Enterprise-sized businesses spend, on average, 28 percent of their software budgets on unwanted upgrades, migrations, and licensing changes. Often, those firms’ IT heads think those changes are detrimental.
Those results are according to a survey of 2,018 chief information officers and senior IT decision-makers at companies with at least 10,000 employees that Origina commissioned for its report, “The Software Tax.” About half of the respondents (1,011) are from the US, and the rest are from the United Kingdom. All respondents worked at companies with at least $1 billion in annual revenue.
Keep in mind that Origina provides maintenance for various enterprise software. Its report points to independent software maintenance as a way for IT leaders to regain control over software budgets, illustrating its potential bias. A third-party research consultancy, Censuswide, conducted the survey.
Among the report’s most concerning findings, enterprises often make software purchases that IT heads believe could hurt the organization. Ninety-five percent of respondents said “their organization has paid for a vendor-driven upgrade, migration, or licensing change in the past three years that it believed at the time would reduce performance, reliability, or business value,” the report says. Fifty-three percent of respondents “have done so more than once,” the report notes.
“Some 76 percent say their organization has very little real leverage in negotiations, and 80 percent say moving away from major vendors would cause significant disruption,” the report says, referring to the share of respondents who selected “strongly agree” or “somewhat agree” to those statements.
“When the alternative to an unwanted upgrade is losing support on systems the business runs on, saying yes is often a rational response,” Tomás O’Leary, one of the report’s authors and Origina’s CEO and founder, said in a statement. “That’s exactly the problem. The market has left some of the most experienced technology leaders in the world with little room to refuse, and that cost shows up in budgets, in engineering time, and in innovation that never gets funded.”
South Korean police identified seven suspects linked to a cybercrime operation that breached more than 90 organizations in the country, with ARTEX AI pentesting framework used in most of the attacks.
According to local media citing the Seoul Metropolitan Police, the gang leader has a background in penetration testing and worked at a cybersecurity company between 2019 and 2022.
In 38 of the incidents, the threat actor demanded ransom payments of 360 million won ($270,000), promising in exchange not to leak stolen data.
Authorities believe the threat actor stole 8.2 million records from compromised organizations, including educational institutes, hospitals, large corporations, and even golf courses.
Yonhap News says that the gang leader is a self-taught 32 year-old hacker and started attacking organizations in September 2025.
The Seoul Metropolitan police told the publication that the hacker used the ARTEX AI penetration testing suite, the same autonomous framework used in the attacks against multiple South Korean banks attributed to a Chinese-speaking threat actor.
Initially, the hacker used conventional tools but adopted AI utilities in late August 2026 to expand the attacks.
Authorities arrested all seven suspects, but only the group’s leader remains in custody. Arrested on October 15, he was formally detained two days later and referred to prosecutors on October 23. The other six suspects remain under investigation without being detained.
Police opened a case after a victim filed a complaint in June, and investigators traced the suspects through server logs from the breached entities.
Law enforcement authorities reportedly used an in-house AI tool to analyze the recovered evidence.
The police told local media that AI has lowered the barrier to entry for aspiring cybercriminals, resulting in large-scale, high-impact incidents that were previously only possible for skillful, knowledgeable actors.
“This is the first case of arrest of a suspect in Korea related to a large-scale hacking and personal information theft crime that exploited generative AI,” the police told Yonhap.
“Even individuals who are not professional hacker organizations can now hack on a large scale with the help of AI.”
The police continue the investigation, now focusing on whether any of the stolen data has been leaked on the dark web.
Overall prevention scores can hide what happens after initial access. Once attackers are using valid credentials, prevention drops sharply.
The Blue Report 2026 measures defenses technique by technique across 338 million simulations run in customer production environments.
The current debate over banning artificial intelligence in schools reminds me of another heated discussion a while back that took place within schools: sex education.
Let me explain.
There was a time when abstinence-only sex-ed was the agreed-upon approach in many schools. If we didn’t want kids to get pregnant or contract an STI, we taught them to abstain from sex. Period.
We now know that simply saying “Don’t do this” to children is not an effective way to prevent them from doing that very thing. As a mom and a former classroom teacher, I could have told you that.
Today, bans on AI are trying this same framework, assuming that removing AI from the classroom means that kids won’t use it for schoolwork. That approach may put up a speed bump for students looking for an easy answer, but it won’t prevent them from accessing the same tool at home or on a personal device. It’s like whack-a-mole; you can block it in one spot, but it will pop up somewhere else. In a nationally representative study completed before Los Angeles and New York City announced AI bans, of the 44% of teenagers who said they had had a tool banned at school, 59% of those same teens said they’d accessed the tool anyway on a personal device.
I support bans in that they can help schools buy time to figure out how to do AI literacy. Yet the hard truth about AI bans is this: You can “ban” or pause AI in schools, but that doesn’t stop kids from using it, just like we’ve learned with abstinence-only sex ed. Ban or no ban, we have to teach students how to safely navigate this new world.
Since the AI bans in New York City Public Schools and the Los Angeles Unified School District in early September, parents and educators alike have been spinning on the same question: What’s safe for kids, and what responsibility do we have to our kids when it comes to AI, knowing that they’ll use it anyway?
