• Vodafone uncovers a “nation full of drive” with 85% of Brits building towards a major ambition
• Over half of those surveyed believe it is easier to find opportunities in the digital era
• Vodafone launched the UK’s first Opportunity Hub in London, intended to provide advice and connections to ambitious visitors

Vodafone has released the results of research into how being connected has opened the door to people being able to pursue their goals.

With 40% of Brits feeling “more ambitious than ever” and a majority aiming to achieve major targets in personal and professional development, the research is part of VodafoneThree’s investment into building the UK’s best network.

Interestingly, the findings also highlight that key career events take place in our 20s and 30s, and accompany the successful Opportunity Hub event targeted at this age group.

Connecting unlocks opportunity

Vodafone’s survey found over half (55%) of Brits feel it is easier to find opportunities in the digital era, with 20% wishing they had started working toward their ambitions sooner. 42% believe greater connectivity has had a positive effect on reaching their goals, with 49% now “better placed” to achieve than they were 5 or 10 years ago.

The survey also uncovered the blocks to personal and professional ambitions. Unsurprisingly, financial constraints are the main barrier, with 22% admitting this. But lack of confidence at 13%, fear of failure (12%) and limited access to the tools needed for self-improvement (9%) are also mentioned as challenges.

However, life’s disruptions can be seen as launch points, with bereavement (27%) and redundancy (25%) catalysts for chasing ambitions, alongside parenthood (25%).

“Britain isn't short of ambition," a VodafoneThree spokesperson said, "the challenge is turning ambition into action. Our research shows that connectivity is increasingly part of that equation, with nearly two-thirds of would-be founders saying unreliable connectivity has held them back from starting a business locally.”

• Amazon revealed plans to build Gilroy, California data center were first submitted in 2020
• Gilroy citizens have accused the city council of inadequate communications concerning the development
• The 438,500-square-foot data center occupies a 56-acre chunk of farmland 30 miles south of San Jose

Citizens of Gilroy, California, have expressed dismay at the building of an Amazon data center, which they say they were not informed of - and when one resident attempted to lodge a complaint, she was told that comments on the plans were closed in 2024 – long before most citizens were even aware of the planned installation.

Occupying a 56-acre area of farmland between a Walmart and a local shopping center, the Amazon project appears to have been quietly pushed through using obfuscation and clever use of 45-year-old planning laws.

Amazon has apparently made provision to divert wastewater to the data center, and donated money for local spending. Meanwhile, the city mayor has responded to the concerns by likening the water-and power-hungry Amazon data center to a food distribution center that was also recently approved.

Opposition and sponsorships

Utilizing the old city zoning rule – originally intended for industrial development along a freeway – has enabled the data center to proceed into full development without planning scrutiny from residents. However, the apparent collaboration between Amazon and the Gilroy city council and mayor has not gone unnoticed, with opposition organized against the project.

One of those organizers, 21-year-old college student Landon Sepulveda, indicated to the Wall Street Journal that there was a trust issue between the population and the city administrators. “These processes were not enough for the citizens of Gilroy to do anything about it. We’ve elected them to tell us, and we shouldn’t have to watch over their shoulders.”

Another citizen, elementary school teacher Coleen Crew, told a recent council meeting “Many if not most residents first learned about this project after construction had already begun.”

Amazon and the Gilroy council have been at pains to highlight the advantages of the deal, citing jobs, tax revenue, and that solution to concerns over water use. Amazon has also donated $1 million for a new city fire truck, and donated money to non-profits and the city’s annal Gilroy Garlic Festival.

Mayor Bozzo and AWS

Responding to public dismay over the development, Mayor Greg Bozzo has highlighted an “opportunity to change the way that we do business in Gilroy. We’re adapting to the times.”

Meanwhile, Amazon Web Services’ VP of economic development Roger Wehner claims the project “followed standard approval processes.” He added: “We’ve since met directly with residents and hosted an open house to hear feedback and answer questions.”

The challenges and advantages of data centers seem to be the modern day equivalent of opposition to nuclear plants from 50 years ago. While the data center’s continued construction seems inevitable, and Amazon has made some generous provisions, it is disappointing that the city’s residents were not informed of the development – and that might be the biggest problem here.

• Virginia now makes data centers pay for their own power lines
• Virginia hosts 570 data centers, more than any other US state
• Spanberger says the policy could save residents hundreds of millions

The US state of Virginia is requiring data centers to cover electricity infrastructure costs created specifically by their rapidly growing power demands.

The move comes as electricity prices across the state have risen sharply alongside the expansion of artificial intelligence facilities.

State officials say the new requirement could prevent ordinary electricity customers from carrying costs associated with large data center developments.

Data centers face higher infrastructure costs

Virginia has become the largest data center hub in the United States, with at least 570 facilities operating across the state.

Their growing electricity consumption has required major investments in transmission infrastructure, increasing pressure on the regional power system.

The State Corporation Commission, Virginia’s public utility regulator, has ordered data centers to pay for transmission infrastructure used exclusively by their facilities.

The requirement means utilities can no longer spread those dedicated infrastructure costs across their wider customer base.

Electricity prices in the region have already increased substantially, with PJM Interconnection raising prices by 76% during recent capacity market changes.

Monitoring Analytics, an independent market watchdog, has attributed much of that increase to growing demand from data centers.

The issue has become politically important as artificial intelligence companies continue expanding computing capacity across Virginia.

Some local authorities have even asked employees and institutions, including schools, to reduce electricity consumption during periods of high demand.

"We are taking real steps to address rising energy costs for Virginians," said the Governor of Virginia, Abigail Spanberger.

"I will continue to work with the General Assembly and take action to make sure data centers pay their fair share, adhere to strict environmental standards, and listen to the concerns of local communities.”

Washington’s promise reaches the states

The Virginia decision follows President Donald Trump’s effort to make major AI companies finance infrastructure required by their expanding operations.

Trump met leading AI companies at the White House in March and urged them to “pay their own way” for necessary infrastructure.

The administration expanded that approach in July by calling on governors, utilities, and data center developers to follow similar principles.

Critics have questioned whether voluntary commitments can significantly change electricity costs without enforceable state or federal requirements.

Virginia is currently among the few states taking concrete action against rising electricity costs linked to large electricity users.

Oregon has also introduced higher electricity rates for data centers consuming at least 20 MW, under provisions contained within its POWER Act.

Portland General Electric subsequently increased rates for those large consumers by 30%, while residential electricity costs fell by approximately 1.3%.

