AMD

ROCm 7.2.4 fixes the quiet layer that decides whether AI teams trust Radeon GPUs

AMD has released ROCm 7.2.4 as a stable update to its open compute stack, focused on performance and stability fixes for AMD GPU users.

Helldivers 2 tests player trust with a performance patch that may split by hardware

Helldivers 2 now has a performance-focused patch, but the real story is not just the upscaler list; it is how consistently those gains land outside ideal configurations.

ByteDance wants a way out of the chip queue slowing its AI plans

ByteDance’s reported move toward in-house processors is not a glamorous chip story, but a blunt AI logistics lesson: if you do not control the silicon, you wait in someone else’s queue.

Qualcomm’s Snapdragon C takes Windows on Arm into the $300 laptop fight

Qualcomm wants to pull Arm Windows laptops out of the premium niche and push them into the $300 range with its new Snapdragon C platform.

Mesa is cleaning up the driver keeping old Radeon cards alive

Twenty-four years after the first R300 hardware, Mesa still is not letting the old Radeon driver disappear into the archive.

Intel wants handheld gaming computers out of Ryzen’s shadow

Intel’s Arc G3 and G3 Extreme chips are an attempt to give handheld gaming PCs a more serious alternative to AMD’s dominance.

Nvidia Vera no longer looks like the weak link in an AI server

Nvidia’s Vera CPU does not topple EPYC and Xeon overnight in early Linux benchmarks, but for a first-generation custom server processor it lands close enough to change the conversation.

Linux 7.2 targets hidden speed lost inside Zen 5 chips

Cache-Aware Scheduling is not a flashy headline feature, but it is the kind of kernel change that can stop expensive Zen 5 systems from losing performance to their own cache topology.

Intel Arc heads for handhelds, with Acer Predator as the sharper comeback story

Intel’s next swing at handheld PC gaming may surface on May 28, not only through a new MSI Claw but also through an Acer Predator portable that has crawled back out of rumor territory.

Nvidia Vera pushes Arm servers closer to the center of AI data centers

NVIDIA’s Vera CPU is not yet broadly ramping, but the first Linux benchmarks already push it beyond the usual category of an interesting Arm experiment.

Radeon targets the waits and hitches players feel before frame counts matter

Advanced Shader Delivery on the Radeon RX 9070 XT targets one of PC gaming’s most irritating problems: waiting and hitching before average FPS even becomes the thing that matters.

Older Radeons get a new software lease for more demanding games

AMD is extending FSR 4.1 to older RDNA 3 Radeon cards, including the RX 7900 XTX, turning AI upscaling into a GPU-longevity issue rather than a new-card-only feature.

A gaming cache trick is moving into workstations for renders and simulations

AMD’s Ryzen PRO 9000 series is the company’s first workstation processor line with 3D V-Cache, and that shifts the debate around what extra cache can do for professional computing.

OpenAI’s new network bet is about keeping costly AI chips busy

Google is preparing a Gemini agent for Mac that organizes local files and asks for sensitive permissions.

Microsoft Opens a Door for AI Graphics Beyond One GPU Brand

Stormgate is losing online play at the end of April because Hathora, its game-server orchestration partner, is winding down gaming services after being acquired by Fireworks AI.

Nvidia is betting that AI’s next bottleneck sits between the chips

The $2 billion checks to Lumentum and Coherent mark Nvidia's largest explicit bet on optical interconnects to date.

AI chips still depend on machines that take months to build and ship

The Dutch lithography giant targets 60 EUV machines in 2026, a 36% jump from 2025.

Intel wants a second lane for AI chips stuck in the packaging queue

EMIB-T matters because Intel is not only trying to build a better process node, but to offer a way around the packaging bottleneck that has squeezed AI chips for years.

Taiwan’s chip giants bet on helium and nukes to dodge supply shocks

TSMC and UMC are quietly preparing for a world where helium is as strategically vital as oil.

