無料セミナー : 2019年11月19日『手術用ロボット:外科医療におけるトランスフォーメーション』 BIS Research共同開催

株式会社グローバルインフォメーション
TEL: 044-952-0102
表紙
市場調査レポート

ディープラーニング向けチップセットの世界市場の分析と予測:CPU、GPU、FPGA、ASIC、SoCアクセラレーター

Deep Learning Chipsets - CPUs, GPUs, FPGAs, ASICs, SoC Accelerators, and Other Chipsets for Training and Inference Applications: Global Market Analysis and Forecasts

発行 Tractica 商品コード 408667
出版日 ページ情報 英文 88 Pages; 151 Tables, Charts & Figures
即納可能
価格
本日の銀行送金レート: 1USD=109.82円で換算しております。
Back to Top
ディープラーニング向けチップセットの世界市場の分析と予測:CPU、GPU、FPGA、ASIC、SoCアクセラレーター Deep Learning Chipsets - CPUs, GPUs, FPGAs, ASICs, SoC Accelerators, and Other Chipsets for Training and Inference Applications: Global Market Analysis and Forecasts
出版日: 2019年03月14日 ページ情報: 英文 88 Pages; 151 Tables, Charts & Figures
担当者のコメント
Tracticaは、時代を先取りした商品ラインナップを展開する市場調査会社です。レポートを分野別にまとめてご購入できる大変お得な『年間契約型レポートパッケージ』もございます。無料のアナリストサービスもご利用いただけますので是非あわせてご検討下さい。
概要

ディープラーニング向けチップセット市場は、2018年の51億米ドルから、2025年には726億米ドルまで拡大すると予測されています。

当レポートでは、世界のディープラーニング向けチップセット市場について調査分析し、CPU、GPU、FPGA、ASIC、SoCアクセラレーターなどを対象に、産業動向、技術問題、市場機会、市場規模と予測、主要企業も含めて、体系的な情報を提供しています。

第1章 エグゼクティブサマリー

第2章 市場問題

  • 背景
  • 市場区分
  • 市場促進要因
  • 市場の障壁と課題
  • 主な市場と用途
  • 地域差
  • AIチップセット企業における新興企業の活動

第3章 技術問題

  • AIとディープラーニング
  • チップセット向け技術パラメーター
  • チップセット向け標準AIワークロードと実装
  • ニューラルネットワーク (NN)
  • チップセットに対する数学の関連性
  • AIワークロードとデータの種類
  • トレーニング vs. 推論のチップセット
  • ディープラーニング向けチップセットアーキテクチャー
  • SoCアクセラレーター
  • 低精度、整数、固定小数点、浮動小数点
  • ディープラーニング開発のフレームワーク
  • フレームワークの相互運用性
  • OpenCLとCUDA
  • 新興のアーキテクチャー
  • 大学・研究機関

第4章 主な業界企業

  • Amazon
  • AMD
  • ARM
  • Cerebras Systems
  • CEVA
  • Facebook
  • Google
  • Graphcore
  • Groq
  • Gyrfalcon Technologies
  • Huawei
  • IBM
  • Intel
  • Mobileye
  • NVIDIA
  • Qualcomm
  • Thinci
  • Wave Computing
  • Xilinx

第5章 市場予測

  • 予測手法と前提条件
  • 市場全体
  • 出荷数:チップセットの種類別
  • 収益:チップセットの種類別
  • 収益:トレーニング vs. 推論
  • 収益:コンピューター容量別
  • 収益:電力消費別
  • 平均販売価格:チップセットの種類別
  • 収益:市場部門別
  • CPU
  • GPU
  • ASIC
  • FPGA
  • SoCアクセラレーター

第6章 ディープラーニング向けチップセットとIP企業

第7章 略語・頭文字リスト

第8章 目次

第9章 図表

第10章 調査範囲、情報源、調査手法、注記

目次
Product Code: DLC-19

The rapid adoption of artificial intelligence (AI) for practical business applications has introduced a number of uncertainties and risk factors across virtually every industry, but one fact is certain: in today's AI market, hardware is the key to solving many of the sector's key challenges, and chipsets are at the heart of that hardware solution. Given the widespread applicability of AI, it is almost certain that every chip in the future will have some sort of AI engine embedded. The engine could take a wide variety of forms, ranging from a simple AI library running on a CPU to more sophisticated custom hardware. The potential for AI is best fulfilled when the chipsets are optimized to provide the appropriate amount of compute capacity at the right power budget for specific AI applications, a trend that is leading to increasing specialization and diversification in AI-optimized chipsets.

