Artikel Arm AGI CPU Targets Intel Xeon and AMD EPYC With 136 Cores pertama kali tampil pada todayinasian.com.
]]>For more than three decades, Arm has primarily supplied CPU architectures that companies such as Qualcomm and MediaTek turn into their own chips. The AGI CPU changes that business model by giving Arm a finished processor aimed at generative AI and agentic AI infrastructure. The shift also reflects a change in how data centers handle AI workloads. GPUs remain crucial for intensive model computation, but CPUs are increasingly responsible for the surrounding tasks, including database access, API calls, orchestration and business applications.
The AGI CPU is manufactured by TSMC using a 3nm process and can feature as many as 136 Neoverse V3 cores running at up to 3.7GHz. Arm is using a chiplet-based approach to build the processor, allowing multiple computing components to work together rather than relying on one enormous piece of silicon. At Hot Chips 2026, further technical details showed that the AGI design uses two chiplets, with each chiplet containing 70 CPU cores connected through Arm’s CMN fabric.
Read More: Intel Xeon Diamond Rapids Packs 256 Cores and 1.28GB of Cache
The scale becomes even more interesting at the rack level. A server configuration can reportedly accommodate as many as 64 AGI CPUs, creating a huge pool of Arm-based processing power for cloud and AI workloads. Jagat Review says Arm claims the platform can achieve up to twice the efficiency of comparable x86 systems, although real-world performance will depend heavily on software, workload and system configuration.
Meta is reportedly the first customer and co-developer of the processor, while OpenAI, Cloudflare and AWS have also shown interest. The AGI CPU is not intended to operate in isolation, either; it can work alongside accelerators and platforms from companies including NVIDIA, Intel and AMD.
Arm has already outlined plans for future generations of the platform, suggesting this is more than a one-off experiment. With AI infrastructure becoming increasingly heterogeneous, Arm’s decision to build complete server silicon could give data-center operators another route beyond the traditional x86 model. The server CPU battlefield just gained a rather formidable new contestant.
Artikel Arm AGI CPU Targets Intel Xeon and AMD EPYC With 136 Cores pertama kali tampil pada todayinasian.com.
]]>Artikel Snapdragon X2 Elite Extreme Crushes Intel and AMD in Tests pertama kali tampil pada todayinasian.com.
]]>Signal65 compared the Snapdragon X2 Elite Extreme with Intel’s Core Ultra X9 388H and AMD’s Ryzen AI 9 465, two processors designed for modern thin-and-light Windows laptops. In Geekbench 7, Qualcomm’s chip was up to 53% faster than Intel’s processor and as much as 83% ahead of AMD’s offering. The gap became even larger in Cinebench 2026 multi-threaded rendering, where the Snapdragon chip reportedly delivered up to 87% higher performance than both rivals.
The testing becomes more interesting when power conditions enter the picture. Signal65 evaluated the Snapdragon system using Windows’ Balanced mode while running on battery, whereas the Intel and AMD machines were plugged in and configured for Best Performance. Even under those conditions, the Snapdragon X2 Elite Extreme remained ahead across the CPU and AI benchmarks tested.
Read More: Lenovo IdeaPad Vibe Takes Aim at Apple’s MacBook Neo
AI performance was another major advantage. Qualcomm’s Hexagon NPU delivered nearly twice the performance of Intel’s system and approximately 2.4 times the performance of AMD’s chip in the Procyon AI Computer Vision benchmark. The difference is partly explained by the Snapdragon’s 80 TOPS NPU, compared with 50 TOPS units in the competing processors, although the performance gap extends beyond the raw TOPS figures.
Signal65’s findings also reinforce Qualcomm’s argument that its ARM-based laptop processors can combine high performance with power efficiency. The X2 Elite Extreme features 18 Oryon CPU cores and is designed for premium Windows laptops where sustained performance and battery life are both important.
Still, benchmark victories do not automatically translate into superiority in every real-world application. Software compatibility, graphics performance, cooling, laptop design and workloads can all influence how a processor behaves outside controlled testing.
Windows Central noted that the Signal65 study essentially validates the impressive performance Qualcomm has been promising for its latest Snapdragon platform. With the X2 Elite Extreme now showing substantial advantages against Intel and AMD in independent testing, the competition for high-end Windows laptops is becoming considerably more interesting.
Artikel Snapdragon X2 Elite Extreme Crushes Intel and AMD in Tests pertama kali tampil pada todayinasian.com.
