We came across a bullish thesis on Nebius Group N.V. on Darius Dark Investing’s Substack. In this article, we will summarize the bulls’ thesis on NBIS. Nebius Group N.V.'s share was trading at $97.78 as of March 4th. NBIS’s trailing P/E was 794.04 according to Yahoo Finance. Stifel Lowers Credo (CRDO) Price Target Despite Tensorwave AI Infrastructure Collaboration Copyright: ralwel / 123RF Stock Photo Nebius Group N.V., a technology company, engages in building full-stack infrastructure to service the g ...

Core Insights - VCI Global is advancing its strategic transformation towards AI-native infrastructure and software, marking a significant milestone with the launch of Malaysia's first NVIDIA-powered AI GPU computing center [1] - The company aims to solidify its position as a leading AI infrastructure owner and operator in Southeast Asia, catering to the increasing demand for AI computing power from government and enterprise clients [1][2] Group 1: AI Infrastructure Development - V Gallant, a wholly-owned subsidiary of VCI Global, successfully launched an AI GPU computing center on March 3, 2026, establishing the company as a pioneer in the Southeast Asian AI computing sector [1] - The center is designed to meet the growing AI computing needs and stringent data governance requirements faced by clients, including government agencies and SMEs [1][2] Group 2: Market Opportunities - The global AI infrastructure and cloud computing ecosystem is experiencing explosive growth, particularly in the Asia-Pacific region, driven by surging demand for GPU-intensive workloads [2] - Malaysia is emerging as a strategic AI hub, with projected investments in AI and data centers exceeding $40 billion by 2030, contributing over $115 billion to the country's GDP [2] Group 3: Revenue Streams - V Gallant is activating multiple revenue streams, including enterprise-level LLM training and fine-tuning, high-performance AI workload execution, and a hybrid AI infrastructure + SaaS monetization model [3][4][5] - The company offers GPU computing power leasing, a subscription-based enterprise-level LLM platform (Intelli-X), and localized AI contracts for regulated industries and public sector clients [5][6] Group 4: Strategic Partnerships - V Gallant has signed memorandums of understanding with key partners to enhance market access, talent development, and deployment channels, reinforcing its ecosystem-driven model [6][10] - These partnerships aim to accelerate commercialization for SMEs, public sectors, and regulated industries [6] Group 5: Product Offerings - Intelli-X, the enterprise-level LLM platform, provides a zero-access private LLM architecture, ensuring sensitive government and enterprise data remains fully controlled [6][8] - The company’s offerings include scalable analytics and insights platforms, pre-configured GPU servers, and collaborative workspaces for on-demand access to cutting-edge computing infrastructure [8]

Core Insights - IREN Limited plans to expand its AI cloud infrastructure by purchasing over 50,000 NVIDIA B300 GPUs, increasing its total GPU fleet to 150,000, which is expected to support over $3.7 billion in annualized run-rate revenues by the end of fiscal 2026 [1][10] Expansion Plans - The additional GPUs will be deployed in phases during the second half of fiscal 2026 and will be installed in existing air-cooled data centers in Mackenzie, British Columbia, and Childress, Texas, allowing for expansion without the need for new data centers [2] - IREN reported approximately $2.3 billion of ARR under contract in the second quarter of fiscal 2026, including a significant agreement with Microsoft and $0.4-$0.5 billion from the Prince George site, indicating a strong starting point for growth [2] Capital Raising Efforts - IREN has raised about $9.3 billion over the past eight months through various means, including customer prepayments and GPU financing, with $3.6 billion secured from Goldman Sachs and JPMorgan in the second quarter [3][10] - The company plans to raise an additional $3.5 billion for new GPU orders in the second half of 2026 to support its AI cloud infrastructure expansion [3] Revenue Projections - The Zacks Consensus Estimate for IREN's fiscal 2026 revenues is $984.53 billion, reflecting a year-over-year increase of 93.3%, while the estimate for fiscal 2027 revenues is $2.70 billion, indicating a year-over-year increase of 173.9% [4] Competitive Landscape - IREN faces significant competition from Applied Digital and TeraWulf in the AI infrastructure sector [5] - Applied Digital is constructing a large AI data center campus, Delta Forge 1, designed to support up to 430 megawatts of power [6] - TeraWulf is expanding its portfolio by acquiring two sites that add about 1.5 GW of power capacity, increasing its total platform size to approximately 2.8 GW [7] Stock Performance and Valuation - IREN's shares have increased by 64% over the past six months, outperforming the Zacks Financial Miscellaneous Services industry's decline of 26.8% [8] - The company is currently considered overvalued, with a Value Score of F, trading at a forward price/sales ratio of 6.80X compared to the industry's 2.70X [12] - The Zacks Consensus Estimate for IREN's fiscal 2026 earnings is 38 cents per share, reflecting a significant year-over-year increase despite a 43.3% decrease over the past 30 days [15]

