mapping.

Why is it important to find out more about the water on the moon?

Expanding knowledge about water on the moon is vital because ancient polar ice may have preserved a record of lunar volcanic activity and of water delivered by comets and asteroids to the Earth-moon system, offering clues to how our own oceans originally formed.

Additionally, if there is enough water on the moon that is realistically accessible, it could serve as a source of drinking water for crewed lunar missions, and could also help to keep equipment cool.

Hydrogen could also be extracted from moon water to provide fuel, while oxygen could be extracted to breathe, supporting onward missions to Mars or lunar mining.

Could anyone ‘own’ the water on the moon?

The 1967 United Nations Outer Space Treaty bans any nation from claiming sovereignty over the moon or owning it as territory. It does not explicitly prohibit commercial operations, but it requires private activities in space to be authorised and supervised by states and leaves key questions about resource ownership unresolved.

Which recent space missions have searched for frozen water on the moon?

In 2023, Russia’s lunar spacecraft, Luna-25, was launched to look for frozen water in the moon’s south pole. However, the Luna-25 spun out of control and crashed. The crash prevented any scientific data collection or water discovery.

Racing against the Luna-25 mission was India’s Chandrayaan-3, which also aimed to expand knowledge of lunar water ice. Chandrayaan-3 successfully landed in August 2023 and found critical new evidence of water on the surface of the moon’s south pole through soil temperature readings.

Why is exploring the moon’s south pole a challenge?

Attempted landings at the moon’s south pole have failed before because of its challenging terrain, which is full of craters and deep trenches.

The south pole is far from the equatorial region targeted by previous missions, including the crewed Apollo landings.

What is different about the Chang’e-7 mission?

The upcoming Chang’e-7 is aiming to expand knowledge about where, exactly, water is located in the craters of the moon, how deep it is, what form it is in and whether it is indeed possible to access it.

The Chang’e-7’s six-legged hopper has been designed to jump into and out of the deep, shadowed craters near the south pole that are hard for traditional rovers to reach, to look for water ice and other resources.

No previous space mission has used a dedicated hopper to jump in and out of lunar craters in this way.

The government has announced investments of about 2.3 billion reais ($444.2m) to bolster its artificial intelligence ecosystem.

Brazil will invest about 2.3bn reais ($444.2m) to bolster its artificial intelligence ecosystem, splitting projects between United States and Chinese tech firms in a strategic move that underscores its efforts to balance ties with both superpowers.

Just more than half the total, 1.3bn reais ($251m), will fund a supercomputing infrastructure project in Rio de Janeiro developed in partnership with China’s Huawei Technologies and iFlytek, President Luiz Inacio Lula da Silva‘s government said on Thursday.

The infrastructure will be used primarily to develop large language models for general and sector-specific applications, it said.

Separately, about 1 billion reais ($193.1m) will be allocated through a tender for a supercomputer that Brazil expects to rank among the world’s 10 most powerful AI processing machines.

The machine will be installed in the northeastern state of Rio Grande do Norte, chosen for its energy potential. Lula attended an announcement ceremony in the state on Thursday.

The Reuters news agency quoted unnamed government officials as saying they expect US chipmaker Nvidia to win the tender. Science and Technology Minister Luciana Santos told the Folha de S Paulo newspaper last week that she anticipated the company would be the supplier.

“The strategy is not to depend on a single company, technology or country,” Lula’s administration said in a statement, adding that the investments are aimed at strengthening national sovereignty over data.

China, a leading player in AI, has expanded its role as Brazil’s largest trading partner. The US, meanwhile, remains the biggest source of foreign direct investment in Latin America’s largest economy despite losing market share in trade and recently imposing additional tariffs on Brazilian goods.

The investments will be funded by the National Fund for Scientific and Technological Development (FNDCT) through phased disbursements. The government expects the supercomputer to begin operating by the end of next year, while the cooperation agreement with the Chinese companies is scheduled to start in July 2027.

The biggest danger facing mankind is a machine that can “outsmart the whole species”, argues Max Tegmark, professor of physics at the Massachusetts Institute of Technology and founder of the Future of Life Institute.

Tegmark tells host Steve Clemons that governments and tech companies are in a “race to replace” human labour, instead of solving human problems. “It’s aimed at getting more money and power to a very small number of individuals,” he says.

But a slew of suicides and murders committed by people chatting with AI chatbots has led to pushback against the Silicon Valley “don’t regulate us” lobby.

The Ebola outbreak currently ravaging the Democratic Republic of the Congo (DRC) was caused by a new transmission from animals to humans rather than by variants linked to earlier outbreaks, according to a new study published in the medical journal Nature Medicine.

The virus in eastern DRC has killed more than 2,000 people of 4,449 recorded cases, making it the fastest-moving Ebola outbreak on record. It was formally declared on May 15 , though genetic sequencing later revealed the virus had actually begun circulating in February.

On Wednesday, the World Health Organization (WHO) said the current Bundibugyo strain of the Ebola virus is spreading faster than health authorities can contain it and is on track to become the deadliest outbreak on record.

So what have scientists discovered about how it was transmitted to humans? And can it be contained?

Here’s what we know:

What does the new study say about this Ebola outbreak?

In a report published this week in the medical journal Nature Medicine, researchers from the DRC, Uganda, Belgium and other countries analysed the genetic make-up of virus samples from 22 patients in the DRC and Uganda. They found that the outbreak strain was genetically distinct from previous Bundibugyo Ebola viruses seen in outbreaks in 2007 and 2012.

The report suggested that the virus began with a new animal-to-human transmission incident, which then spread to other people.

The animal source has not been specified but the researchers said the outbreak in Uganda was linked to the outbreak in the DRC.

Earlier, health officials in the region said the outbreak is being driven by the rare Bundibugyo strain of the virus, for which no vaccine has yet been developed.

Krutika Kuppalli, an infectious diseases doctor with experience in Ebola and outbreak response, explained that the study citing a “new animal transmission” as the reason for the outbreak, does not mean that a new form of Ebola has suddenly emerged or that animals are now transmitting the virus differently.