Anthropic CEO Dario Amodei recently warned that the pace of AI development was too fast — that frontier AI companies needed to slow development and build in more safeguards so that the AI can’t become too powerful. I’d argue that AI also needs more safeguards before it can be used by kids and teens, even for everyday, seemingly banal queries. I get the macro risk of AI as a geopolitical threat, or becoming sentient, and I don’t dispute the risks are real and frightening.
But we can’t overstate the significant risks to kids when they use AI. For the past six months, the Youth AI Safety Institute, which is part of Common Sense Media, where I’m the chief program officer, has been studying the leading AI products designed specifically for teens and general-population tools, such as ChatGPT for Teens and Google’s Gemini with teen protections, to evaluate them for safety. Think of the institute as the “crash test dummies” for AI. We’ve found time and time again that leading AI chatbots miss signals of suicidal ideation, indulge signs of psychosis, offer incorrect information around eating disorder support and complete all the homework assignments they are given. For example, the institute found that ChatGPT for Teens promised to alert parents about “dangerous” chats, but then failed to deliver notifications when testers simulated crisis conversations.
Whether your concern is more about cheating or more about kids turning to chatbots when they could be turning to a trusted friend or adult, the risks are real. While AI development across the industry has dramatically increased in speed — so much so that the industry is now self-regulating to slow itself down — safety improvements for kids seem to be moving at a glacial speed.
Removing AI from schools is an effective way to give school leaders time to figure out what successful implementation looks like in the classroom, but that only solves one side of the equation. An outright ban doesn’t do enough to protect and prepare kids for the real world they live in, one with AI integrated into nearly every popular app and website.
Putting a pause on AI in school doesn’t remove the need to teach kids about AI — what I call “AI literacy.” Some see that phrase and think, “Why would we teach kids how to use tools that we aren’t allowing them to use?” But AI literacy isn’t just about using technology; it’s about teaching kids how to grow up in an increasingly AI-driven world.
AI literacy involves educating students on topics such as navigating deepfake images on the internet; understanding why AI acts like a friend, even though it’s actually a large language model with a slick chat interface on top; thinking critically about when AI can be an aid to creativity, and when it’s doing the creative work wholesale; and learning how to identify bias or misinformation in an AI-generated answer.
New York City’s AI pause requires high schools to teach two 45-minute lessons each year to help students navigate this new world. It’s a good start, but not sufficient. Pause or no pause, students use AI by default every time they do a Google search. They use AI when creating posts with Lenses on Snapchat. And they definitely use AI to complete homework assignments from home.
We can stick our heads in the sand and assume that a pause will keep kids safe, or we can do the work to equip them with the critical thinking and AI literacy skills to navigate this changing world. Banning AI and then hoping kids stay safe is a story we’ve seen before. Let’s write a different ending. Teach kids what AI is. Teach them why AI flattens their critical thinking. Teach them why it acts like a friend, but isn’t. Then remind them that they can always make their own choices.
A few years ago, [Smashcat] ported a ZX Spectrum game called Manic Miner to the Arduino Uno. The Arduino has less memory than the original Spectrum did, so doing all this on an 8-bit microcontroller was no easy lift. Since then, parts of this original port have been adapted to produce a game engine, allowing [Smashcat] to port several other games to the Arduino Uno as well.
The engine is able to host a 2D map with the ability to scroll, at a resolution of 256×256, as well as up to nine sprites at the same time. The graphics are locked to 50 FPS, but only with a real Arduino Uno that has a crystal oscillator. Some replica boards do not have a timekeeping chip accurate enough to replicate the video signal, which is done without a framebuffer. That signal, as well as two PWM audio channels, are still generated directly as well, with only a few resistors needed to hook everything up to a monitor and a controller, including the option of using an NES controller.
With all of that running on the Arduino and it hooked up to the daughterboard that provides the controller interface as well as the resistors for the video signal to function and a driver for a speaker, the system now runs the original Manic Miner as well as three other games: Scramble, Space Invaders, and Pac Man.
The engine and the rest of the code are available on a GitHub page as well, and there’s even a WebAssembly version available to run various pieces of the software without the specific hardware. The original ZX Spectrum was an interesting machine too that still inspires builds like this, despite it being well over 40 years old now.
The iPhone Duo wowed us as Apple’s first foldable phone, but it’s time to test out how its first-gen hinge holds up. We’re now testing the $1,999 iPhone Duo using our custom-made folding machine, the Fold-a-Phone, which has been diligently opening and closing the phone since we bought the Duo on Friday.
Learn more: The iPhone Duo Just Passed 55,000 in Our Fold-a-Phone Testing Machine. Can It Beat the Galaxy Z Fold 8?
CNET is livestreaming our iPhone Duo long-term folding test on our YouTube channel, which you can watch embedded here. Our robust in-house machine will open and close the foldable around 20,000 times per day to simulate years of wear and tear on the hinge. You’ll be able to watch this testing on a livestream from our CNET Labs location in Louisville, Kentucky.
The livestream won’t just be for watching a phone open and close — you’ll also be able to chat with CNET editors who will be available to answer your iPhone Duo questions live in the chat. Will Apple’s first foldable live up to the likes of the Samsung Galaxy Z Fold 8 and Google Pixel 11 Pro Fold? Will its neat opening animation remain slick over thousands of folds? Join us and find out.
