Samsung’s six affiliates announced a combined $1 billion commitment to Helix Digital Infrastructure on September 29, 2026, pushing the KKR-backed AI infrastructure company’s total secured capital past $11 billion. The deal, confirmed by Samsung Newsroom, Reuters, CNBC, the Wall Street Journal, and Business Wire, is one of the largest single-day infrastructure commitments made by an Asian industrial group to the AI buildout and marks a structural shift in how enterprise AI capacity gets funded and built.

Power, it turns out, is the new GPU.

What Happened

Helix Digital Infrastructure, launched in June 2026, is a purpose-built company that acquires, develops, and operates AI-enabling infrastructure: hyperscale data centers, power generation and transmission assets, fiber-optic networks, and the logistics that connect them. It is led by Adam Selipsky, the former CEO of Amazon Web Services, and backed with long-duration capital from KKR’s global infrastructure platform.

On September 29, Samsung announced a $1 billion commitment through its long-duration capital fund. Samsung Electronics is contributing $500 million; Samsung C&T, Samsung SDS, Samsung SDI, Samsung Life Insurance, and Samsung Fire & Marine Insurance account for the rest.

The round brings Helix’s total secured capital to more than $11 billion. Founding anchor investors include KKR, the Kuwait Investment Authority (KIA), NVIDIA, and Vistra, one of the largest power producers in the United States.

“Samsung’s commitment is a strong vote of confidence in Helix’s strategy and further deepens the long-term capital base we’ve built to meet the scale of AI infrastructure demand,” said Adam Selipsky, CEO and Co-Founder of Helix.

Why Power Is Now the Governing Constraint

The conventional story about AI infrastructure focuses on GPU scarcity. That story is dated. As of mid-2026, the leading constraint on hyperscale AI deployment is power availability, not chips.

Helix’s pitch rests on this diagnosis. By targeting both the compute facility and the underlying power infrastructure simultaneously, Helix bets that whoever controls power will control where AI runs. Vistra’s presence as a founding investor is not incidental: Vistra operates roughly 41 gigawatts of generation capacity across the U.S. and is already partnered with hyperscalers on large-scale power procurement.

Samsung’s investment reinforces the thesis. Samsung SDI manufactures high-capacity UPS and battery backup units (BBUs), critical for maintaining uptime in high-density GPU clusters where even a two-second power interruption can corrupt multi-hour training runs. Samsung SDS operates the Korea AI Computing Center (KOACC) with some of the lowest PUE (Power Usage Effectiveness) levels in the industry.

For enterprise AI teams, the implication is practical: every workload they run on a hyperscale cloud ultimately traces back to a physical power connection. The companies that control those connections will determine which AI capabilities are affordable, available, and compliant with data-residency requirements.

Samsung’s Full-Stack Role

Samsung’s investment is not a passive financial bet. Each affiliate brings operational capability Helix can deploy:

Samsung EntityRole in AI Infrastructure
Samsung Electronics DSSemiconductors: HBM, DRAM, advanced memory for GPU clusters
Samsung Electronics DXData center cooling: air and liquid CDUs via FläktGroup (acquired 2025)
Samsung C&TEngineering, procurement, and construction (EPC) for large data centers and power plants
Samsung SDSData center design, build, and operations; GPUaaS; ultra-low PUE operations
Samsung SDIUPS and battery backup units for continuous-uptime data centers
Samsung Life / Fire InsuranceLong-duration capital and risk underwriting for infrastructure assets

This full-stack involvement means Samsung is not merely writing a check. It is committing to be a supply-chain partner, operator, and financier across the most capital-intensive layers of AI infrastructure.

The KKR-Helix press release notes: “Helix expects to explore opportunities to leverage Samsung’s capabilities across areas such as advanced technology, construction, energy storage, and cooling to support the development and delivery of AI infrastructure at scale.”

Adam Selipsky and the Cloud-to-Infrastructure Pipeline

The choice of Adam Selipsky as CEO matters. Selipsky ran AWS from 2021 through 2024, presiding over the period when the cloud platform became the dominant channel for enterprise AI deployment. He understands hyperscaler procurement cycles, enterprise SLA requirements, and the operational complexity of running infrastructure at global scale.

