Total general computing power will see a tenfold increase and reach 3.3 ZFLOPS
Computing 2030
In the near future, the data we produce annually will soon be calculated in yottabytes. As this happens, our processing power will also become more energy efficient. We will transcend our human limitations, and uncover the secrets of all the cosmos, from the tiniest quarks, to the largest nebulas. We will use our newfound digital power to achieve carbon neutrality, and turn the digital age, into the intelligent age.
Download Report
Predictions
Pervasive diversified computing
AI computing power will increase by a factor of 500, to more than 100 ZFLOPS
Global data will be growing by 500-fold to reach one yottabyte every year
Cloud services as 87% of total enterprise application expenditure
Key Features
Computing will become both more open and more secure. Every person and every organization will be given equal opportunity to build a more innovative computing industry and share in its value
Cognitive Intelligence
From Weak AI to Strong AI: Human-like Cognition
-
All-scenario, multi-modal, and multi-task models for general intelligence
-
Autonomous AI without human intervention
-
Brain-like intelligence based on digital neurons and neuro dynamics
-
Creative generative AI
-
Cross-discipline knowledge computing
Intrinsic Security
From Add-on Security to Intrinsic Security: Self-adaptive Throughout the Process
-
Security is built into chips, firmware, hardware, and software
-
Security must be ensured throughout the entire data processing lifecycle
-
Open-source security services built into application architecture
Green, Integrated Computing
Chip-level, DC-level, and Cross-region Computing: Low-carbon Computing
-
2.5D/3D chip packaging, co-packaged optics, and higher chip performance with minimal power consumption
-
Integrated DCs: PUE approaching or even less than 1
-
Computing networks that adapt to different latency and cost requirements, with globally optimal energy efficiency
Diversified Computing
From Computing-centric to Data-centric: 10x Efficiency
-
Peer-to-peer interconnection, near-data computing, computing using memory
-
From board-level to DC-level memory-semantic buses
-
HPC, AI, and big data increasingly converged
Multi-dimensional Collaboration
Cubic Computing & Integration of Reality and Virtuality: On-demand Distribution
-
Vertical collaboration between cloud, edge, and devices
-
Horizontal collaboration between storage, computing, and network devices
-
Collaboration between the physical and digital worlds
Physical Layer Breakthroughs
New Materials, New Media, and New Computing: Exponential Growth in Density
-
Analog computing: Quantum computing & optical computing
-
Non-silicon-based computing: 2D materials & carbon transistors
-
Ferroelectrics and phase change materials commercially used; new media like optical storage and DNA storage to be explored