Core Viewpoint - Panmnesia and SK Telecom have formed a strategic partnership to develop a next-generation AI data center architecture based on Compute Express Link (CXL) technology, aiming to enhance cost efficiency and performance in AI data centers [1][2]. Group 1: Partnership and Objectives - The partnership was announced at MWC26 in Barcelona and focuses on creating a CXL-based AI data center architecture [1]. - The collaboration aims to address the rising costs associated with GPU deployments in large-scale AI services by improving the utilization of existing computing resources [2]. Group 2: Challenges in Current AI Data Center Architectures - Current AI data centers face limitations due to fixed ratios of CPUs, GPUs, and memory, leading to inefficiencies and increased costs when resources are underutilized [4]. - The conventional architecture requires additional GPUs to be deployed when memory capacity is insufficient, which lowers GPU utilization rates and raises operational expenditures [4]. Group 3: Proposed Solutions - SKT and Panmnesia propose a disaggregated architecture that separates computing resources by type, allowing for flexible composition and minimizing resource waste [5]. - The new architecture will utilize a CXL Fabric Switch to interconnect resources at the rack level, enabling dynamic allocation based on workload requirements [5]. Group 4: Enhancements in Computational Efficiency - The collaboration aims to improve computational efficiency by replacing traditional network-based interconnects with CXL, eliminating the need for data copies and software intervention [7][8]. - The architecture will feature a Link Controller that facilitates direct communication over CXL, enhancing processing efficiency and allowing GPU-to-GPU and GPU-to-memory communication without software intervention [9]. Group 5: Implementation and Future Plans - SKT will lead the design of the architecture, leveraging its expertise in AI data center construction and operational management [11]. - Panmnesia will implement the CXL-based AI Rack, extending the link architecture beyond individual servers to the rack level [12]. - The companies plan to validate the architecture by running real AI models and evaluating performance metrics by the end of the year, followed by proof-of-concept deployments [13]. Group 6: Industry Impact - The collaboration is expected to enhance the competitiveness of AI data centers by addressing the "Memory Wall" bottleneck and optimizing system-level performance [14]. - Companies utilizing Panmnesia's link technology in their devices are anticipated to strengthen their market position in the AI data center sector [17].

SAN DIEGO, March 03, 2026 (GLOBE NEWSWIRE) -- Robbins Geller Rudman & Dowd LLP announces that purchasers or acquirers of Fermi Inc. (NASDAQ: FRMI): (i) common stock pursuant and/or traceable to the registration statement and prospectus issued in connection with Fermi’s October 2025 initial public offering (the “IPO”); and/or (ii) securities between October 1, 2025 and December 11, 2025, both dates inclusive (the “Class Period”), have until Friday, March 6, 2026 to seek appointment as lead plaintiff of the F ...

Management attributes the delay in signing new customer contracts to the temporary halt in hyperscaler engagement during previous merger discussions, which have since restarted with 500 megawatts currently under exclusivity. The company is transitioning its portfolio toward 100% colocation within three years, driven by the higher predictability and margins of AI infrastructure compared to the volatile Bitcoin mining sector. Operational performance is defined by the 'energization-to-billing' cycle, wit ...

Core Viewpoint - A class action lawsuit has been filed against Fermi Inc. and its executives due to significant stock drops attributed to potential violations of federal securities laws [1][3]. Group 1: Lawsuit Details - Investors have until March 6, 2026, to request to lead the case, which includes claims of securities fraud under Sections 10(b) and 20(a) of the Securities Exchange Act of 1934, as well as Sections 11 and 15 of the Securities Act of 1933 [3][10]. - The lawsuit is pending in the U.S. District Court for the Southern District of New York, titled Lupia v. Fermi Inc., et al., No. 1:26-cv-00050 [3]. Group 2: Company Background - Fermi Inc. is an energy and AI infrastructure company aiming to build large-scale nuclear reactors to support grid-independent data centers for AI companies, with its flagship project being Project Matador [4]. - The company completed its IPO in October 2025, claiming strong demand for Project Matador and securing a 20-year lease agreement with an investment-grade-rated tenant [5]. Group 3: Allegations and Stock Impact - Allegations state that Fermi overstated tenant demand for Project Matador and misrepresented the agreement with the First Tenant [6]. - Following the announcement of the First Tenant's termination of the funding agreement on December 11, 2025, Fermi's stock dropped by $5.16 per share, over 33%, from $15.25 to $10.09 [7][10].