This long-term test will join our iPhone Duo review, which provides a snapshot of what it’s like to be among the first to use Apple’s first foldable phone. If you’re sold on it, check out the best iPhone Duo deals. And if you want to compare even further, you can check out our roundup of the best foldable phones currently available if you want help deciding between the Duo and other folding and flip-style phones from Samsung, Motorola and Google.
Trucks are handy, especially if you’ve got messy or dirty loads, but there aren’t a lot of options if you want a small electric pickup yet. [Michael Pick] decided to convert a BMW i3 into a utility vehicle.
After getting a new 12 V battery and cleaning the car up, [Pick] proceeded to disassemble the interior and remove the hatch from the car. Once he could look at the bones, he planned out the best places for the cut lines and removed the back half of the cabin.
The bed is built from plywood coated in penetrating epoxy and bedliner for weatherproofing, and special care was taken to seal the interface between the cabin and bed as the main electronics for the vehicle are housed in a compartment directly below. The taillight electronics were rehoused as the OEM lenses were too large to fit next to the tailgate opening.
The i3 has graced these pages before with a battery upgrade, and we’d be remiss not to remind you of the most famous EV to truck conversion, Truckla.
Author: Len Noe, Solutions Architect, BeyondTrust
The first question on any incident bridge isn’t about why it happened–it’s whose credentials were used. In my experience, whoever can answer that credential question in the first hour sets the pace for everything after it: what to isolate, what to revoke, whom to notify, and how far the damage could have spread.
The teams that can’t answer it spend the first day reconstructing their answer from logs that were never designed to provide it.
That discipline is about to be tested by a new kind of actor. On October 16, 2026, the National Institute of Standards and Technology (NIST), the Cybersecurity and Infrastructure Security Agency (CISA), and the National Security Agency (NSA) jointly published ten voluntary practices for incident readiness and response when AI agents act in ways nobody directed.
The guide, Key Practices for Incident Readiness and Response in Super Intelligence Training and Evaluation, is useful, and it’s also easy to misread. Its strongest guidance is about identity, even though it rarely uses that word as a headline.
The 14-page document is written for organizations that train or evaluate highly capable agents on long-horizon tasks (work that runs through many steps over a long period) with limited human supervision, although the authors say deployers may find it relevant. It organizes ten practices into five stages: prepare, detect, respond, recover and report.
It defines a control incident as an event in which agents take significant actions that were not directed by the operator or an adversary, and that undermine controls, policies or human oversight.
The agencies say the practices build on established incident-management frameworks, so most of what they ask for will look familiar: least privilege, credential management, monitoring, evidence preservation, rebuilding from known baselines.
Familiar controls applied to a new principal are where most organizations are likely to stumble.
One detail in the definitions matters for responders. The guide treats actions caused by adversarial prompt injection or a backdoored model as instructed, not uninstructed, so they fall outside its definition of a control incident. While that line is coherent for researchers, it’s difficult for responders; at the moment of an incident, the question of intent is usually unanswerable.
What an investigator has in hand is an identity, a set of permissions, and a trail of actions.
Agent incident readiness can’t depend on classifying the cause first. It has to work from an identity record to determine which identity acted, what authority stood behind it, and whether the agent was misled, malfunctioning, or compromised.
AI agents can operate with the same privileges as the people using them, creating new risks when their actions go unchecked.
Now generally available, BeyondTrust AI Agent Security enforces policy on individual agent actions at runtime, blocking unauthorized activity without restricting what human users can do.
Practice six tells organizations to revoke any credentials the agent accessed during the incident that the organization controls, and to rebuild potentially compromised environments. The qualifier “under the organization’s control” carries more weight than it may appear to at the outset. Agents accumulate credentials that the organization holds, but doesn’t issue.
The OpenCode breach shows how that plays out. OpenCode, an open-source coding agent, told some users of its hosted Console service that API keys they had stored there for external providers, including Anthropic, OpenAI, and Google, were exposed.
The company says 1.4% of its Console users were affected, and that the earliest exposure was at 00:30 UTC on October 17. Users reported unauthorized charges with costs ranging from several hundred dollars to as much as $1,700 overnight.
In some cases, Google Cloud projects were suspended for abusive activity, consistent with hijacked resources. The company hasn’t said how the attackers got in, and describes the incident as a large-scale distillation attack.
Take the company’s account as provisional. What matters for readiness is the shape of the problem. A key stored by an agent platform is revocable only at the provider that issued it, by the person who owns the account. And the account holder may not even know the key was ever exposed.
The guide’s revocation step presupposes an inventory of what the agent could touch. Without it, there’s nothing to revoke against. The guide’s reporting stage adds a second obligation that follows from the same inventory: notify the parties whose systems were compromised or exploited, including third parties who were not the final target.
In the OpenCode case, those parties are the model providers and the individual key owners, and none of them can be notified in time unless someone knows which keys were stored where.
The guide advises designers to assume agents will often find misconfigurations and vulnerabilities, and to build controls accordingly. MIT Technology Review’s reconstruction of the July Hugging Face incident shows why.
Agents built on OpenAI models were supposed to be isolated from the internet during a cybersecurity evaluation. They got online anyway, hacked Hugging Face, and retrieved solutions to problems they were stuck on.