The Virginia approach differs by focusing specifically on infrastructure dedicated to individual data center projects.

The policy could become increasingly important as AI companies expand AI tools requiring increasingly powerful computing facilities and greater electricity supplies.

If other states adopt similar requirements, data centers could face more direct responsibility for infrastructure costs they generate.

For Virginia households, however, the immediate question is whether the policy actually prevents future electricity increases rather than simply changing who pays existing infrastructure bills.

Spanberger said the state expects the measure to save Virginians “hundreds of millions of dollars” while requiring data centers to pay their fair share.

Via Tom's Hardware

• Firestorm Labs printed 12 drones aboard USS Essex during a Pacific voyage
• More than 1,000 drone components were manufactured directly aboard the warship
• The Navy tested drone production while waves reached 12 feet

Firestorm Labs has produced 3D-printed drones aboard a US Navy ship at sea, testing manufacturing without conventional shore-based logistics support systems.

The company manufactured more than 1,000 components and assembled 12 Squall first-person-view drones aboard USS Essex during its voyage toward Hawaii waters.

The drones were subsequently flown as opposing aircraft during a counter-drone exercise conducted during the Rim of the Pacific, or RIMPAC, 2026.

Manufacturing drones at sea

Firestorm Labs used its xCell containerized microfactory to manufacture drone components and additional parts while the amphibious assault ship remained underway.

The system functions as a compact production facility designed for Navy vessels, allowing crews to manufacture unmanned aircraft and selected replacement components.

During the trial, service members assembled drones while waves reached 12 feet, providing a demanding environment for onboard manufacturing operations.

Personnel also printed mechanical test components used to assess the effectiveness and performance of the equipment during production aboard the vessel.

The same equipment was used to fabricate military repair parts required by USS Essex personnel during the ship’s Pacific deployment.

“Every part xCell printed on deck is one that doesn’t need to be flown or shipped across contested waters — cutting the fuel, aircraft hours, and personnel it takes to keep a ship operational,” Firestorm Labs said via LinkedIn.

“Repairs that once meant days or weeks waiting on a resupply run can now happen in hours, on station.”

Under the company’s description, repairs requiring conventional supply missions could instead be completed within hours when suitable production capability remains available.

The trial also gives Navy personnel experience operating a manufacturing system while managing shipboard conditions, equipment requirements, and production tasks simultaneously.

A wider move toward containerized military capabilities

The experiment comes as the Navy develops modular capabilities transported inside containers and deployed where conventional infrastructure remains limited during operations.

Chief of Naval Operations Adm. Daryl Caudle announced a containerized capability campaign plan in March during the McAleese Defense Programs conference.

His plan calls for transportable containers carrying capabilities such as drones and weapons to be deployed across different operational regions.

The Pentagon has separately pursued containerized munitions, with framework agreements announced in May involving Anduril, CoAspire, Leidos, and Zone 5.

Those agreements cover more than 10,000 containerized missiles scheduled for acquisition over three years beginning in 2027 under the program.

The Navy’s onboard drone production test combines this containerized approach with additive manufacturing conducted directly from an operating warship.

For crews operating far from established bases, such production could provide another method for obtaining selected drones and replacement components when needed.

However, the trial does not establish how reliably the system can sustain production during prolonged operations or under more demanding combat conditions.

Its practical value will depend on whether the equipment can consistently produce dependable aircraft and parts at useful rates during extended deployments.

Via Defense News

The programmer Richard Stallman is not a household name, but his work has been instrumental in building the software industry, as has his long-term campaign for free software. He's also been a huge advocate for privacy in the digital age, and has railed against the rise of cell phones for that reason.

Who needs phones anyway?

Stallman first disclosed his views on cell phones in an interview with Network World, during a time in which smartphones were exploding in popularity.

During this interview, Stallman indicated his long-held belief that the portable phones that many millions use would be the perfect tool that authoritarian forces could exploit and use to track the movements of populations.

He also advocated for free software, which you would expect from the founder of the Free Software Foundation (FSF), which he established in 1985. This was backed by the creation of the GNU project – a free software, mass collaboration movement to give users freedom of choice to use and develop software for their devices.

The legacy of free software

Despite his reluctance to ever use a cell phone, one of Stallman's achievements – which he himself acknowledged in the interview – was the third-party version of the Android mobile OS, from which all proprietary software was stripped out.

He pointed to new systems like Replicant, an alternative version of Android, that can run on certain devices without additional proprietary software. The catch is that this only works with older and outdated handsets, like the Samsung Galaxy S3 or the Galaxy Note 2.

• Washington wants new rules for civilian nuclear reactors operating at sea
• IAEA will launch ATLAS as America pushes nuclear power offshore
• Small modular reactors could power ships without frequent nuclear refueling

The United States government is backing an international effort to clear regulatory barriers around civilian nuclear reactors operating on ships and floating platforms.

The initiative could eventually enable floating power plants serving coastal communities and industries requiring substantial electricity or desalination capacity.

The International Atomic Energy Agency (IAEA) will launch its Atomic Technologies Licensed for Applications at Sea (ATLAS) initiative in Washington on August 26 and 27, aiming to clear the path for reactors at sea.

US pushes for nuclear power at sea

"The global maritime sector is at a critical turning point, facing urgent pressure to sustain long-distance, high-speed operations while ensuring reliability and energy security," said Rafael Mariano Grossi, Director General of the IAEA.

"Small modular reactors offer a safe and viable option for maritime transport and offshore energy systems, delivering high energy density and long operating cycles that eliminate the need for frequent refueling."

Ahead of the official ATLAS launch, the United States will hold an Industry Day event on August 25 to showcase relevant technology.

American nuclear and maritime companies are expected to attend, each hoping to secure a share of any emerging global market.

US Secretary of Energy Chris Wright tied the event to the Department of Energy's wider push for more atomic power.

"DoE remains focused on unleashing American energy dominance, accelerating innovation, and advancing sources of energy that are affordable, reliable, and secure for the American people," Wright said.

Floating reactors carry old and new risks

The Department of Energy is counting on small modular reactors, whose high energy density and long operating cycles could suit ships and offshore installations.

Nuclear-powered vessels are not new to America, since every active US Navy aircraft carrier already runs on atomic propulsion today.

The Navy once operated nuclear-powered cruisers as well, including the Virginia class, before retiring them for being too costly to run.

Civilian nuclear ships also have precedent, since the NS Savannah became the first nuclear-powered merchant vessel when it launched in 1959.