Financial Data and Key Metrics Changes - The company reported a record full-year revenue of $101.4 million, up 35.7% year-on-year, marking the first time it has crossed $100 million in annualized revenue [3] - IFRS operating loss narrowed to approximately $13.7 million from $66.9 million last year, representing a 79.6% reduction [4] - IFRS net loss improved to about $11.3 million from $64.8 million last year, an 82.6% improvement [4] - Adjusted EBITDA was around $19.1 million, with adjusted net income at about $19.9 million, leading to an adjusted basic EPS of $0.89 and an adjusted diluted EPS of $0.88 [4] Business Line Data and Key Metrics Changes - The fourth quarter revenue was approximately $35.6 million, exceeding market consensus of $34.75 million, with an adjusted EPS of roughly $0.37, beating the consensus of $0.30 by about 22% [5][6] - The company maintained a strong underlying profitability while executing a turnaround, with adjusted EPS for the full year at $0.89, a 6% beat versus consensus [6] Market Data and Key Metrics Changes - The Asia-Pacific data center investment is projected to grow from roughly $30 billion in mid-2026 to about $90 billion by 2030-2031, with installed capacity expected to double from about 29,000 MW today to about 63,000 MW by the end of the decade [9] - Southeast Asia's data center investment is expected to grow from the low teens of billions to roughly $30 billion by 2030 [9] - India is rapidly scaling its data center capacity from just over 1 GW of installed IT load today to multiple GW by 2030 [9] Company Strategy and Development Direction - The company is advancing its AI infrastructure and data center build strategy across Malaysia, Thailand, Indonesia, Singapore, and other regions, including India and the Middle East [10] - Plans include exploring opportunities to buy or build data center assets to gain more control over delivery and create recurring revenue streams [11] - The company aims to be cash flow positive in 2026, focusing on disciplined delivery, overhead control, and cash collection [14] Management's Comments on Operating Environment and Future Outlook - Management emphasized that AI spending is becoming a structural demand rather than discretionary, with a shift from training-led to inference-led cycles, broadening demand for AI compute [6][7] - The company is witnessing a significant increase in demand for distributed compute across regional data centers and edge environments, which is expected to drive growth [7] - Management expressed confidence in the pipeline, with a $7 billion revenue opportunity and a strong balance sheet to fund growth initiatives [20] Other Important Information - The company finished 2025 with a cash balance of $104.8 million and reduced total debt to about $13.8 million, a 35.6% decrease from the previous year [13] - The company has spent over $11 million on share buybacks, indicating confidence in its valuation relative to performance [14] Q&A Session Summary Question: Update on the timing for the Freyr partnership phases - Management confirmed that they are on track with the Freyr contract, with slight delays due to market forces but no significant issues [27] Question: Will the execution on the Freyr contract hold back other agreements? - Management stated that the pipeline is strong and growing, with no delays in signing new contracts despite the Freyr project [34] Question: Recruitment needs and market conditions - Management highlighted rapid hiring across regions, with significant needs at the senior executive level to support growth [38] Question: Update on gross margin and revenue modeling - Management explained that the revenue mix shifted in 2025, affecting gross margins, but they expect improvements moving forward [48][52] Question: Update on operations in the Middle East amid geopolitical tensions - Management reported no material impact on operations, continuing to monitor the situation closely [93]

Crossroads Capital LLC, an investment management company, released its fourth-quarter 2025 investor letter. A copy of the letter can be downloaded here. The fund rose 2.7% net of fees and expenses in Q4 2025, bringing its annual return to 37.7% net. Since its inception, the Fund compounded at the rate of 21.2% gross and 17.1% net.  By the end of December 2025, the Fund's overall non-delta adjusted gross and net exposure were recorded at 89.7% and 72.1%, respectively. After a strong Q3, November revealed ho ...

Core Viewpoint - A class action lawsuit has been filed against Fermi Inc. and its executives due to significant stock drop attributed to potential violations of federal securities laws [1][4]. Group 1: Lawsuit Details - Investors have until March 6, 2026, to request to lead the case, which includes claims of securities fraud under Sections 10(b) and 20(a) of the Securities Exchange Act of 1934, and Sections 11 and 15 of the Securities Act of 1933 [4][10]. - The lawsuit is pending in the U.S. District Court for the Southern District of New York, titled Lupia v. Fermi Inc., et al., No. 1:26-cv-00050 [4]. Group 2: Company Background - Fermi Inc. is an energy and AI infrastructure company aiming to build large-scale nuclear reactors to support grid-independent data centers for AI companies, with its flagship project being Project Matador [5]. - The company completed its IPO in October 2025, claiming strong demand for Project Matador and securing an agreement with a tenant for funding construction costs [6]. Group 3: Allegations and Stock Impact - Allegations state that Fermi overstated tenant demand and misrepresented the agreement with the First Tenant [7]. - Following the announcement of the First Tenant's termination of the funding agreement on December 11, 2025, Fermi's stock dropped by $5.16 per share, over 33%, from $15.25 to $10.09 [8][10].