From Leo Chen, Corporate Senior Vice President, President of Enterprise Sales, Huawei:
“At last year’s summit, we explored the opportunities and challenges of industrial intelligent transformation. As we revisit this topic today, rapid AI development has significantly changed the context surrounding related discussions, and several major trends have emerged.
First, expectations for AI continue to rise. Although 2024, hailed as the “The First Year of AI Application at Scale”, is still a recent memory, conversations in 2025 have already shifted to how we can accelerate toward Artificial General Intelligence(AGI).
Second, AI’s evolution path is becoming increasingly clear[1]. AI agents, or agentic AI, are becoming the focus in 2025. There is already hope that AI agents will take GenAI “from answers to actions”[2].
Third, the threshold for AI application is gradually lowering. DeepSeek’s groundbreaking open-source model (R1), released in January this year, has reshaped the AI landscape in three key ways:
- It has shifted industry focus from merely increasing computing power to optimizing its use.
- The scale of AI models is no longer the sole factor that determines their intelligence.
- And, open-source AI models are gradually becoming more competitive.
These changes have shown us the possibilities of an intelligent future. However, we must remain grounded. The true measure of AI’s value lies in its ability to enhance productivity across industries, and drive a new wave of social and economic development.
Key challenges in digital & intelligent transformation
And to further drive industrial intelligent transformation, we must answer several key questions:
The first is, how can we fully integrate advanced intelligent technologies into industry-specific scenarios? For example, data silos are a key challenge we need to address – A Gartner survey suggests that 83% of data-focused projects stumble due to challenges like this.
The second question is, can the ICT infrastructure for intelligent transformation across industries handle the future exponential growth of AI workloads? Currently, 52% of existing enterprise systems cannot directly connect to intelligent platforms – this means ICT infrastructure needs to be upgraded.
The third is, how can we translate the concept of ‘inclusive AI adoption across all industries’ into a reality? Gartner predicts at least 30% of generative AI (GenAI) projects will be abandoned after proof of concept (PoC) by the end of 2025, with high costs being a key cause.
And the fourth question is, do industries have sufficient ICT talent to drive continuous breakthroughs in intelligent transformation?
Vision to action: Accelerating industrial intelligence
Throughout its 38 years in ICT, Huawei has always drawn upon its expertise to make words a reality. As we are committed to addressing these questions and challenge, we believe that four key areas are essential to accelerating intelligent transformation across industries.
First, to integrate technologies into industry scenarios, we must build a target ICT architecture for industrial intelligent transformation based on industry requirements, pain points, and development stages.
Second, we need to build advanced and AI-oriented ICT infrastructure to address the exponential growth of AI workloads.
We must develop high-performance AI products that seamlessly integrate with open-source models, enhance AI development toolchains, and collaborate with industry partners. These efforts will drive accessibility, cost-effectiveness, and widespread adoption, making AI as universal as water and electricity for all industries.
Fourth, we must train ICT talent in a more targeted manner.
As for the four areas, I would like to share our practices.
Diving deep into scenarios to define industrial ICT architectures
Regarding integrating technology with industry scenarios, we have established eight integrated teams for more than 20 industries targeting 255 scenarios. Today, we are proud to have supported the transformation of over 200,000 global enterprises. Through this experience, we have identified four key objectives we must achieve:
First, strategies must be actionable. We must align investment in technological and data architectures with each company’s strategic goals.
Second, in any company, the ICT systems of each business module must work collaboratively to eliminate information silos.
Third, risks must be controllable. We need to establish clear security frameworks to ensure subsystems align with overall designs.
Fourth, innovation must be scalable and integrate emerging technologies, so that we can rapidly incubate and validate innovative industrial solutions.
In addition to these objectives, we must consider the unique demands of each industry. In more than 20 sectors, we have planned target architectures for ICT infrastructure tailored to specific needs and pain points.
