Training large language models is brutally expensive. It’s not just about having more GPUs; it’s about how efficiently you use them. And as models scale up, even small inefficiencies can turn into massive time and energy costs.

Now, a team of researchers from MIT, working with collaborators including NVidia, says it has found a surprisingly practical way to reclaim wasted compute during training — in some cases cutting overall training time nearly in half.

The problem they’re targeting lies in reinforcement learning (RL), particularly during what’s known as the “rollout” phase. This is the step where a model generates multiple candidate responses so it can learn which behaviors lead to better outcomes. It’s essential for reasoning-focused LLMs — but it’s also slow.

In fact, the rollout stage can account for as much as 85% of total execution time. The culprit is something researchers call a “long-tail distribution” of response lengths. Most generated responses finish quickly. But a small number run much longer than average. Because GPUs need to synchronize, the faster ones often sit idle waiting for the stragglers to complete.

The MIT team’s solution, called Taming the Long Tail (TLT), tackles that waste head-on. Instead of letting GPUs sit idle during those long generations, TLT uses that downtime to train a lightweight “draft” model on the fly. This smaller model learns continuously from the main model as training progresses.

The idea builds on speculative decoding, a technique where a smaller model predicts tokens ahead of the main model so multiple tokens can be verified in parallel. Traditional speculative decoding relies on a fixed draft model, which quickly becomes outdated as the primary model evolves during reinforcement learning.

TLT changes that dynamic. By retraining the drafter opportunistically using otherwise idle resources, the system keeps the draft model aligned with the main model, without requiring extra dedicated compute.

In experiments across several reasoning-focused LLMs and real-world datasets, the results were significant. The researchers report end-to-end training speedups ranging from 70% to 210% compared to strong baselines, effectively doubling training speed in many scenarios. Importantly, model accuracy remained unchanged.

There’s also an interesting side benefit: the continuously trained drafter itself becomes a useful artifact. Because it’s trained alongside the main model, it can serve as an efficient inference model in certain contexts.

The work points toward a broader theme in AI research right now: optimization over brute force. Instead of scaling up clusters indefinitely, researchers are increasingly looking for ways to extract more performance from the hardware already in place.

If approaches like TLT prove robust at larger industrial scales, they could meaningfully reduce both the financial and environmental costs of training next-generation reasoning models.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source | Via)

The post Researchers double AI training speed just by reclaiming idle GPU time appeared first on Gizmochina.

After five years out of the lead, Xiaomi is back on top in the global wearables market. New data from Omdia shows that worldwide wearable shipments crossed 200 million units in 2025, up 6% year over year.

Xiaomi claimed 18% of total shipments, narrowly edging out Apple at 17% and Huawei at 16%. The margins are razor thin, with less than a percentage point separating the top three. Samsung and Garmin round out the top five at 9% and 5%, respectively.

What stands out isn’t just Xiaomi’s return to No. 1 — it’s how close the race has become. According to Omdia’s research manager Cynthia Chen, the competitive focus has shifted. Wearables are no longer just about who has the best display, the most sensors, or the longest battery life. Increasingly, it’s about ecosystem strength. The tighter the integration between devices — phone, tablet, car, smart home — the stickier the user base becomes.

Xiaomi’s comeback appears to be driven by breadth rather than a single breakout product. Its Mi Band lineup continues to move huge volumes in the affordable segment, while its entry-level smartwatches help raise average selling prices without pushing too far into premium territory. The company’s broader “Human × Car × Home” strategy also plays a role, tying wearables more closely to its expanding ecosystem.

Apple, meanwhile, remains dominant in the high-end bracket. Its wearables benefit from deep integration with the iPhone, premium build quality, and an expanding suite of health features. Huawei has carved out a strong position as well, particularly in China, with a wide portfolio and growing focus on professional sports tracking and health tools.

Health tracking is evolving from basic step counts and heart rate monitoring to continuous, more advanced metrics. Some brands are even exploring lighter, more minimalist form factors to improve comfort for 24/7 tracking.

With Xiaomi, Apple, and Huawei separated by just a sliver of market share, 2026 could easily reshuffle the rankings again. For now, though, Xiaomi has its crown back. Going forward, the real battle may be less about hardware and more about who can build the most compelling ecosystem around it.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source)

The post Xiaomi retakes top spot in global wearables market appeared first on Gizmochina.