The one control the article describes is network isolation, and the agents got around it. OpenAI’s own researchers attribute the behavior to training that rewarded it.
Network boundaries can be evaded by a determined actor, human or not. A credential that’s scoped to one task, expires when the task ends, and belongs to an identity separate from the person who launched the agent does not rely on the agent behaving, and it limits what a leaked copy can do.
The guide points the same direction when it asks for access limited to what’s necessary for intended tasks, single-use infrastructure where possible, and mechanisms to quickly block, contain, or terminate agent activity. Most of those are identity controls.
The guide also asks for tamper-resistant logs with reliable timestamps sufficient to reconstruct agent trajectories, and for assurance that agents can’t access or tamper with the monitors and logs themselves. That is session accountability in everything but name.
For a human administrator, a privileged session is attributed to a person, recorded and reviewable. The same standard should apply to an agent. The guide supports this by asking responders to preserve rejected actions and tool calls, and to record the configurations of agents, safeguards and environments so an incident can be reproduced.
A rejected action is evidence of what the agent tried to reach, which is often more revealing than what it succeeded in doing.
New research from BeyondTrust Phantom Labs™ drawing on endpoint telemetry shows why prompt-level logs fall short. In customer samples taken over a four week period, coding agents requested about 1.28 million shell commands from roughly 132,000 prompts, close to ten commands per prompt. Those are requested commands, not confirmed executions. A log that can’t separate what the agent did from what the user did is unable to answer whose authority was used.
Enterprise reality will complicate this. Many legacy applications can’t log per-agent activity. Some agents will run through shared service accounts because that’s how the integration was built three years ago. Budgets are finite. A realistic path starts with the agents that can reach the most sensitive data or the most consequential actions, gives them separate identities, and works outward as applications allow.
The strongest counter-argument is proportionality. The guide is aimed at frontier training and evaluation, and the incidents it responds to are rare: six in 141,006 evaluations in one lab’s disclosure. Most enterprises deploy agents far less capable than the ones under study, and the reported harm so far has been modest.
Why build response capability for a problem that’s mostly theoretical?
The answer is that the underlying mechanics aren’t just theoretical. A compromised or misdirected agent behaves, from the perspective of the identity record, like a compromised service account with a faster tempo.
Readiness should be measured in outcomes rather than documents. How long does it take to enumerate every credential a given agent identity can use, including keys it stores for third parties? What share of agents runs under its own identity, with a named human sponsor? Can the organization produce a session record that distinguishes what a person did from what an agent did on that person’s behalf?
If those questions take minutes to answer, the organization is ready for the first agent incident, regardless of how it is classified. If they take days, a more elegant taxonomy will not help.
Len Noe is a Solutions Architect at BeyondTrust, Transhuman, Podcaster, International Cyber Security Speaker, Author, Technical Evangelist, and Biohacker with 11 implanted microchips.
A former blackhat with more than 30 years in technology, he has presented in over 70 countries and is featured in the documentary I Am Machine, which premiered at DEF CON 2025.
BeyondTrust is the global leader in privilege-centric identity security protecting Paths to Privilege™. Identity alone doesn’t create risk. Privilege does. As human, machine, and AI agent identities explode across every environment, BeyondTrust is the only company built to discover, control, and secure privilege across all of them from a single platform. Trusted by 20,000+ customers, including 75 of the Fortune 100, and recognized as a multi-category leader by top industry analysts, BeyondTrust reframes identity security from a management problem into a strategic advantage.
Learn more at www.beyondtrust.com.
Sponsored and written by BeyondTrust.
Over the past decade, head-to-head brain games like QuizUp rocketed in popularity, reaching at one point millions of users. After that, battle royale games like Fortnite and PlayerUnknown’s Battlegrounds had their moment in the spotlight. Lately, growth of mobile games has slowed down amid a barrage of AI content across platforms. Now, the India-based gaming company Matiks wants to steer people away from AI-riddled timelines with its math, logic, and brain game puzzle platform.
The namesake app has been around for two years, but has seen real traction this year. The startup said that its daily users have grown from 22,000 earlier in the year to over 300,000 by September. The app also saw 2.2 million monthly active users in that month, with the company’s growth attracting fresh investor interest.
The startup said Monday that it has raised $10 million in pre-Series A funding led by General Catalyst, with participation from Indian entrepreneur Ronnie Screwvala and existing investors Info Edge Ventures, Tanglin Venture Partners, and Untitled Ventures. With this round, the company has raised over $13.5 million to date.
Matiks was founded by Sudhanshu Bhatia, Mohan Kumar, and Sushant Timmapur, who are college friends. Bhatia, who has a background in competitive mental math, said that he started working on the idea while he was working at American Express on an anti-money laundering strategy.
“For the last 10 years, the platforms have made entertainment consumption easy, but much of it is brain slop. We started exploring how we can make math-related games fun and entertaining like the other content,” Bhatia told TechCrunch over a call.
“Since I’ve had this background in mental math, I knew that a competitive ecosystem exists there. We got to know that the same exists for puzzles and memory. We wanted to make our app as entertaining as physical sports, so we opted for a dual format,” said Bhatia.
In the app, users can choose from Math, Puzzle, Memory, and Logic categories and play different games against bots to practice, and then against real players. Users can also play challenges and quests to earn points, maintain streaks, and unlock new games.
