Why it matters
The strategic read
Turns the earlier Argentum / Boosteroid / DL Invest 300 MW framework into a much larger European GPU-capacity lane. A Super Micro-backed independent cloud platform is trying to aggregate power, real estate, financing, high-density colocation, cooling, networking and GB300 supply into a single non-hyperscaler AI factory, while the reported $7.8B contract stack shows that enterprise and AI-native demand is still spilling into specialist infrastructure operators outside the U.S. hyperscaler core.
Source-supported terms
What was disclosed
- 01
Barron's reported on June 23, 2026 that Argentum AI told the publication it had signed two major long-term AI infrastructure deals worth a combined $7.8B.
- 02
The reported contracts cover deployment of about 47,000 NVIDIA GB300 chips at a 300 MW data center in Poland.
- 03
Barron's reported one agreement is worth $4.1B over five years and the other is worth $3.7B, with computing capacity expected to begin coming online in 2026.
- 04
Barron's also reported that Super Micro Computer has invested $100M in Argentum AI through a convertible note.
- 05
Data Center Dynamics separately reported the $4.1B contract with an unnamed leading AI company on June 23, saying it covers long-term access to around 27,000 NVIDIA GB300 GPUs delivered through Argentum's cloud platform.
- 06
DCD reported Argentum will roll out that capacity in phases starting in 2026, while details about where the capacity will be deployed were not shared.
- 07
DCD also tied the update back to the prior Boosteroid / DL Invest Group project, reporting that Argentum will deploy institutional-scale GPU infrastructure within a new 300 MW-plus data center based in Bielsko-Biala, Poland.
- 08
Argentum AI's own site highlights the Barron's June 23 coverage and describes the company as an independent enterprise compute platform aggregating global GPU and power capacity under long-term contracts.
- 09
Argentum AI previously announced on May 15, 2026 a 300 MW AI data-center infrastructure agreement valued at approximately $2.5B with Boosteroid and DL Invest Group.
- 10
The May 15 announcement said the agreement establishes a long-term framework for hyperscale AI infrastructure supporting enterprise AI workloads, advanced model training and next-generation inference.
- 11
Argentum AI said the facility would include high-density colocation infrastructure, advanced cooling, redundant power architecture and enterprise-grade network connectivity optimized for AI compute.
- 12
The May agreement was expected to support tens of thousands of next-generation GPUs, including future NVIDIA GB300 systems.
- 13
Reuters separately reported the May 15 deal, citing Argentum AI and describing it as one of the largest independent AI infrastructure projects in the region.
Disclosure boundary
What is known—and what is not
Barron's market report citing Argentum AI for the June 23 $7.8B / 47,000 GB300 update, supported by same-day DCD reporting on the $4.1B / 27,000 GB300 contract, Argentum AI's own site linking the Barron's coverage, and the company's May 15 PR Newswire announcement for the 300 MW Boosteroid / DL Invest Group project. Public materials disclose the aggregate reported value, two-contract framing, reported GPU count, reported Poland / 300 MW-plus Bielsko-Biala data-center framing, Super Micro backing, prior 300 MW framework parties and the high-density colocation / cooling / redundant power / network scope. Public materials do not disclose the customer names for the June 23 contracts beyond DCD's unnamed leading AI company description, full contract text, cancellation rights, SLAs, delivery milestones, financing package, purchase orders, utility agreements, site coordinates, construction/EPC awards, permits, final GPU allocation schedule or whether all reported capacity is legally unconditional.