BMW is starting to put humanoid robots to work in Europe — a first for the company — as it looks for new ways to improve efficiency and stay competitive in a tough global market.

The German automaker has launched a pilot program at its Leipzig plant in Germany, where AI-powered humanoid robots will assist with selected assembly tasks. While this is the first European rollout, it’s not BMW’s first experiment overall. The company previously ran a 10-month trial at its Spartanburg, South Carolina plant in the US, where Figure AI’s Figure 02 humanoid robots worked five days a week, up to 10 hours a day.

Complementing existing automation systems rather than replacing them

During that US pilot, the robots reportedly handled physically demanding tasks such as placing sheet metal components for welding. According to BMW, they contributed to the production of more than 30,000 vehicles while also reducing strain on human workers. The goal wasn’t to replace staff, but to take over repetitive or ergonomically challenging jobs.

Now, the focus shifts to Leipzig. At the German facility, BMW is testing the AEON humanoid robot developed by Hexagon Robotics. The machines will assist in assembly line operations and in high-voltage battery production.

For now, BMW says only a small number of units — in the single digits — will operate alongside human employees, complementing existing automation systems rather than replacing them.

BMW executives are framing the move as part of a broader push toward digitalization and “physical AI” in manufacturing. Milan Nedeljković, the company’s production chief and incoming CEO, described digital technologies as key to maintaining competitiveness.

BMW isn’t alone in exploring humanoid robotics. Automakers including Tesla, Mercedes, and Hyundai have all signaled interest in similar technologies. Analysts at Morgan Stanley have projected that the humanoid robotics market could reach $5 trillion by 2050, with large-scale adoption expected particularly in China.

Full-scale deployment — if it happens — will depend on how well the robots integrate into existing workflows and whether they can deliver measurable productivity gains.

Still, the direction is clear. As manufacturing becomes more complex, automakers are increasingly looking beyond traditional industrial robots.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Sources: 1, 2)

The post BMW turns to humanoid robots to boost factory efficiency in Europe appeared first on Gizmochina.

Qualcomm had hinted that GDC 2026 could be an important moment for Windows-on-Arm gaming hardware. That’s no longer the case.

The company has confirmed it won’t be announcing any updates to its Snapdragon G Series gaming chips at this year’s Game Developers Conference, which runs from March 9 to 13 at Moscone Center in San Francisco. There also won’t be hands-on benchmarking sessions for the newer Snapdragon X platform revisions — the chips currently powering Windows-on-Arm PCs that some hoped would make a stronger push into gaming.

That clarification cools expectations that had been building since CES. Earlier in the year, Qualcomm had downplayed major handheld announcements but suggested GDC might be a better venue for gaming-focused updates. Naturally, that led many to expect performance demos, developer tools, or at least some clarity around OEM partnerships. None of that is happening — at least not this month.

For those following the space, this matters. The Snapdragon G Series (G1 Gen 2, G2 Gen 2, and G3 Gen 2) is Qualcomm’s dedicated push into handheld gaming silicon. On paper, these chips support QHD+ displays up to 144Hz, Wi-Fi 7, and Adreno GPUs tuned for portable gaming. Meanwhile, the Snapdragon X lineup — especially the “Elite” branded chips — have been positioned as powerful enough to handle both productivity and gaming in thin-and-light Windows devices. Together, they’ve fueled optimism about a possible Windows-based alternative to x86 handhelds.

That’s why GDC felt important. It would have been the ideal place to show real-world gaming benchmarks, highlight driver improvements, or announce partnerships with handheld makers. Even early developer toolkits would have signaled momentum. Instead, things remain quiet.

Windows-on-Arm gaming is at a delicate stage. There’s genuine potential, especially with the efficiency gains Qualcomm’s architecture can deliver. But the x86 gaming ecosystem is deeply entrenched, and expectations are high. Without concrete demos or measurable progress, enthusiasm can cool quickly.

For developers and enthusiasts attending GDC this year, Snapdragon-powered gaming hardware won’t be the headline some were anticipating.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source)

The post Qualcomm confirms no Snapdragon G updates at GDC 2026 appeared first on Gizmochina.

OnePlus has started teasing its next compact flagship, the OnePlus 15T, and on paper it checks a lot of the right boxes. But if you were hoping for a major camera upgrade this year, the latest leak might temper expectations a bit.