During the past 2 years, the deep learning chipset market has experienced a dramatic period of evolution, led by NVIDIA and Intel. Yet, the chip companies are somewhat behind in their delivery schedule. Smaller chips aimed at the edge (embedded) market are shipping, but larger chips aimed at the enterprise market are seeing delays. Meanwhile, market validation has already begun for the edge market and should begin for the enterprise market in 2019. A ramp-up in deep learning chipset volumes will likely occur in 2019 and 2020, and the winners will begin to emerge. Tractica forecasts that the market for deep learning chipsets will increase from $5.1 billion in 2018 to $72.6 billion in 2025. The edge computing market, where AI computation is done on the device, is expected to represent more than three-quarters of the total market opportunity, with the balance being in enterprise/data center environments.

This Tractica report assesses the industry dynamics, technology issues, and market opportunity surrounding deep learning chipsets, including CPUs, GPUs, FPGAs, ASICs, and SoC accelerators. The report provides market sizing and forecasts for the period from 2018 through 2025, with segmentation by chipset type, compute capacity, power consumption, market sector, and training versus inference. The study also includes 19 profiles of key industry players.

Key Questions Addressed:

  • What chipset types are being used for deep learning today, and how will they change during the next 10 years?
  • What are the power consumption and compute capacity profiles of chipsets used for deep learning applications?
  • What is the market opportunity for deep learning chipsets in enterprise/data center environments versus edge devices?
  • Which market sectors and industries will drive demand for deep learning chipsets?
  • What is the state of technology development for deep learning chipsets, and which companies are the key industry players driving innovation?
  • What are some of the emerging architectures for deep learning chipsets?
  • What are the key performance matrices for deep learning chipsets?
  • What are some of the use cases for deep learning chipsets in different application markets?

Who Needs This Report?

  • Semiconductor and component manufacturers
  • Service providers and systems integrators
  • End-user organizations deploying deep learning systems
  • Industry associations
  • Government agencies
  • Investor community

Table of Contents

1. Executive Summary

  • 1. Introduction
  • 2. Market Overview
  • 3. New Design Starts and Startup Activity
  • 4. Technology Parameters

2. Market Issues

  • 1. Background
  • 2. Market Segmentation
    • 1. Segmentation by Architecture
    • 2. Segmentation Based on Training vs. Inference
    • 3. Segmentation Based on Compute Capacity
    • 4. Segmentation Based on Power Consumption
    • 5. Segmentation Based on the Market: Enterprise/Data Center
    • 6. Segmentation Based on the Application: Edge Market
  • 3. Market Drivers
    • 1. Availability of Large Datasets
    • 2. Improved AI Algorithms and Deeper Neural Networks
    • 3. Multiple AI Pipelines
    • 4. Complexity of Training
    • 5. Growth in Enterprises
    • 6. Cost per Inference
    • 7. Latency and Throughput Requirements for Inference
    • 8. Computer Vision
    • 9. Embedded Devices
  • 4. Market Barriers and Challenges
    • 1. Development Costs
    • 2. Availability of Expertise
    • 3. Time to Market
  • 5. Key Markets and Applications
    • 1. Data Center and Enterprise
    • 2. Consumer
    • 3. Industrial: Robotics, Drones, and Quality Assurance
    • 4. Automotive and Transportation
    • 5. Smart Camera and Surveillance
    • 6. Mobile Devices and Smartphones
    • 7. Government and Defense
    • 8. Other
  • 6. Regional Differences
  • 7. Startup Activity in AI Chipset Companies

