GPU & Compute Supply

Duos Technologies has agreed five-year hosting orders with neocloud firm Axe Compute covering 55 MW of AI data centre capacity across multiple US sites.

Castrol's direct-to-chip fluids have passed NVIDIA's technical review for AI factory infrastructure, covering thermal stability, corrosion resistance and supply-chain

NVIDIA has released a 30-billion-parameter open model and an open source routing library aimed at multi-agent AI deployments across edge, cloud and

Lenovo's Q1 FY2026/27 revenue rose 43% year-on-year to an all-time group record, with AI-related sales growing 60% to $9

ExxonMobil says machine learning and seismic imaging have surfaced four new exploration targets in Guyana as the country eyes 1.3m bpd output by 2027.

Firebird has launched what it calls the CIS region's largest AI factory in Armenia, with NVIDIA set to invest and a 2 GW global

The NASDAQ-listed wafer-processing equipment maker hit $292.9m in Q2 revenue and lifted its 2026 guidance to up to $1.175bn.

The EDA and TCAD software vendor posted $17.8m in Q2 revenue, its first non-GAAP operating profit in nearly two years, backed by Micron

Sandisk showcased next-generation NAND and High Bandwidth Flash at FMS 2026, targeting AI inference workloads in data centres and cloud.

AMD has agreed to acquire Toronto-based Taalas, a specialist in AI inference silicon, to strengthen its Instinct GPU platform and full-stack AI portfolio.

Oxylabs analysis of 850,000 US job postings finds AWS in 30% of listings, with cloud platforms cited in 42% of all tech roles.

KAYTUS says it cut a typical 180-day deployment cycle by 60%, delivering a high-density compute cluster for a Japanese cloud provider in 72

Lovable will run latency-sensitive workloads on Cerebras' Wafer-Scale Engine as the two firms jointly explore faster AI software-creation experiences.

The Australian neocloud said the agreement, with an unnamed global AI platform, will see revenue begin in Q1 2027 and lifts contracted capacity to 120MW.

AMD's data centre segment hit $6.7bn in Q2 2026, up 107% year-on-year, as EPYC and Instinct GPU demand drove record group

The two semiconductor firms demonstrated a PCIe Gen 6 switch-and-SSD architecture at FMS 2026, targeting AI, HPC and cloud data centre workloads.

Tower Semiconductor beat its own records in Q2 2026 and is guiding for a further 13% sequential rise as silicon photonics demand surges.

Alibaba Cloud's Qwen3.8-Max activates 95 billion of its 2.4 trillion parameters via sparse MoE architecture, targeting efficiency at frontier scale.

The IP licensing firm also renewed its Google deal and signed RPX to cover 10 new e-commerce customers in Q2 2026.

The Nasdaq-listed Finnish quantum firm reported H1 2026 revenues of EUR 8.9m and initiated full-year guidance of EUR 42m to EUR 47m.