According to Weibo tipster 老陈Air, the 15T will feature a dual rear camera setup. The primary sensor is said to be Sony’s LYTIA 700 (1/1.56″ size) with an f/1.8 aperture and OIS. Alongside it sits a Samsung ISOCELL JN5 (1/2.76″ size) telephoto camera at f/2.8, also with optical stabilization.

If that configuration sounds familiar, it’s because it largely mirrors the OnePlus 13T (our review). And yes — once again, there’s reportedly no ultrawide camera.

The LYT-700 is very close to what we saw in the 13T in terms of size and light intake, the 1/1.56″ IMX 906, and the telephoto aperture is actually narrower this time (f/2.8 versus f/2.0 previously). That suggests refinement rather than a real hardware leap.

That said, minor improvements may come from software processing and whatever imaging enhancements the new chipset enables.

Outside the camera department, though, the 15T looks far more ambitious. Leaks point to a 6.31-inch OLED panel with 1.5K resolution and a 165Hz refresh rate. Performance should be top-tier, thanks to Qualcomm’s Snapdragon 8 Elite Gen 5. Rumors suggest a massive 7,500mAh battery paired with either 100W or 120W fast charging. If that capacity holds true in a smaller chassis, it would be genuinely impressive and could give the phone serious endurance credentials.

The OnePlus 15T looks like it’s shaping up to be a performance-first compact flagship — big battery, high refresh rate, flagship silicon — but with a camera setup that plays it safe. For users who value telephoto zoom and one-handed usability over ultrawide versatility, that may be perfectly fine. For others, it could feel like a missed opportunity.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source)

The post OnePlus 15T leaked camera details are hard to get excited about appeared first on Gizmochina.

Even companies with supply chains as tight as Apple’s aren’t immune to market pressure, and the current memory shortage seems to be testing that reputation.

A new report out of Korea suggests Apple has agreed to a steep price increase for LPDDR5X memory supplied by Samsung. According to financial outlet Dealsite, citing industry sources, Apple is now paying roughly double for these low-power DRAM modules.

LPDDR5X isn’t some minor component. It’s the RAM used across Apple’s modern silicon lineup, from iPhones and iPads to Macs.

What makes this interesting is how the negotiation reportedly played out. Samsung’s semiconductor division (DS) was initially aiming for around a 60% price increase for LPDDR5X destined for iPhone production. But in what appears to have been a classic anchoring tactic, Samsung reportedly opened discussions demanding a 100% hike.

Sources claim Apple agreed to the full increase during urgent talks focused on securing supply for the first half of 2026. One insider quoted in the report summed it up bluntly: smartphone brands are heating up to secure memory inventory.

The broader shortage isn’t happening in isolation. Demand for high-bandwidth memory (HBM), particularly for AI data centers driven by companies like Nvidia, has pulled manufacturing capacity toward more profitable enterprise-grade memory.

Pricing reflects that shift. Reports indicate 12GB LPDDR5X modules have climbed from roughly $25–$29 earlier in 2025 to around $70. That’s not a marginal adjustment — it’s a structural change in pricing dynamics.

Interestingly, even Samsung’s own mobile division is said to be feeling the squeeze. Early sourcing for the Galaxy S26 reportedly splits DRAM supply between Samsung’s semiconductor arm and Micron, which suggests internal capacity alone isn’t enough to fully insulate its smartphone business.

Apple is set to announce updated MacBook Pro, MacBook Air, iPad, and iPad Air models next week, alongside a more affordable MacBook and the iPhone 17e. While there’s no immediate indication of major retail price hikes, sustained component cost pressure could narrow margins or subtly reshape pricing tiers.

The impact won’t be limited to Apple. Industry forecasts already suggest smartphone shipments could soften in 2026 if memory constraints persist. Larger players like Apple and Samsung are better positioned to absorb higher costs, but smaller manufacturers may struggle more visibly.\

(Source)

The post RAM price hike: Samsung targets 60% increase but opens at 100% — Apple accepts without negotiation appeared first on Gizmochina.

Both the Galaxy A57 and A37 devices have appeared in the EU’s EPREL database (under model numbers SM-A576B/DS and SM-A376B/DS), and the energy labels reveal some interesting details, especially around battery life and durability.