Bartlett Lake’s hidden desktop potential—if Intel lets it happen

Custom BIOS patches tricking a Z790 motherboard into recognizing Intel’s Core i9-273PQE as a Raptor Lake chip took one modder just days to develop—yet the CPU itself remains locked behind OEM contracts.

Nvidia Invests $2B in Marvell

Nvidia's $2 billion investment in Marvell Technology Group marks a significant step towards expanding its AI infrastructure capabilities globally.

AutoKernel’s LLM agent loop: GPU optimization or hype repackaged?

PyTorch models may soon get their GPU kernels written by an LLM agent loop—if RightNow AI’s AutoKernel delivers on its 1.8x speedup claims without hallucinating CUDA syntax.

PS3 Emulator Breaks Through

Developers behind the RPCS3 emulator have made a significant breakthrough in emulating the PS3's Cell Broadband Engine processor, which improves performance across all games.

NVIDIA’s robot hype vs. the reality of physical AI

NVIDIA’s latest robotics push hinges on a claim its own partners can’t consistently prove: that virtual training translates to real-world performance.

Dell’s 50 TOPS micro PC isn’t just small—it’s a power play

A 50 TOPS AI desktop that sips 100W via USB-C and fits in a drawer wasn’t supposed to exist—until Dell’s forced a rethink of office hardware tradeoffs.

Nvidia’s NTC demo: 85% VRAM cut or just clever repackaging?

Nvidia’s GTC demo cut a 6.5GB texture set to 970MB using neural decompression—a trick that sidesteps traditional compression’s fidelity tradeoffs.

Intel’s Core Ultra 200S rewrites the rules for mid-range CPUs

Intel’s Core Ultra 270K Plus outperforms AMD’s Ryzen 9700X by up to 20% in AI and rendering tasks, according to early benchmarks.

Intel’s $200 18-core CPU just broke AMD’s budget gaming math

Intel’s new **18-core, $199 Core Ultra 5 250KF Plus** doesn’t just match AMD’s Ryzen 5 7600 in gaming—it embarasses it in productivity by up to 40%.

Arm’s quiet coup: 90% of AI servers by 2029?

Omdia’s projection that 90% of custom AI servers will run Arm by 2029 hinges on a quiet revolt: hyperscalers are now designing chips like they write software—iteratively, aggressively, and with zero sentimentality for x86.

Memory Spend Surge: The Hidden Cost of AI's Appetite

Nvidia’s behind-the-scenes memory deals could reshape cloud economics for years, analysts warn.

U.S. AI chip whiplash: Who’s left holding the bag?

Chris McGuire, the ex-Trump NSC director now at the Council on Foreign Relations, calls the latest AI chip restrictions *‘a policy written in erasable ink’*—and the ink’s smudging fast.

Nvidia’s 288-GPU flex hides the real AI benchmark war

AMD and Intel’s MLPerf submissions quietly abandoned the GPU arms race—leaving Nvidia’s 288-H100 cluster as the lone monument to raw, unaffordable speed.

Nvidia’s $2B Marvell bet: Locking in AI’s plumbing

Marvell’s stock jumped 12% on the news—because $2 billion buys more than chips; it buys Nvidia a direct line to the data center’s spine.

Windows 11’s NPU transparency: AI PCs get real usage data

Microsoft’s Task Manager now quantifies NPU usage down to the millisecond, ending years of guesswork over AI workloads on Windows 11.

Nvidia Loses Ground

Nvidia's market share in China has fallen to 55%, a significant drop from its previously claimed high of 95%.

Nvidia’s background shader hack cuts game load times—no magic, just math

Nvidia’s new *Auto Shader Compilation* beta silently preps shaders post-driver update, trimming *several minutes* off titles like *Alan Wake 2* and *Cyberpunk 2077*—but first-time installs remain a bottleneck.

OneComp: The One-Line Promise That Might Actually Deliver

OneComp’s open-source framework promises to cut AI compression workflows from days of manual tuning to a single command—but the real test lies in production.