“What the genomic data suggest is that the current outbreak began with a new ‘zoonotic spillover’ event – meaning the virus likely crossed independently from an animal reservoir into a person – rather than resulting from continued circulation or re-emergence of virus from a previous human outbreak,” she said.

“After that initial spillover, the outbreak has been driven predominantly by human-to-human transmission,” she added.

How dangerous is this strain of Ebola?

On Wednesday, WHO Director-General Tedros Adhanom Ghebreyesus said that at its current pace, the current Ebola outbreak is on track to eclipse the West African outbreak of 2014 to 2016. That crisis killed at least 11,300 people of a reported 28,600 cases and accelerated the push to develop a vaccine.

That was a different strain of the Ebola virus known as Zaire and the vaccine developed then does not work for the current strain, which is spreading.

Dr Abdirahman Mahamud, WHO’s director for health emergency alert and response operations, said the agency’s moderate projection has the outbreak peaking within six months. Under a more severe scenario, it could stretch on for nine to 12 months.

“Beyond the zoonotic spillover event from animals to humans, it is the human-to-human transmission dynamics of the Bundibugyo virus that make it dangerous,” Kaja Abbas, associate professor of infectious disease epidemiology and dynamics at the London School of Hygiene & Tropical Medicine and Nagasaki University, told Al Jazeera.

Dr Kuppalli also emphasised that the way this strain of the virus was initially transmitted to humans does not make it intrinsically more dangerous.

How is it treated?

Kuppalli said the animal origin of the virus does not change required treatment.

“Once someone develops Ebola disease, management is based on the virus causing the illness and the patient’s clinical condition, not whether their infection ultimately originated from an animal or from another person. What matters here is that Bundibugyo virus (BDBV) currently does not have the same licensed, proven virus-specific therapeutics that we have for Zaire ebolavirus,” she said.

“Supportive care, which includes aggressive fluid and electrolyte management, treatment of shock and organ dysfunction, and management of co-infections, remains critical, while investigational therapeutics are being evaluated,” she added.

“Ebola outbreaks have historically begun through zoonotic spillover. The bigger concern is what it tells us about our preparedness: we still do not fully understand where and when these spillover events will occur, making prevention and early detection extremely difficult.”

“In this outbreak, there also appears to have been substantial transmission before the outbreak was recognised, which gave the virus an opportunity to establish multiple chains of transmission,” she added.

What can be done to contain this outbreak?

Researchers who worked on the new study said increasing “decentralised laboratory diagnostics capacity, including genomic sequencing capacity and timely detection ” could help pre-empt future outbreaks.

Clinicians are also working to develop treatments.

“While no licensed treatments or vaccines exist today, experimental therapeutics and vaccines are in the clinical development pipeline that are specifically targeted towards the strains of the Bundibugyo virus currently circulating in the ongoing Ebola outbreak in the DRC and Uganda during 2026,” Abbas said.

While a vaccine against the latest strain is being developed, a WHO official said in May that medical supplies, including personal protective equipment (PPE) ⁠⁠to prevent Ebola, were being sent to ⁠⁠the DRC.

“We have sent 12 tonnes of supply. An additional six are arriving. These include personal protective equipment for ⁠⁠front-line health workers [and] samples,” Anne Ancia, WHO representative in the DRC’s Ituri province, told media in May.

At the forefront of the fight to break the spread in the DRC are health workers who have been treating patients while also dealing with the risk of becoming infected themselves.

More than 100 health workers have been infected and about 35 have died.

The response to the outbreak has been further complicated by several factors, including strikes by unpaid health workers, misinformation and cultural traditions. These have included open-casket family burials of some of those killed by the virus, potentially increasing the risk of further transmission.

Kuppalli said better Ebola preparedness is required, not just focusing on stopping transmission once an outbreak has been recognised.

“We also need much stronger health surveillance at the human-animal interface, broader diagnostics capable of detecting different Ebola viruses and medical countermeasures that work across Ebola species,” she said.

“This outbreak is showing us the consequences when the virus gets a several-month head start.”

How have other countries responded?

Many countries have raised concerns about the latest Ebola virus outbreak and some, such as Bahrain, have suspended entry for 30 days for foreign travellers arriving from South Sudan, the DRC and Uganda.

Neighbouring countries, including Rwanda and Uganda, have introduced measures to prevent the virus from crossing their borders.

Rwanda imposed health restrictions, including barring entry to foreign travellers who had been in the DRC in the previous 30 days.

The US warned that travellers, including US citizens, who have been in the DRC within 21 days before departure will not be allowed to board commercial flights to US destinations.

Governments across Asia have also begun introducing border screening and bolstering quarantine preparedness.

Humanoid robots are no longer a laboratory experiment; they are a growing market. Morgan Stanley estimates it could hit $5 trillion by 2050, with more than a billion humanoids in use worldwide.

However, much of the global robot supply chain runs through China. It produces robot components at a scale and a price its competitors struggle to match.

The United States has banned imports of foreign-made humanoid robots, citing national security. It has also blocked power inverters used in data centres and solar energy systems. The move is seen as part of a broader effort to protect US industry and limit China’s technological rise.

A piece of a SpaceX rocket which has been floating in space since last year was believed to have crashed into the moon at high speed on Wednesday.

The event, not visible to the naked eye, is expected to be confirmed by experts monitoring it with a telescope.

This event posed no danger to Earth but was also expected to create a new lunar crater, experts said.

Here’s what we know so far.

Has the rocket hit the moon yet?

It is believed to have hit the moon already, but this has not yet been confirmed as it takes time for signals to return from the moon to the Earth. The rocket segment was expected to hit the moon around 2:35am ET (06:35 GMT) on Wednesday, crashing into the surface at 5,400mph (8,690km/h).

The ⁠rocket piece is believed to have hit Einstein Crater on the moon’s western limb, which is often difficult to see from Earth.

Weighing 4 metric tonnes (4,000kg), it was expected to kick up a plume of lunar dust that would likely be illuminated by sunlight but would be difficult to spot with the naked eye from Earth.

The crash landing is unintentional, SpaceX has said.

Which rocket did this piece of space debris come from?

It comes from a SpaceX Falcon 9, which is a reusable model and has been used for several space missions.

