CONNECT WITH US
NEWS TAGGED TRAINING
Friday 8 May 2026
Analysis: Anthropic's xAI deal puts Musk's GPU efficiency problem in focus

Elon Musk's latest move to team up with Anthropic in a major computing-capacity agreement captures the logic of "the enemy of my enemy...

Friday 8 May 2026
Compal partners with Verda to supply liquid-cooled GPU servers for sovereign AI deployments
Compal formed a strategic partnership with European AI cloud provider Verda to accelerate deployment of next-generation AI infrastructure across Europe and the Asia-Pacific region,...
Friday 8 May 2026
Column: The sim-to-real problem—why robots that pass every test still fail on the floor
Simulators are robotics' most seductive shortcut. Spin up a virtual environment, generate millions of training trajectories at near-zero cost, tune the weather, reposition the obstacles,...
Thursday 7 May 2026
MediaTek opens AI R&D data center in Taiwan with Nvidia DGX SuperPOD
MediaTek announced the opening of a new research and development data center at the Tongluo Science Park in Miaoli, Taiwan, aimed at supporting growing demand for edge AI and cloud...
Tuesday 5 May 2026
Tesla AI5 dual sourcing may not mean equal weight for Samsung
Tesla CEO Elon Musk has confirmed that the company's in-house AI5 chip has completed design tape-out and entered a critical pre-production validation stage. The move has drawn renewed...
Tuesday 5 May 2026
South Korea eyes memory-led AI order against Nvidia
As AI shifts from training to inference and from single-task use to multi-agent collaboration, South Korea's semiconductor industry is seeking to recast the market around memory rather...
Thursday 30 April 2026
Google's split TPU chips signal shift from universal to specialized AI accelerators
Google has unveiled its eighth-generation Tensor Processing Units (TPUs) by splitting them into two distinct chips: the training-focused TPU 8t and the inference-optimized TPU 8i....
Thursday 30 April 2026
Unisplendour, H3C gain from AI buildout, monetization in focus

China's AI infrastructure race is entering a phase where scale alone is no longer enough. As large models move from training to real-world...

Thursday 30 April 2026
OpenAI accelerates Stargate build, surpasses 10GW US goal, and expands community-focused approach
OpenAI announced it has exceeded its initial Stargate commitment by securing more than 10GW of AI infrastructure in the US ahead of schedule, adding over 3GW in the past 90 days. Stargate...
Thursday 30 April 2026
China photonics chipmaker Lightelligence lists in HK, CPO commercialization in focus
As generative AI drives a sharp rise in computing demand, traditional electrical interconnect architectures are increasingly constrained by power consumption and density limits. On...
Wednesday 29 April 2026
China's GPU IPO wave meets its hardest test: cluster stability
China's GPU startups, including Moore Threads, MetaX, and Biren, are gaining ground in the AI accelerator market. But as demand shifts from training to inference, cluster stability,...
Wednesday 29 April 2026
Osense develops baseball hawk-eye system to capture US$10 billion sports tech market
Taiwanese artificial intelligence (AI) software company Osense Technology is targeting the rapidly growing sports technology market — valued at more than US$10 billion —...
Wednesday 29 April 2026
Nvidia's LPX cabinet and Foxconn's supply lead reshape inference-era AI infrastructure
The shift from model training to real-time inference, driven by open-source agent applications, is reshaping global data center design and supplier dynamics, with implications for...
Wednesday 29 April 2026
CPUs reclaim the core of AI architecture as multicore trend tightens substrate supply
Workloads are shifting from training to inference. In this transition, CPUs are regaining a central role in coordinating diverse computing tasks, significantly boosting their importance...
Tuesday 28 April 2026
Indian startup targets AI inference opportunity with full-stack compute platform
While global AI infrastructure investment remains concentrated around massive GPU clusters for training frontier models, Indian startup Turiyam.ai is betting on a different commercial...