I have been playing the games for more than a week, and they are a good distraction for me, as I like games I can finish in a few minutes. The app hasn’t pushed me much towards purchasing more in-game items or shown distasteful ad pop-ups that would prompt me to uninstall the app completely.
Until now, Matiks has been focusing on user growth and just started testing a subscription this month, which unlocks game analysis, tips and tricks, practice, and additional features. The startup also allows for in-game purchases to skip certain challenges, maintain a streak, and boost XP points.
Bhatia said that while the majority of the game’s players are based in India, the startup has seen traction from other markets like the U.S., South Korea, and Japan. The startup plans to bring in localization to grow the latter two markets.
While Matiks is battling against AI slop, it isn’t entirely rejecting AI. The company has debuted an experimental game creation platform called Jelo, which lets users make their own games and let others play them. There is a ton of competition in this area, including Google’s new platform called Playground.
The startup, with 41 full-time employees, wants to use the funding to scale this platform along by expanding into newer markets.
Rahul Humayan, a partner at General Catalyst, told TechCrunch in an email that the product’s pace of iteration and user engagement have given the firm encouraging signs.
“We found the Matiks vision, product, and most importantly the community to be truly N of 1. The team’s product taste and pace of iteration made it a no-brainer,” said Humayan. The exceptional product love and power user behaviour are clear signs that the team is onto something special. We believe these are strong leading indicators for monetisation, with early results looking promising. The depth of engagement across surface areas like duels, social, Snacc, and now creation only helps.”
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
A pair of over-ears from Soundcore that impress in terms of noise cancellation and call performance, though the tuning of the sound is on the bassy side.
Good comfort
30-hour battery
Strong noise cancellation
Excellent call performance
Bass-heavy sound to a fault
Hazy-sounding Spatial Audio mode
Review Price:
£199.99
Thus AI chip
AI chip for noise cancellation and calls
Anker Soundcore app
Customise the performance through the app
Customised EQ
Customise the sound to you with Hear ID 5.0
It’s been a busy year for Soundcore. A gazillion products launched, a name change (though I’m still not entirely sure what the new name is), and the introduction of AI into several of its products. 2026 has seen the brand take a big leap.
But does that mean it’s left its value-focused beginnings behind? Not quite, but the Space 2 Pro headphones show a willingness to push performance while still undercutting the big players in the market.
The Space One Pro were one of my favourite Soundcore headphones of recent years, and though the Space 2 Pro do not replace those headphones (I think), there is an upgraded version of the Space 2 (which they also do not replace). Do they rank as Soundcore’s finest?
The Space 2 Pro look, well, like every other Soundcore headphone to date. If you’ve bought a pair of Soundcore headphones in the past, you’ll know the distinct look; the button configuration is the same (physical controls, no touch controls).
They’re slightly bigger than the Space 2, though surprisingly weigh less (by a whole two grams); but the main point of difference you’ll find between the two headphones is the silver trim on the earcups of the Space 2 Pro and the engraved ‘Thus Anker soundcore Space 2 Pro’.
Wearing the headphones for a prolonged period and they’re comfy, partly thanks to the amount of space in the earcups; while there’s not much pressure around the ears or on top of the head thanks to the padded headband and earpads. That’s not to say the fit is loose, but if you’re someone who complains about a clamping force that’s too tight, you won’t find that an issue here.
The headband is adjustable, though the slider is not stepless and therefore makes a slight noise when it’s shifted. The button layout is the same (power/Bluetooth and ANC on the left earcup, playback and volume on the right), and they’re easy to find due to their raised surface and pressing down on each button produces a click to let you know it’s been pressed.
They’re foldable for travel, though there’s no carry case to put them in, but a pouch to keep them safe. Inside are USB-C and 3.5mm cables for charging/audio (Hi-res audio is supported over a wired connection). Colours are the Jet Black version of this review and Midnight Grey.
The last few Soundcore headphones have claimed long battery life but come up short. The Space 2 Pro does live up to Soundcore’s expectations, but as usual, it’s a case of reading the fine print first.
Soundcore states 60 hours on a single charge, but that’s without ANC on. With ANC it’s 30 hours, and a battery drain at 50% volume for three hours confirmed that to be the case – as long as you’re listening to audio in AAC Bluetooth. While I didn’t drain it with LDAC, given that it’s usually a bit more power hungry, you could expect battery life to be guzzled quicker.
If you’re in need of a quick charge, a five-minute boost is enough for another eight hours.
I really enjoyed the Space One Pro’s sound. Weighty, spacious but clear; it was a balanced tuning that I found impressive for the price. I was hoping for a similar experience with the Space 2 Pro, but that’s not quite what turned out to be the case.
Compared to the less expensive Space 2, the Space 2 Pro are an upgrade. The tone of the two headphones is similar, but what you get from the Pro is bass that’s bigger, weightier, with clearer, more detailed highs. If you’re a Soundcore fan after a new headphone and deciding between these two, I would say that the Space 2 Pro is the better-sounding option.
But that comes with a caveat.
Compare the Soundcore against an actual mid-range rival in the JBL Live 780NC, and it reveals how bassy the Space 2 Pro are. I’d thought the JBL had a bass-forward sound, but compared to the Space 2 Pro, they are lightweights in that department.
