GPUSmith
Articles (English) - Page 1 / 3
Llama 4 Hardware Requirements: VRAM, Nodes and Costs (2026)
Covers Llama 4 Scout and Maverick VRAM needs by quantization, multi-node GPU setups, per-token API pricing across seven providers, and 2026 cloud GPU rental costs.
DGX vs HGX vs NVL72 vs MGX: NVIDIA Platform Comparison 2026
A 2026 analyst comparison of NVIDIA DGX, HGX, GB200/GB300 NVL72, and MGX covering specs, pricing, MLPerf benchmarks, hyperscaler deployments, and how to choose.
NVIDIA SK Group $500 Billion AI Deal Explained (2026)
Explains the July 2026 NVIDIA SK Group $500 billion AI deal, covering SK hynix HBM4 supply, SK Telecom's 2GW AI factory, market reaction, and what it means for GPU buyers.
NVIDIA Vera CPU Price and Specs: What to Know
A 2026 pricing and specifications guide to NVIDIA's Vera CPU: 88 Olympus cores, $5,000 to $20,000+ per-chip pricing, Vera Rubin NVL72 rack specs, Grace CPU comparison, and release timeline.
AWS GPU Instance Pricing 2026: P6-B300, P6-B200, P5 Costs
Complete 2026 guide to AWS GPU instance pricing covering P6-B300, P6-B200, P5, P5e, and P5en rates, the 2026 price hikes, Capacity Blocks, and Azure/GCP/Oracle cost comparisons.
AMD Anthropic Deal: Inside the $5 Billion MI450 Partnership
A 2026 analyst breakdown of the AMD Anthropic deal: the 2GW MI450 rollout, AMD's up to $5B equity stake, MI450 vs Nvidia Rubin specs, and Anthropic's multi-vendor compute strategy.
Intel Crescent Island AI GPU Explained: Xe3P and LPDDR5X
A 2026 explainer on Intel's Crescent Island AI GPU covering Xe3P architecture, 160GB to 480GB LPDDR5X memory, 350W design, release timeline, and comparisons to Nvidia Rubin and AMD MI400.
AMD Instinct MI455X Specs, Price, and Azure Deal Explained
Full 2026 breakdown of AMD Instinct MI455X specs, Helios rack architecture, pricing, MI400 series lineup, Nvidia Vera Rubin comparison, and the Microsoft, Oracle, OpenAI, Meta and Anthropic deals.
Hyperscaler AI Capex 2026: Amazon, Microsoft, Google, Meta
A 2026 data-report on Amazon, Microsoft, Alphabet and Meta AI capex, covering Amazon's $200 billion guidance, Microsoft's $190 billion fiscal year plan, and Goldman Sachs, Morgan Stanley and IEA forecasts through 2028.
H100 GPU Cloud Pricing Comparison: AWS vs GCP vs Azure
A 2026 pricing comparison of NVIDIA H100 GPU cloud rentals across AWS EC2 P5, Google Cloud A3, and Azure ND/NC H100 v5, plus CoreWeave, Lambda, Oracle, and other neoclouds.
NVIDIA DGX B300 Price: Specs and Lead Times Guide
A 2026 pricing guide to the NVIDIA DGX B300, covering the $300,000-$350,000 base system price, DGX B200 comparisons, GB300 NVL72 rack costs, cloud rental rates, and 8-20 week lead times.
Why AI GPU Prices Are Rising in 2026: HBM Shortage Explained
A 2026 analysis of why AI GPU prices are rising, covering the HBM memory shortage, TrendForce and IDC price data, Samsung, SK Hynix, and Micron sold-out capacity, and RTX 5090 pricing.
NVIDIA HGX B300 Price, Specs, and Lead Times Explained
2026 pricing guide to the NVIDIA HGX B300: baseboard, OEM server, and DGX B300 prices from $420,000 to $714,000, cloud rates from $2.63/hr, specs versus HGX B200 and GB300 NVL72, and lead times.
What Is NVIDIA Kyber? Rack-Scale System Explained (2026)
Explains what NVIDIA Kyber is, how it differs from the Oberon rack, Rubin Ultra GPU specs, 800 VDC power architecture, and the SemiAnalysis-reported 2028 delay as of July 2026.
CUDA Compute Capability by GPU: The Complete sm_75-sm_121 Chart
A 2026 data report mapping every NVIDIA GPU to its CUDA compute capability, from Turing sm_75 through Blackwell sm_121, with PyTorch and TensorFlow arch-flag guidance.
Eviden Bull: French State Ownership Explained for AI Buyers
How the French State acquired Bull (formerly Eviden Advanced Computing) from Atos for up to 404 million euros in 2026, what HPC, quantum, and AI assets transferred, and what stayed with Eviden.
NVIDIA GPU Export Restrictions: Current US Chip Controls 2026
A July 2026 analysis of NVIDIA GPU export restrictions to China: the H20 and H200 licensing rules, the Blackwell ban, the 15% revenue share deal, Entity List gaps, and smuggling cases.
HBM3e vs HBM4: What Changes for AI GPU Memory in 2026
Compares HBM3e and HBM4 memory specs, bandwidth, power, and 12-hi/16-hi capacity, covering JEDEC's standard, NVIDIA Rubin, AMD MI400, and 2026 pricing and supply data.
NVIDIA GPU Form Factors Explained: SXM vs PCIe vs NVL72 vs MGX
A 2026 analyst guide to NVIDIA GPU form factors: SXM vs PCIe power and bandwidth differences, HGX vs DGX vs MGX platforms, GB200 NVL72 rack architecture, and Grace Hopper/Blackwell superchips.
NVLink 5 vs NVLink 6: Bandwidth, Domains and Cluster Design
A 2026 analyst comparison of NVLink 5 and NVLink 6 covering per-GPU bandwidth, NVL72 rack topology, InfiniBand scale-out, Rubin platform adoption, and financial data.
Which Server OEMs Design Their Own NVIDIA Servers?
A 2026 analyst breakdown of which NVIDIA server OEMs (Dell, Supermicro, HPE, Lenovo) design their own hardware versus Taiwan ODMs like Foxconn and Quanta, with IDC market-share data, revenue figures, and named case studies.