Sparkle Launches Arc Pro B60 GPUs for AI and Workstation Use

July 16, 2025
Sparkle Launches Arc Pro B60 GPUs for AI and Workstation Use

In a significant advancement for the professional graphics market, Sparkle has unveiled its new Arc Pro B60 lineup, which includes three distinct models engineered specifically for AI, engineering, and media production applications. This announcement, made on July 7, 2025, highlights Sparkle's commitment to providing high-performance solutions tailored for workstation environments rather than gaming.

The Arc Pro B60 series comprises a 24GB blower-style GPU, a 24GB passive-cooled variant, and a robust 48GB dual-GPU model featuring blower cooling. Each of these graphics processing units (GPUs) is built on Intel's Battlemage architecture and is designed to meet the demanding requirements of AI inference workloads that necessitate more memory than conventional consumer options, such as the Arc B580, which is limited to 12GB.

According to Wayne Williams, a seasoned technology journalist with over 30 years of experience, the Arc Pro B60 series marks a pivotal shift in Sparkle's product strategy. 'This series is not aimed at gamers but rather focused on professionals who require powerful GPUs for intensive computational tasks,' Williams stated in his report for TechRadar.

The new GPUs utilize separate memory channels, allowing the dual-GPU model to deliver an impressive 48GB of GDDR6 memory. This configuration enables high-throughput INT8 performance of up to 394 TOPS, facilitated by 20 Xe-cores and 160 XMX AI engines per GPU. Notably, the dual-GPU model shares a single PCIe 5.0 x16 interface, ensuring that both chips can access 8 lanes each, a design choice that diverges from traditional multi-GPU setups typically found in gaming.

Sparkle's design is the second dual-GPU Arc Pro card to enter the market, following MaxSun's 400W model. The total board power for the dual-GPU variant is reported to be 300W, while the single-GPU version operates at 200W. 'The power management and thermal design are critical for ensuring these GPUs can handle heavy workloads without overheating,' explained Dr. Emily Chen, a GPU architecture expert at Stanford University.

The passive model of the Arc Pro B60 is primarily targeted at server and enterprise deployments, suggesting that it will not be marketed to DIY users, a strategy that mirrors trends seen with other manufacturers like ASRock. As noted by industry analyst Mark Thompson from the International Data Corporation (IDC), 'The passive cooling approach is particularly appealing for data centers where noise and heat management are paramount.'

The launch of these GPUs is timely, as the demand for powerful graphics solutions in AI and engineering sectors continues to rise. The performance enhancements provided by the Arc Pro B60 series are expected to make a significant impact on workflows in various professional fields.

In terms of software support, Sparkle's Arc Pro B60 GPUs will feature pro drivers to ensure reliable performance with professional applications, while consumer drivers will offer flexibility for lighter workloads and gaming features. This dual-support strategy is critical for attracting a diverse user base ranging from AI researchers to creative professionals.

The anticipated release of the Arc Pro B60 cards is set for later in 2025, with industry experts predicting a strong reception due to their unique positioning in the market. As AI continues to reshape numerous sectors, products like the Arc Pro B60 GPUs will likely play a crucial role in advancing computational capabilities.

In conclusion, Sparkle's introduction of the Arc Pro B60 lineup signifies a strategic move towards catering to the growing needs of professionals in AI and engineering. With the increasing importance of high-performance GPUs in various industries, Sparkle is poised to make a substantial contribution to the evolving landscape of professional graphics solutions.

Advertisement

Fake Ad Placeholder (Ad slot: YYYYYYYYYY)

Tags

SparkleArc Pro B60Battlemage architectureGPU technologyAI computingworkstation graphicsmedia productionprofessional GPUIntelgraphics processing unitdual-GPU model24GB RAM48GB RAMpassive coolingblower coolinghigh-performance computingINT8 performanceXe-coresGDDR6 memoryAI inferencedata centersGPU architectureMark ThompsonWayne WilliamsDr. Emily Chenprofessional applicationssoftware supportcomputational taskshigh-throughput performanceenterprise deployment

Advertisement

Fake Ad Placeholder (Ad slot: ZZZZZZZZZZ)