June 2026’s roundup: SK Group, GS Group, and Naver unveil a $650 billion, 10-year AI infrastructure megaproject, Google shifts part of its TPU manufacturing from TSMC to Intel, Nvidia diversifies its HBM4 supply chain, and AI data center capacity keeps expanding into new global markets from Indonesia to Paraguay.
At a glance – Digital Infrastructure Insights for June 2026
- SK Group, GS Group, and Naver will execute a two-phase, 10-year AI infrastructure megaproject with an estimated budget of $650 billion — the largest single commitment covered in this series to date.
- Google reportedly placed a 2028 order with Intel to manufacture more than 3 million custom Tensor Processing Units, shifting part of its TPU manufacturing away from TSMC.
- Nvidia approved Samsung, SK Hynix, and Micron as HBM4 suppliers for its Vera Rubin AI platform, with production underway and first shipments planned for Q3.
- A 360MW data center in Batam, Indonesia is under construction to house up to 170,000 Nvidia accelerator chips, while Taiwan committed $200M to build a diplomatically-backed data center in Paraguay.
- Apple raised MacBook and iPad prices by roughly 20%, citing AI-driven memory component shortages, while SK hynix filed for a Nasdaq listing targeting a roughly $29B raise and $1.2 trillion valuation.
- Alibaba Cloud opened a new France cloud region with plans for enterprise-grade agentic AI services in Europe, as the European Commission moved to designate Microsoft and Amazon cloud services as Digital Markets Act gatekeepers.
1. South Korea’s AI infrastructure buildout reaches historic scale
South Korea’s combined AI infrastructure commitments this month dwarf any single country’s activity covered elsewhere in this series:
- SK Group, GS Group & Naver: will execute a two-phase megaproject over ten years with an estimated budget of $650 billion — by far the largest single AI infrastructure commitment covered in this series to date.
- Nvidia: announced a multi-year technology partnership with SK Hynix to advance next-generation memory for AI data centres, cooperation with Naver and Doosan to build AI data centres, and confirmed SK Telecom‘s plans to deploy a gigawatt-scale AI cloud using Nvidia technology with the first data centre targeted for 2027; Doosan expects its energy solution to integrate with Nvidia data centre platforms and contributes materials used in Nvidia’s Blackwell chips — deal values were not disclosed.
- SK hynix: filed with the US SEC to list American Depositary Receipts on Nasdaq, planning to issue 17.79 million new shares to raise about KRW45.5 trillion (roughly $29.4B), implying a company valuation near $1.2 trillion.
- LG Uplus: designated its forthcoming 200MW Paju AI data center as the centerpiece of its AI data center (AIDC) infrastructure strategy, prioritizing inference-driven AI workloads.
Why it matters for buyers: South Korea’s AI infrastructure commitments this month — a $650B megaproject, Nvidia partnerships spanning three separate companies, a ~$29B IPO, and a dedicated 200MW facility — collectively dwarf any single country’s AI infrastructure activity covered elsewhere in this series. Buyers with APAC supply chain or infrastructure dependencies should treat South Korea as a primary AI infrastructure hub on par with the US, not a secondary market. → Model South Korea-scale infrastructure economics with TeckNexus ROI/TCO tools.
2. Global AI data center capacity expands outside traditional hubs
AI data center capacity landed in genuinely new geographies this month, for reasons beyond pure economics:
- Batam 360MW Data Center: under construction in Batam, Indonesia to accommodate up to 170,000 Nvidia accelerator chips, with operations expected by Q1 2027.
- Taiwan Government: committed $200 million to build a data center in Paraguay, targeting 10MW computing capacity with operations expected by end-2027; the facility will be constructed on government land and is reported as a diplomatic initiative, with backing cited from the Trump administration.
- Meta Platforms: expanded its partnership with Reliance Industries — Reliance will build a 168MW AI-enabled data centre in Gujarat, India, which Meta will lease to scale its AI infrastructure in India.
- PLDT: filed with the Philippine SEC to rebrand its data center subsidiary as Vitro REIT and plans to raise about $395 million via a REIT initial public offering.
Why it matters for buyers: AI data center capacity is now landing in genuinely new geographies for reasons beyond pure economics — Taiwan’s Paraguay facility is explicitly framed as diplomatic infrastructure, not a cost-optimization decision, while PLDT’s REIT structure shows data center assets themselves becoming a distinct investable asset class in emerging markets. → Compare emerging-market data center economics with TeckNexus ROI/TCO tools.