Arm’s 136-core AGI CPU lands in China—hype or hardware edge?

Arm’s Neoverse V3-based AGI processor—136 cores, no US export restrictions—just cleared a path to China’s data centers, where Nvidia’s A100s still dominate.

Ollama’s MLX speedup: Apple Silicon’s AI edge or just hype?

Apple’s MLX framework now powers Ollama’s local models—yet the release omits benchmarks, benchmarks, or even a hint of AMD/Intel support.

DLSS 4.5’s MFG 6X is AI’s sneakiest power move yet

MFG 6X doesn’t just double down on AI frames—it sextuples them, turning NVIDIA’s DLSS from a performance boost into a full-blown rendering revolution.

Japan’s 1.4nm AI chip: Hype or real semiconductor independence?

Rapidus’ first 1.4nm customer isn’t a smartphone giant or hyperscaler—it’s Fujitsu, betting on an AI inference chip Japan’s own fabs can’t yet mass-produce.

Lab-grown retinas reveal how cones fight degeneration—with limits

Twenty thousand lab-grown human retinas—each a cluster of cells no wider than a sesame seed—just rewrote a key chapter in how cone photoreceptors resist degeneration.

AMD’s 208MB cache gambit leaves Intel scrambling

Intel’s cancellation of the Core Ultra 9 290K Plus leaves AMD’s 208MB cache monster without a high-end rival.

NVIDIA’s ProRL: Rollout-as-a-Service or Just Another Bottleneck Fix?

Leaked docs show Anthropic’s next model boasts scores 30% above Opus—but details on real-world use remain scarce.

Nvidia’s AI art war: Why players are sharpening the pitchforks

New Blood Interactive’s CEO just turned Nvidia’s DLSS 5 into a referendum on AI’s role in game art.

TurboQuant: Google’s 6x AI memory shrink is real—but Pied Piper isn’t

Google’s TurboQuant shrinks AI memory by 6x in lab tests, but the real test is whether it escapes the demo stage.

Arm breaks its own rules with 136-core AI chip

Arm debuts 136-core AI chip, shifting from licensing to silicon.

Intel’s 32GB Arc Pro GPUs: AI’s New Budget Workhorse

Intel’s Arc Pro B70 and B65 GPUs pack 32GB of RAM—double the capacity of earlier Battlemage teases—while undercutting rivals on price for AI inference workloads.

LG’s battery-saving laptop screens are a rare win for users

LG's new LCD panel saves juice by dynamically adjusting refresh rates down to 1Hz.

Arm’s first solo chip: hype meets hardware reality

Meta will deploy Arm’s first in-house CPU in its AI datacenters before year-end, marking the chip designer’s shift from licensing to production.

SK hynix’s $8B ASML bet reshapes memory for AI and PCs

SK hynix’s $8 billion ASML order is the largest in semiconductor history, targeting 30 EUV machines for HBM and advanced DRAM by 2026.

DLSS 5’s Quiet Revolution: Why Players Should Care

Sony shuts Bluepoint Games, silencing a 20-year legacy.

Arm moves into AI hardware and tests the trust that built its chip empire

Arm Holdings, for decades synonymous with licensing processor architectures, is now manufacturing its own semiconductors — a direct assault on the very business model it helped create.

Nvidia is betting AI agents will live inside software, not replace it

Jensen Huang called the claim that AI will destroy software 'ridiculous' during a public address.

AMD’s 2026 data center push: specs vs. real-world muscle

AMD’s Helios platform and MI500 GPUs aim to unseat NVIDIA’s AI dominance by 2027, but the battle hinges on software, not just silicon.

Gimlet Labs Solves AI Bottleneck

Gimlet Labs' $80 million Series A funding round is a significant development in the AI industry, with the company's technology enabling AI inference to run simultaneously across multiple hardware platforms.