]]>Artikel Peking University Unveils 1 nm Ultra-Low Power Chip Breakthrough pertama kali tampil pada todayinasian.com.
]]>Led by Qiu Chenguang and Peng Lianmao from the School of Electronics, the research team introduced an innovative “nanogate ultra-low-power ferroelectric transistor” that operates at an astonishingly low voltage of just 0.6 volts, thanks to a novel nanogate-induced electric field concentration effect built into the device structure. At such a diminutive scale — roughly ten thousand times thinner than a human hair — the transistor’s energy consumption drops to 0.45 fJ/μm, positioning it at the frontier of low-power memory technology and heralding pathways toward sub-1 nm logic and memory integration.
Ferroelectric transistors, prized for their inherent ability to retain data without power due to reversible polarization states, have long been eyed as a leading candidate for non-volatile computing-in-memory architectures. These architectures merge storage and computational functions, offering a solution to the longstanding “memory wall” that hampers traditional systems by slowing data transfer between memory and processing units.
Read More: Volvo EX60 Sparks End of Plug-In Hybrid Era, Redefines EV Future
Yet conventional ferroelectric designs have been hamstrung by higher operational voltages — often above 1.5 volts — which limited their compatibility with mainstream logic circuits and stalled broader adoption.
By shrinking the gate electrode to an unprecedented nanometer scale and exploiting the resulting concentrated electric fields, the Peking University team bypasses this fundamental limitation. This nanogate structure effectively lowers the energy barrier for polarization switching, enabling stable memory operation at voltages far below those once thought possible. Such ultralow-voltage operation not only enhances energy efficiency but also aligns ferroelectric memory with modern chip logic levels, making the technology far more practical for integration into future computing architectures designed for high-performance artificial intelligence workloads.
The implications ripple beyond academia: this transistor could catalyze a new generation of chips that blend memory and compute seamlessly, shrinking device footprints while cutting both energy use and latency. It may help to propel post-Moore innovations where traditional scaling slows, hinting that the next frontier in computing may not just be smaller circuits but smarter, more efficient ones.
Artikel Peking University Unveils 1 nm Ultra-Low Power Chip Breakthrough pertama kali tampil pada todayinasian.com.
]]>Artikel Nvidia Halts New Gaming GPUs as It Doubles Down on AI pertama kali tampil pada todayinasian.com.
]]>The report, citing industry sources, says Nvidia has temporarily shelved its plans for new mainstream gaming graphics cards. The decision was not caused by technical limitations, but by a broader shortage of DRAM and graphics memory across the semiconductor sector. With supply constrained, the company has chosen to divert precious memory resources toward AI chips, which currently deliver far higher financial returns than gaming hardware.
Read More: Unitree G1 Humanoid Sets Record Snow Walk in Extreme Cold
One casualty of this strategy is a project known internally as “Kicker,” an updated version of the RTX 50 series. Although the design was reportedly completed, it has been postponed indefinitely because the cost of memory makes mass production less viable. At the same time, manufacturing of the existing RTX 50 lineup has been reduced, a move that is pushing up prices and making gaming cards harder to find at official retail levels.
The slowdown may stretch even further. Sources suggest that the next generation of GPUs, code-named Rubin or RTX 60, could be delayed until late 2027. If that happens, it would create the longest gap between gaming GPU generations in the company’s history.
Financial data underscores the transformation. Nvidia’s data center division has generated more than US$51 billion in revenue, while gaming now accounts for less than 8 percent of the company’s total earnings. In effect, the firm that once thrived on gamers is evolving into what observers describe as an “AI empire,” where gaming is no longer the main engine but a legacy segment.
Artikel Nvidia Halts New Gaming GPUs as It Doubles Down on AI pertama kali tampil pada todayinasian.com.
]]>Artikel Qualcomm Unveil Snapdragon X2 Plus, Redefines Midrange PC Chips pertama kali tampil pada todayinasian.com.
]]>The heart of this new platform is the third-generation Qualcomm Oryon CPU, which Qualcomm says boosts single-core performance by up to 35 % compared with the last generation while slashing energy consumption by as much as 43 %. That means laptops powered by X2 Plus chips could deliver noticeably snappier responsiveness for tasks like data analysis, creative design and long video calls — while still sipping far less energy.