3. Technology Issues

  • 1. AI and Deep Learning
  • 2. Technology Parameters for Chipsets
    • 1. Performance
    • 2. Power
    • 3. Performance per Watt
    • 4. Programmability
    • 5. IP and Ecosystem
    • 6. Development Tools
  • 3. Classic AI Workloads and Implementations for Chipsets
    • 1. Classification
    • 2. Regression
    • 3. Transcription
    • 4. Machine Translation
    • 5. Anomaly Detection
  • 4. Neural Networks
    • 1. Feedforward Networks vs. Recurrent Networks
    • 2. Convolutional Neural Networks
    • 3. Long Short-Term Memory
    • 4. Other Neural Networks
  • 5. Relevance of Mathematics to Chipsets
    • 1. Processing Elements and Arithmetic Logic Units
    • 2. Memory
    • 3. Connectivity
    • 4. Off-Chip Connectivity
  • 6. AI Workloads and Data Types
    • 1. Vision: Image and Video
    • 2. Audio and Speech
    • 3. Text/Natural Language Processing
    • 4. Search
  • 7. Training vs. Inference Chipsets
  • 8. Chipset Architectures for Deep Learning
    • 1. Central Processing Units
    • 2. Graphics Processing Units
    • 3. Field-Programmable Gate Arrays
    • 4. Application-Specific Integrated Circuits
  • 9. System-on-a-Chip Accelerators
  • 10. Low Precision, Integer, Fixed-Point, and Floating-Point Mathematics
  • 11. Deep Learning Development Frameworks
  • 12. Framework Interoperability
  • 13. OpenCL and Compute Unified Device Architecture
  • 14. Emerging Architectures
  • 15. Universities and Research Institutions

4. Key Industry Players

  • 1. Amazon
  • 2. AMD
  • 3. ARM
  • 4. Cerebras Systems
  • 5. CEVA
  • 6. Facebook
  • 7. Google
  • 8. Graphcore
  • 9. Groq
  • 10. Gyrfalcon Technologies
  • 11. Huawei
  • 12. IBM
  • 13. Intel
  • 14. Mobileye
  • 15. NVIDIA
  • 16. Qualcomm
  • 17. Thinci
  • 18. Wave Computing
  • 19. Xilinx

5. Market Forecasts

  • 1. Forecast Methodology and Assumptions
  • 2. Overall Market
  • 3. Unit Shipments by Chipset Type
  • 4. Revenue by Chipset Type
  • 5. Revenue by Training vs. Inference
  • 6. Revenue by Compute Capacity
  • 7. Revenue by Power Consumption
  • 8. Average Selling Price by Chipset Type
  • 9. Revenue by Market Sector
  • 10. Central Processing Units
  • 11. Graphics Processing Units
  • 12. Application-Specific Integrated Circuits
  • 13. Field-Programmable Gate Arrays
  • 14. System-on-a-Chip Accelerators

6. Deep Learning Chipset and IP Companies

7. Acronym and Abbreviation List

8. Table of Contents

9. Table of Charts and Figures

10. Scope of Study, Sources and Methodology, Note

List of Charts, Figures, and Tables

Charts

  • Deep Learning Chipset Revenue, World Markets: 2018-2025
  • Deep Learning Chipset Year-on-Year Revenue Growth Rates, World Markets: 2019-2025
  • Deep Learning Chipset Unit Shipments by Type, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Type, World Markets: 2018-2025
  • Deep Learning Chipset Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning CPU Unit Shipments and Revenue, World Markets: 2018-2025
  • Deep Learning GPU Unit Shipments and Revenue, World Markets: 2018-2025
  • Deep Learning ASIC Unit Shipments and Revenue, World Markets: 2018-2025
  • Deep Learning FPGA Unit Shipments and Revenue, World Markets: 2018-2025
  • Deep Learning SoC Accelerator Unit Shipments and Revenue, World Markets: 2018-2025
  • Deep Learning CPU Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise CPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge CPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise CPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge CPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise CPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge CPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning GPU Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise GPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge GPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise GPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge GPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise GPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge GPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning FPGA Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge FPGA Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge FPGA Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge FPGA Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning ASIC Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge ASIC Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge ASIC Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge ASIC Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning SoC Accelerator Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Region, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Type, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue, Inference vs. Training, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Type, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue, Inference vs. Training, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Device Category, World Markets: 2018-2025

