[121 Pages Report]The data center accelerator market was valued at USD 1.60 billion in 2017 and is expected to reach USD 21.19 billion by 2023, at a CAGR of 49.47% during the forecast period. In this report, 2017 has been considered as the base year, and the forecast period is from 2018 to 2023.
The data center accelerator market, by processors type was valued at the highest CAGR during the forecast period. The growth of the market for FPGA is attributed to the increasing adoption of FPGAs for acceleration of enterprise workloads. With the exponential data growth, data center operators have to establish a balance between the need for the performance at scale and the operational efficiencies. To boost performance and power efficiencies, data centers are most widely adopting Intel Xeon Scalable processors to support data-intensive performance requirements. For instance, in December 2015, Intel Corporation (US) acquired Altera Corporation (US), a provider of FPGA technology to offer new product categories in the high-growth data center and IoT market segments.
The data center accelerator market has been segmented on the basis of application into deep learning training, public cloud inference, and enterprise inference. The market for enterprise inference is likely to reach the highest CAGR during the forecast period. Data center accelerator market for enterprise inference is expected to gain traction in the coming years on the back of the hyper-scale cloud-based companies such as Google Inc. (US), Facebook, Inc. (US), Amazon.com (US), which are more focused on digital transformation to build similar types of cloud-native applications. The cloud-native application development requires a new set of tools, methodologies, and underlying IT infrastructure. With the availability of both open-source and commercial software, the cloud-native architecture has been widely adopted by enterprises.
The data center accelerator market in APAC is estimated to grow at the highest CAGR during the forecast period. The organizations in APAC have more preference for deploying a hybrid cloud. The organizations are adopting a mix of on-premise, third-party, co-location, private cloud, hosted cloud, and public cloud—depending on nature of workloads, legacy decisions made by the team, budgets, technology maturity within the organization, etc.
Apart from chipmakers making accelerators, the companies that make data centers, including Dell EMC (US) and HP (US), are also emphasizing on integrating deep learning-based accelerators into their HPC data centers.
The growth in consumer generated data and rising use of AI-based services have led to the increased demand for AI-centric data centers. AI provides personalized services by understanding customer behavior data generated from CRM systems, product reviews, and media comments. NVIDIA’s (US) reported GPU sales to data centers in Q3 2016 was worth USD 75 million, a large portion of which was attributed to HPC data centers. Later at 2017 Investor day presentation, NVIDIA provided an estimate that the market for accelerators aimed at data centers could touch USD 30 billion by 2020. This staggering number is 19 times of NVIDIA’s current revenue from data centers, and it also shadows Intel’s USD 17 billion in data center revenue from 2016. Though the forecast from NVIDIA looks overestimated in terms of market value, the company’s impressive growth in the data center business segment in recent years gives a clear indication that the market is expected to witness an exponential growth in the coming years.
AI is a complex system, and for developing, managing, and implementing AI systems companies require personnel with certain skill sets. For instance, people dealing with AI systems should be aware of technologies such as cognitive computing, ML and machine intelligence, deep learning, and image recognition. In addition, integrating AI solutions with existing systems is a difficult task that requires well-funded in-house R&D and patent filling. Even minor errors can translate into system failure or malfunctioning of a solution, and this can drastically affect the outcome and desired result.
Professional services of data scientists and developers are needed to customize existing ML-enabled AI processors. Due to AI being a technology that is still in its early stage of life cycle, a workforce possessing in-depth knowledge of this technology is limited. The impact of this restraining factor will likely remain high during the initial years of the forecast period.
The Moore’s law states that the number of transistors per square inch on integrated circuits will double about every eighteen months until at least 2020. In April 2015, Intel stated that it can keep the Moore’s law going for another few years by developing 7nm and 5nm fabrication technologies. However, moving forward it would be hard to further reduce the size of processors as doing so would also reduce the space between electrons and holes, culminating into problems such as current leakage and overheating in ICs These problems would lead to slower performance, high power consumption by ICs, and reduced durability. Thus, the need to find an alternate way to increase the computational power of chips has fueled the development of accelerators or coprocessor chips.