3. Chip supply diversifies beyond Nvidia and TSMC
The largest AI buyers are building direct manufacturing relationships rather than relying solely on Nvidia and TSMC:
- Google: reportedly selected Intel to fabricate more than 3 million custom Tensor Processing Units (TPUs) for delivery in 2028 after months of process validation, shifting part of its TPU manufacturing from TSMC; the TPUs support AI training and inference for Google Cloud.
- OpenAI & Broadcom: introduced “Jalapeno,” a custom inference processor co-developed with Broadcom, tailored to OpenAI‘s inference systems and positioned as the first release in a jointly developed multi-generation compute platform.
- Nvidia — HBM4 diversification: approved Samsung, SK Hynix, and Micron as HBM4 suppliers for its Vera Rubin platform, with production in progress and first shipments expected in Q3 — diversifying supply for scarce high-bandwidth memory and addressing AI data center bottlenecks, coinciding with a new R&D center in South Korea.
- Nvidia — China alternative: told Chinese customers its Vera CPU for AI data centers could be available in August, positioning it as an alternative after H200 GPU shipments were stalled by US export restrictions, per Reuters.
Why it matters for buyers: Google’s TPU shift from TSMC to Intel and OpenAI/Broadcom’s custom inference chip both point to the largest AI buyers building direct manufacturing relationships rather than relying solely on Nvidia and TSMC — a genuine structural shift in who controls AI compute supply, not just a diversification headline. → Compare chip supply diversification strategies with TeckNexus Technology Selector.
4. Memory scarcity and next-gen chip design reshape pricing and roadmaps
AI-driven demand is now flowing through to non-AI consumer hardware pricing:
- Apple: implemented approximately 20% price increases on MacBook and iPad lines, attributing the move to rapid rises in memory component costs linked to AI infrastructure demand; iPhone pricing remains unchanged.
- IBM: disclosed a sub-1nm (0.7nm) silicon design employing a 3D stacked architecture that fits nearly 100 billion transistors in a fingernail-sized area, claiming roughly double the density of its 2nm demonstration from 2021, targeting continued performance and efficiency gains as physical scaling nears atomic dimensions.
Why it matters for buyers: Apple’s price increase is a concrete, consumer-facing data point showing AI-driven memory demand is now flowing through to non-AI consumer hardware pricing — a cost dynamic buyers should factor into any hardware refresh cycle, not just AI-specific procurement. → Track memory and chip pricing dynamics with TeckNexus ROI/TCO tools.
5. Physical AI data center infrastructure and edge computing mature
This month’s infrastructure announcements span the full physical stack, from modular construction to edge distribution:
- Vertiv: announced SmartRun, a modular prefabricated overhead infrastructure solution intended to speed the build-out of high-density AI data centers.
- CoreWeave: announced it has brought a fully validated and operational Vera Rubin NVL72 system online in its AI cloud, positioning itself as the first AI cloud provider to do so.
- Hewlett Packard Enterprise: introduced AI data center networking and routing updates, Agentic AIOps enhancements, and a unified AI-native SASE platform — adding support for HPE Networking CX wired access switches in the HPE Mist platform, expanding Marvis insights and self-healing in Aruba Central, and applying agentic reasoning for faster root-cause analysis; HPE also launched HPE Juniper Networking QFX switches optimized for inference and scale-up architectures, integrated into the HPE AI Data Center Solution and managed via HPE Networking Data Center Director.
- Equinix: launched the first phase of an AI-ready data center in Hong Kong, providing 1,000 cabinets with an initial investment of US$124 million.
- Telefónica: finished rolling out 17 distributed edge computing nodes across Spain, creating a low-latency compute layer integrated with its network to support digital services and data residency requirements.
- Megaport: announced plans for a globally distributed AI inference cloud built around an on-demand GPU pool funded by A$350 million, offered to enterprise customers via contracted and consumption-based pricing models.
Why it matters for buyers: this month’s infrastructure announcements span the full physical stack — modular prefab construction (Vertiv), validated next-gen compute (CoreWeave), networking/AIOps (HPE), colocation (Equinix), edge distribution (Telefónica), and inference-specific cloud (Megaport) — suggesting AI data center infrastructure is maturing into distinct, specialized layers rather than one undifferentiated buildout. → Plan multi-layer AI infrastructure with TeckNexus Network Planning tools.