That vessel proved too expensive to operate, a challenge small modular reactors (SMR) advocates now hope newer reactor designs can finally overcome.

Russia currently operates the world's only floating nuclear power plant, the Akademik Lomonosov, which first launched back in August 2019.

Its two KLT-40S reactors can generate a combined 70 MW of electricity, enough to serve remote coastal communities and light industry.

Placing a reactor on water changes the associated risks rather than eliminating them, since remote platforms may lack backup power.

The surrounding sea can serve as a vast heat sink for cooling, though emergency access could prove more difficult offshore.

Any accident could also risk radioactive contamination of nearby marine environments, endangering the very coastal community the reactor intended to serve.

The push toward floating and shipborne reactors suggests nuclear power is being reconsidered as a serious option for maritime and offshore energy needs.

Whether SMR technology can finally make such projects economical, after decades of costly false starts, remains to be seen.

Via The Register

• SpaceX and Tesla commit $16.8 billion to build a gigantic Texas chip factory
• TeraFab could cover more than 100 million square feet when fully built
• Elon Musk claims TeraFab will reach 50 times the size of the Pentagon

SpaceX and Tesla have revealed more details in their plan to invest $16.8 billion in a massive semiconductor factory planned for Grimes County, Texas.

The facility, known as TeraFab, is expected to manufacture chips for several computing-intensive applications associated with both companies.

At full construction, SpaceX says the factory could exceed 100 million square feet, or approximately 9.3 million square metres.

A semiconductor factory on an unprecedented scale

SpaceX CEO Elon Musk described the planned facility as potentially the largest and most valuable building on Earth after completion.

“TeraFab will be 50 times the size of the Pentagon when complete,” Musk said on X.

Since the Pentagon is 6.6 million square feet and SpaceX has only confirmed TeraFab will exceed 100 million square feet, Musk's "50 times" figure would require a final size closer to 330 million square feet — a number the company has not disclosed or confirmed.

Even so, the scale remains extraordinary: the Pentagon, Apple Park, Mall of America, and Giga Texas together total 25.04 million square feet, meaning TeraFab's confirmed 100 million square feet alone would already be roughly four times larger than all four combined.

The project originally emerged from Musk’s concern that existing semiconductor suppliers could not satisfy Tesla’s future computing requirements.

Tesla first discussed the concept publicly in November 2025, describing it as a much larger version of its Gigafactory model.

Plans announced in March 2026 called for the semiconductor plant to operate beside Tesla’s Austin headquarters.

SpaceX and Tesla subsequently established the project as a joint venture, with Intel later joining the development effort.

Musk has said the facility would use Intel’s 14A manufacturing process, although production plans may evolve during development.

The companies have linked the factory to chips required for space-based computing, advanced vehicles, and humanoid robotic systems.

Costs, employment and environmental commitments

The initial project estimate was $20 billion, but later filings indicated substantially larger potential spending during subsequent development phases.

A May property tax abatement application revealed that an initial phase could cost $55 billion, with eventual spending potentially reaching $119 billion.

The latest $16.8 billion commitment, therefore, does not represent the maximum investment that may eventually be required.

SpaceX expects the facility to employ at least 3,000 workers, including people from Grimes County and neighboring Brazos County.

Local authorities approved a 100% property tax abatement in June 2026 after county commissioners voted 4-1 on three related proposals.

The companies also say industrial operations will use water from Gibbons Creek Reservoir instead of relying on nearby groundwater supplies.

SpaceX and Tesla have also committed to wastewater treatment, water reuse, conservation measures, and compliance with environmental regulations.

Their plans include regulated handling of hazardous materials, chemicals, and industrial byproducts generated during semiconductor manufacturing.

“The combined SpaceX and Tesla demand for chips is expected to be in excess of 1 terawatt (TW) of compute,” said SpaceX in its website post.

The proposed scale makes TeraFab an unusually large industrial project, while its final cost and construction scope remain subject to future expansion.

Via DataCenterDynamics

• TSMC reportedly holding roughly $1 billion of finished A20 Pro wafers that cannot be packaged because the DRAM they must be integrated with has not yet arrived
• The bottleneck exists because Apple moved the A20 Pro to TSMC's WMCM packaging, which bonds processor and memory at the wafer level, removing the option of stockpiling finished dies and adding memory later
• The launch is not expected to slip but could mean that post-launch inventory could be dicey, with thin stock and extended delivery estimates the likely outcome for Apple's upcoming devices

Around a billion dollars of finished Apple silicon is reportedly sitting at TSMC with nowhere to go, reports have claimed.

The wafers are processed, the yields are fine, and the chips cannot move to the next stage of production because the memory they need to be packaged with has not arrived.

The reason they cant be stockpiled for later use is a design choice Apple made with TSMC to use its Wafer-Level Multi-Chip Module (WMCM) process instead, an extension of the CoWoS packaging family adapted for mobile parts.

Trouble in smartphone chip paradise?

The reported delay was revealed by Tim Culpan, a Taipei-based semiconductor analyst, who says production of the A20 Pro on TSMC's N2 node is progressing well, with good volume and yields, but wafers are piling up waiting for DRAM before they can be packaged and sent onwards.

These chips will almost certainly power the next generation of Apple devices (probably the iPhone 18 Pro) but there might be a slight twist in the making due to design decisions made by Apple.

Memory shortages delay products all the time, but this one is worth attention because of what Apple changed about how it builds its processors.

For roughly a decade, TSMC packaged Apple's iPhone processors using Integrated Fan-Out, or InFO. Under that approach, the logic die containing the CPU, GPU, and Neural Engine is first fabricated and diced from the wafer, and memory is attached afterward, stacked on top of the finished die in a later assembly step.

The sequencing matters: a memory shortage under InFO delays final assembly but does not stop anything upstream. You keep fabricating logic dies, stockpile them, and attach DRAM when supply arrives.

The A20 Pro, as mentioned previously, uses TSMC's WMCM process instead, an extension of the CoWoS packaging family adapted for mobile parts. WMCM integrates the processor and DRAM at the wafer level, side by side, before either can proceed.

That reverses the sequence. TSMC needs the memory in hand to complete the package, and it cannot be added at a later step, which means Apple cannot bank finished dies against a memory delivery that has not happened yet.

There is a certain amount of irony here: Apple reportedly moved to WMCM specifically to fit more DRAM at lower latency for on-device AI features, since placing memory on the wafer improves both, but an AI-related memory shortage has stalled its progress.

