Alpha Compute: $21 Million NTM Sales Outlook Shows Scale-Up Potential – Quarterly Update Report

CRWV ALP SMCI NBIS IREN Download the Complete Report Here $32.2 million AI lab contract validates ALP’s pivot to AI GPUaaS and confidential compute, creating $16.1 million of annual contracted revenue. $21 million NTM revenue outlook anchors the scale-up trajectory, expanding the revenue base beyond the initial AI lab contract through expected GAMEE contribution. ALPHA-01 is now live with 504 NVIDIA B200 GPUs, while ALPHA-02 targets 576 B300 GPUs in Sweden in 3Q26. GAMEE adds consumer-platform optionality, with 119 million registered users, 10 billion gameplays, and 56% y/y 1Q revenue growth. ALP trades at 0.41x NTM sales versus the peer average of 11.85x, despite early contracted AI compute revenue visibility. Alpha Compute Corp.’s (ALP) mid-quarter update points to a material inflection from concept-stage AI infrastructure buildout toward contracted, revenue-generating operations. ALP secured a $32.2 million, two-year agreement with a leading AI research firm, translating to $16.1 million of annual contracted revenue, up from ~$30,000 in 1Q26. The contract supports ALPHA-01, the company’s inaugural enterprise-scale deployment comprising 504 NVIDIA B200 GPUs in Canada. ALP is also advancing ALPHA-02, a 576-GPU NVIDIA B300 cluster in Sweden targeted for 3Q26 deployment. Importantly, ALP now expects $21 million in NTM revenue, including contracted AI compute revenue and expected contribution from the pending GAMEE acquisition, alongside a $200+ million qualified sales pipeline across AI research, enterprise, and sovereign compute customers. The NTM outlook is a key scale-up indicator, extending beyond the initial AI lab contract to reflect broader infrastructure monetization and anticipated contribution from GAMEE. The rebrand from AlphaTON Capital to Alpha Compute formalizes the company’s strategic shift from Telegram ecosystem exposure to AI GPUaaS and confidential compute infrastructure. The company changed its name and Nasdaq ticker from ATON to ALP on April 20, 2026, explicitly positioning itself as an AI GPU-as-a-service and confidential compute platform rather than primarily a TON ecosystem investment vehicle. The new platform is focused on confidential AI compute, sovereign AI infrastructure, scalable GPU deployment, and privacy-native AI ecosystems, with binding agreements for the deployment of over 1,000 Blackwell-generation GPUs and strategic relationships across Telegram, Animoca Brands, Midnight Network, and GAMEE. The Cocoon software stack deployment is in pilot phase as Telegram begins scaling applications, while the GAMEE acquisition is nearing completion pending final audits. Strategically, ALP is positioning for rising demand from regulated AI workloads through hardware-level encryption and TEE-based confidential compute powered by Blackwell-generation GPUs. We believe the $32.2 million AI lab contract provides the clearest commercial validation of ALP’s AI compute strategy, establishing a measurable revenue baseline for the platform. The two-year definitive lease agreement with a leading frontier artificial intelligence laboratory has a total contract value of $32.2 million and delivers $16.1 million of annual contracted revenue, with an expected $7.5 million upfront payment securing the reservation of compute capacity for two years. The agreement covers a dedicated 504-GPU NVIDIA B200 cluster in ALP’s Canadian data center, powered by 100% hydroelectric energy, and provides the customer with exclusive high-performance compute access for next-generation model development. The B200 architecture is designed for large-scale AI training and inference workloads, making the deployment ALP’s first enterprise-scale customer agreement and a meaningful commercial foundation for serving AI laboratories, sovereign entities, and enterprise customers seeking dedicated GPU infrastructure outside traditional hyperscaler environments. ALPHA-01 validates customer demand, while ALPHA-02 and the expansion ROFRs create a visible roadmap from a 504-GPU base toward a multi-cluster Blackwell footprint. ALPHA-01 is live for its first enterprise customer as of May 2026, with the Canada cluster consisting of 504 NVIDIA B200 GPUs powered by 100% hydroelectric energy. ALP also holds a ROFR to expand the Canada site to more than 1,000 B200 GPUs under ALPHA-03. ALPHA-02 is under construction at Equinix’s AtNorth datacenter in Sweden and is expected to include 576 NVIDIA B300 GPUs, also powered by 100% hydroelectric energy, with go-live targeted for 3Q26. ALP also holds a ROFR to expand the Sweden facility to more than 1,000 B300 GPUs under ALPHA-04. The planned ALPHA-01 through ALPHA-04 footprint is projected to support approximately $72 million of annual revenue, or $6 million per month, if fully operational. The roadmap is meaningful because the already-live 504-GPU cluster supports $16.1 million of annual contracted revenue, while the company’s NTM sales outlook captures the first broader revenue step-up and

Alpha Compute: $21 Million NTM Sales Outlook Shows Scale-Up Potential – Quarterly Update Report

Key takeaways

More in business