Let’s look at the financial industry as an example. The global financial sector is moving toward “real-time payments,” requiring banks to offer “always-on services” and “rapid risk interception.” These developments call for a next-gen ICT infrastructure featuring “Four Zeros”: Zero Trust for in-depth security, Zero Wait for better user experience, Zero Downtime for always-on services, and Zero Touch for automated maintenance.
By examining the unique needs of the financial sector, Huawei developed the RAAS target architecture for banking ICT infrastructure. RAAS stands for Reliability, Availability, Autonomy, and Security, to achieve the “Four Zeros” goal.
In Thailand, Kasikorn Bank (KBANK) has adopted Huawei’s resilient data center solution which is based on the RAAS architecture. By employing Huawei’s next-gen all-flash storage Dorado and intelligent autonomous driving network, KBANK established a high-availability service zone and achieved more robust transactions. This has tripled its system performance, with zero data loss and an annual transaction growth rate of over 40%.
Innovating to consolidate the foundations for intelligent transformation
It is important to recognize that every industry has its own unique challenges and pain points when it comes to intelligent transformation. Generally, an industry’s intelligent transformation goes through three key phases: informatization, digitalization, and intelligence.
Huawei comprehensively considers customer priorities at different stages of intelligent transformation while also considering how we can protect their investment, and achieve continuous evolution, seamless data migration, ecosystem integration, and more.
For example, Meralco, the Philippines’ largest power distributor, is partnering with Huawei to drive its digital and intelligent transformation through broadband, cloud, and intelligence.
In the broadband phase, Meralco adopted Huawei’s wireless private network solution, achieving four-nines stability, a fivefold capacity increase, and a 50% reduction in OPEX.
Now, in the cloud phase, Meralco is deploying Huawei’s Advanced Meter Infrastructure (AMI) and a power distribution monitoring solution on its private cloud. This has led to a 30% improvement in resource utilization, a 40% increase in system reliability, and a 20% boost in O&M efficiency, laying a solid foundation for full digitalization.
Looking ahead, Huawei and its industry partners will collaborate to develop intelligent substation solutions, enabling minute-level data collection and increasing inspection efficiency by 50%, further accelerating Meralco’s intelligent grid transformation.
Leading products and solutions with AI-oriented infrastructure
In terms of creating AI-oriented technology and product solutions, Huawei plays on its unique full-stack strengths in network, storage, computing, and cloud to constantly set industry benchmarks.
For example, through the synergy of our full-stack capabilities, Huawei has developed an industry-leading Multilayer Ransomware Protection (MRP) solution. This solution uses AI to accurately identify viruses. With six layers of protection, it prevents data leakage and ensures service recovery just seconds after an incident.
In connectivity and data transmission, Huawei has built a high throughput data center network (DCN) using the industry’s first 100-terabit DC switch. Additionally, with our innovative NSLB (network scale load balance) algorithm and StarryLink optical modules, AI training efficiency is improved by 10%.
In data storage, Huawei’s next-gen OceanStor Dorado storage system introduces an innovative data-control separation architecture, boosting IOPS[3] performance by 3 times. This architecture also achieves unprecedented “eight nines” reliability.
To meet both general and intelligent computing needs, we have developed Kunpeng for general-purpose computing, and Ascend for intelligent computing. openEuler is an open-source operating system, supporting different processor architectures, such as Kunpeng, x86, and RISC-V. To date, we have more than 6.35 million developers for Kunpeng and Ascend.
Finally, in cloud technology, Huawei Cloud has introduced Model as a Service (MaaS), enabling AI models to be used out of the box. Huawei Cloud also offers high-performance distributed GaussDB databases that support up to one billion daily active users.
Empowering inclusive AI adoption in every industry
In terms of inclusive AI adoption, it is driven by three key trends: First, open-source large models lower access barriers, making AI more widely available. Second, synthetic data combined with small industry-specific datasets enhances model performance while reducing the need for large volumes of raw industry data. Thirdly, efficient architectures and distillation techniques reduce computational demands during the inference phase, enabling more flexible deployment on edge devices and AI hardware.