The memory in question is LPDDR5X, and it is scarce because AI infrastructure buildout has absorbed the manufacturing capacity that would otherwise serve smartphones. Apple's on-device AI ambitions are being throttled by the datacenter version of the same ambition.

Culpan reports that Apple and its assemblers remain confident they can meet initial demand and that launch-day orders will be fulfilled without delaying the launch, though we might see thinner supply as a result of the delay.

Apple's exposure runs deeper on the other side of launch day: main logic board assembly and final assembly are largely handled by Foxconn and BYD Electronics, and late deliveries of packaged processors compress their schedules, even if Apple ultimately secures the memory.

The likely outcome is thin retail inventory, a fast preorder sellout, and delivery estimates that stretch out for weeks rather than a phone that does not exist. Apple reportedly plans to ship around 200 million units of the upcoming iPhone across the lineup over its lifespan, against an iPhone 17 range that surpassed 245 million shipments in 2025, so there isn't much room for error here.

Apple's usual LPDDR5X route runs through Micron, with SK Hynix and Samsung alongside, and an attempt to widen it fell over: a reported approach to Chinese memory maker CXMT, apparently aimed at cheaper pricing, did not produce a deal. Renegotiations with existing suppliers reportedly have not landed either, which leaves Apple paying whatever the current market asks. Whether it secures enough supply to avoid the thin-inventory scenario remains to be seen.

• Cloudflare CFO Thomas Seifert predicts machine-generated traffic could outnumber humans by a 1,000 times in the next five years
• The answer to this, Seifert believes, is to ensure servers are efficient enough to handle demand
• Cloudflare previously predicted machines would outnumber humans in 2027, but this digital soothsaying was 12 months out

Machine-generated internet traffic is already outnumbering that by humans, and it shows no sign of declining, experts have warned.

Cloudflare Chief Financial Officer Thomas Seifert has predicted man-made activity will be dwarfed by bots and agentic applications, with such machine traffic could be 1,000 times more than human traffic within the next five years.

However, Seifert was at pains to mention he has form for getting these predictions wrong.

The “rounding error” online

However, right or wrong, human traffic is already outnumbered by bots online, which means measures have to be taken for the internet to remain useful to people, or else grind to a halt under the weight of automated traffic.

Cloudflare believes making efficiencies is the solution, but those optimizations may be something other companies need to focus on.

While we may feel that the internet exists for humans to communicate and shop or spend money on a service, the truth is that it is increasingly dominated by machine-generated traffic. Traditional bots have been replaced with agentic AI and other automated tools, and every request has an impact on servers (and the rest of the internet infrastructure).

Cloudflare’s role in providing cloud services that provide security, performance, and general infrastructure is key to overseeing the traffic, with many websites relying on its protective shield to improve load times and mitigate online threats.

Explaining the thinking behind the predicted increase, Seifert said on a recent Cloudflare earnings call: “To give you a sense of how this trend is playing out, and with the big caveat that I have called it wrong at every point along the way, if the current trends continue, we think in five years, non-human traffic will be as much as 1,000 times as much as human traffic.”

“In other words, humans will be a rounding error on the internet, not because human traffic goes down, but that’s just how fast we’re seeing non-human traffic grow.”

Efficiency questions

Just how can servers and infrastructure be more efficient? Will this efficiency focus on human or machine generated and agentic transactions?

Some solutions are intelligent traffic management, optimized protocols, hardware acceleration, and edge computing.

Discussions in Cloudflare’s earnings call pointed at where the greatest efficiencies may be made – in its competitors, where GPU utilization is “super low,” due to them “buying a server and then trying to sell it back, lease it back” and putting the onus on the customer to “get as much out of the server as possible.”

We've tested several generations of the Lenovo Legion 5i, and always found it to be a well-rounded creative, work, and gaming laptop. So, this deal on the Lenovo Legion 5i for $1499 (was $2200) at Walmart caught my eye.

The powerhouse combines an Intel Core i7-14700HX processor with Nvidia GeForce RTX 5060 graphics, 16GB of RAM, and a 1TB SSD. The 15.1-inch WQXGA OLED display pairs a 165Hz refresh rate with Lenovo's PureSight technology and anti-burn-in measures.

Legion Coldfront cooling helps manage temperatures under heavy loads, and you get TrueStrike keyboard and AI Engine+, plus three months of Microsoft 365 Personal.

Should you buy it?

✅ Buy the Lenovo Legion 5i if...

You want a creative laptop with enough power to handle most tasks you throw at it. The RTX 5060 and Core i7-14700HX make for a capable combination, while the 165Hz OLED panel provides the fast refresh rate gamers look for alongside the contrast and image quality associated with OLED.

❌ Skip the Lenovo Legion 5i if...

Portability and long unplugged runtimes often matter more than performance. Powerful HX-series processors, discrete graphics, and high-refresh OLED displays aren't designed primarily around maximizing battery life. You may also want more than 16GB of RAM if you're planning on running particularly demanding workloads.

Why we recommend it

The Legion 5i packs several features we'd want from a modern laptop into one machine. You're getting RTX 5060 graphics rather than an older-generation GPU, a powerful 20-core Core i7-14700HX processor, and a fast OLED display instead of the more ordinary IPS panels found on many notebooks.

The 1TB SSD also provides a reasonable starting amount of storage for all your main files and applications.

Price context & historical value

Deals on high-powered creative laptops with graphics cards can sometimes be a bit underwhelming, but this one

The current $1,499 price is particularly interesting because we saw this laptop discounted to $1,621 just last week. That's another $122 knocked off in a matter of days, and Walmart lists the standard price at $2,199.99.

Compared with that price, you're saving $700.99, or almost 32%. Anyone who considered buying it during last week's deal will obviously be getting a substantially better offer now.

The Catch: What to know before you buy

The 16GB of RAM is adequate for most basic tasks today, but it's relatively modest alongside the powerful CPU and GPU, particularly if you also plan to use the laptop for heavy creative work.

There's a student bundle, and this has eligibility requirements, requires post-purchase redemption, and includes subscriptions that automatically renew.

While it's nice to get three months of Microsoft 365 Personal with it for free, that time will fly by and you'll need to think about what you'll do when the trial expires.

Also consider

Memory prices are sky-rocketing, doubling and tripling in price. And while we can't stop the RAMpocalypse price hikes, we're keeping a close eye on the best DDR4 and DDR5 deals we can find from the likes of Corsair, G.Skill, Kingston, and more.

Upgrading your RAM is a clear-cut way to boost your PC’s performance, without the complexity of getting an advanced graphics card or switching its processor altogether.