Challenge: Unreliability of AI algorithms
AI is implemented through machine learning using a computer to run specific software that can be trained. Machine learning can help systems process data with the help of algorithms and identify certain features from that dataset. However, a concern associated with such systems is that it is unclear as to what is going on inside algorithms; the internal workings remain inaccessible, and unlike humans, the answers provided by these systems are un-contextualized. In July 2017, researchers at the Facebook AI Research (FAIR) lab found that the chat bots they created had deviated from their predefined script and were communicating in a language created by themselves, which humans could not understand. While one of the important goals of current research is to improve AI-to-human communication, the possibility that an AI system can create its own unique language that humans cannot understand could be a setback. Moreover, several scientists and tech influencers, such as Stephen Hawking, Elon Musk, Bill Gates, and Steve Wozniak, have already warned that future AI technology could lead to unintended consequences
Report Metric |
Details |
Market size available for years |
2016–2023 |
Base year considered |
2017 |
Forecast period |
2018–2023 |
Forecast units |
Million/Billion (USD) |
Segments covered |
Processor type, Application, Type, and Region |
Geographies covered |
North America, Europe, APAC, RoW (Brazil and Others) |
Companies covered |
NVIDIA (US), Intel (US), Alphabet (US), Advanced Micro Devices (AMD) (US), Achronix Semiconductor (US), Oracle (US), Xilinx (US), IBM (US), Hewlett Packard Enterprise (HPE) (US), Dell (US) |
The research report categorizes the Data Center Accelerator Market to forecast the revenues and analyze the trends in each of the following sub-segments:
NVIDIA (US), Intel (US), Alphabet Inc. (US), Advanced Micro Devices (AMD), Xilinx (US), Oracle (US), IBM (US), Achronix Semiconductor (US), Hewlett Packard Enterprise (HPE) (US), and Dell (US)
NVIDIA is among the world leaders in visual computing business. It has a well-established geographic footprint and deals with major OEMs or ODMs. The company continues to lead the development of new products for the data center accelerator market. Recently, NVIDIA has witnessed significant jump in its data center revenue, and the company is posing a tough competition to Intel in the market. The company claims that its data center business has started ramping up. It has also won support from Microsoft, Amazon, Google, Alibaba, Baidu, Tencent, and Oracle for Volta. Volta would be available for their internal use in deep learning as well as for external public cloud services. With the Moore’s Law slowing down, GPU accelerators are expected to witness significant success in deep learning and training and inference applications.
To speak to our analyst for a discussion on the above findings, click Speak to Analyst
Table of Contents
1 Introduction (Page No. - 13)
1.1 Study Objectives
1.2 Definition
1.3 Study Scope
1.3.1 Markets Covered
1.3.2 Geographic Scope
1.3.3 Years Considered
1.4 Currency
1.5 Limitations
1.6 Stakeholders
2 Research Methodology (Page No. - 17)
2.1 Research Data
2.1.1 Secondary Data
2.1.1.1 Secondary Sources
2.1.2 Primary Data
2.1.2.1 Key Data From Primary Sources
2.1.2.2 Key Industry Insights
2.1.2.3 Breakdown of Primaries
2.2 Market Size Estimation
2.2.1 Bottom-Up Approach
2.2.2 Top-Down Approach
2.3 Market Breakdown & Data Triangulation
2.4 Research Assumptions
3 Executive Summary (Page No. - 25)
4 Premium Insight (Page No. - 30)
4.1 Attractive Opportunities in Market
4.2 Market, By Type
4.3 Market for Cloud Data Center Accelerator, By Country
4.4 Market in APAC in 2018, By Country & Application
4.5 Market for Deep Learning Training, By Processor Type
5 Market Overview (Page No. - 33)
5.1 Introduction
5.2 Market Dynamics
5.2.1 Drivers
5.2.1.1 Growth of Cloud-Based Services
5.2.1.2 Growing Demand for AI in HPC Data Centers
5.2.1.3 Focus Toward Parallel Computing in AI Data Centers
5.2.2 Restraints
5.2.2.1 Premium Pricing of Accelerators
5.2.2.2 Limited AI Hardware Experts
5.2.3 Opportunities
5.2.3.1 Rising Market for FPGA-Based Accelerators