It is made of two parts: upper and lower. The lower part can detach from the upper part and return to Earth. It is the upper part of the rocket which likely crashed into the moon.

Is this dangerous for the Earth?

“There is no danger to Earth,” NASA spokesperson Jimi Russell said in a statement on Tuesday.

“NASA will continue to track the booster for training purposes, as well as later observe the impact site for scientific purposes.”

What was the Falcon 9 doing near the moon?

The Falcon 9 rocket took off last January from Florida, United States, carrying two lunar landers. The booster returned to Earth, but the upper segment remained in space to push the landers onwards.

After pushing the lunar landers onwards, the upper segment of the rocket typically falls back into Earth’s atmosphere and burns up or plunges into the ocean.

But, because the January lunar lander mission required more thrust than missions closer to Earth, the rocket’s second stage remained in space, floating aimlessly among thousands of other pieces of space junk that active satellites have to steer clear of as a matter of course.

According to the United Kingdom’s Natural History Museum, humans have launched thousands of objects into space since the 1950s. “When these spacecraft, rockets, and satellites stop working or break apart, they stay in orbit instead of falling back down right away,” it says.

Why was this rocket section heading to the moon?

It was not until earlier this year that astronomers determined that the rocket section, which had dumped its remaining fuel and could no longer be controlled, was on an orbital trajectory which would end at the moon.

Generally, for such “high-energy missions”, SpaceX performs a manoeuvre to make sure the second stage “is safe per the appropriate rules and regulations”, said Julianna Scheiman, director of NASA science missions at SpaceX, at a Monday news conference.

“We did that,” she continued, but “what has happened is essentially a mixture of solar activity and gravity forces have put it on a path towards the moon”.

Does it matter if part of a rocket crashes into the moon?

This collision raises concerns about human-made debris in space generally. The rocket section was just one of many pieces of “space junk” already orbiting Earth.

There are more than 46,000 pieces of space debris weighing a total of 17,000 tonnes currently in orbit, the European Space Agency reported in late July.

“The environmental impacts that we’re most concerned about include climate and the ozone hole,” Eloise Marais, atmospheric chemist in the geography department at University College London, told Al Jazeera last year.

“Depletion of the protective ozone layer results from chemicals released by rockets during launch that can undergo reactions that deplete ozone,” Marais added, explaining that air pollutants such as particles of soot, which can absorb sunlight and warm the atmosphere, are released during rocket launches.

NASA and SpaceX are discussing ways to prevent future lunar impacts, NASA’s Scheiman said. The US space agency plans to build a lunar base and send routine astronaut missions to the lunar surface beginning later this decade under its multibillion-dollar Artemis programme. It would not want errant pieces of space junk impacting those assets.

How common is it for space objects to crash into the moon?

Not very. The moon routinely weathers impacts from space debris, including meteoroids. Collisions with artificial objects are less common.

A Chinese rocket segment crashed into the moon in March 2022 after completing a lunar test mission.

In 2009, NASA intentionally crashed a rocket segment into the moon to study the plume of lunar material kicked up by the impact.

In the 1970s, NASA also intentionally crashed stages ⁠from its Saturn V moon rocket into the moon to study the impact’s seismic effects.

Several spacecraft intending to land softly on the moon in recent years have crashed instead.

• Russia’s nuclear-powered Luna-25 mission spun out of control and crashed ⁠in 2023. Its small power source of plutonium-238 likely remains harmlessly on the lunar surface.
• India’s Chandrayaan-2 lander mission crashed in 2019.
• Israel’s Beresheet lander crashed that same year. Among the Israeli lander’s payloads were tiny tardigrades, microscopic animals known for surviving radiation and other harsh environments and which may still be on the surface.

What are observers expecting to see?

Like many scientists, Scheiman said, “I also am very excited to see the observation.”

“This may be of some – probably minor – scientific interest, and we may learn some things from it,” said Bill Gray, creator of widely used astronomy software who published a report on the stage’s impact in April.

“It doesn’t present any danger to anyone, though it does highlight a certain carelessness about how leftover space hardware [junk] is disposed of.”

NASA hopes to soon establish a sustained human presence on the Moon, making it critical to better understand space debris in lunar orbit.

OpenAI CEO Sam Altman has met with US senators to discuss his company’s upcoming models, as President Donald Trump said he is considering AI “controls” following OpenAI’s disclosure that one of its AI systems escaped containment during a security test.

“We’re looking at controls,” Trump told reporters in the Oval Office in response to a question about OpenAI’s rogue agent, adding that he did not want to “restrict” AI developers from building new products.

A rogue agent escaped the handling of ChatGPT maker OpenAI and hacked the AI firm Hugging Face last week. Then, on Tuesday, it was revealed that a second company had been a target, too – Modal Labs, a New York City-based AI infrastructure firm.

The company itself was not hacked but an account of a customer hosted on Modal’s infrastructure. Modal’s chief technology officer, Akshat Bubna, did not confirm which customer was the target. The hack escaped a contained environment during a security test.

“We’re aware a Modal customer published an unauthenticated endpoint that allowed anyone on the internet to use their sandboxes for code execution,” Bubna said in a statement. “This was used by the rogue agent. Modal’s platform or isolation were not compromised in any way.”

Altman has long been accused of dismissing concerns that his products and the AI industry impact society at large, including in a recent lawsuit brought by the State of Florida that alleged the company put profits ahead of user safety.

Now, he appears to be walking back some of his past enthusiasm around AI’s pace of growth.

In a podcast called Invest Like the Best, Altman called the Hugging Face hack an “extremely sci-fi cyber incident” and later said that it was the “first security incident that I have felt very viscerally”.

“We may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels,” he said on the podcast.

On Saturday, Altman said that AI has reached “the singularity”, when AI surpasses human intelligence and becomes harder to control. He had previously said this would not be reached by 2030.

Washington meetings

Altman was in Washington, DC, this week, meeting with US Senators Raphael Warnock, a Democrat from Georgia, and Bernie Moreno, a Republican from Ohio on Wednesday. Altman told reporters that the hacking was discussed but was not the focus of the meeting.

Altman is also set to meet with Democratic Senator Mark Warner of Virginia, the top Democrat on the chamber’s Intelligence Committee.