Beyond raw CPU power, the Snapdragon X2 Plus packs serious AI muscle thanks to Qualcomm’s Hexagon NPU, rated at 80 TOPS (trillions of operations per second). This on-chip AI engine is built to handle complex generative and “agentic” AI tasks locally — without leaning heavily on cloud computing — enabling smooth multitasking and enhanced AI-based features embedded into Windows 11 Copilot+ experiences.
Read More: U.S.-Taiwan Chip Pact Slashes Tariffs, Spurs $250B Tech Investment
At the CES unveiling, Qualcomm’s Senior Vice President Kedar Kondap summed up the company’s vision for the platform “Modern professionals and creators want to do more, create more, and push the limits of generative AI and all-day performance. Snapdragon X2 Plus platform delivers the power, efficiency and intelligence to surpass their ambitions, making each experience more responsive and personal.”
The Snapdragon X2 Plus family includes both 10-core and 6-core models, giving OEMs flexibility to design devices that balance cost, performance and thermal efficiency. Connectivity hasn’t been overlooked either: the platform supports Wi-Fi 7, optional 5G, Bluetooth 5.4, and Qualcomm’s Snapdragon Guardian security suite, aimed at keeping data safe in mobile and enterprise environments.
Qualcomm clearly hopes this platform will accelerate adoption of Arm-based PCs running Windows 11 Copilot+. With its blend of power, AI capabilities and all-day battery life, the Snapdragon X2 Plus chips are expected to start appearing in consumer and professional laptops in the first half of 2026 — a pivotal moment that could reshape the competitive landscape against traditional x86 processors from Intel and AMD.
Artikel Qualcomm Unveil Snapdragon X2 Plus, Redefines Midrange PC Chips pertama kali tampil pada todayinasian.com.
]]>Artikel Samsung Unveils Exynos 2600, World’s First 2nm Mobile Chip pertama kali tampil pada todayinasian.com.
]]>At the core of the Exynos 2600 lies a 10-core CPU based on Arm’s v9.3 architecture, comprising a high-performance C1-Ultra prime core alongside multiple performance and efficiency cores. Samsung claims this configuration delivers up to 39 % higher CPU performance compared with its predecessor, the Exynos 2500, while maintaining power efficiency that helps extend battery life.
Graphics performance also gets a significant boost, thanks to the Xclipse 960 GPU, which offers roughly twice the compute power and up to 50 % better ray tracing performance than older Exynos GPUs. To enhance gaming visuals further, Samsung has integrated Exynos Neural Super Sampling (ENSS) — an AI-driven feature that uses resolution upscaling and frame generation for smoother, more detailed gameplay.
Read More: Snail-Inspired Swarm Robots Work Together in Tough Terrain
On the AI front, the Exynos 2600 sports a powerful neural processing unit (NPU) with 32K MAC capability, delivering more than 113 % improvement in generative AI performance compared with previous designs and enabling on-device intelligence for tasks like image editing and advanced assistance. Camera and multimedia systems are also enhanced: devices featuring this chip will support sensors up to 320 MP, advanced real-time scene analysis, deep-learning noise reduction and 8K video recording at 30 fps or 4K at 120 fps. These capabilities reflect Samsung’s push to bring flagship-grade imaging to its own silicon ecosystem.
Thermal performance, historically a weak point for Exynos chips, has been addressed with a new Heat Path Block (HPB) design that improves heat dissipation, lowering thermal resistance and helping sustained performance under heavy workloads. Memory and storage technologies such as LPDDR5X RAM and UFS 4.1 storage support ensure the chip meets demands of modern apps and media.
The Exynos 2600 is widely expected to debut in select Galaxy S26 series models in early 2026, especially in regions like Europe and Asia, offering a homegrown alternative to competitor processors. With its blend of architectural upgrades, AI prowess and efficient design, Samsung aims to reclaim leadership in premium mobile chip technology — a challenge that’s been ongoing amid fierce competition from Qualcomm and Apple.
Artikel Samsung Unveils Exynos 2600, World’s First 2nm Mobile Chip pertama kali tampil pada todayinasian.com.
]]>Artikel Apple’s A19 Becomes World’s Fastest Single-Core CPU pertama kali tampil pada todayinasian.com.
]]>What makes this feat even more impressive is its reported efficiency: the A19 is estimated to draw only ~4 watts under single-thread load, far less than the 44W consumed by the Intel 285K or 56W by the EPYC, according to PassMark’s estimates. The comparison becomes dramatic when one considers that those competing CPUs rely on active cooling solutions, while the A19 operates passively.