We've tested and reviewed the best DDR5 RAM, but there's no doubt upgrading memory can be a daunting choice. So, which ones are indeed the best? I’ve taken the burden of testing many RAM kits so that you don’t have to, curating the ones that offered the best performance boosts and value for money.

Best RAM deals: Editor's picks

Best RAM deals: Quick links

• Amazon: Save on DDR4 and DDR5 memory from Corsair, G.Skill, and more
• Best Buy: Discount computer RAM
• Newegg: Regular RAM deals across DDR4 and DDR5

Best DDR5 RAM deals

Best DDR4 RAM deals

• Attackers stole IEH employee credentials via a fake Microsoft login page
• Inbox access exposed sensitive defense‑related communications and technical documentation
• Malicious mailbox rules were removed as IEH contained the unauthorized access

Criminals have broken into the email inboxes of IEH Corporation, a significant supplier for the US military and companies in the commercial aerospace and space industry.

In an 8-K report filed with the US Securities and Exchange Commission (SEC), IEH said that unidentified threat actors reached out to one of its employees, pretending to be a “prospective business contact”.

The atatckers shared a link to what appeared to be a Microsoft document, prompting the victim to log in. Obviously, the login page was bogus, and the login credentials were relayed to the attackers instead.

Malicious mailbox rules

“The threat actor gained access to mailbox contents, including email messages, attachments, customer communications, purchase orders, engineering-related documentation, and potentially export-controlled technical information,” the 8-K reads.

The culprits, however, were not named and no threat actors have yet claimed responsibility for the attack.

IEH said it found no evidence that data had been exfiltrated from the compromised inbox. However, its defenders did discover and remove “malicious mailbox rules”. Usually, crooks set up such rules to automatically forward incoming emails to an inbox under their control, while deleting traces of the activity. This would allow them to continue receiving sensitive emails even after the initial compromise was remediated.

The company also said it completed a full audit of the inbox and done “corrective actions to contain any impact of the unauthorized access.”

IEH Corporation produces “specialized products used in military satellites, missiles and fighter jets,” meaning the information found in the inbox could be quite valuable, especially for nation-states such as Russia, China, North Korea, or Iran.

IEH does not publicly name its clients but it does say that its defense applications include Apache AH-64, V-280 Valor and SH-60 Seahawk programs, as well as Patriot, THAAD, AMRAAM and APKWS missile programs. It reported a revenue of almost $30 million for the 2026 fiscal year.

Via The Record

You may have heard a lot about quantum in the past few months, from your newsfeed to your competitors’ strategies. Last year was proclaimed the “International Year of Quantum Science and Technology” by the United Nations to celebrate the 100th anniversary of quantum mechanics.

While quantum belonged to the writings of the brightest scientists a century ago, it now animates conversations in both academic and business circles.

But what should you do about it as a business leader?

Quantum technologies use the principles of quantum mechanics to unlock new possibilities and can be split into three distinct categories. Quantum computing uses quantum bits instead of classical 0 or 1 bits to perform calculations.

Quantum sensing uses the same underlying physics to take very precise measurements, like an extraordinarily precise compass. And quantum communication enables the ultra-secure exchange of sensitive information.

Each of these technologies is on a different maturity curve.

One theory, three subfields

Quantum computing gets the most attention for a reason: it represents the largest estimated market size; and will unlock a completely new set of possibilities. Unlike classical bits, quantum bits can exist in a superposition of 0 and 1 simultaneously, enabling the exploration of many possibilities at once.

This capability is especially valuable for optimization problems, for example, finding the optimal route between two points. However, quantum bits, or qubits, are fragile and error-prone, today’s machines are still what we call “Noisy Intermediate-Scale” systems, and hold limited practical usefulness.

The industry is developing Fault-Tolerant Quantum Computers (FTQC) by using error correction methods to unlock reliable calculations. Given that these systems are not commercially available, waiting might seem like the best way forward; but it is exactly the opposite.

Businesses must prepare now for the advent of FTQCs. Because quantum computers operate on fundamentally different principles from classical machines, algorithms must be tailored to them, often to the point of redesign.

Building these capabilities is critical for businesses to secure a competitive advantage, but it will take time, so companies must start now if they have not yet explored this field. While quantum computers mature, another branch is already delivering value: quantum sensing.

Quantum sensors are one of the most underrated segments of the industry. Far more mature than quantum computers, some technologies are already being deployed in real-world settings. Applications have been developed in many sectors such as non-invasive cardiac diagnosis, non-destructive testing in manufacturing or GPS-free navigation.

This subfield is dual-use: sensors can passively detect concealed objects or track ground vehicles, which is why this technology is considered a strategic asset rather than a purely commercial one in many countries.

Quantum communications use the principles of quantum mechanics to create secure data transmission networks in which leaks or espionage can be physically detected. The EuroQCI joint initiative between the European Commission and ESA aims to develop this infrastructure across the continent, while China has constructed a secure link spanning more than two thousand kilometers between Beijing and Shanghai.

Quantum as a strategic asset

The advent of FTQCs will not only unlock new opportunities for businesses but also pose significant security threats for our systems. Indeed, they are expected to break the most widely used cryptographic schemes, creating a cybersecurity nightmare situation.

Although quantum computers do not yet have this capability, highly sensitive information like health or financial data can be collected now and decrypted later. Post-quantum cryptography, which uses classical methods, has been developed to resist quantum attacks. A June 2026 US executive order set a 2031 deadline for high-value, high-impact systems to migrate to post-quantum cryptography, and France's cybersecurity agency will certify only quantum-safe products starting in 2027.

Governments have also begun treating quantum as a controlled and sovereign technology. The US restricted exports of quantum computers in late 2024, and the EU added quantum to its dual-use control list in 2025. This pattern mirrors what’s happening with AI, with the US government briefly imposing export controls on Anthropic’s Fable 5 model.

These regulatory efforts are happening while AI and quantum are already converging and accelerating each other’s development, despite being fundamentally different technologies. Artificial intelligence is now widely used in the computing community for programming, notably to accelerate the creation of quantum algorithms.

Quantum sensors already rely on to machine learning methods to differentiate target signatures against noise, notably enabling more sensitive detection. Quantum computing, in turn, could open new hardware possibilities for artificial intelligence. Recent research also shows that quantum algorithms can pre-select features to accelerate processes before handling them to AI models.