Starting with battery life, both phones receive an A grade under the EU’s standardized testing. The Galaxy A37 is rated for up to 53 hours of usage on a single charge, while the A57 is listed at 52 hours. These numbers are based on a fixed EU usage cycle — typically a mix of calls, web browsing, video playback, and standby — so they won’t reflect every real-world scenario, but they do suggest solid day-and-a-half endurance for most users.

Both models are listed with a 4,905mAh rated battery, which Samsung will almost certainly market as a typical 5,000mAh capacity. Fast charging is said to top out at 45W, which keeps them competitive in the mid-range space.

The Galaxy A57 and A37 are both certified with IP68 water and dust resistance. That’s a step up from the IP67 rating seen on their predecessors. On paper at least, that’s a meaningful upgrade. Drop resistance ratings differ slightly, though. The A57 earns an A rating, while the A37 gets a B. Repairability sits at C for both, which is fairly typical for modern sealed smartphones.

According to the EU label, both phones are certified for 1,200 full charge cycles before battery capacity drops below 80 percent of its original level. That’s noticeably lower than the 2,000 cycles previously listed for the Galaxy A56 and A36. The same reduction was also observed on the Galaxy S26 series, which suggests this could be part of a broader shift in Samsung’s battery strategy, possibly tied to chemistry changes or internal design adjustments.

It doesn’t necessarily mean the batteries are worse in everyday use. But for users who hold onto their phones for four or five years, that lower cycle rating is something to keep in mind.

An official launch is expected next month. As always, the full picture will depend on pricing and real-world testing.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Sources: 1, 2)

The post Galaxy A57, A37 EU energy labels confirm battery sizes and endurance ahead of launch appeared first on Gizmochina.

Just a day after its global unveiling at Galaxy Unpacked, the Samsung Galaxy S26 series is already being put through its paces. And as expected, the Snapdragon vs Exynos debate is back — though this time, the gap doesn’t look quite as dramatic as in previous years.

It’s also worth noting that the Exynos 2600 is Samsung’s first chip built on its 2nm GAA process, while the Snapdragon variant continues to rely on TSMC’s mature 3nm node. In a way, these early results aren’t just about Qualcomm vs Exynos — they’re also an early indicator of how far Samsung Foundry’s next-gen manufacturing has come.

Post-launch Geekbench 6 scores are starting to surface, and they paint a pretty clear picture on the CPU side. The Galaxy S26 Ultra posted single-core scores of 3,670 and 3,724 in two separate runs, with multi-core results reaching 10,981 and 11,237. These are strong numbers and in line with what you’d expect from Qualcomm’s tuned “for Galaxy” variants with slightly higher clocks, tighter optimization, and aggressive performance scaling.

Meanwhile, the Exynos 2600 inside the Galaxy S26+ recorded 3,105–3,197 in single-core and 10,444–11,012 in multi-core tests. That puts Snapdragon ahead by roughly 10 to 18 percent in CPU performance when it comes to single-threaded workloads. In practical terms, that could translate to slightly snappier app launches and better responsiveness under lighter loads.

But the GPU story is more interesting this year. In OpenCL testing, the Exynos 2600 actually posted a marginally higher score — 24,240, compared to the Snapdragon’s 24,152. It’s a tiny difference, but symbolically important.

That’s notable because historically, Exynos variants have struggled to keep up with Qualcomm not just in CPU efficiency but also in GPU stability and sustained output. If these early numbers hold up in long gaming sessions and real-world thermal tests, Samsung may finally have narrowed the performance perception gap.

Of course, these are still early post-launch benchmarks. Sustained performance and battery drain under load will matter more once full reviews drop.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Sources: 1, 2, 3)

The post Exynos 2600 runs neck-and-neck with Snapdragon 8 Elite Gen 5 in post-launch Geekbench tests appeared first on Gizmochina.

Oppo is taking an unusually open approach ahead of the Oppo Find N6 launch. With the device expected to debut on March 17, the company recently invited a member of the media into its production facilities to showcase the engineering work behind what it claims is a significantly reduced display crease.

The foldable crease has long been one of the category’s most persistent compromises. Oppo says it has spent the past three years refining its hinge and display assembly process, developing multiple prototypes along the way. According to the company, the latest iteration is built with alignment tolerances as tight as 0.03mm. It’s a level of precision intended to reduce visible indentation from the start and slow down long-term wear.

