- Samsung’s $1 billion Helix commitment combines capital with industrial capabilities
- Its businesses span chips, cooling, construction, data-center operations and backup power
- • • Most rivals would need partnerships to assemble a comparable infrastructure offering
Samsung can make your fridge and your phone. Now it wants a larger role in building your AI data center. But how many other technology companies can fund that infrastructure and contribute to so many parts of it?
Samsung Electronics and its affiliates have committed $1 billion to Helix Digital Infrastructure, a company launched in June 2026 by private-equity firm KKR and backed by Nvidia, among others. Helix is designed to span the infrastructure value chain: hyperscale data-center development and operations; baseload and flexible power generation; electricity transmission and distribution; and fiber-optic networks. It also aims to shorten the process of bringing AI data centers into operation.
Samsung said the move would help it shift from a traditional hardware-component supplier to an active architect of the global AI infrastructure ecosystem. Rather than simply selling components to data-center builders, it is backing a company that plans to build the infrastructure, with Samsung’s components and expertise contributing along the way.
The South Korean group has several businesses that can contribute. Samsung Electronics’ Device Solutions division makes DRAM, flash memory and other memory chips used in GPU clusters. Its Device eXperience division supplies data-center cooling technology through FläktGroup, acquired in 2025.
Samsung C&T’s Engineering & Construction Group can deliver large infrastructure projects as an engineering, procurement and construction contractor. Samsung SDS designs, builds and operates data centers and has entered the GPU-as-a-service business, allowing customers to rent GPU computing capacity rather than buy the hardware outright.
Samsung SDI supplies battery technology for uninterruptible power supplies used in data centers.
Who else could assemble the pieces?
Huawei is an obvious contender. It can provide Ascend chips and other semiconductors, storage, AI networking and interconnectivity, security, and data-center power and cooling technology.
SoftBank is another contender, with semiconductor capabilities through Arm, Ampere Computing and Graphcore. The group is also developing a major AI data-center project in Ohio.
Nokia has AI data-center network fabrics, IP routing and switches. It has attracted investment from Nvidia, but would probably need to partner with the chipmaker and others to offer the broader infrastructure model Samsung is pursuing.
Ericsson appears further from assembling that complete offering. Its work with SoftBank on AI-RAN focuses on AI architectures for cellular radio networks rather than the full AI data-center infrastructure model.
Samsung’s breadth is the distinction. Outside Samsung, Huawei and SoftBank, most technology companies would need partnerships to bring together a comparable combination of capital, components and infrastructure expertise.