Equinix and FuriosaAI Target Europe With AI Data Center Push
TECHWORLD ·
✦ AI Summary
Equinix and FuriosaAI are targeting the European market based on cooperation in AI semiconductors and data center infrastructure.
FuriosaAI began mass production of its second-generation AI inference semiconductor this year and is carrying out a proof of concept (POC) in Europe.
Equinix is supporting FuriosaAI's local market entry and additional collaboration based on its globally standardized data center operations experience and infrastructure support.
Equinix and FuriosaAI are seeking to expand into the European market based on collaboration in AI semiconductors and data center infrastructure. FuriosaAI is a domestic AI semiconductor design company that is pursuing global business expansion, starting with Europe. Equinix is supporting FuriosaAI's local market entry by providing globally standardized data center infrastructure and operational experience.
Against this backdrop, the two companies have recently been carrying out a proof of concept (POC) for the European deployment of FuriosaAI's AI inference semiconductor solution through their cooperation. FuriosaAI began mass production of its second-generation AI inference semiconductor this year and has operating experience in the domestic market. Based on that experience, FuriosaAI is exploring new business opportunities in Europe.
A media briefing was held on the 15th at Equinix's office in Yeouido, Seoul. The briefing featured Equinix CEO Jang Hae-deok and Product Manager Yoon Ji-hoon, who explained the background to the partnership, the ongoing proof of concept (POC) in Europe, and plans for future global business expansion. Anthony Ho, Equinix's APAC Director of Product Management, gave a presentation on the networking impact of agentic AI.
Jang Hae-deok, CEO of Equinix, said that FuriosaAI is an AI semiconductor design company and Equinix is a global data center infrastructure provider, and that he believed the two companies could generate synergy because their areas of strength are clearly distinct.
FuriosaAI's current operating base for AI inference infrastructure is in South Korea. However, entering global markets requires an understanding of local data center environments and customer requirements.
As sovereign AI, which emphasizes national data sovereignty and infrastructure self-reliance, has emerged as a major topic in the AI industry, securing real-world implementation experience in Europe has become an important task. Accordingly, FuriosaAI has chosen Europe as its first base for global business expansion.
FuriosaAI is carrying out POC activities in Europe. Its European strategy goes beyond product sales and focuses on actually operating AI semiconductor solutions in local data center environments, through which it will verify both technological competitiveness and business viability.
Yoon Ji-hoon, Product Manager at Equinix, said FuriosaAI's challenges in the European market include securing real data and validating its products in local customer environments, and explained that using Equinix's global infrastructure can make preparation for entering new markets more efficient.
He said that in the European market, the location of data centers, data center operating environments, and the stability of power supply directly affect business expansion. He explained that this is because high-performance servers, stable power, cooling, and network environments are essential for large-scale AI infrastructure computation, and data centers provide the foundation equipped with these elements.
Jang said demand in the AI market continues to expand and demand in the data center market continues to expand as well, and that the importance of local infrastructure is rising accordingly. He added that understanding and using infrastructure conditions by region is important for expanding business in global markets.
FuriosaAI has strengths in AI semiconductor design and development. However, applying and operating products in actual server environments in global markets requires expertise in data center infrastructure.
Equinix's role goes beyond providing data center space to advising on the overall process of infrastructure construction and operation. Its support is built on experience operating globally standardized data centers.
Equinix's support base also includes a range of products and solutions. On that basis, Equinix can provide support across multiple stages, from contract discussions to actual deployment and operations.
Yoon, the Product Manager, said Equinix can support in many ways from the contract discussion stage through actual deployment and operations. He also said that the ability to smoothly support detailed processes based on globally standardized data center infrastructure is an advantage of this partnership.
The speaker said FuriosaAI's core strength lies in chip design, while Equinix's strength lies in infrastructure operations. He explained that the roles of FuriosaAI and Equinix are divided between design and operations.
He added that Equinix has experience operating multi-region data centers globally, and that Equinix's power management know-how is a major help in actual server operations. He also said Equinix has a broad lineup of data center products and solutions.
The speaker said that when AI semiconductor companies enter global markets, operational efficiency and stability matter in addition to product performance. He explained that the efficiency and stability of actual data center operations are competitive factors.
He also said server power management, cooling, and network connectivity are key factors that determine the cost and performance of AI inference services. Considering these points, he explained, the scope of cooperation between the two companies is likely to expand beyond a single POC.
The two companies plan to begin by establishing a foothold in Europe. They also plan to identify additional collaboration opportunities in global markets in the future, with FuriosaAI expanding its global business based on AI semiconductor technology and Equinix supporting data center infrastructure and operational capabilities in each region.
Jang said that through this partnership, he expects a blueprint for future global business expansion that goes beyond a short-term project.
Jang said that Equinix's diverse product and solution lineup could create multiple collaboration opportunities.
Yoon said he is considering additional collaboration possibilities with FuriosaAI based on Equinix's ability to provide a wide range of global infrastructure environments.
A presentation session followed, during which Anthony Ho, Equinix's APAC Director of Product Management, spoke on the networking impact of agentic AI.
Equinix conveyed the message that network AI speed evolution is needed in the AI era. Director Ho said AI's impact on the networking environment is expanding rapidly, and projected a 85% compound annual growth rate for new AI network traffic over the next 10 years. He also projected that the growth rate in enterprise networking for agentic AI would reach 175% annually by 2029.
The limits of existing network operations methods were also pointed out. Network operations have remained tied to manual work based on ticketing, scripts, and specialist dependency, and such manual operations have caused delays in service deployment. In complex environments, there have also been problems with slower fault detection and response.
Director Ho said changes in network operations methods are needed to support AI workflows. He explained that passive management methods delay AI adoption and innovation, and that distributed telemetry can reduce network visibility.
With this awareness of the issue, Equinix is promoting a distributed AI hub strategy. The core of the strategy is 'Equinix Fabric Intelligence,' whose purpose is to connect AI infrastructure and networks. The strategy focuses on connecting private and public data, connecting open and closed AI models, and connecting diverse ecosystems such as cloud service providers and neoclouds.
Equinix is also promoting a shift from existing manual ticket-based methods to automated deployment through natural language-based network provisioning and the use of AI agents. The company said its Fabric Super Agent can shorten network deployment times. The reduction can range from weeks to minutes.
Director Ho said network and infrastructure operations must change to match AI usage environments. He explained that AI systems and complex network connections are needed, that natural language-based infrastructure design and deployment environments are needed, and that a transition to AI networks at AI speed is necessary.
In line with this, Equinix said it plans to provide 'Fabric Insights.' 'Fabric Insights' was presented as a tool that supports early detection of anomalies based on near-real-time telemetry analysis, early identification of capacity risks and performance risks, and automated troubleshooting.
Source: TECHWORLD · Park Kyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406958
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.