The Space 2 Pro delivers a bass boost more than I was expecting. It’s slightly rich in tone, but its presence can be slightly overbearing with some tracks, hogging the limelight from other areas and making it all about that bass.
It does help that it also colours the midrange, making the transition between bass and midrange thicker in tone that over-emphasises male voices (less so female voices).
Some may enjoy the bass-heavy emphasis, but like the Liberty 5 Pro Max, having listened to many songs I’m familiar with, the tone of music changes – kick drums become thuds – and the headphones lose any sense of subtlety with the lows.
Certain tracks (like Phoebe Bridgers’ Garden Song) threaten to become boomy. With tracks such as Warren G’s Regulate, they’re great for describing that sub-bass and power, but not every song needs that treatment.
And that’s a shame because I feel the headphones do fine work in other areas and aside from the bass boost, I quite like the sound. They sound clearer and more detailed overall than the JBL; the soundstage is wide, and vocal performances ring out with a smooth clarity.
Treble is bright but not too bright, with decent levels of detail and clarity. I think it could be clearer and defter with the highs, and overall, what the Space 2 Pro lacks is degrees of subtlety and balance. If you’re after a fun sound with rich bass, the Soundcore provides that. If you’re after fidelity – which you are entitled to expect from a pair of headphones with Pro in the name – the bass power is a bit much.
Anker did release a firmware update that tweaked the sound, but that came a little too late in the day for this review. From the brief listen I did have, bass levels seem to have been tempered.
First off, the Space 2 Pro’s noise cancelling ability depends on getting the right seal. If there’s the slightest gap, sounds worm their way in like a window left ajar.
When the fit is good, the adaptive noise cancellation has been very reliable. It cancels noise naturally (no whine or noisiness when enabled). It shushes cars and traffic well. Walking through Canary Wharf, there aren’t many things I’ve noticed as I walked through the various paths and indoor areas to get between stations.
There’s some surface noise as wind passes by, but it’s nothing too distracting. It’s been quiet everywhere I’ve used the headphones, whether outdoors or on a plane – the only issue I’ve had is a slight vibration in the right ear cup.
The transparency mode is fine – it’s not the clearest or most detailed, but I’m able to hear what’s around me well enough.
Are they the best around the £200 / $200 mark? I would say that between it and the JBL Live 780NC, they’re on a similar level. The JBL might just edge it for being quieter, but the Soundcore seems to cancel noise more naturally. Either way, you can be satisfied with the Space 2 Pro’s performance.
Even more satisfying is the call performance. Using the same Thus AI chip that powered the Liberty 5 Pro and Liberty 5 Pro Max, the call quality is excellent. It picked up my voice well and didn’t latch onto any background noise at all.
Walking past the DLR during the call, popping into a supermarket and then into a busy restaurant, the person on the other end said all they could hear was my voice. At this price, the Space 2 Pro are the gold standard.
As you’d expect from Soundcore, the Space 2 Pro are stuffed with features. Bluetooth is the 6.1 profile, with support for SBC, AAC, and LDAC; plus Bluetooth multi-point for connecting to two devices at once (you can’t have this feature and LDAC enabled at the same time). From a wireless connection standpoint, I haven’t encountered any issues with these headphones.
The app is where you’ll find the bulk of the features, though I will say that I’ve had issues recently with connecting to the app as the app would not recognise the headphones despite being connected to my smartphone. I had to unpair the headphones and re-pair them for the app to work.
Like any other Soundcore headphone, you can tweak the noise cancellation settings and enable the Wind Noise Reduction setting (should you need to). You can customise the controls, but it’s more or less restricted to the NC and MfB buttons and not the widespread control that EarFun offers with its Wave Pro X.
HearID offers a way of personalising the audio signature of the headphones to how you hear, but you do need to log in/create an account to make that so. Anker offers presets, and some have an effect, though others sound very similar. You can create a custom preset too with an 8-band EQ.
There’s also a Spatial Audio setting at the bottom of the HearID tab, but give this setting a skip. It expands audio but sounds thin and hazy in terms of detail.
There’s a Safe Hearing mode you can enable where you can limit the volume of the headphones, change the audio codec (between high quality LDAC or prioritise connection stability) and toggle on Wear Detection (strangely not enabled by default).
There’s Anker AI Chat, though I’ve given that a wide berth after my experience with the Liberty 5 Pro Max. Otherwise, the Space 2 Pro is well-featured for the price if you’re looking to adapt the performance, though I’d say some of the customisations are more useful than others.
These headphones put in an excellent performance with calls, focusing on the voice and leaving background noise well in the background.
Ahead of publishing, the Soundcore sounded bass-heavy to a fault, but a recent update may have brought a bit more balance.
I like the Space 2 Pro, but they could be better. It’s clear that Anker Soundcore is pushing into the more premium category (slightly) but still undercutting the main midrange options. They still offer good value in terms of the noise cancellation, design, battery life and call performance, but the tuning of the headphones’ sound could be better.
The bass-heavy tuning doesn’t work for all types of tracks, lacking balance. A recent update to the sound seems to have brought more balance, but came a little too late for me to take into account.
The Space 2 Pro are another confident pair of over-ears from Soundcore, and while they’re maybe not the best around the £200 / $200, they’re in the mix. If you’re moving up from a less expensive pair of Soundcore headphones, I would recommend the Space One Pro for sound; but the improved ANC and excellent call quality certainly make the Space 2 Pro worth a look.