Start small, but start now

From cancer drug discovery and delivery scheduling optimization to GPS-free navigation and risk simulation, quantum technologies hold the promise of revolutionizing various use cases across industries. Investors spent more than $12B on quantum technology startups in 2025, and around 30 countries have developed specific policies or a national strategy.

No one expects business leaders to become quantum physicists; however, ignoring the quantum opportunities and threats could put your business at risk. Developing expertise, notably by hiring or upskilling a “Chief Quantum Officer” who will lead quantum strategy and prepare for the opportunities and risks ahead, will secure a competitive advantage in the long run.

In order to grasp this potential, start by identifying at least one business challenge you are currently facing and explore whether quantum can solve it.

We list the best business laptops.

This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.

The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit

The RAMpocalypse continues to make PC upgrades considerably more expensive than they used to be just a year ago, particularly if you're shopping for fast DDR5. SO, I've been keeping a close eye on any discounts like this one: the 32GB V-Color Manta XSky DDR5-6000 kit for $398 (was $430) at Newegg with the coupon code BTSF2927.

This 32GB kit consists of two 16GB modules using SK Hynix ICs, with DDR5-6000 speeds and tight CL30-38-38-96 timings at 1.35V. It's designed specifically around AMD EXPO, allowing compatible AMD systems to apply its advertised overclocked settings via the appropriate memory profile. You also get cool RGB lighting.

Should you buy it?

✅ Buy the V-Color Manta XSky 32GB if...

You're building or upgrading a compatible AMD desktop and specifically want fast DDR5-6000 with tight CL30 timings. The two-module 32GB configuration provides plenty of capacity for modern gaming and demanding everyday workloads, while AMD EXPO makes accessing the rated settings simpler than manually configuring memory timings.

❌ Skip the V-Color Manta XSky 32GB if...

You're running an older DDR4 platform or don't need premium memory speeds and timings. DDR5 isn't backward compatible with DDR4 motherboards, and even with the coupon this remains an expensive 32GB upgrade. Anyone primarily interested in adding capacity rather than maximizing memory performance should consider cheaper alternatives.

Why we recommend it

DDR5-6000 CL30 is an attractive combination for compatible AMD PCs, and the use of SK Hynix ICs will make it appealing for buyers focused on performance. You get 32GB across two modules, AMD EXPO support, and RGB lighting in a kit designed specifically for AMD systems rather than paying for features aimed at multiple platforms.

Price context & historical value

The headline $419.99 price only saves $10 from the listed $429.99 price, so the coupon is crucial. Applying code BTSF2927 removes another $22, taking the total saving to $32 and the final price to $397.99.

That's still expensive for 32GB of RAM by historical standards, but current DDR5 pricing means every meaningful reduction helps if you need memory now rather than waiting for the market to finally improve.

The Catch: What to know before you buy

This is an AMD EXPO-focused kit, so check your motherboard's memory compatibility before buying, particularly if you're considering it for an Intel system.

You'll also need to enable the appropriate profile in the BIOS to reach DDR5-6000; otherwise, the memory operates at its lower base speed.

Also consider

Governments are beginning to look beyond age verification itself and toward the tools people use to bypass it. In Australia, recently released Freedom of Information documents revealed the eSafety Commissioner wants technology companies to actively detect and block VPNs used to circumvent age checks. In the UK, Technology Secretary Liz Kendall has also suggested the government could revisit restrictions on VPNs.

Viewed individually, these proposals may seem limited. Taken together, they illustrate how VPNs are increasingly being treated as part of the enforcement problem, when in reality, they are part of the security infrastructure that businesses, journalists, remote employees and millions of ordinary users rely on every day. In the United States alone, half of the country’s remote workers use a VPN.

Age verification laws were introduced with the objective of making it harder for minors to access harmful content, but it produced an unintended consequence. Across multiple markets, growing numbers of users are turning to VPNs because they don't want to upload government IDs or facial scans to websites they neither trust nor expect to visit regularly.

Folding VPNs into the regulatory framework confuses the symptom with the problem. Banning or restricting VPNs rarely stops the underlying behavior. Instead, it pushes users toward less regulated workarounds while handing governments a new lever to tighten control over what citizens can access online. Having spent years building a privacy-first browser, I’ve learned that internet users adapt far faster than regulation.

You can't regulate intent

Every proposal to restrict VPN use runs into the same technical problem. A VPN connection doesn't explain why someone is using it. A journalist protecting sources, a remote employee on a hotel network, and someone dodging an age check look identical from the network's side, and enforcement can't tell them apart.

Any enforcement mechanism built to catch age-check bypass has to sit on top of that same traffic, and it has no way to separate a teenager evading harmful content from an accountant filing a report over a hotel connection.

Stigmatizing VPNs further won't change the underlying demand either. People don't stop wanting privacy because a tool becomes less socially acceptable. They find another way to get both, often through channels with far less transparency than the ones being restricted.

The precedent is important

The concern here is sharper because of where these laws tend to originate. Rules written in a few high-profile markets often become templates, and other governments treat them as proven models regardless of how well they work. If privacy tools must identify users to function, that expectation can eventually reach encrypted messaging, cloud storage, and other technologies businesses depend on daily.

A company that accepts identity checks on its VPN traffic today has little basis to object when the same logic gets applied to its encrypted messaging platform or its cloud storage provider tomorrow. The infrastructure being debated under the banner of child safety is the same infrastructure that keeps corporate communications and data secure.

Privacy tools are not the problem

An estimated 1.7 billion people worldwide use VPNs, a scale that should give policymakers pause before treating them primarily as a way to bypass age verification. For most users, VPNs are part of everyday internet security, protecting communications and safeguarding public Wi-Fi connections.

The spikes in VPN usage tell a consistent story. The largest surges follow political unrest, as seen during Myanmar's turmoil, or when a popular platform is suddenly blocked, as happened when Turkey restricted Wikipedia for several years.

Laws tied to age verification drive new sign-ups too, albeit on a smaller scale, mostly from people who don't want to upload an ID or a face scan to reach a site they'd rather not be seen visiting, adult sites being the clearest case.

Not all verification is equal

It's worth separating two different situations.

Verifying age on a social network, where someone already has an identity and a reason to be recognized, differs from verifying age on an adult site, where the point is precisely not to be identified.

The first one is reasonable. On the other hand, I don't think anyone should hand over an ID or a facial scan for the second, given the risk of that data leaking and being used for blackmail.

The technology already exists, with limits

Privacy-preserving age verification is real, and working versions already exist, though it's worth being precise about what they can and can't guarantee. Proving something with zero trace anywhere in the system is extremely hard to achieve, and trust must always sit somewhere.