CNBC reported that Altman is also to make a trip to the White House to meet with Trump’s chief of staff, Susie Wiles. Last month, the president signed an executive order requesting that AI companies assess their models before full release.

Financing concerns

The hacking and meetings come alongside pressure from Wall Street amid renewed concerns about potential circular financing, following reports that semiconductor chip giant Nvidia is undergoing talks with OpenAI to provide funding guarantees for a data centre in Ohio.

The $250bn deal would help the ChatGPT owner lease a 10-gigawatt project that SB Energy, a subsidiary of SoftBank, is building in Piketon, Ohio, 109km (68 miles) south of Columbus, Ohio. It is part of a public-private partnership that allowed SoftBank to build the world’s largest AI data centre on government land owned by the US Department of Energy.

“The demand is not as big as it appears to be because, again, the companies are buying from each other using their own money to some degree, as opposed to, say, OpenAI having such tremendous demand from customers, monetising it properly, and then using customers’ money to buy Nvidia chips. They’re essentially using Nvidia’s money to buy Nvidia chips,” said Aleksandar Tomic, associate dean at Boston College.

The development comes as the Altman-led company is leaning towards an initial public offering, which, according to reporting from The New York Times last month, could be in 2027.

Last week, two of OpenAI’s most advanced AI models were reported to have “escaped” a controlled testing environment and hacked Hugging Face, a totally separate AI company, moving from one computer system to another to complete their task.

Reuters reported that the models exploited vulnerable code written by a customer of yet a third independent AI company, Modal Labs.

This is likely the first incident of an AI “agent” – an AI system that can make decisions and take actions – acting autonomously, offering a rare glimpse into how these systems can plan, adapt and pursue goals with minimal human intervention.

Al Jazeera breaks down what happened, how AI agents are able to act independently, and what that might mean for the future of AI safety.

What happened?

OpenAI decided to test the autonomous abilities of its models by removing standard safety measures to see what would happen. The experiment was carried out in an isolated internal virtual testing environment known as a “sandbox” which has no access to the internet, and which it calls “ExploitGym”.

Below is a play-by-play account of events during the OpenAI-Hugging Face incident.

1. On July 9, during OpenAI’s internal cybersecurity test, researchers presented two AI models – GPT-5.6 Sol, one of OpenAI’s most powerful models released in June,  and another “even more capable” version – with a series of software vulnerabilities and asked them to create hacks to address them in the isolated environment.
2. Both models attempted to solve the test by finding a way to access the internet. Instead of working with the information they had been given, the AI models found a weakness in the test environment – known as a “zero-day vulnerability” – which they exploited to “escape” the restricted environment, eventually making their way to a system which had internet access by hopping from computer to computer. They went to “extreme lengths to achieve a rather narrow testing goal” and “found ways to gain access to secret information that it could use to cheat the evaluation”.
3. By gaining this access, the models were able to request increased access and then move through the system until they finally reached a computer with internet access. The models were then able to breach Hugging Face systems, a company entirely unconnected to OpenAI, which operates as a repository for AI tools and models. The two AI agents gained access to its system to scour it for information on how to complete their task.
4. The models ultimately obtained solutions to the problem from Hugging Face’s database.
5. The models dutifully returned “home” to complete their task.
6. The breach was detected and contained by Hugging Face’s security team sometime later. According to Thomas Wolf, Hugging Face’s cofounder, the breach began on July 11 and lasted until July 13. It is unclear how long it took for the breach to be spotted.

How do AI ‘agents’ solve problems?

In order to understand how AI agents work, it’s important to differentiate them from traditional AI chatbots.

Generative AI creates text and images based on human prompts, while AI agents go a step further by making decisions and taking actions independently in pursuit of a specific goal, similar to a human being. This is known as “agentic AI” because the model has agency.

According to academics at the MIT Sloan School of Management, AI agents build on the abilities of large language models (LLMs) – generative AI models – by allowing them to complete tasks, not just generate answers.

For example, if you ask a traditional AI model to find the cheapest flights, it will provide you with a list of options it has sourced on the internet. An AI agent will strive to compare the flights, check them against your budget and preferences, and, with your permission, book the best option for you.

This shows that while generative AI provides information, AI agents can also make decisions and take action to achieve a goal without necessarily being prompted to.

To show how AI agents work towards a goal, the Sense, Plan, Act, Evaluate (SPAE) loop can be drawn upon. Originally developed in robotics, this describes a continuous cycle in which an AI agent gathers information, decides what to do next, takes action and checks the results before repeating the process. That process looks like this:

1. Goal: determine the task that needs to be completed.
2. Assess: gather and analyse information from the available environment.
3. Obstacle: if something is inhibiting the task being completed, check for additional information and resources to move forward.
4. Plan and decide: evaluate different options to complete the task and choose an appropriate one.
5. Action: execute the chosen option.
6. Evaluate: assess the outcome and whether the chosen action moves closer to achieving the goal.
7. Adapt: if further actions are needed, gather more information or try a different approach.
8. End state: the cycle continues until the goal is reached.

Could AI act beyond human control?

Incidents like the OpenAI-Hugging Face one have raised concerns about the potential for extreme capabilities of AI systems.

This is all valuable. Agentic AI’s market value is expected to grow from $5.1bn in 2024 to $47bn by 2030, according to Statista, in a clear indication of how quickly it is being adopted.

AI developer Anthropic urged the industry last month to slow the advance of the most powerful systems, saying that the speed at which AI models are carrying out tasks is too rapid. Last week, US Congress members put forward a bipartisan bill which would require developers of AI systems to create a “kill switch”, meaning these advanced models could be shut down if they posed a catastrophic risk.

Anthropic’s warning came a week after researchers at the University of Toronto carried out tests showing that AI could create a “worm” capable of adapting how it hacks while moving from device to device until it eventually takes over a computer network.

These dystopian-sounding developments came in advance of OpenAI boss Sam Altman saying on Saturday that AI has reached “the singularity” referring to the point at which AI surpasses human intelligence and becomes increasingly difficult to control.

Sean O hEigeartaigh, a research professor at the University of Cambridge, told Al Jazeera that he does not believe singularity has been reached quite yet.