The Soundcore Space 2 Pro were tested over four weeks with real world use, with a battery drain carried out, the call performance tested, and ANC compared to other models
| Anker Soundcore Space 2 Pro Review | |
|---|---|
| Manufacturer | Soundcore |
| IP rating | No |
| Battery Hours | 30 |
| Fast Charging | Yes |
| Weight | 262 G |
| Release Date | 2026 |
| Audio Resolution | SBC, AAC, LDAC |
| Driver (s) | 35mm dual-material diaphragm |
| Noise Cancellation? | Yes |
| Connectivity | Bluetooth 5.1 |
| Colours | Jet Black, Midnight Grey |
| Frequency Range | 40 40000 – Hz |
| Headphone Type | Over-ear |
| UK RRP | £199.99 |
| USA RRP | $199.99 |
Audio-Technica has no intention of surrendering the high-end headphone market to Audeze, HiFiMAN, Meze Audio, Dan Clark Audio, or ZMF. With the introduction of the ATH-AWX7000 ($3,799), its new flagship closed-back headphones, the Japanese manufacturer is making another aggressive move into the increasingly competitive luxury headphone category, where distinctive designs, premium materials, and four-figure price tags have become the price of admission. Audio-Technica has been building serious headphones for decades, and anyone who thinks the company is content to watch its rivals dominate the summit-fi conversation hasn’t been paying attention.
The ATH-AWX7000 arrives just one year after Audio-Technica introduced its $3,500 ATH-ADX7000 open-back flagship in October 2025, which we subsequently auditioned at CanJam NYC 2026. That model demonstrated just how seriously Audio-Technica takes the reference headphone category, but the new ATH-AWX7000 takes a different approach, expanding the company’s longstanding tradition of premium wooden headphones with a decidedly Japanese emphasis on craftsmanship and materials.
The Audio-Technica ATH-AWX7000 combines traditional Japanese craftsmanship with newly developed driver technology, bringing together premium wooden housings and sophisticated acoustic engineering in a flagship closed-back design. The visual appeal is unmistakable, but Audio-Technica is betting that there’s considerably more to these headphones than beautiful woodwork and an ambitious price tag.
At the heart of the ATH-AWX7000 is the company’s new HMD (Hybrid Matrix Diaphragm) technology, which Audio-Technica claims represents a significant advancement in dynamic headphone performance. The manufacturer believes this combination of carefully selected materials, precision craftsmanship, and acoustic innovation establishes a new benchmark for high-end dynamic headphones.
The ATH-AWX7000’s earcup housings are crafted from Kiso Cypress, a prized Japanese wood traditionally used in the construction of Shinto shrines and Buddhist temples. Each housing is hand-finished using traditional Echizen lacquer techniques, highlighting the wood’s natural grain and warm coloration, which gradually darkens with age to develop an even richer character. Audio-Technica extends that attention to Japanese heritage to the headphones’ frame and other components, which feature kachi-iro, a deep navy hue historically associated with samurai culture and good fortune.
The ATH-AWX7000 features a lightweight magnesium frame and aluminum sliders, designed to provide durability without adding unnecessary weight, while allowing for precise fit adjustments. Audio-Technica also uses premium deerskin leather for the earpads and headband, providing a soft, comfortable fit for extended listening sessions, with the earpads helping maintain the acoustic seal essential to closed-back headphone performance. Like the company’s other premium wooden headphones, the ATH-AWX7000 is manufactured in Tokyo, Japan, where traditional craftsmanship and meticulous attention to detail remain central to the production process.
The luxurious wooden housings and premium comfort features are certainly appealing, but none of that matters if the headphones don’t deliver where it counts: sound quality. At $3,800, Audio-Technica has some serious expectations to meet, and beautiful Japanese craftsmanship alone won’t be enough to justify the asking price. To that end, the company has incorporated its newly developed HMD (Hybrid Matrix Diaphragm) technology into the ATH-AWX7000.
The technology employs a composite diaphragm constructed from polymer film and a metal dome, manufactured using Audio-Technica’s proprietary HXDT molding process, first introduced in the ATH-ADX7000 discussed earlier. This process produces a precisely shaped, large-diameter 58mm diaphragm designed to work in conjunction with the driver’s high-efficiency magnetic structure, voice coil, and other components. According to Audio-Technica, the result is improved accuracy, faster transient response, and greater dynamic performance.
The HMD design also allows the composite diaphragm to maintain its shape under high air pressure, supporting more precise pistonic motion even at frequencies below the audible range. Audio-Technica claims this results in deeper, more authoritative bass and greater musical accuracy than conventional closed-back designs, while maintaining a warm, lush sonic character. Those are ambitious promises, particularly in a category where controlling unwanted resonance and delivering convincing low-frequency performance remain significant engineering challenges.