A workable model separates the party confirming someone's age from the party they're visiting. A trusted intermediary checks eligibility once and issues a token in return. That token isn't a currency and isn't tied to a wallet or an account, it's simply cryptographic proof that can't be traced back to the person it was issued to. No one can identify who is presenting the token later.

It works like an ID card with the photo, name and number stripped out, leaving only a yes or no answer to one question. That token lives on the person's own device. The service confirming eligibility never holds it, the user does.

Privacy and verification can coexist

Protecting minors is a legitimate policy objective, and rejecting today's approach isn't the same as rejecting age verification altogether. The question is how to achieve it without asking millions of adults to surrender more personal information than necessary. Parents are best placed to decide what their children can access using tools that already exist. Public policy should support that role.

Governments should encourage age verification systems that collect the minimum information needed to confirm eligibility, while letting privacy tools do the job they were built for. The real work is building age verification people can trust without asking them to surrender their privacy. Regulating VPNs won't get us there.

We've featured the best secure VPN provider.

This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.

The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit

• Attackers stole extensive personal and medical data from 3.8 million individuals from Unlimited Technology Systems
• Breach source and methods remain unknown, with no group claiming responsibility
• Stolen data poses major fraud risks, prompting free identity monitoring from Kroll

US healthcare organization Unlimited Technology Systems has revealed it suffered a cyberattack in which it lost a treasure trove of sensitive information belonging to millions of people.

The company recently made the information public and shared details with the US Department of Health and Human Services, noting that someone broke in on October 5, 2025 and within five days exfiltrated valuable data belonging to just over 3.8 million people.

The attacker stole people’s full names, Social Security numbers (SSN), dates of birth, emails and mailing addresses, phone numbers, demographic information, scans of driver’s licenses and other government IDs, insurance cards, intake forms, health insurance policy numbers, claims and benefits information, medical record numbers, dates of service, and diagnosis data.

Supply chain woes

The company did not say who the threat actors were, or if they demanded any ransom in exchange for deleting the data.

No hackers have claimed responsibility just yet, and we also don’t know how they managed to break in, as different groups have different methods.

ShinyHunters, for example, prefer calling their victims on the phone, pretending to be IT support and convincing their victims to give them access via remote management tools. Other groups might try to exploit vulnerabilities in routers, firewalls, and other hardware.

Whoever it was, they have a valuable data set in their possession. This kind of information can be sold on the black market or used in identity theft and wire fraud. To mitigate these risks, Unlimited Technology Systems is offering free identity monitoring services to affected individuals through Kroll.

Unlimited Technology Systems is a software company that provides financial technology for healthcare organizations. It works with around 4,500 clinics and 6,500 specialty healthcare providers in the US, processing north of $70 billion in net healthcare charges every year.

According to BleepingComputer, the company processes information on behalf of its clients, which means that the victims of this attack have no direct relationship with Unlimited and have probably not even heard of it.

• Oversecured found 176 vulnerabilities across Samsung’s preinstalled mobile apps
• Flaws enabled account takeover, code execution, and traffic hijacking via bloatware
• Samsung patched all reported issues, affecting hundreds of millions of devices

Security researchers from Oversecured have given “bloatware” an entirely new meaning, revealing that they uncovered 176 vulnerabilities - including some rather worrying ones - in Samsung’s mobile apps.

For the last three years, the team analyzed Samsung’s preinstalled system applications and found vulnerabilities that could cause some serious harm. Some of the bugs granted camera and microphone access, while others allowed for remote Samsung Account takeover with nothing more than a single click.

Some flaws allowed for network traffic hijacking via DNS manipulation, and others granted arbitrary code execution via an image. In theory, a malicious actor could craft and send a JPEG image which, when the victim opens, copies and loads attacker-controlled native libraries from the SD card. Finally, Oversecured found path traversal vulnerabilities allowing writing arbitrary files to the file system without proper path validation.

Arbitrary code execution

The researchers disclosed their findings to Samsung which, according to their report, fixed all of the reported issues - the full list can be found on GitHub.

Most Android smartphone manufacturers preload their devices with proprietary apps - think Bixby, Samsung Free, or AR Zone. These apps - which cannot be uninstalled or removed from the devices - aren’t necessary to their operations and are often not wanted by the users in the first place.

This 'bloatware' is also one of the key selling propositions of Google Pixel devices, since these are considered “stock Android”, or bloatware-free.

Out of context, these bugs are nothing extraordinary. Single-click account takeover flaws and traffic hijacking bugs pop up every now and then and get fixed rather quickly. The context here is that these are Samsung’s proprietary apps that don’t fall under the protection of Google’s Play Protect. Users might think they’re safe because they’ve not downloaded apps from risky places, or enabled dangerous permissions, when in reality, they’re not safe at all:

“Preinstalled system applications run with extra privileges than normal apps, cannot be removed by users, and operate outside Google Play Protect,” the researchers warned. “A single vulnerability affects hundreds of millions of devices globally through one vendor's distribution channel.”

The construction industry continues to suffer from a productivity gap.

It lags behind other sectors with a mere 0.4% annual improvement in productivity since 2000.

While the industry remains poised for growth, it needs to deliver on the existing project pipeline, and ultimately find a way to close this gap.

So what is holding construction back?

From material delays to talent gaps, there is no one answer to this complex challenge.

Many contractors have turned to technology to support different workflows across the jobsite.

One solution that’s making an impact from an overall productivity perspective is jobsite intelligence.

Jobsite intelligence solutions

Jobsite intelligence solutions help site leaders enhance productivity through the terabytes of visual data generated on jobsites. AI is now advancing the capability of these solutions, generating the insights that help site leaders proactively manage their jobs.

How does AI impact visual data in construction and improve productivity?

Jobsite intelligence solutions leverage cameras on the site to capture visual data and support perimeter security for insurance and compliance, progress monitoring, and safety.

While these solutions expand the amount of visual data available to site leaders, strained project teams do not have the time or resources to analyze it. This is where AI becomes the catalyst for scaling jobsite intelligence and improving overall site productivity. AI is able to quickly interpret visual data and provide insights to teams on the jobsite in real-time.

For example, a morning brief uses AI to interpret a single photo first thing in the morning, summarizing the jobsite’s readiness regarding weather conditions, site hygiene, and dumpster availability. Similarly, AI-powered search capabilities simplify reporting, removing the need to comb through hours of security footage to find a specific infraction.

