"Since 2016, we have been emphasizing the importance of algorithms, data, and computing power. Eight years later, the sales of AI products have accounted for nearly half of Inspur's overall sales."
This was stated by Mr. Peng Zhen, Chairman of Inspur, during his speech at the IPF2024 Inspur Ecological Partner Conference on April 17. He pointed out that today, the impact generated by artificial intelligence is more rapid than any other technology before, bringing about fundamental changes to the entire social productivity.
From April 17th to 18th, at the Beijing National Convention Center, the IPF2024 Inspur Ecological Partner Conference with the theme "Intelligent Genesis, Creative Vision, Joint Action" was successfully held. Inspur announced its comprehensive expansion and promotion of its "AI+" strategy.
During the two-day conference, Inspur not only unveiled the new year's ecological partner business strategy and a series of heavyweight products but also held multiple thematic forums to explore various cutting-edge technologies and industry trends such as digital economy, generative AI, large-scale model technology, AI for Science, and business intelligence.
In the era of "AI+"
At the main venue of IPF2024, facing the technologically immersive blue giant screen and bathed in the interplay of dazzling lights, every participant could vividly feel the changes that technology has brought to human life.
But the real revolution goes beyond that.
In the broader field beyond the conference venue, artificial intelligence (AI) is sweeping through the traditional patterns of social productivity like a gust of wind, bringing unprecedented and profound changes to the global economy.
Peng Zhen, Chairman of WaveInfo, pointed out that AI has changed the three elements of production relations, transforming labor from traditional "human power" to "human+AI," turning production materials from tangible entities into intangible data, and extending labor tools from human limbs to "intelligent entities."
According to the ARK Invest report, from 2023 to 2030, the contribution of artificial intelligence to the global GDP will exceed 4.5 times the impact of revolutionary technologies like steam over the past 80 years. The efficiency improvements brought about by AI technology in industries are significant, such as a 2.2-fold increase in coding efficiency, a 10-fold increase in video editing, and a 3-fold increase in new drug development. While replacing certain traditional occupations, artificial intelligence is also projected to create a brand-new market worth up to $25 trillion.
Surprising cases abound. Sora can now generate detailed scene videos based on verbal descriptions, Figure AI robots can learn to make coffee independently after observing human activities for 10 hours, and Copilot can quickly generate complex programming test scripts.
"A certain future seems to be rapidly approaching us—where AI intelligence may surpass that of humans, bringing irreversible changes," Peng Zhen envisioned.
Hu Leijun, General Manager of WaveInfo, believes that the significant change brought about by the era of intelligence is that "all computations are AI," with AI computing power permeating into every computing device. The paradigm of computing power for artificial intelligence is continuously evolving, and all computing devices in the future will need to have AI capabilities.
Application-oriented with systems at its core
WaveInfo proposed at the conference that to develop "AI+", it is necessary to adhere to an approach of "application-oriented with systems at its core," focusing on stimulating innovation, expanding innovative pathways, accelerating innovation implementation, and promoting artificial intelligence to become the productivity of hundreds of industries.
Wu Shaohua, Director of AI Software Development at WaveInfo, pointed out that large models have greatly expanded the application boundaries of AI, unleashing countless innovative possibilities.
Since the launch of the "Source 1.0 Large Model" in 2021, WaveInfo has been exploring the application of large model technology. However, the application of large models still faces many challenges, such as outputting seemingly reasonable but practically worthless information due to a lack of professional depth, and the need for technical support for deployment and operation in a rapidly updating data environment.
To better address these issues, WaveInfo officially launched "EPAI (Enterprise-level AI)" at this conference—an end-to-end software platform for enterprise-level large model development. This platform not only supports the entire process from data preparation to model training and knowledge retrieval but also efficiently schedules diverse computing power and multi-mode algorithms, helping enterprises develop and deploy generative AI applications efficiently.
Wu Shaohua told Chinese Science Bulletin that for the illusion problem of large models, EPAI provides over a hundred million pieces of basic knowledge data, and through automated data processing tools, helps industry partners and enterprise customers organize industry data and professional data, generating high-quality fine-tuning data and industry/enterprise knowledge bases to create exclusive data assets for enterprises, thereby enhancing their understanding and adaptability to industry scenarios. In addition, combined with Retrieval-Augmented Generation (RAG) technology, EPAI can perform targeted knowledge retrieval when problems arise, providing relevant professional references, significantly improving the accuracy and reliability of answers.
The future is already here
The world of the future will be built on technology.
In addition to EPAI, WaveInfo has made a series of innovative releases, such as jointly releasing general-purpose intelligent servers with Intel, which can support the inference of large models with hundreds of billions of parameters, building an application-oriented intelligent computing foundation; releasing the distributed all-flash storage AS13000G7 to address challenges in large model training data; and launching the domestic first super AI Ethernet switch X400 to accelerate large model training and inference.
Mao Bailin, General Manager of Channel Promotion Department at WaveInfo, stated that in 2024, WaveInfo will strengthen AI empowerment and implement four major transformative actions: driving transformation with EPAI, accelerating industry AI application, deepening the transformation of distribution value, and building a "Partner Intelligent Hub," comprehensively enhancing the AI+ innovation capabilities of its ecosystem partners, jointly promoting the intelligent transformation of industry customers, and achieving win-win development.
In the future, the world will be driven by openness.
"Open source" and "openness" have always been two keywords for WaveInfo. At the conference, Wu Shaohua emphasized the importance of open cooperation, particularly noting that open source is a major driving force for advancing AI technology innovation and application, reducing the threshold for research and development, and accelerating the market promotion of solutions.
In terms of internationalization, Liu Jun, Senior Vice President of WaveInfo, described WaveInfo's global layout to Chinese Science Bulletin, including research and development centers, manufacturing centers, delivery centers, etc. WaveInfo will continue to serve global customers, continuously promote the construction of an open cooperation ecosystem.