His move to Helix reflects a broader trend: the most experienced cloud operators are now the most valuable assets in physical infrastructure, because AI infrastructure requires the same reliability engineering, capacity planning, and enterprise contract management that hyperscale cloud platforms have spent two decades learning.

Helix is not a real estate company that happens to host servers. It is being built as a managed infrastructure platform that sells guaranteed capacity to hyperscalers, with the operational DNA of a cloud provider wrapped around physical assets.

What the $11B Capital Base Signals

The $11 billion committed to Helix is structured as long-duration capital, the kind typically used in infrastructure funds that hold assets for ten to twenty years. This is not venture capital looking for a three-year exit. It is patient money designed to match the amortization curve of a data center or a power plant.

For enterprise AI buyers, the strategic read is this: a new class of infrastructure operator is emerging that sits between the traditional colocation market (Equinix, Digital Realty) and the hyperscale clouds (AWS, Azure, GCP). These operators aim to deliver hyperscale-grade compute with the geography flexibility, sovereign AI compliance, and power integration that enterprise buyers increasingly require.

The Kuwait Investment Authority’s presence alongside NVIDIA and Samsung signals that sovereign wealth capital is now a meaningful funding source for AI infrastructure. That matters for enterprises in regions with strict data-residency requirements: sovereign-backed infrastructure providers may offer more predictable long-term compliance guarantees than hyperscalers operating under U.S. company law.

Competitive Landscape

Helix is not the only new entrant in this space. In July 2026, the EU announced its InvestAI program targeting 30 billion euros for AI gigafactories across member states (covered in EU AI Gigafactory announcement). Crusoe, which raised $3 billion in September 2026, is building AI cloud infrastructure on stranded and renewable power. Positron AI closed its $875 million Series C earlier this month, targeting memory-first inference hardware.

The pattern is consistent: AI infrastructure is attracting capital at a pace that suggests the market does not yet trust that existing cloud capacity can absorb the coming demand from agentic AI workloads.

The difference with Helix is the breadth of its supply chain integration. Most infrastructure plays focus on one layer: compute, power, or construction. Helix, with Samsung’s full affiliate network, is attempting a vertically integrated model spanning all three.

What Enterprise AI Leaders Should Do Now

The Samsung-Helix deal has several practical implications for enterprise AI strategy:

1. Map your infrastructure dependencies. Most enterprise AI teams know which models they use but not which data centers those models run in, who owns the power feeding those data centers, or what the concentration risk looks like. That mapping exercise is now strategic, not just operational.

2. Sovereign AI requirements are becoming infrastructure requirements. The Kuwait Investment Authority’s participation in Helix signals that Gulf sovereign wealth funds are betting that AI infrastructure with regional ownership structures will command a premium from enterprise buyers with data-residency requirements. Enterprises in regulated industries should monitor whether Helix or similar platforms offer compliant alternatives to U.S. hyperscalers for sensitive workloads.

3. Power costs will affect AI unit economics. As power becomes the governing constraint on AI scaling, the cost of power will increasingly flow through to inference and training prices. Enterprises building long-term AI budgets should model scenarios where compute costs track power costs, not just chip availability.

4. Long-horizon infrastructure signals long-horizon AI demand. KKR and Samsung do not commit long-duration capital to ten-year infrastructure bets unless they believe the underlying demand is durable. Their commitment is a strong external signal that enterprise AI adoption is not a cyclical trend but a structural shift.

For enterprises already deep into agentic AI deployments, the Samsung-Helix deal is confirmation that the infrastructure layer is commoditizing upward. The constraint is no longer whether AI models exist for a given task. The constraint is whether the infrastructure to run them at enterprise scale, with acceptable cost, latency, and compliance, is available. Helix is a bet that it will be.

Sources

Enera helps enterprise teams build AI-native GTM and operations systems. If your team is navigating infrastructure decisions for production agentic deployments, book a call. You can also explore our thinking on autonomous GTM systems and AI-native transformation maturity.