Cloud Engine-Cost-effective AI GPU Rental Platform

 A100 pcie40G - $ 200/card/month;BMS:8*Ascend 910B $ 3000/month  8*4090D - $ 1000/month  8*4090 - $ 1200/month  8*A100 pcie40G - $ 1800/month  8*A100 pcie80G - $ 4000/month  8*A100 nvlink80G - $ 5400/month  8*A800 nvlink80G - $ 5400/month  8*H20 - $ 6000/month  8*L20 - $ 2000/month  8*L40 - $ 2400/month  8*L40S - $ 2500/month  8*H100 - $ 13000/month  8*H200 - $ 20000/month  8*B200 - $ 30000/month  8*B300 - $ 48000/month 

Global GPU Computing News【20260607】

  1. NVIDIA launches RTX Spark chip, officially entering the PC market. At the GTC Taipei conference on June 1, NVIDIA CEO Jensen Huang unveiled the RTX Spark superchip for Windows PCs, featuring a Blackwell architecture GPU with 6,144 CUDA cores. This move marks NVIDIA's direct entry into the PC processor market, competing head‑on with Intel and AMD.


  1. NVIDIA's next‑gen AI platform Vera Rubin enters full production. At the same conference, Huang announced that the Vera Rubin AI platform has entered full production. Built on TSMC's 3nm process and designed for AI agents, the platform is 1.8 times faster than x86 architecture processors and is expected to hit the market this fall.


  1. Google and SpaceX sign a roughly 30billioncomputingpowerdeal.AccordingtoanagreementformallysignedonJune5,Googlewillleaseabout110,000NVIDIAGPUsfromSpaceXoverthenextseveralyears.StartinginOctoberthisyearthroughJune2029,GooglewillpaySpaceXupto920 million per month, bringing the total contract value to approximately $30 billion.


  1. LG Group places massive NVIDIA GPU order. South Korea's LG Group announced plans this week to procure 10,000 NVIDIA Blackwell architecture GPUs, with total investment in supporting equipment and infrastructure expected to reach trillions of Korean won. It is the largest single NVIDIA GPU purchase in LG's history, with all funds directed toward its AI transformation strategy.


  1. NVIDIA partners with multiple companies to expand edge AI ecosystem. Aptiv announced an expanded collaboration with NVIDIA, based on the Jetson platform to accelerate edge AI mass production. Additionally, robotics company Unitree appeared at GTC and showcased collaborative achievements in embodied AI with NVIDIA, further extending AI applications at the edge.


  1. Intel unveils multiple AI infrastructure achievements, mounting a full‑scale counterattack. At Computex 2026, Intel demonstrated rack‑scale AI infrastructure for disaggregated inference and the Xeon 6+ processors. These moves form a clear counter‑offensive against NVIDIA's entry into the PC market, aiming to solidify Intel's influence in AI and data centers, signaling an all‑out escalation of competition among chip giants.


  1. Domestic GPU computing power ecosystem accelerates development. In domestic computing power construction, KEHUA Data announced this week that it has established deep strategic partnerships with leading domestic GPU chipmakers including M‑Thread, Biren, Enflame, and Tsing Micro. Meanwhile, Haizhi Technology signed an agreement with Hygon to jointly develop a full‑stack solution combining “domestic computing power + graph‑model intelligence,” promoting the deployment of localized AI technologies in key industries.


  1. China's intelligent computing power enters a unified monitoring system. Media reports this week noted that 137 PFLOPS of intelligent computing power has been incorporated into the national unified monitoring system, accounting for about 72% of the country's total intelligent computing capacity. At the same time, local governments such as Chengdu have introduced policies to accelerate the construction of the national integrated computing network (Chengdu‑Chongqing hub) and the Tianfu data center cluster, strengthening the computing infrastructure foundation.


  1. Intensifying global semiconductor supply chain competition. TSMC Chairman C.C. Wei recently stated that customer demand is astonishingly high and that the company's capacity has reached its limit. Meanwhile, NVIDIA CEO Jensen Huang traveled to South Korea this week to meet with partners including Samsung and SK Hynix, promoting diversified cooperation in the chip supply chain to alleviate capacity bottlenecks.


  1. HBM memory supply remains tight, driving high prices for high‑end computing power. NVIDIA CEO Jensen Huang this week explicitly denied rumors of “cutting HBM usage,” stating that the company will use a “substantial amount” of HBM memory while acknowledging supply constraints. Driven by supply‑demand imbalances, the one‑year lease price of H100 GPUs has risen 38.2% in the past six months, while B300 cloud lease prices have surged 73.8%, indicating that high‑end computing resources remain scarce.


Sales:+8619886543278 (WhatsApp)
Email:yqtxben@163.com
Address:2nd Floor, Block A, Garden City Digital Building, Nanshan District, Shenzhen City, China.
Contact US