6. Cloud sovereignty, regulation, and GPU-as-a-Service models emerge
A genuine regulatory inflection point landed alongside new sovereignty and consumption-based cloud models:
- Alibaba Cloud: opened a fully operational cloud region in France with two availability zones in Paris, adding to existing European infrastructure in Germany and the UK; the region offers enterprise cloud services across compute, storage, containerization, networking, security, databases, and developer tools, and supports a forthcoming rollout of enterprise-grade agentic AI services in European markets later this year.
- European Commission: following probes into Amazon Web Services and Microsoft Azure, issued a preliminary decision to classify their parent companies as Digital Markets Act gatekeepers for cloud computing, signaling forthcoming interoperability, data portability, and anti-self-preferencing obligations pending final designation.
- Verizon Business: disclosed a partnership to trial a GPU-as-a-Service proof of concept aimed at supporting AI workloads, per comments from SVP and Chief Product Officer Scott Lawrence; partner identity and technical specifications were not disclosed.
- Tech Mahindra: is implementing AI-first operations with cloud-native orchestration and infrastructure-as-code frameworks as part of an engagement with Telefónica Germany to modernize operations.
Why it matters for buyers: the EU’s DMA gatekeeper designation for AWS and Azure is a genuine regulatory inflection point — it would impose interoperability and anti-self-preferencing obligations on the two dominant cloud providers specifically, right as Alibaba Cloud is expanding European capacity and agentic AI services. Buyers building multi-cloud or EU-based AI strategies should track how this designation plays out over the coming months. → Compare sovereign cloud and GPU-as-a-Service models with TeckNexus Vertical Intelligence.
Every June item, with full source detail, is on the curated Digital Infrastructure Monthly Insights page →
A note on scope: this roundup covers data center infrastructure, power/cooling, chip manufacturing and supply chain, and physical AI infrastructure specifically. AI models, agents, and applications are covered in the companion AI & Automation Insights series; wireless/spectrum/satellite infrastructure is covered in Advanced Connectivity Insights. Five archive items were excluded from this draft for lacking any identifiable company, deployment, or technical detail: a generic “shift to agentic AI workloads” commentary piece, an unnamed “sovereign AI project,” an unnamed “AI Factory Operator” model shift, a generic note on generative AI vendors entering India, and a fragment that itself referenced an unrelated sour gas plant listing. Two Verizon GPU-as-a-Service entries covering the same story were consolidated into a single mention.
What this means if you’re planning digital infrastructure
June’s throughline is that AI infrastructure has entered a phase of genuine structural diversification — geographically (South Korea’s $650B commitment, new capacity in Indonesia, Paraguay, and India), in chip supply (Google/Intel, OpenAI/Broadcom, Nvidia’s multi-supplier HBM4), and in regulatory structure (the EU’s DMA gatekeeper move). Five moves follow directly from the month:
- Treat South Korea as a primary AI infrastructure hub in your APAC planning — this month’s combined commitments (megaproject, Nvidia partnerships, IPO, dedicated facility) are unmatched by any other single country covered in this series.
- Watch which AI data center locations are driven by diplomatic or geopolitical logic (Taiwan’s Paraguay facility) versus pure cost/latency optimization — the underlying risk profile differs.
- Build vendor diversification into your own chip supply assumptions — Google’s Intel TPU order and OpenAI’s Broadcom chip both show even the largest AI buyers are no longer defaulting to a single manufacturing partner.
- Factor AI-driven memory pricing into non-AI hardware refresh budgets — Apple’s 20% price increase shows this cost pressure has already reached consumer devices.
- Track the EU’s DMA gatekeeper designation for AWS and Azure closely if your cloud strategy touches Europe — the interoperability and anti-self-preferencing obligations could reshape vendor lock-in dynamics.
Catching up? Last month’s edition is here: Washington Takes a Stake in Intel as AI Infrastructure Heads to Orbit: Digital Infrastructure Insights, May 2026 →
→ Start with the TeckNexus Intelligence Platform — independent, buyer-neutral tools for spectrum, TCO, architecture, and RFP decisions.
This analysis is drawn from TeckNexus’s full curated Digital Infrastructure Monthly Insights for June 2026. See every deployment, product, and partnership update