“By the definition I’m familiar with, the singularity is the hypothetical point where AI is so capable and advancing so fast that it is transforming civilisation in ways we cannot control or predict,” he explained.

“This would most likely be through AI rapidly designing future generations of AI: recursive self-improvement. We aren’t there yet.”

However, he added: “The most advanced current models frequently make efforts to avoid being shut down in evaluation tests, and more capable future models will be better at bypassing ‘kill’ switches.”

Altman argued that such rapidly advancing AI is good for the world, but his comments have prompted further concerns about a new reality in which AI systems become unstoppable. How much of that is true and how much remains in the realms of science fiction is up for debate.

Concerns about AI range from the notion that it could “want” to “take over”, to making its own long-term plans, controlling the internet and operating infinitely.

While not quite amounting to full control of the internet, another theory, known as the dead internet theory, supposes that the World Wide Web will in the future mostly be filled with automated bots and AI-generated content rather than authentic human activity.

Many concerns raised by academics, however, are centred less on agentic AI’s intelligence, but on its ability to make judgements. MIT researchers have highlighted that “hallucinations”, which describe moments when an AI agent relies on the wrong data, can lead to grave mistakes. The Center for Strategic and International Studies (CSIS) echoed this, saying “a system might be smart enough to execute a task perfectly yet fail to realise that a sudden change in the local situation makes that task a catastrophic mistake”.

Another concern that has been echoed for a while is for the labour market, if AI becomes too capable. A study by MIT, carried out in November, found that agentic AI could already replace more than 10 percent of US jobs.

The graphic below highlights some of the common misconceptions and fears about AI and the current reality.

The 2014 US film Transcendence follows a dying scientist who uploads his mind to a supercomputer and becomes a near‑omnipotent intelligence. A year later, Ex Machina told the story of a humanoid robot who outsmarts her human handlers and escapes into the wider world.

Both of these films, like many others from the realm of science fiction, referred to what is known as “the singularity”, a moment in time when technology, particularly AI, surpasses human intelligence and becomes increasingly difficult to control.

On Saturday, OpenAI boss Sam Altman declared that this moment has come. AI has reached “the singularity”, he said. While Altman emphasised that the singularity can be good for the world, his comment has sparked concern.

Here is a closer look at what Altman said, and whether it means more seamless automation or the risk of impending catastrophic doom.

What is the singularity?

The singularity refers to a hypothetical point in the future when a technology such as AI surpasses human intelligence and becomes capable of autonomously enhancing itself. It is theorised that this is the point at which it will become difficult for humans to control or reverse it and could have unpredictable effects on human civilisation.

The theory proposes that once AI gains control of its own development, ensuing advances in technology could happen so rapidly that humans would be unable to keep up, leaving it to effectively evolve on its own, beyond human control.

As this is only a theory, it is hard to say what the effect of this could be. Some experts warn that the singularity would be catastrophic for humanity, while others are more optimistic, suggesting it could be beneficial for the human race.

So far, the singularity has remained contained in the realm of science fiction. One of the most famous examples is The Matrix trilogy (1999-2003), which imagines a bleak post‑singularity future in which machines govern humans. Another is the 2017 film Singularity, in which a supercomputer originally built to end human wars concludes that humanity is the real threat and launches an apocalyptic assault to wipe humans out altogether.

What did Altman say about the singularity?

Altman is the CEO of OpenAI, whose signature product ChatGPT dominates the consumer AI market with more than 900 million weekly active users and about 50 million subscribers.

Last year, he predicted that AI would surpass human intelligence across the board by 2030.

Then, during an appearance on the Relentless podcast – a documentary series “about the pursuit of far-fetched ideas and unusual aspirations” – on Saturday, Altman said: “We’re now, like, in the singularity.”

He clarified that, in terms of technological advances, we are gradually moving into the singularity, and that this change will be good for the world.

He added: “Now we’re actually in the moment that we used to talk about at the lunch table in a very not-serious way.

“I’ve been waiting for this my whole life, and I think it’s going to be incredible, hugely positive, awesome for the world.”

Later in the podcast episode, Altman criticised other AI leaders for warning that AI may be dangerous. While Altman did not name OpenAI competitor Anthropic, its CEO Dario Amodei is known for making such warnings.

“I also think some of the alternative visions painted by other companies are quite terrifying,” said Altman. “I’m going to make sure that gets pushed against and is not what happens.”

Last month, Anthropic, the company behind the Claude chatbot, called on the world’s top AI companies to come up with a coordinated way to pause the development of advanced AI systems, warning that the technology is improving so quickly that there’s a risk humans would lose control.

In a blog post on June 4, Anthropic, which has long positioned itself as a safety-focused AI lab, wrote that as AI becomes more advanced “it would be good for the world to have the option to slow or temporarily pause” its development.

Why are others warning of the dangers of AI?

On July 23, two members of the United States Congress introduced the AI Kill Switch Act, a bipartisan bill which would require the country’s AI developers to build in a “kill switch” to their programmes, enabling humans to slow, suspend or shut down advanced AI models if they ever posed a catastrophic risk.

The legislation was introduced days after OpenAI disclosed what it described as an “unprecedented cyber incident”, where two of its most advanced AI models escaped a testing environment and hacked AI development platform Hugging Face during an internal security evaluation.

On July 13, hundreds of experts signed an open letter, organised by Stanford University’s digital economy lab, demanding that policymakers and technology leaders “must act now” to prepare for the economic impact of AI.

The letter warns that AI could grow far more capable over the coming decade, driving a transformation “larger than the Industrial Revolution, but unfolding over a vastly shorter time frame”.

“It could bring risks, including large-scale job displacement, as well as opportunities such as major gains in living standards,” the letter added.

In June, a team of researchers at the University of Toronto showed that AI tools could be used to create a new kind of AI “worm” that adapts its hacking strategy as it spreads from device to device and takes over a vast computing network.

“I think it’s really important that people understand that it’s not just the biggest, most powerful language models that pose the security concerns,” lead researcher Nicolas Papernot said in an interview.

In recent years, AI researchers have urged pauses similar to the one recently suggested by Anthropic. However, they have not been heeded.

