Huawei's vision for the future of data storage in the AI era

In an exclusive interview, Xie Qiangqiang, Vice President of Huawei Data Storage Product Line and President of Huawei Data Storage Carrier Domain shared insights into how the company is evolving its storage strategies to meet the demands of an AI-powered world during MWC 2025 in Barcelona.

With 50% of his focus on the Chinese market and the other half dedicated to global operations, Xie provided a comprehensive view of the challenges, innovations, and opportunities in this rapidly shifting landscape.

AI has become an integral part of business and technology discussions worldwide, and data storage is no exception. "Before last year, when I went to Brazil, carrier customers were asking the same question: ‘What do we need to do to be AI-ready?’" Xie remarked.

He explained that while many industry players emphasise AI-ready infrastructure, Huawei believes that "AI-ready should be data-ready first."

One of the biggest challenges for carriers is data sharing. "The problem is not that they are not ready for AI; it's that their data is not ready," Xie said.

Carriers still struggle with siloed data stored across different applications and data centres, making it difficult to leverage AI effectively. To address this, Huawei has introduced the AI Data Lake, a solution designed to support customers in becoming AI-ready.

Beyond AI readiness, performance remains a critical challenge. Xie noted that while GPU cards are powerful, their utilisation rates remain disappointingly low, often less than 50%.

"That’s a big waste of investment," he said. Huawei aims to optimize storage solutions and provide OceanStor A Series High-Performance AI Storage to reduce GPU wait time and avoid repeated inference, increasing utilisation rates by up to 10%-30% and cutting costs by 40%.

One of the key industry trends making waves at Mobile World Congress 2025 (MWC) this year was the shift from telecommunications companies (telcos) to technology companies (techcos).

Traditionally, storage solutions were considered a background element in the telecom industry. "Before, when we talked about storage, carriers didn’t care very much—it was hidden between applications and networks," Xie explained. "But now, even CIOs and CEOs recognise the importance of monetising data."

This shift is crucial because telcos are now repositioning themselves as ICT (Information and Communications Technology) solution providers.

By integrating AI into their business models, they can expand beyond connectivity and offer AI-powered services. "The business driver is no longer just connectivity; it’s about developing emerging AI services based on AI-Ready data storage," Xie emphasized.

Innovation is central to Huawei’s approach to data storage. Xie stressed that Huawei does not just focus on storage itself but offers a comprehensive AI-Ready data storage.

"We provide the DCS AI solution, which integrates compute, storage, virtualisation, container, big data and AI itself," he said.

A key area of innovation is the speed and efficiency of AI cluster deployment. "Before, it might take two months to install an AI cluster," Xie noted. "Now, with our solution, it can be done in just seven days." This level of efficiency is game-changing for enterprises looking to accelerate their AI adoption.

Use cases for DCS AI solution extend beyond telecoms. "We’ve collaborated with hospitals in China to integrate AI into medical data processing, improving efficiency for pathology implementation," Xie said. Other use cases include AI call centres and smart manufacturing.

With AI becoming increasingly data-intensive, protection and resilience concerns have grown. "Data is the foundation of AI," Xie stated. "No data, no AI."

Last year, a staggering 92% of ransomware attacks targeted not just production sites but also backup sites, highlighting the need for robust protection solutions. Huawei addresses this with end-to-end data protection solution that includes air-gap technology, multilayer ransomware protection, 4 DC multi-active solution and isolated recovery zones. "At any time, our customers need clean data for the fastest recovery," Xie explained. Huawei’s data protection approach focuses on three key aspects:

1. Understanding what is happening within the network.

2. Reducing the impact of ransomware attack or disaster incidents.

3. Enabling rapid business recovery.

Huawei operates in a diverse range of markets, each with distinct data storage requirements. "In China, the focus has shifted from supporting AI clusters to supporting inference scenarios," Xie said. Meanwhile, regions such as the Middle East and Europe have different priorities.

"Regardless of the market, the core principles of data strategy remain the same," he added. Previously, storage solutions were evaluated based on three dimensions: performance, reliability, and energy efficiency. However, the industry has evolved, and Huawei now considers six dimensions, adding scalability, easy management, and data paradigm.

Energy efficiency is a growing concern for carriers, and Huawei is actively working on reducing the Total Cost of Ownership (TCO) for storage solutions. "We’ve launched the new-gen OceanStor Pacific scale-out storage that stores twice the data volume with the same space and energy consumption," Xie revealed. "This means that in just two rack units, we can provide four petabytes of storage. Before, achieving this capacity required an entire rack."

In addition to hardware innovation, Huawei is improving data deduplication and compression technology. "Our Data Management Engine (DME) not only manages devices but also manages the data itself, optimising energy usage and reducing costs," Xie said.

As AI continues to reshape industries, data storage solutions must evolve to meet new demands. Huawei is positioning itself at the cutting edge of this transformation, with a focus on AI readiness, data resilience, innovation, and sustainability.

RELATED STORIES

Huawei’s Li Peng: AI will drive 5G-A growth, boosting network monetisation

Huawei on accelerating industrial digitalisation and intelligence