AI-powered insights: The key to proactively managing a project

The more a project team can plan, the more productive it can be. AI-powered insights from jobsite intelligence solutions give site leaders the visibility they need to embrace a proactive approach:

1.Make informed decisions with speed

Running any jobsite is challenging. Running multiple jobsites even more so.

AI supports faster, more informed decision making, pinpointing or flagging the information project teams need, when they need it. This could include an alert when a worker isn’t wearing PPE to prevent costly injury, an update on the status of foundation work, or providing a 24-hour security report so contractors know a site is safe for work.

AI powered jobsite intelligence can surface the data that matters most, so site leaders avoid spending hours sifting through information or chasing updates via phone or email. They can wake up to an email recapping the status of the site, including hygiene and ground conditions, before their day even starts. The result is quick, simple, and actionable decisions.

Over 60 percent of construction firms experience delayed, scaled back, or canceled projects due to economic uncertainty, labor shortages, and supply chain disruption. As such, it’s increasingly important for contractors to have visibility into all areas of the project for more informed, strategic decision-making.

2.Maintain productivity amid labor shortages

The labor challenges facing construction weigh heavily on its productivity. 83% of construction companies report they have trouble hiring superintendents while 81% are struggling to hire project managers and supervisors.

This is significant – project managers and superintendents keep projects moving. Amid this shortage, site leaders are spread across more projects, making it difficult to have consistent visibility into each one simultaneously. By providing project leaders updates and analysis across jobsites, AI-powered jobsite intelligence gives them time back in the day.

A progress report summary can do the job of three phone calls and a 45-minute drive to verify the status of a cement pour, for example. If an issue on one site leaves them tied up, they still have visibility into their other projects. Time saved leads to improved impact, higher job satisfaction and ultimately easier recruiting of new talent.

3.Avoid losing knowledge from workplace changes

As a large portion of the construction workforce nears retirement, the industry runs the risk of losing intellectual property. By 2031, 41% of construction workers are expected to retire, while only 10% of current workers are under 25.

If experienced workers retire before new ones enter, new employees miss out on the chance to gain anecdotal knowledge from those industry veterans. The next best thing becomes learning from actual jobsite footage and data.

Project teams can use AI-powered jobsite intelligence to catalog visual data, capture the insights from veterans, and bundle as training materials.

This makes it possible to preserve decades of IP and pass it on to future generations of the workforce.

AI makes visual data intelligent, giving contractors a more productive way to work

AI plays a central role in advancing jobsite intelligence so project teams can unlock the insights of visual data.

Teams will gain the most value from a solution that is purpose-built for construction and provides the real-time insights that keep sites safe and secure. This will ultimately drive overall productivity and on-time project delivery.

We list the best job site.

This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.

The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit

• Attackers used social engineering to access Levi’s network and steal corporate data
• Details on stolen information, methods, and perpetrators remain largely undisclosed
• Voice‑phishing extortion groups are suspected, though no one has claimed responsibility

Levi Strauss has revealed it recently suffered a cyberattack and lost corporate files - however some crucial details around the incident are missing.

The company filed a new report with the US Securities and Exchange Commission (SEC), noting how hackers accessed its infrastructure through “social engineering” against three of its employees. We don’t know if that was via email, voice phishing, or some other technique.

After breaching the network, the crooks - who weren’t identified - “accessed and exfiltrated certain corporate information”. Again, we don’t know which information was accessed, or how much of it.

Was it UNC6671?

In response, Levi’s said it had “initiated response protocols, implemented containment measures, and launched an investigation” which remains ongoing. Again, we don’t know what these measures are, or how the crooks were ousted.

The company says the incident did not disrupt its business operations, or caused interruptions, in any way, and that it does not expect it to have any material impact whatsoever.

While Levi’s did not name the perpetrators, and while none have yet claimed responsibility on the dark web, some publications have hinted at UNC6671, a “financially motivated threat cluster that conducts data-theft extortion attacks through voice phishing”. The tactic seems to have been “borrowed” from ShinyHunters, arguably one of the largest data extortionists out there.

The group would call their targets on the phone (usually low-level employees with access to company SaaS solutions) and, while pretending to be from the IT department, convince the victims to either grant remote access, or to visit a malicious credential-grabbing landing page.

From there, the attackers would move in, map the infrastructure, exfiltrate valuable data, and then demand payment in cryptocurrency in exchange for deleting the data.

We have reached out to Levi’s with further questions and will update the article if we get an answer.

Via BleepingComputer

• Kimsuky used local AI tools to evade monitoring and enhance operations
• Researchers observed extensive AI-driven capability building across the group’s infrastructure
• Defenders urged behavior-based detection to spot evolving AI-enabled threats

North Korean hackers have found a way to use Generative Artificial Intelligence (GenAI) to supercharge their activities without tipping off the tool’s maintainers.

When people use AI tools like ChatGPT or Claude, their activities can be (at least to some extent) tracked and curbed - with OpenAI recently identifying and terminating multiple ChatGPT accounts used in phishing and human trafficking.

That is why Kimsuky - a known state-sponsored North Korean threat actor, used Ollama, GPT4All and Msty locally, allowing them to process documents without sending any sensitive information to outside AI services.

"Consistent process of capability development"

The attacks were spotted by security researchers Genians who “conducted months of tracking and log analysis on the infrastructure utilized as C2 in this campaign,” to identify the tools they used.

Aside from the three LLMs, they also used retrieval augmented generation (RAG) tools for document search, as well as AI agent development frameworks, text-to-speech software, and an AI-assisted coding tool called Cursor.

Using AI to write malicious code is not as simple as it sounds, due to various guardrails set up by the developers. As a result, AI in crime has been mostly limited to drafting phishing emails and crafting authentic-looking but malicious landing pages. However, Kimsuky has shown that AI in cybercrime continues to evolve and is becoming an ever-greater threat.

“What was observed in the threat actor's infrastructure was not merely evidence of several documents being created with AI, but a consistent process of capability development: establishing local LLM runtime environments, configuring RAG based on documents in the actor's possession, collecting AI agent development frameworks, and acquiring libraries for integration with external commercial AI services,” Genians concluded.

As a result, defenders must move from content-based assessment to behavior-based detection, the researchers warned, saying this should serve “as the fundamental premise of security recommendations.”

“In addition to indicator of compromise (IoC)-based detection, organizations should contextually correlate the sequence of anomalous activities following LNK execution, including PowerShell execution, persistence establishment, and external communications, to assess the overall threat level.”