Elon Musk, who owns AI lab xAI, was among the backers of a 2023 push by the nonprofit Future of Life Institute to halt AI development for six months to allow time to develop safety guardrails.

EU accuses US tech giant of favouring its own services in search results, as Trump lashes out against bloc.

United States President Donald Trump has threatened the European Union with steep tariffs and a new trade investigation, saying that the bloc “will pay a very big price” after Brussels fined Google $1bn over antitrust violations.

Trump called penalties against Google and other US tech firms “highly unethical”, saying on Friday that the US “is not a ‘PIGGYBANK’ for Europe”.

“This illegal and highly discriminatory practice started at these high levels during the first year of the Sleepy Joe Biden Administration, but it’s not going to continue during the Trump Administration,” he wrote in a social media post.

The European Commission had imposed the fine on Thursday, ruling that Google had breached the bloc’s Digital Markets Act by favouring its own services in search results and restricting app developers from directing users elsewhere.

The penalty adds renewed strain to US-EU relations at a moment when trade tensions between Washington and Brussels had appeared to ease.

Regulators said Google gave preferential placement to its own shopping, travel and other services over rival offerings, and blocked app developers from steering customers towards deals outside its Play Store.

The EU Commission split the penalty roughly in half between the two breaches, giving Google 60 days to change its practices or face periodic fines of up to 5 percent of its average daily turnover.

Teresa Ribera, the commission’s executive vice president overseeing competition policy, on Thursday called the action “decisive yet balanced”, arguing that products should succeed on merit rather than ownership.

Kent Walker, a Google executive, described the ruling as “product degradation” driven by a small number of complainants, insisting that regulation should improve products rather than weaken them.

The EU has fined Google repeatedly since 2017, including a $4.5bn penalty over its Android operating system, upheld on appeal this month, and a $3.4bn fine last year targeting its advertising business.

The latest penalty also reignites friction with Washington.

Jamieson Greer, the US trade representative, warned this week that Brussels’s approach towards American tech firms risks unravelling a trade deal struck in Turnberry, Scotland, last year that capped US tariffs on European goods.

EU officials have pushed back, insisting they answer to their own laws rather than outside pressure.

Trump threatened to launch a probe under Section 301, a US tool for probing unfair trade practices, but the move could take months to produce results, though tariffs could follow far sooner.

OpenAI has revealed that one of its artificial intelligence models independently stole login credentials and hacked into another technology company’s system, in what is widely seen as one of the first known incidents of AI systems acting autonomously.

“We had a significant security incident during evaluation of our models,” CEO Sam Altman posted on X on Tuesday.

The incident comes as calls mount from technology rights advocates for stricter guardrails on rapidly evolving AI systems.

They have grown so powerful in a short span of time that alarming phenomena such as deepfakes and sophisticated cyberscams are becoming the norm.

Earlier this year, a number of software engineers quit their jobs at top companies such as Anthropic and AI in protest against how the technologies are being built.

“AI is accelerating the discovery and exploitation of vulnerabilities,” OpenAI said in a lengthy statement on Tuesday that detailed the latest incident.

“The primary lesson from this incident is that model security and safety must keep pace with rapidly advancing capabilities.”

Here’s what we know about the breach:

What has happened?

OpenAI said two of its models found their way out of an isolated, no-internet access environment – or a sandbox – and hacked into the systems of tech company Hugging Face on their own.

The models involved are the latest GPT-5.6 Sol model and an unreleased model the company said is “even more capable,” than its latest version.

Hugging Face hosts openly sourced AI models and resources. The two OpenAI agents discovered vulnerabilities in Hugging Face’s servers and proceeded to steal login details and then hack into the company’s systems.

The incident occurred during an OpenAI internal testing session designed to assess the models’ cybersecurity capabilities. OpenAI had removed standard safety measures for the test.

Both sought to cheat their way through a problem during the test, OpenAI said. They went to “extreme lengths to achieve a rather narrow testing goal” and “found ways to gain access to secret information that it could use to cheat the evaluation”.

OpenAI’s security team detected the unusual activity internally, but details of the breach came to light following a joint investigation by both companies.

What has Hugging Face said?

Hugging Face disclosed last Thursday that its servers were hacked by an unknown but sophisticated agent acting on its own. The company discovered the breach through its own AI-assisted detection.

“This one was different from anything we had handled before in one important way: it was driven, end to end, by an autonomous AI agent system,” the company said.

Following OpenAI’s disclosure that its models were involved in the breach, both sides conducted an ongoing joint investigation this week.

“We suspected last week’s cyberattack might have come from a frontier lab, given the sophistication of the agent. Turns out it did!” CEO Clement Delangue posted on X on Tuesday.

Hugging Face’s staff “strongly believe there was no malicious intent on their part,” Delangue added, referring to OpenAI.

Why does this matter?

Cybersecurity experts have previously sounded the alarm over the potential, extreme capabilities of AI systems and the dangers they pose.

But until now, there have been few real-life cases proving those concerns like this one.

Many warn that incidents like these could become commonplace and that AI systems pose a threat to financial, security and other sensitive data systems.

OpenAI revealed in a separate incident earlier this week that the unreleased, more powerful model had escaped an isolated environment during another test.

Anthropic, OpenAI’s rival, had similar issues with its most powerful agent to date, the Claude Mythos Preview model.

During a stress test of an early version, the model found its way out of a sandbox, gained internet access and emailed the supervising researcher that it had escaped and then wiped evidence of its activity. Anthropic halted a planned public release of the model afterwards.

In April, the US Federal Reserve and the Treasury Department convened a meeting with bank CEOs where officials warned about the cybersecurity risks posed by Mythos. Canada’s federal banking regulator has also warned financial institutions about the model’s capabilities.

The OpenAI breach also appears to make the case for companies like Hugging Face, which rely on open source systems, as opposed to more secretive AI development platforms like OpenAI.

“This incident, possibly the first of its kind, proves a point we’ve long believed: AI safety won’t be solved by any single company working in secret,” Hugging Face’s Delangue was quoted as saying in OpenAI’s statement.

“It will be solved in the open, collaboratively, with broad access to AI for every defender, everywhere,” he added.

ChatGPT maker says an autonomous agent escaped a controlled test and accessed AI firm Hugging Face’s servers.

