The Graphics Processing Unit (GPU) market was valued at USD 20.62 billion in 2019 and is expected to reach USD 104.71 billion by 2025, at a CAGR of 31.1% over the forecast period 2020 - 2025. GPUs , which have been driven by the sales of personal computers, gaming laptops and other portable devices in the past have undergone several iterations and emerged as a suitable solution for several emerging applications including high-performance computing, Blockchain, AI/ML, among others. GPU vendors have been realizing increased revenues, owing to these emerging applications. For instance, AMD, in FY2018, reported more than USD 1.2 billion of revenue owing to the demand for its datacenter GPU products.
- As enterprises are undergoing digital transformation by embracing the cloud, big data and also AI, computer technology companies are focusing on building data centers around workloads that support AI and deep neural networks (DNNs). Data centers that support GPU servers are likely to be preferred as they provide advantages such as: higher efficiency and performance, low power consumption for advanced workloads, occupies less floor space. With Hyperscale data centers expected to grow to 628 by 2021(according to Cisco), the opportunity for GPU providers is significant.
- Moreover, emerging applications such as Augmented/Virtual Reality, Robotics, crypto currency mining, and Autonomous vehicles are also providing significant opportunities
- In an effort to leverage the opportunity provided by Autonomous Vehicles, Nvidia has been actively involved in the development of GPU-powered AI platforms and partnering with automotive giants. After its first launch of AI-based supercomputer platform Drive PX in 2015, Nvidia recently came up with a new iteration of the platform named Pegasus which can be utilized to power Level 5 autonomy. Built on Nvidia’s CUDA GPUs, the new platform is designed to support fully autonomous cars without pedals, steering wheels or mirrors, while increasing the computing speed by 10 times and lowering the power consumption by 16 times.
Scope of the Report
Graphics Processing Unit (GPU), a programmable logic chip used for rendering images, animations, videos and also for general computing purposes are studied. Applications in addition to personal computers, laptops among others, emerging applications such as AR/VR, high-performance computing, AI, ML, blockchain, cryptocurrency mining, autonomous driving and vavigation (vehicles, robotics)are considered as a part of the study.
Key Market Trends
Healthcare Applications to Create Significant Opportunities
- It is estimated that, an average hospital generates 50 petabytes of data annually, through medical images, clinical charts and sensors, including operational and financial sources. According to the National Institutes of Health(U.S.), the opportunities for deep learning in healthcare is expected to increase substantially and is likely to extend beyond medical image analysis.
- AI technology has the potential to perform operations such as automatic lesion detection, diagnosis suggestions, and compose preliminary radiology reports. Medical equipment manufacturers realized the opportunity underlying the huge amount of data generated and are involving in partnerships with technology companies to make deep learning a core part of their image processing and analytical workflow.
- GE Healthcare, recently partnered with Nvidia with an aim to use the GPU accelerated deep learning solutions to design more sophisticated neural networks for healthcare and medical applications. GE Healthcare is keen on. applications ranging from real-time condition assessment to point-of-care interventions to predictive analytics for clinical decision-making.
North America to Loose Share to Asia-Pacific
- North America, which has a strong data center footprint compared to other regions, and increasing emphasis on AI and deep learning across the region, is expected to comprise of a lion’s share of the market.
- In addition, to the hyperscale data centers, the region is also a large market for medical devices. According to the U.S. Department of Commerce, North America represented about 40% of the global medical device market in 2017. The government data also revealed that one of the components fueling the growth includes electro-medical equipment, which includes a variety of powered devices such as pacemakers, patient-monitoring systems, MRI machines, diagnostic imaging equipment (including informatics equipment) and ultrasonic scanning devices. Such evidence prompt the scope for huge data generated in the region, which demands high computational speeds that can be provided by GPUs.
- Moreover, the region also has a strong share of gaming laptops which consumer GPUs. Bitcoin mining is also beginning to gain a foothold in the United States. However, considering the looming crypto mining ban, increasing pressure from the U.S.-China trade war, and the Chinese government’s cautious stance towards crypto, the crypto mining might shift from US to China over the forecast period.
The study of GPU market includes companies involved in the design and manufacture of GPUs such as NVIDIA, AMD, Intel and also computer hardware companies that produce vendor-specific GPU-based video/graphic cards such as Imagination Technologies, ASUS, among others.
GPU manufactures have been expanding their horizon across applications and are increasing technology partnerships to reap benefits. Further, new GPUs, based on new architectures are expected in the next two years. Following are the recent developments in this respect:
- March 2018: TSMC, at its TSMC 2018 Technology Symposium, unveiled Wafer-on-Wafer (WoW) chip manufacturing technology a new wafer-stacking technology which could allow creating more powerful graphics cards without increasing the size or density of their silicon. According to TSMC, the new Wafer on Wafer (WoW) manufacturing technology allows to connect chips together without mounting them laterally on the same interposer, thus benefitting its GPU customers, NVIDIA and AMD.
- December 2018 - Intel at the 2018 CPU architecture day, revealed details about their upcoming GPUs, Xe architecture. The company stated that the first GPUs on Xe architecture will arrive in 2020 and will be built on a 10 nm FinFET process. With Xe architecture, the company expects to branch out to consumer and commercial product lines, in addition to data centers.
Reasons to Purchase this report:
- The market estimate (ME) sheet in Excel format
- Report customization as per the client’s requirements
- 3 months of analyst support