5.2.3.2 Rising Need of Coprocessors Due to Slowdown of Moore’s Law
5.2.4 Challenge
5.2.4.1 Unreliability of AI Algorithms
5.2.4.2 Creating Models and Mechanisms of AI in Cloud
6 Market, By Processor Type (Page No. - 38)
6.1 Introduction
6.2 CPU
6.3 GPU
6.4 FPGA
6.5 ASIC
7 Market, By Type (Page No. - 48)
7.1 Introduction
7.2 Cloud Data Center
7.3 HPC Data Center
8 Market, By Application (Page No. - 54)
8.1 Introduction
8.2 Deep Learning Training
8.3 Public Cloud Inference
8.4 Enterprise Inference
9 Geographic Analysis (Page No. - 61)
9.1 Introduction
9.2 North America
9.2.1 US
9.2.2 Canada
9.2.3 Mexico
9.3 Europe
9.3.1 UK
9.3.2 Germany
9.3.3 The Netherlands
9.3.4 Rest of Europe
9.4 APAC
9.4.1 China
9.4.2 Japan
9.4.3 Singapore
9.4.4 India
9.4.5 Australia
9.4.6 Rest of APAC
9.5 RoW
9.5.1 Brazil
9.5.2 Others
10 Competitive Landscape (Page No. - 81)
10.1 Market Ranking Analysis: Data Center Accelerator Market
10.2 Competitive Situations and Trends
10.2.1 Product/Solution Launches
10.2.2 Partnerships/Collaborations
10.2.3 Acquisitions
10.2.4 Expansions
11 Company Profiles (Page No. - 85)
(Business Overview, Products Offered, Recent Developments, SWOT Analysis, and MnM View)*
11.1 Key Players
11.1.1 NVIDIA
11.1.2 Intel
11.1.3 Alphabet
11.1.4 Advanced Micro Devices (AMD)
11.1.5 Achronix Semiconductor
11.1.6 Oracle
11.1.7 Xilinx
11.1.8 IBM
11.1.9 Hewlett Packard Enterprise (HPE)
11.1.10 Dell
11.2 Other Key Players
11.2.1 Lenovo
11.2.2 Fujitsu
11.2.3 Cisco
11.3 Key Innovators
11.3.1 Leap Motion
11.3.2 Algolux
11.3.3 Clarifai
11.3.4 Ditto Labs
11.3.5 Mad Street Den
*Details on Business Overview, Products Offered, Recent Developments, SWOT Analysis, and MnM View Might Not Be Captured in Case of Unlisted Companies.
12 Appendix (Page No. - 114)
12.1 Insights of Industry Experts
12.2 Discussion Guide
12.3 Knowledge Store: Marketsandmarkets’ Subscription Portal
12.4 Available Customizations
12.5 Related Reports
12.6 Author Details
List of Tables (50 Tables)
Table 1 Market, By Processor Type, 2016–2023 (USD Million)
Table 2 Market for CPU, By Application, 2016–2023 (USD Million)
Table 3 Market for CPU, By Type, 2016–2023 (USD Million)
Table 4 Market for CPU, By Region, 2016–2023 (USD Million)
Table 5 Market for GPU, By Application, 2016–2023 (USD Million)
Table 6 Market for GPU, By Type, 2016–2023 (USD Million)
Table 7 Market for GPU, By Region, 2016–2023 (USD Million)
Table 8 Market for FPGA, By Application, 2016–2023 (USD Million)
Table 9 Market for FPGA, By Type, 2016–2023 (USD Million)
Table 10 Market for FPGA, By Region, 2016–2023 (USD Million)
Table 11 Market for ASIC, By Application, 2016–2023 (USD Million)
Table 12 Market for ASIC, By Type, 2016–2023 (USD Million)
Table 13 Market for ASIC, By Region, 2016–2023 (USD Million)
Table 14 Market, By Type, 2016–2023 (USD Million)
Table 15 Cloud Data Center Accelerator Market, By Processor Type, 2016–2023 (USD Million)
Table 16 Cloud Market, By Region, 2016–2023 (USD Million)
Table 17 HPC Data Center Accelerator Market, By Processor Type, 2016–2023 (USD Million)
Table 18 HPC Market, By Region, 2016–2023 (USD Million)
Table 19 Market, By Application, 2016–2023 (USD Million)
Table 20 Market for Deep Learning Training, By Region, 2016–2023 (USD Million)
Table 21 Market for Deep Learning Training, By Processors Type, 2016–2023 (USD Million)
Table 22 Market for Public Cloud Inference, By Region, 2016–2023 (USD Million)
Table 23 Market for Public Cloud Inference, By Processors Type, 2016–2023 (USD Million)
Table 24 Market for Enterprise Inference, By Region, 2016–2023 (USD Million)
Table 25 Market for Enterprise Inference, By Processors Type, 2016–2023 (USD Million)
Table 26 Market, By Region, 2016–2023 (USD Million)
Table 27 Market for HPC in North America, By Country, 2016–2023 (USD Million)
Table 28 Market for Cloud in North America, By Country, 2016–2023 (USD Million)
Table 29 Market in North America, By Processors Type, 2016–2023 (USD Million)
Table 30 Market in North America, By Application, 2016–2023 (USD Million)
Table 31 Market in North America, By Type, 2016–2023 (USD Million)
Table 32 Market for HPC in Europe, By Country, 2016–2023 (USD Million)