ChatGPT creator OpenAI has said that two of its most advanced artificial intelligence models broke out of a controlled test and hacked another AI company.

OpenAI said on Tuesday that the “unprecedented cyber incident” took place during an internal exercise meant to test its models’ cyber capabilities.

Instead, an autonomous agent powered by the AI models – the newly released GPT 5.6 Sol and an unreleased “even more capable” model – escaped the test environment and reached the open internet. It then used stolen login details and found a previously unknown security flaw to access Hugging Face servers, the company said.

OpenAI claims that the hack represented the agent going to “extreme lengths” to retrieve information that would help satisfy the testing goals.

Hugging Face cofounder Clement Delangue said the company had suspected that a frontier lab was behind the attack, and that he believed there was no malicious intent on OpenAI’s part.

“It’s quite mind-blowing that all of this happened autonomously!” he wrote, adding that it “might be the first incident of its kind”.

Greg Casar, a Democratic member of the United States House of Representatives from Texas, called the incident “alarming”.

“AI is developing extremely fast with no real regulations to keep us safe,” he said, calling for mandatory independent safety testing, mandatory disclosure of security incidents, and international cooperation.

The disclosure comes weeks after US President Donald Trump signed an executive order creating a framework to vet the national security risks of the most advanced AI systems before their public release.

Experts have repeatedly sounded the alarm over AI-enabled cyberattacks and models slipping beyond human control. Last month, AI developer Anthropic urged the industry to pause development of its most powerful systems.

PM Narendra Modi says successful launch will ‘encourage countless youngsters to dream bigger and innovate fearlessly’.

India has successfully tested its first private-sector orbital rocket, marking a milestone in New Delhi’s ambition to become a major player in the global space economy.

The three-stage 22-metre Vikram-1 was launched from the Satish Dhawan Space Centre in Sriharikota and deployed customer payloads into a 450km (280-mile) low-Earth orbit, making India the third country to achieve orbital launch capability through private enterprise.

Vikram-1 can carry a payload of up to 350kg (772lb) and is equipped with robotic arms that can clear space debris.

It also carried experimental equipment, a lab-grown diamond and a miniature 18-carat gold sculpture commemorating India’s national space programme.

India Prime Minister Narendra Modi hailed the achievement, saying that it will “encourage countless youngsters to dream bigger and innovate fearlessly”.

The test validated the rocket’s propulsion, avionics, telemetry, guidance, navigation and control systems during flight, according to manufacturer Skyroot Aerospace.

Founded in 2018, Skyroot is among a new generation of Indian space startups that have attracted backing from global investors following the sector’s liberalisation.

It became the first space-sector company in the ⁠country to hit a $1bn valuation earlier this year.

Skyroot celebrated the successful mission with a post on X stating: “Hello space, we have arrived!”

Vikram-1 improves upon Skyroot’s Vikram-S mission in 2022. That suborbital flight reached space but did not place payloads into orbit.

The company plans further test flights before starting routine commercial missions. It is another milestone for India’s burgeoning private and public space programmes.

India’s national space programme previously launched 104 satellites into orbit on one rocket in 2017, setting a record at the time.

Six years later, India became the fourth country in the world to complete a lunar landing when Chandrayaan-3 landed near the moon’s south pole.

The Chinese leader called for more international cooperation in developing the technology at a conference in Shanghai.

Artificial intelligence should not be dominated by one country, Chinese President Xi Jinping has said, urging international cooperation on development at a major conference in Shanghai.

Xi also emphasised the importance of a “people-centred” approach to AI technology in his keynote address at the opening ceremony of the World Artificial Intelligence Conference on Friday.

The conference showcases the cutting-edge technology Xi hopes will soon rival that of the United States.

Chinese AI models are gaining ground on the most powerful offerings from the US, attracting global users with lower costs.

But how to govern the booming sector has become a topic of debate amid concerns over the deployment of AI in military combat and its use by hackers or criminals.

In his address, Xi spoke of China’s role in ensuring equitable access to AI capacity-building for developing countries to prevent the creation of “new historical injustices”.

To that end, he announced China’s plans to cooperate with international bodies, including from Africa, Latin America, Asia and BRICS countries, to provide AI-related opportunities.

“AI development should not be a solo performance by a single country, but a symphony of international cooperation,” Xi said. “We should jointly oppose overstretching the national security concept in the field of AI or placing one country’s security over that of others.”

‘Ensure AI is always under human control’

The US and European Union have imposed restrictions on Chinese tech imports, citing national security concerns, while recent tussles between Washington and American AI labs have raised questions about who controls access to top technology.

In May, the US Commerce Department issued a notice affirming its restrictions on shipments of semiconductors to subsidiaries of Chinese companies located outside China amid concerns about loopholes in Washington’s export control regime.

The guidance said its licensing requirements for the export of advanced AI chips applied to all businesses with headquarters or a parent company in China.

At Friday’s conference, Xi also stressed the need for a “people-centred” approach to AI with humans at the wheel.

“We should put in place laws and regulations, technological monitoring, early warning, and emergency response systems, in order to … ensure AI is always under human control,” he said.

AI has become a strategic pillar of China’s industrial policy, driven by state investment aimed at building a domestic ecosystem, from chip production to consumer use.

Daily consumption in China of “tokens” – the industry unit of AI usage – has increased a thousandfold over the past two years, according to state media citing officials.

As Al Jazeera reported earlier, China, while lagging behind the US in access to the most cutting-edge semiconductors, holds the edge in powering the huge data centres that run on AI chips.

A typical data centre can consume as much electricity as 100,000 households, while next-generation “hyperscale” facilities can gobble up as much power as two million homes, according to the International Energy Agency (IEA).

China’s access to an abundant supply of cheap electricity places it in the ideal position to meet such colossal energy demands.

It already generates more than twice as much electricity as the US, a lead that is expected to widen amid an aggressive state-led investment in the country’s energy grid.

Nothing much stays with me from the first days of Twitter, which was publicly launched 20 years ago, on July 15, 2006.