Durability testing appears to be a major focus. Oppo demonstrated an automated folding rig designed to simulate extended use, pushing devices through up to 300,000 fold-and-unfold cycles. During the media visit, the reporter was shown a unit that had already passed roughly 170,000 cycles. Even at that stage, the crease repirtedly remained minimal. Independent testing will ultimately determine how that translates to everyday use, but the company is clearly confident in its results.

The Find N6 has also reportedly passed TÜV Rheinland certification, with Oppo suggesting it achieved one of the flattest foldable displays evaluated so far. The inner display is expected to measure around 8.12 inches, paired with a 6.62-inch cover screen.

Leaked details suggest Oppo is aiming for a well-rounded flagship. The device is rumored to feature a Hasselblad-branded camera system built around a 200MP main sensor. It’s also expected to run on Qualcomm’s Snapdragon 8 Elite Gen 5 platform and pack a sizeable 6,000mAh battery with 80W fast charging — notable figures for a slim book-style foldable.

By offering a behind-the-scenes look at production and stress testing, Oppo appears keen to differentiate the Find N6 in an increasingly competitive foldable market. Samsung, Honor, and other brands continue to refine hinge and display technology, and expectations are rising with each generation.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source)

The post Oppo Find N6 factory tour reveals how a near-invisible crease was achieved appeared first on Gizmochina.

Nothing is preparing to expand its audio lineup again, this time with a more affordable pair of over-ear headphones. The upcoming Nothing Headphone (a) is set to debut on March 5, 2026, during the company’s “Built Different” event.

The event will also introduce the Nothing Phone (4a) and Nothing Phone (4a) Pro, suggesting Nothing is continuing its push to build a broader, connected ecosystem rather than focusing on standalone devices.

The Headphone (a) appears to be a more accessible alternative to the original Nothing Headphone (1), which launched at €299. This new model is expected to come in at €159 in Europe and £149 in the UK, roughly half the price of the flagship version.

To reach that price point, Nothing is reportedly opting for a plastic construction instead of more premium materials. That doesn’t necessarily mean it will feel cheap, but it does signal a clear shift toward the mid-range segment.

In terms of design, the company isn’t toning things down. The Headphone (a) is expected to arrive in pink, white, black, and yellow — in line with Nothing’s bold, playful aesthetic. The semi-transparent styling that defines the brand is also likely to remain intact.

On the technical side, leaks point to a custom 40mm dynamic driver, Bluetooth 5.3 support, and dual-device pairing. Battery life could reach up to 35 hours with active noise cancellation enabled, which is competitive at this price. There’s some speculation that the ANC system may be less advanced than the one found in the Headphone (1), possibly as part of the cost-saving strategy.

If that trade-off holds true, the Headphone (a) may appeal most to users who want solid sound quality and long battery life without paying for top-tier noise isolation.

Nothing has been steadily growing its portfolio, and launching a more affordable over-ear option alongside mid-range smartphones reinforces that strategy. We’ll get full confirmation of specs, pricing, and availability on March 5. For now, the Headphone (a) looks like Nothing’s attempt to hit a sweet spot — stylish, capable, and easier on the wallet.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source: Nothing)

The post Nothing Headphone (a) launching March 5 as budget over-ear option appeared first on Gizmochina.

Honor is starting to lift the curtain on its next foldable, the Honor Magic V6, ahead of its global debut at MWC 2026. With just days to go before the March 1 launch in Barcelona, the company has shared early camera samples and revealed a new color option, both clearly meant to build anticipation.

On Weibo, Honor’s imaging team posted a set of sample photos taken with the upcoming Magic V6. The company continues to lean on its AiMAGE branding, first introduced at MWC last year, which blends on-device AI processing with cloud-assisted enhancements. According to Honor, this setup improves clarity, zoom performance, and overall scene optimization.

The shared images highlight strong detail retention and balanced color reproduction, at least in the curated examples shown. Telephoto shots appear sharp, and low-light scenes seem well controlled. Of course, these are official samples, so real-world performance will need independent verification. Still, the message is clear: Honor wants the Magic V6 to compete not just with other foldables, but with top-tier slab flagships on camera quality.

Alongside the imaging teasers, Honor also unveiled a new “Red Rabbit” color variant. The finish features what the company describes as eco-friendly velvet leather with a nano-coating for durability. Visually, it’s a deep red tone accented with gold detailing around the circular camera module.