Table 33 Market for Cloud in Europe, By Country, 2016–2023 (USD Million)
Table 34 Market in Europe, By Processors Type, 2016–2023 (USD Million)
Table 35 Market in Europe, By Application, 2016–2023 (USD Million)
Table 36 Market in Europe, By Type, 2016–2023 (USD Million)
Table 37 Market for HPC in APAC, By Country, 2016–2023 (USD Million)
Table 38 Market for Cloud in APAC, By Country, 2016–2023 (USD Million)
Table 39 Market in APAC, By Processor Type, 2016–2023 (USD Million)
Table 40 Market in North America, By Application, 2016–2023 (USD Million)
Table 41 Market in APAC, By Type, 2016–2023 (USD Million)
Table 42 Market for HPC in RoW, By Country, 2016–2023 (USD Million)
Table 43 Market for Cloud in RoW, By Country, 2016–2023 (USD Million)
Table 44 Market in RoW, By Processors Type, 2016–2023 (USD Million)
Table 45 Market in RoW, By Application, 2016–2023 (USD Million)
Table 46 Market in RoW, By Type, 2016–2023 (USD Million)
Table 47 Product/Solution Launches, 2017–2018
Table 48 Partnerships/Collaborations, 2017–2018
Table 49 Acquisitions, 2018
Table 50 Expansions, 2016–2018
List of Figures (45 Figures)
Figure 1 Segmentation of Market
Figure 2 Research Design
Figure 3 Process Flow of Market Size Estimation
Figure 4 Data Triangulation
Figure 5 GPU to Hold Largest Size of Market During Forecast Period
Figure 6 Market for Enterprise Inference to Grow at Highest CAGR Between 2018 and 2023
Figure 7 Market for HPC Data Center Accelerator in APAC to Grow at Highest CAGR During 2018–2023
Figure 8 Market for Cloud Data Center Accelerator in APAC to Grow at Highest CAGR During Forecast Period
Figure 9 Market in APAC Expected to Grow at Highest CAGR During 2018–2023
Figure 10 Rising Need of Coprocessors in Data Centers Due to Slow Down in Moore’s Law as Potential Opportunity for Market
Figure 11 Market for HPC Data Center Accelerator to Grow at Higher CAGR During Forecast Period
Figure 12 Cloud Data Center Accelerator Market in Canada to Grow at Highest CAGR During Forecast Period
Figure 13 Deep Learning Training to Hold Largest Share of Market in APAC in 2018
Figure 14 Market for ASIC in Deep Learning Training to Grow at Higher CAGR Between 2018 and 2023
Figure 15 Market: Drivers, Restraints, Opportunities, and Challenges
Figure 16 Slow Down in Moore’s Law
Figure 17 Market, By Processor Type
Figure 18 Market for CPU in APAC to Grow at the Highest CAGR During Forecast Period
Figure 19 Cloud Data Center to Hold Largest Size of Market for GPU During Forecast Period
Figure 20 FPGA Market for Public Cloud Inference to Grow at Highest CAGR During Forecast Period
Figure 21 ASIC Market for Enterprise Inference to Grow at Highest CAGR During Forecast Period
Figure 22 Market, By Type
Figure 23 Cloud Data Center Accelerator Market in APAC Expected to Grow at Highest CAGR During Forecast Period
Figure 24 GPU to Hold Largest Size of HPC Data Center Accelerator Market During Forecast Period
Figure 25 Market, By Application
Figure 26 Deep Learning Training: Market for ASIC to Grow at Highest CAGR During Forecast Period
Figure 27 North America to Hold Largest Size of Market for Public Cloud Inference During Forecast Period
Figure 28 Enterprise Inference: Market for FPGA to Grow at Highest CAGR During Forecast Period
Figure 29 Market Segmentation, By Geography
Figure 30 Geographic Snapshot: Market in APAC Expected to Grow at Highest CAGR During Forecast Period
Figure 31 Market in India Estimated to Grow at Highest CAGR From 2018 to 2023
Figure 32 North America: Market Snapshot
Figure 33 Europe: Market Snapshot
Figure 34 APAC: Market Snapshot
Figure 35 Key Growth Strategies Adopted By Top Companies, 2015–2017
Figure 36 Ranking of the Top 5 Players in Market, (2017)
Figure 37 NVIDIA: Company Snapshot
Figure 38 Intel: Company Snapshot
Figure 39 Alphabet: Company Snapshot
Figure 40 Advanced Micro Devices (AMD): Company Snapshot
Figure 41 Oracle: Company Snapshot
Figure 42 Xilinx: Company Snapshot
Figure 43 IBM: Company Snapshot
Figure 44 HPE: Company Snapshot
Figure 45 Dell: Company Snapshot
Benchmarking the rapid strategy shifts of the Top 100 companies in the Data Center Accelerator Market
Request For Special Pricing
Growth opportunities and latent adjacency in Data Center Accelerator Market