Figures

  • Performance Requirement for Popular Neural Networks
  • Evolution of Artificial Intelligence
  • Number of Companies Working with NVIDIA on Deep Learning
  • Neural Network Zoo
  • AI Workloads Running on NVIDIA Chipsets
  • Training vs. Inference Illustrated
  • System Considerations When Choosing a Hardware Platform
  • Suitability of Different Chipsets for Training and Inference
  • Error in ImageNet Competition
  • Google's Tensor Processing Unit
  • Popularity of Open-Source Deep Network Repositories in GitHub
  • Neural Network Exchange Format Explained
  • Xeon CPU Training Time vs. NVIDIA V100 GPU Training Time
  • Diagram Depicting Various Technologies in AI
  • Feedforward Network vs. Recurrent Neural Network
  • A Convolutional Neural Network Used for Image Recognition

Tables

  • Key Players in Different Deep Learning Chipsets
  • Comparison of Deep Learning Chipset Parameters
  • Summary of Chipset Requirements for Training and Inference
  • Overview of Deep Learning Frameworks
  • Deep Learning Companies
  • IP Companies
  • Deep Learning Chipset Revenue, World Markets: 2018-2025
  • Deep Learning Chipset Year-on-Year Revenue Growth Rates, World Markets: 2019-2025
  • Deep Learning Chipset Revenue by Type, World Markets: 2018-2025
  • Deep Learning Chipset Unit Shipments by Type, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Chipset Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning CPU Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning CPU Shipments by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise CPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge CPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise CPU ASP, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge CPU ASP, World Markets: 2018-2025
  • Deep Learning Enterprise CPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge CPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise CPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge CPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning GPU Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning GPU Shipments by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise GPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge GPU Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise GPU ASP, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge GPU ASP, World Markets: 2018-2025
  • Deep Learning Enterprise GPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge GPU Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise GPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge GPU Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning FPGA Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning FPGA Shipments by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge FPGA Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA ASP, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge FPGA ASP, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge FPGA Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise FPGA Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge FPGA Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning ASIC Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning ASIC Shipments by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge ASIC Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC ASP, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge ASIC ASP, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge ASIC Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise ASIC Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge ASIC Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning SoC Accelerator Revenue by Market Sector, World Markets: 2018-2025
  • Deep Learning SoC Accelerator Shipments by Market Sector, World Markets: 2018-2025
  • Deep Learning Enterprise SoC Accelerator Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator Revenue, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Enterprise SoC Accelerator ASP, Training vs. Inference, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator ASP, World Markets: 2018-2025
  • Deep Learning Enterprise SoC Accelerator Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Enterprise SoC Accelerator Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Edge SoC Accelerator Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Region, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Year-on-Year Revenue Growth Rates, World Markets: 2019-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Type, World Markets: 2018-2025
  • Deep Learning Chipset Unit Shipments by Type, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Chipset Enterprise Sector Revenue, Inference vs. Training, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Year-on-Year Revenue Growth Rates, World Markets: 2019-2025
  • Deep Learning Chipset Edge Sector Revenue by Type, World Markets: 2018-2025
  • Deep Learning Edge Sector Chipset Unit Shipments by Type, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Power Consumption, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Compute Capacity, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue, Inference vs. Training, World Markets: 2018-2025
  • Deep Learning Chipset Edge Sector Revenue by Device Category, World Markets: 2018-2025
Back to Top