While the A19 leads in single-core workloads, it naturally lags behind in multi-core performance due to its limited number of cores. In multi-threaded benchmarks, desktops and server chips with higher core counts remain dominant. Still, the A19’s performance in the single-core domain underscores how far mobile SoC (system-on-a-chip) engineering has evolved—and how aggressively Apple is pushing boundaries.
Read More: Snapdragon X2 Elite Series Unveiled with 80 TOPS NPU
The architecture of the A19 was announced alongside the iPhone 17 series, with the A19 and A19 Pro built on TSMC’s N3P (3nm class) process. The chip uses a 6-core configuration, divided into 2 performance cores and 4 efficiency cores, with differences in caching and clock speeds distinguishing the regular and Pro variants. The A19 Pro is also reported to have enhancements in microarchitectural front-end bandwidth, branch-prediction logic, and increased last-level cache on the efficiency cores.
These gains not only propel Apple’s mobile chips ahead of many desktop alternatives but also reflect a paradigm shift in expectations for handheld devices. As SoCs become more powerful, the line between high-end mobile experience and desktop responsiveness narrows. The A19’s record suggests that future smartphones may handle tasks once reserved for PCs with increasing ease—and with substantially better energy efficiency.
That said, benchmarks such as PassMark measure specific workloads and should be interpreted carefully. Real-world performance depends on software, system integration, thermal limits, and sustained loads. Yet the A19’s achievement in single-thread performance is a milestone in silicon design—highlighting both Apple’s engineering ambition and broader shifts in how we think about computing power.
Sources: Tom Hardware
Artikel Apple’s A19 Becomes World’s Fastest Single-Core CPU pertama kali tampil pada todayinasian.com.
]]>Artikel Snapdragon X2 Elite Series Unveiled with 80 TOPS NPU pertama kali tampil pada todayinasian.com.
]]>Cited from Jagat Review, At the heart of these new chips is a third-generation Oryon CPU architecture, produced using a 3 nm process node. Jagat Review also said, that, Qualcomm claims that this allows the new X2 Elite series to achieve 43 percent greater efficiency compared to the previous generation. A notable leap is the integration of a Hexagon NPU rated at 80 TOPS, which positions these chips among the most powerful in their class for on-device AI tasks.
The X2 Elite Extreme (X2W-96-100) represents the flagship model. It features an 18-core CPU configuration, comprised of 12 “prime” cores clocked at up to 4.4 GHz and 6 performance cores at 3.6 GHz, with dual-core boost reaching 5.0 GHz. The chip also includes 53 MB of cache for high throughput workloads. On the graphics side, it carries an Adreno X2-90 GPU operating up to 1.85 GHz, with support for ray tracing and upgrades in performance per watt. Its memory subsystem supports LPDDR5x with bandwidth up to 228 GB/s.
Read More: The New Ryzen 7 9800X3D Can be Overclocked to 7.3 GHz
Meanwhile, the standard X2 Elite line is offered in two SKUs: X2E-88-100 (18 cores) and X2E-80-100 (12 cores). The 18-core variant shares the same core layout as the Extreme model but runs at lower clock speeds (e.g. prime cores up to 4.0 GHz). The 12-core SKU divides into 6 prime and 6 performance cores, with reduced cache (34 MB) and a slightly scaled GPU (Adreno X2-85) compared to its higher sibling.
On the GPU front, the X2 Elite series achieves a 2.3× improvement in performance per watt compared to previous-generation Adreno architecture. Support is extended to modern graphics APIs like DirectX 12.2 Ultimate, Vulkan 1.4, and OpenCL 3.0. Qualcomm markets the new chips as capable of “multi-day” battery life, although real-world performance will depend on device integration and workload.
Connectivity is robust: the chips support 5G, Wi-Fi 7, and Bluetooth 5.4. Qualcomm expects the first wave of laptops using X2 Elite and X2 Elite Extreme to arrive in the first half of 2026, with major OEMs like ASUS, Lenovo, HP, Acer, and Samsung likely among early adopters.
Analysts see the X2 Elite series as Qualcomm’s continued push into the Windows-on-ARM space—a strategic effort to challenge Intel, AMD, and Apple’s dominance in the PC sector. With AI on-device becoming an increasingly central differentiator, Qualcomm’s emphasis on NPU performance and energy efficiency may help sway OEMs and developers alike.