More cores, less cache: Cloudflare shows the new cost of a faster edge

Cloudflare's new server generation is not a space story, but an architecture lesson: more cores matter only if the request stack can live with less cache.

Intel's Arrow Lake Refresh Forces AMD to Sweat

Intel has finally found AMD's weak spot: the assumption that performance justifies premium pricing.

PlayStation’s AI frames won’t fix your 30fps blues

Sony’s lead PlayStation architect just confirmed AI frame generation is coming—but don’t expect it to turn 30fps into 60fps overnight.

AMD’s Agentic AI: The PC’s next killer app or just another hype train?

AMD’s latest AI push arrives with a new buzzword—*Agentic*—but zero concrete details on hardware or release timelines.

FSR 4.1 turns AMD’s ray tracing from ‘meh’ to ‘PS5 Pro who?’

AMD’s latest FSR update quietly closes the gap between PC and PS5 Pro ray tracing without a single hardware change.

Qualcomm’s ARM ambush: Is Intel’s laptop crown slipping?

Qualcomm’s Snapdragon X2 Elite Extreme smokes Intel’s Core Ultra X9 388H in Geekbench—ARM’s boldest laptop play yet.

The home lab is leaving the server closet for small computers and cloud glue

Nvidia is selling AI a trillion-dollar future before Rubin reaches the test bench

Jensen Huang didn't just unveil new silicon at GTC — he set a trillion-dollar order target through 2027, a benchmark that will determine whether Nvidia maintains dominance or faces its first serious friction between ambition and reality.

Nvidia wants AI buyers to stop counting chips and start buying whole racks

Nvidia's GTC 2026 didn't unveil a new star chip — it unveiled an entire galaxy: seven distinct processors in a modular system that shifts focus from individual GPUs to complete AI factories in racks.

Project Helix’s FSR Diamond: Xbox’s Raytracing Gamble Pays Off?

Microsoft’s neural rendering gambit with FSR Diamond could double raytracing—if devs care more about silicon than gamers’ framerates.

The Xbox One Jailbreak: A Victory for Digital Preservation

Xbox One's boot ROM yielded to voltage glitching after 13 years, turning a piracy target into a preservation victory.

AI’s copper problem just got a 3.2Tb/s optical fix—if it works

Meta’s AI training clusters already spend 40% of their budget moving data between GPUs—now they’re betting optical cables can cut that waste.

Quantum hardware finally meets its Möbius moment

Quantum chips finally model a twisted Möbius molecule—proving they can crack chemistry’s toughest simulations.

Nvidia wants AI data centers to move data with light, not copper

PETRUSHKA is the first mental-health AI to prove itself in a randomized clinical trial: patients were 40 percent less likely to drop their antidepressant regimen within eight weeks.

Apple’s laptop chip now looks desktop-class, but only inside a benchmark

An 18-core CPU configuration in a 16-inch MacBook Pro just clocked 29,233 in Geekbench 6 multi-core tests.

AMD’s $999 combo deal: fast gaming CPU meets market reality

MSI’s MAG X870 motherboard—typically a $220 part—now ships with AMD’s gaming-dominant Ryzen 7 9850X3D and 32GB of DDR5-6000 RAM for $999 total, undercutting DIY builds by 16%.

Apple’s M5 chips: GHz flash, real-world fizz

Apple’s M5 Pro and M5 Max chips squeeze out a 4.61 GHz Super core, but benchmarks and pricing remain conspicuously absent.

Arm laptops just got a big number, but games still need real proof

Snapdragon X2 appears to post a major single-core jump in early Geekbench results, but that number does not yet prove real Windows gaming performance.

Micron’s 2TB RAM Push: The Quiet Upheaval in AI Server Costs

Micron’s first 256GB LPDDR5X samples arrive as NVIDIA’s GB200 and AMD’s Turin push RAM needs higher.

OCuLink eGPU docks double speed — but at what cost?

Early adopters report 30–40% faster renders, but most laptops still lack OCuLink ports.