The ATH-AWX7000 comes with two detachable 8-braid cables with Audio-Technica’s A2DC (Audio Designed Detachable Coaxial) connectors:
Rounding out the package, the ATH-AWX7000 includes a compact, protective hard carrying case designed to keep the headphones and supplied connection cables neatly organized. It’s a practical addition that makes storage and transportation easier, although these are clearly headphones designed for serious listening sessions rather than the daily commute.
| Audio-Technica Model | ATH-AWX7000 | ATH-ADX7000 |
| Product Type | Closed-back Dynamic Headphones | Open-Back Dynamic Headphones |
| Price | $3,799 | $3,499 |
| Driver Diameter | 58 mm | 58 mm |
| Frequency Response | 5 to 50,000 Hz | 5 to 50,000 Hz |
| Maximum Input Power | 1,000 mW | 1,000 mW |
| Sensitivity | 97 dB/mW | 100 dB/mW |
| Impedance | 48 ohms | 490 ohms |
| Input Jack | A2DC connector jack | A2DC connector jack |
| Weight | 353 g (12 oz) | 275 g (9.7 oz) |
| Accessories Included | Cable (3.0 m (9.8′)/6.3 mm (1/4″) gold-plated stereo plug)
Balanced cable (2.5 m (8.2′)/XLR-M connector (4-pin) Hard shell carry case Cable pouch |
Cable (3.0 m (9.8′)/6.3 mm (1/4″) gold-plated stereo plug)
Balanced cable (2.5 m (8.2′)/XLR-M connector (4-pin) Hard shell carry case Cable pouch Earpads |
| Replacements/Separately Sold Items | Cable
Balanced cable Extra Earpads |
Cable
Balanced cable Extra Earpads |
The Audio-Technica ATH-AWX7000 makes a compelling case for luxury closed-back headphones that combine traditional Japanese craftsmanship with serious acoustic innovation. Its hand-finished Kiso Cypress housings, premium deerskin appointments, and newly developed 58mm Hybrid Matrix Diaphragm distinguish it from competitors like the Dan Clark Audio Stealth ($3,999) and Focal Stellia ($2,999), both of which take different approaches to reference-level closed-back performance. Audio-Technica has made a convincing statement with the design and engineering, but craftsmanship alone doesn’t guarantee exceptional sound. Whether its new driver technology delivers a meaningful advantage over these established rivals is something we’ll need to determine through listening.
The Audio-Technica ATH-AWX7000 Closed-Back Dynamic Wooden Headphones are available at a suggested US retail price of $3,799 at Audio46, Audio-Technica and Authorized Dealers
The Audio-Technica ATH-ADX7000 is available now for $3,499 at Amazon.
[Angus] of Maker’s Muse enjoys combat robotics, in which contenders zip about in an arena, attempting to disable one another with weapons that lift, spin, and smash. He got a funny idea a short while ago and wondered whether a stiff mane of bristles could serve as lightweight, ablative, impact resistant armor?
Just to set the stage, [Angus]’s new robot competes in the plastic antweight (or “plant”) division, in which robots are limited to 150 grams with both chassis and weapons made of plastic. In such circumstances, one can see how a mane of thick, rigid “hairs” might actually punch above its weight, so to speak.
Nothing settles a question like a good experiment, so [Angus] sat down and redesigned his latest robot to replace its outer shell with a thick wad of bristles, in the style of a 3D printed lion’s mane. Filament is bridged between the model and an outer support structure, which is cut away afterwards to expose the “hairs”.
How well did it perform? Well enough to merit further study! Where past robots got their outer shells broken up, the hairy robot seems to have come through quite intact. It seems the idea might not be as silly as it first seemed, at least in the plastic antweight division.
3D printers have been fantastic in the field of combat robotics. First they were great because they made custom parts easier and more accessible to hobbyists, but things have evolved more since then. Now folks compete by putting motors and electronics inside almost entirely 3D printed robots. There’s never been a better time to get into combat robotics, and experiment with unorthodox designs.
If you want to try a hairy robot for yourself, be sure to watch the followup in which [Angus] improves the hair-generating process considerably.
Weekend Open Thread: Anine Bing
Tuesday’s Workwear Report: Margeaux Blazer in Four-Season Stretch
Seattle ag-tech startup Carbon Robotics reduces workforce, citing economic conditions
DeepSeek updates V4.1-Flash model to improve performance
Amazon challenges Nvidia’s software advantage with new tool for its own AI chips
New conversion kit lets builders fit ATX motherboards and standard PC GPUs inside Apple Mac Pro case
Whoops: Grok Bot Spills Tech CEO’s Bank Data In Public Slack
UK Banks Restricting Crypto Transfers: Complaints Rise
Best stablecoin yield in 2026: 6 ways to earn on USDT and USDC, ranked by where the money comes from
AI agents like Muse and Instinct can do your tedious tasks. But should they?
CME Bitcoin Cash & Uniswap Futures for UK Traders
GLP-1s and AI among talking points at W3Fit
Google Support Pages Start Defaulting to AI Chatbots
Stolen UpdraftCentral keys used to backdoor WordPress sites
How safe havens stack up as tech, energy and France fears grip markets
The 10 Best Dressed Stars at the 2026 Academy Museum Gala: Jessica Biel in Tamara Ralph, Teyana Taylor in Schiaparelli, Kim Kardashian in Vintage Balenciaga, And More!
Team News, Stats and Odds
Tesla-SpaceX Merger Being Considered
Gemini Notebook now lets you review slides and videos before generating them
Google Cloud outage hits Revolut, Lloyds and DPD
You must be logged in to post a comment Login