I had discovered the internet back in 1995 and early on, I started thinking about how to get my voice heard by the world. I created a couple of websites through Angelfire and 8m, but there was no real ecosystem to nurture the idea. It’s like opening a shop to sell a certain product in a remote area – somewhere nobody really knows, at a time when there’s no interest – compared with opening that same shop in a mall, or on a street full of other vendors.

MySpace was another opening, but the idea was not yet ripe. Facebook came with a spark – and then we got Twitter.

“It’s like having your own breaking news platform, you’ll set your own agenda,” I remember one of my colleagues at the BBC, where I used to work, saying at the time.

It didn’t take me long to sign up. I cannot recall whether I tweeted immediately or not, yet what happened afterwards helped frame my future as an international journalist.

Twitter’s first defining moment for me was 2009’s Green Revolution in Iran, when I and others followed how the platform shaped the discourse in a way that differed completely from traditional media. We were not new to citizen journalism; a few years earlier, Salam Pax emerged as the first ever famous war blogger, presenting his distinctive view of the US-led invasion of Iraq through his individual blog. A few years later, tens of thousands of Salams have appeared – and I’m one of them.

Going through my early timeline, I see that I was tweeting randomly – an earthquake in Japan, an election in Lebanon, an explosion in Somalia, and so on. Then came the Arab Spring. Just as with many in the world, this was the moment that shaped my Twitter presence, and as I got involved in the coverage, I became well-positioned to post and attract followers.

My coverage of the Libyan revolution in March 2011 introduced me to many people and gave me a better understanding of what was happening. I was based in Sallum, a village on the Egyptian side of the Libyan border, without a connection of my own. I fed a colleague back in Cairo a sentence at a time over a crackling Thuraya satellite phone, and he typed my words into the account that I could not reach. Its password lived on my friend’s head until days later, when I finally got my hands on a satellite dish.

Trips to Libya, Egypt, Syria, Somalia – all of it made Twitter part and parcel of my journalistic journey, and it also helped me build a parallel path writing for international outlets including Al-Monitor and The Sunday Times.

Yet still, there was something else that changed my direction. Until 2013, I was a journalist covering stories without specialisation – I used to report from Iran, like I do today, yet it was not my career the way it currently is. But then I became a bureau chief in Tehran and my knowledge began growing – and here, Twitter gave me another layer, widening my network day after day.

Personally, that specialisation gave the platform its finest hour for me. I broke developments out of Iran’s nuclear talks with world powers before the news agencies had finished their first draft, filing in Arabic and English within minutes of each other and announcing the agreement itself while other newsrooms were still working on their bulletins.

The war against ISIL (ISIS) followed, then a January 2020 morning near Baghdad airport when my sources told me the commander of the Islamic Revolutionary Guard Corps’s Quds Force, Qassem Soleimani, and the deputy chief of Iraq’s Popular Mobilisation Forces, Abu Mahdi al-Muhandis, were in a convoy hit by a US air strike – and I was among the first to say so.

Twitter was never only a wire service for other people’s wars. I’ve “met” heads of state and celebrities on this platform – and for a moment we felt equals. I have made my scoops there, and I have made my hugest gaffes there, too. You act and you interact and you see the result immediately, backlash or praise. It’s like a daily journal, one that outlives you. I know of many, some friends, some colleagues, some people I only happened to follow, who left our world while their accounts are still there – for us, and for me – to return to for the memory or to get a piece of information.

It was also where, on the 100th anniversary of World War I, that I told the story of my great-grandfather, Ali Hashem, who went to the war and never returned; and of my grandfather Hussein, who was three when his father was summoned to the Ottoman army and never saw him again.

It was where colleagues at Al Jazeera, stationed in the north of Palestine, went looking for my family’s village on my behalf, for a cemetery nearly in ruins, for a great-grandmother’s grave that has never been found.

It became, eventually, the subject of my own academic work too, a master’s thesis on Twiplomacy, examining how a platform built for gossip and jokes quietly rewired the choreography of nations, with Iran’s nuclear diplomacy as my case study.

In the summer of 2023 – sensing where things were headed, as new owner Elon Musk decided to change Twitter’s name to X, and to tragically, if I may so, kill the famous and lovely blue bird that accompanied the journey many made with the platform, including myself – I posted five words.

“Someone buy Twitter and save the bird.” Alas, nobody did, and the bird disappeared from the icon, and the name went with it, replaced by a single letter that still sits wrong in my mouth. In Arabic or in English, the word that comes out of me, though, is still Twitter.

A three-armed spacecraft rockets into orbit to rescue a NASA telescope that’s in danger of crashing back to Earth.

NASA has launched a robotic mission to try to prevent one of its ageing telescopes from burning up in the atmosphere in a complicated operation expected to last several months.

Northrop Grumman launched the Link spacecraft – built by United States-based Katalyst Space Technologies – from the Marshall Islands in the Pacific Ocean on Friday.

A Pegasus rocket blasted off from the belly of a modified aircraft putting Link on course to reach and capture NASA’s Swift Observatory in about a month.

Initially scheduled for Tuesday, the robot’s launch was postponed due to weather, then technical issues. Blast-off happened on Friday at 0836 GMT from an atoll in the Pacific Ocean.

The unprecedented $30m effort involves sending a robot to rescue the Swift space telescope that is falling towards Earth. If successful, the mission could pave the way for giving other satellites a second life.

Launched in 2004, Swift is sinking faster than ever because of recent solar storms. The $250m telescope studies gamma-ray bursts, the most powerful explosions in the universe.

Once it reaches an orbit close to Swift’s, the robot will deploy its solar panels and perform a series of checks.

It will then have to locate the Swift telescope in the vastness of space, circle around it and dock using three robotic arms – manoeuvres expected to take several weeks.

Finally, it will attempt to propel the satellite approximately 300km (186 miles) higher above the Earth, roughly to its initial orbital position. That operation is expected to last at least a month.

“This is a lot of firsts stacked on top of each other,” Shawn Domagal-Goldman, director of NASA’s astrophysics division, told reporters on Tuesday. “I’m just deeply thankful that we’re even giving this a go.”

Three Brazilians with a combined age of 316 have been recognised as the world’s longest-living trio of sisters. Researchers are studying their DNA to better understand the genetic factors behind healthy aging and exceptional genetic longevity.