As these chips become available, performance benchmarks and real-world usage will be closely watched. The Snapdragon X2 Elite series could redefine expectations for union between AI, battery life, and sustained compute in future PCs.
Artikel Snapdragon X2 Elite Series Unveiled with 80 TOPS NPU pertama kali tampil pada todayinasian.com.
]]>Artikel The New Ryzen 7 9800X3D Can be Overclocked to 7.3 GHz pertama kali tampil pada todayinasian.com.
]]>The platform enabling the milestone was the Gigabyte X870 Aorus Tachyon, a board tailored for extreme overclocking. It features robust voltage delivery, optimized memory traces, and onboard tuning controls—essential assets when pushing silicon to its limits. To stabilize the system under such high frequency, voltage was raised to 1.631V, and the bus clock was tuned to 101.57 MHz. HiCookie also disabled Simultaneous Multithreading (SMT), keeping all eight physical cores active but reducing thread count to maintain integrity under high stress.
While 7.313 GHz is extraordinary, it was recorded under very light load (effectively idle conditions), rather than during sustained, real-world workloads. Tech4Gamers emphasizes that such frequencies won’t be maintainable during demanding tasks; the system would likely crash under heavy execution.
Read More: Ryzen 7 9800X3D Overclocked to 7.3 GHz in Record Run
Even so, the overclocked chip was subjected to Cinebench 2024 testing (at a reduced frequency of ~6.64 GHz), where it achieved 1,733 points, a substantial uplift compared to stock results (~1,350 points).
When contrasted with other top overclock records, this result narrows the gap with the Ryzen 9 9950X3D, which holds a record above 7.3 GHz. But for the Ryzen 7 X3D line, this new mark is notable—it highlights how far enthusiasts can push high-end desktop CPUs when combining elite components, cooling mastery, and patience.
This overclocking achievement serves as both a technical showcase and a competitive benchmark for future silicon. It underscores the importance of collaboration between motherboard engineers, cooling experts, and overclocking talents. While these results are far from everyday use, they illustrate the upper boundary of what modern CPU architecture and craftsmanship can deliver.
Artikel The New Ryzen 7 9800X3D Can be Overclocked to 7.3 GHz pertama kali tampil pada todayinasian.com.
]]>Artikel Vikram Chip Debut Marks India’s Semiconductor Breakthrough pertama kali tampil pada todayinasian.com.
]]>Developed by ISRO’s Vikram Sarabhai Space Centre and fabricated at its Semiconductor Laboratory (SCL) in Mohali, Vikram-32 is a 32-bit microprocessor designed to operate reliably under extreme conditions making it ideal for space missions, defense applications, and advanced automotive systems. Its robust architecture handles demanding environments, high temperatures, and launch vibrations, making it a trusted choice for mission-critical systems.
Insight into its performance reveals that the Vikram-32 operates at 100 MHz, supports floating-point operations, and includes in-house software tools like compilers and simulators, built for space application programming in Ada, with C support underway. Its design and testing lineage traces back to ISRO’s legacy systems and embodies the country’s stride toward semiconductor self-reliance.
This milestone is part of the broader India Semiconductor Mission, launched in 2021 to foster a complete domestic chip ecosystem from design to packaging. The government has earmarked over ₹65,000 crore (~US$8 billion) under its ₹76,000 crore PLI (Production-Linked Incentive) scheme to build fabrication units and OSAT (Outsourced Semiconductor Assembly and Test) infrastructure.
Read More: GPT-5’s “Literary” Output Reads Like AI Lingua Franca
The new OSAT pilot line in Sanand, Gujarat, is expected to produce up to 500,000 chips per day, signaling readiness for scale-up. Additionally, six semiconductor manufacturing units are already operational, and multiple new projects have received approval projection plans showing India is rapidly transforming into a semiconductor hub.
This landmark launch symbolizes the country’s evolution from being a tech consumer to becoming the creator. With Vikram-32 and the expanding domestic semiconductor strategy, India is laying the foundations for digital sovereignty, supporting autonomy in critical sectors, and positioning itself as a global innovator in microelectronics.
Looking ahead, India’s push into semiconductor design and fabrication promises not only technological independence but also the potential to supply trusted, high-reliability chips to global markets, particularly in areas where rugged performance matters most.
Artikel Vikram Chip Debut Marks India’s Semiconductor Breakthrough pertama kali tampil pada todayinasian.com.
]]>