Every story tagged GPU Architecture, curated for CIOs and IT leaders — ranked by source credibility, engagement, and freshness.
2 stories · open in the command center
Nvidia has increased the mobile RTX 5070's VRAM from 8GB to 12GB, addressing a critical performance bottleneck for AI and graphics-intensive workloads, but at a prohibitive cost—a 71.5% price premium ($1,199 vs $699) driven by ongoing memory supply constraints. This pricing situation signals that enterprise IT leaders should expect significant budget inflation for GPU-based computing initiatives, whether for AI model deployment or graphics workstations, and may need to explore alternative architectures or multi-GPU strategies. The constrained supply and pricing power of memory manufacturers will likely persist across the industry, making GPU procurement planning increasingly complex for organizations modernizing their infrastructure.
Google's upcoming Tensor G6 chip for the Pixel 11 will feature improved CPU performance with latest Arm C1 cores, but will reportedly use an outdated 2021 PowerVR GPU architecture similar to its predecessor, potentially limiting gaming performance and computational workloads critical for enterprise and mobile-first organizations. This performance gap raises concerns about Google's hardware competitiveness and could impact IT purchasing decisions for organizations standardizing on Pixel devices for productivity and AI-enabled workflows. The reliance on older GPU technology, compounded by outdated drivers lacking Vulkan 1.4 support, may necessitate workarounds for organizations leveraging graphics-intensive applications or mobile ML inference.