With these three trends, enterprises will need easily accessible AI products and hardware, easy-to-use AI application development tools, and a supportive developer community and ecosystem.
And so, Huawei has introduced an AI appliance that integrates highly. This product supports over 50 mainstream foundation models. This will make it a lot easier for industry customers to apply AI, speed up AI deployment, and accelerate the transition toward greater intelligence.
In Shanghai, China, Huawei helps Ruijin Hospital launched RuiPath. Within just two months, it had been trained on one million pathology slides, covering the types of cancers that account for 90% of the annual number of cancer patients in China.This model provides doctors with an intelligent assistant for early detection, diagnosis, and treatment.
Regarding the ecosystem, Huawei supports the open-source AI MindSpore community. This community helps developers quickly solve technical problems during AI application development. Currently, the community has 3.15 million developers from 130 countries and regions, 11 million downloads, and 1,600 AI courses.
ICT + industry scenarios to accelerate talent cultivation
In terms of ICT talent cultivation, Huawei has already trained 8 million ICT talent globally. Furthermore, we plan to nurture an additional 10 million ICT talent by 2030.
In 2025, we will focus on two priorities during our ICT talent development efforts:
The first is launching the Industry Elites in the ICT Classroom Program for enterprise customers. By training professionals in every industry, we can strengthen their understanding of technology application, enabling them to better plan and drive the intelligent transformation within their organizations.
Second is launching the Leading ICT Talent Cultivation Program. This program will teach university students about the most advanced ICT technologies, allowing them to fulfil the needs of the industry as soon as they join the workforce.
Official Release of 10 major solutions for industrial intelligence
None of the above goals can be achieved without the valuable support of our partners. Our “Huawei + Partners” collaboration system is essential to our mission of effectively serving customers in various industries worldwide.
A highlight of our partnerships is the many joint innovations we’ve created together across the globe. Today, we are honored to unveil new solutions for ten industries that we have developed alongside our partners. They are Digital Village Solution, Public Services Digitalization Solution, Digital Training Solution, Financial Data Center Resilience Solution, Intelligent Distribution Solution 2.0, Smart Railway Yard & Station Solution, Intelligent Multi-level Port Operation Management Solution, Intelligent Central Processing Facilities Solution, Intelligent Chemical Solution, and Smart Retail Solution 2.0.
I’d like to thank all of our partners for their unwavering support and contributions. Looking ahead to 2025, I am confident these solutions will be adopted in more countries and regions, and we will continue to roll out even more innovative solutions.
Official release of global showcases for industrial digital & intelligent transformation
Additionally, we believe that seeing is believing, so we have created 83 industrial digital-transformation showcases for 71 scenarios across the globe. These showcases can serve as “lighthouses”, offering best-practice references for our customers.
I would also like to thank our customers for their continued trust in Huawei. We would also like to invite our customers to visit our showcases around the world, where we hope they will gain more insights for their transformation journeys.
AI and digital and intelligent transformation for the greater good
Ladies and gentlemen, It is our privilege to be a part of industrial intelligent transformation. This transformation will not only enhance the productivity of every industry, but also contribute to the greater good of humanity in three ways:
First, with industrial intelligent transformation, technological benefits will not be limited to the IT elite, but will be enjoyed by a broader range of people, reaching all industries.
Second, the broader application of AI will also unlock greater vitality and creativity within society, and free humanity to realize its full value. This is because AI can now be directed to do what we no longer have to do, so that we can do the things we want to do.
Third, industrial intelligent transformation will enhance interactions between AI and the physical world, catalyzing the evolution of artificial intelligence to new heights.
As we move to realize these missions, Huawei is committed to working alongside all partners to accelerate the intelligent transformation journey for a brighter future.”
[1] Responsible AI Institute: From GenAI to AI Agents: Preparing for the Next Evolution in Artificial Intelligence (February 18, 2025,Nicole McCaffrey)
[2] 安永：How AI agents will take GenAI from answers to actions（15 Nov 2024）
[3] IOPS：input/output operations per second