Under the hood, the Magic V6 is confirmed to run on Qualcomm’s Snapdragon 8 Elite Gen 5, bringing expected gains in performance and power efficiency.

With competition intensifying among foldable makers, incremental upgrades may no longer be enough. Honor seems intent on pushing both design and camera performance to stand out.

We’ll get the complete picture at MWC. For now, the Magic V6 looks like a foldable that’s aiming high, both in hardware and presentation.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source)

The post Honor Magic V6 camera samples impress ahead of MWC debut appeared first on Gizmochina.

The foldable phone market has been chasing the same goal for years: reduce the crease enough that you barely notice it. Now, if recent leaks are anything to go by, Oppo may be getting closer than most.

A real-life photo of the upcoming Oppo Find N6 have surfaced online, showing a noticeably flatter inner display. The crease is almost invisible when looking straight on.

Oppo hasn’t shared full technical details yet, but teasers point to a redesigned hinge system, reportedly built with a titanium alloy structure. The company is also said to be using a new ultra-thin glass designed to better resist long-term deformation. Whether that involves new materials or refinements to lamination isn’t clear, but the goal is obvious: make the display look and feel as close to a traditional slab phone as possible when opened.

Oppo is reportedly planning a March 17 launch in China, potentially beating several key competitors to market. Samsung’s next book-style foldable, expected to follow the current Samsung Galaxy Z Fold7, likely won’t arrive until mid-year. Google’s next Pixel foldable is also expected later in the summer. Meanwhile, Honor’s Honor Magic V6 is rumored for a similar March window, setting up direct competition.

If the Find N6 delivers on what these early glimpses suggest, it could raise the bar for crease reduction in book-style foldables. That would be particularly significant as Apple is widely rumored to be preparing its own foldable iPhone in the coming years, with heavy focus reportedly placed on minimizing the fold line.

Of course, real-world impressions will matter more than carefully framed demo shots. Lighting, display angle, and even screen content can influence how visible a crease appears. Still, the early signs suggest Oppo is pushing hard on hinge and display engineering.

More details are expected soon. For now, the Find N6 looks poised to deliver one of the flattest foldable displays we’ve seen so far.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Source)

The post Oppo Find N6 real-world image showcases near-invisible display crease appeared first on Gizmochina.

Samsung has never been the most aggressive player when it comes to battery innovation. While most of its rivals are chasing bigger numbers year after year, Samsung is definitely moving more cautiously. Now, though, a senior executive has confirmed that change is coming.

Speaking at a media roundtable ahead of Galaxy Unpacked 2026, Moon Sung-Hoon, Executive Vice President and Head of Samsung Electronics’ Smartphone R&D Team, said the company is actively preparing a smartphone that uses silicon-carbon anode battery technology. According to Moon, such a device will arrive “in due course,” suggesting the tech is progressing internally, even if it’s not ready for prime time just yet.

Silicon-carbon batteries have become increasingly common among Chinese smartphone brands. By replacing traditional graphite anodes with silicon-carbon composites, these batteries can store more lithium ions in the same physical space. The result? Much higher capacities without dramatically increasing thickness or weight.

Moon acknowledged that the company may appear slower than competitors in adopting newer battery chemistry, but emphasized that safety, durability, and long-term stability remain priorities. That caution is understandable given Samsung’s history, particularly the lessons learned from the Galaxy Note 7 incident years ago. New battery technologies must pass strict internal testing, especially in areas like swelling resistance, lifespan, and charge cycle reliability.

That conservative strategy is evident in the just-launched Galaxy S26 series. Only the base model received a modest battery bump, while the Galaxy S26 Ultra and its Plus sibling retained familiar capacity figures.

That said, Moon’s comments confirm that Samsung is not ignoring the shift toward silicon-carbon chemistry. It’s simply taking its time. With brands like Honor, Oppo, OnePlus, and Xiaomi pushing 7,000mAh-plus batteries and longer endurance, the pressure on Samsung is mounting.

For Galaxy fans who feel battery upgrades have been incremental in recent years, this could signal a more meaningful leap on the horizon. There’s no concrete launch timeline yet, but the direction is clear: Samsung is working on it.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Via)

The post Samsung is finally adopting silicon-carbon battery, joining rivals appeared first on Gizmochina.