The call, moderated by the VP of Investor Relations, includes forward-looking statements about Mars Inc.'s growth plans, liquidity, and financial performance. It highlights non-GAAP financial measures and references the shareholder letter. An analyst interview with management is conducted following prepared remarks, focusing on Q3 2025 earnings.

Mara discusses transitioning from Bitcoin mining to a vertically integrated digital infrastructure, leveraging abundant low-cost energy to create value in finance and intelligence. By focusing on AI inference and private cloud solutions, Mara aims to maximize the utility of every megawatt, integrating Bitcoin and AI on a single platform. Initiatives with mplx and acquisition of xion underscore Mara's commitment to expanding its AI and HPC infrastructure capabilities, positioning itself as a leader in secure, localized inference capacity for enterprises.

Mara emphasizes transitioning from Bitcoin mining to a comprehensive digital infrastructure model, leveraging energy generation and compute capabilities. The company highlights operational excellence, strategic execution, and energy ownership as key pillars. Mara has expanded its Bitcoin holdings and compute capacity significantly, while also exploring international opportunities and optimizing capital allocation. The dialogue underscores Mara's vision to capitalize on power as a critical constraint in AI growth, positioning itself as a leader in the digital energy and infrastructure sector.

Discusses the strategic benefits of vertical integration, emphasizing access to low-cost, reliable energy and the flexibility to expand data center capacity. Highlights the importance of controlling power in digital infrastructure, with a focus on AI and Bitcoin mining. Addresses the cost efficiency and technological advancements in ASICs, contrasting with GPU depreciation, and predicts the market advantage of operating at the lowest cost per token.

Discusses the evolution from CPUs to ASICs in computing, highlighting the dramatic increase in compute power per joule of energy. The conversation explores the implications of this trend on capital expenditure risks, particularly for hyperscalers investing in long-term data center operations. It suggests that while hardware procurement carries significant risk due to rapid technological obsolescence, building large data centers also presents substantial challenges. The dialogue emphasizes the growing importance of energy efficiency and cost-effective hardware in mitigating these risks.

Discusses various business models in GPU hosting, emphasizing inference AI powered by ASICs over GPUs, highlighting cost efficiency and future market potential in AI inference.

Xion, an EDF subsidiary, operates four data centers, three in France and one in Canada, supporting nuclear fleet data processing. It offers AI and traditional data center services, houses quantum technology for research, and provides software solutions for fully encrypted, private data storage.

Discusses the acquisition of Exxon's data center operations, highlighting their tier 3 and 4 data centers, encryption security, and customer key control. Emphasizes leveraging existing expertise for global expansion. Addresses regulatory hurdles for gas-fired power plant construction and grid attachment, noting confidence in securing necessary permits. Highlights a staged payment structure reflecting current business value and future growth potential.

The dialogue emphasizes the strategic advantage of vertically integrating gas-fired power plants with large-scale data centers, offering reliable, cost-effective power solutions. It discusses modular, containerized data center designs that can adapt to varying needs, from traditional Bitcoin mining to advanced AI applications. The conversation also highlights the growing preference for on-premises or near-premises AI solutions among corporations due to concerns over data security and operational risks, underscoring the importance of private cloud environments for mission-critical operations.

Discusses the necessity of engaging with customers who will fund the project, emphasizing understanding their needs and aligning project scope, possibly referring to a small-scale sample or pilot for validation.

Discusses the potential of small micro data centers as prototypes for enterprise AI, highlighting examples from oil drilling, financial trading, defense, healthcare, and media industries. Emphasizes the importance of considering future market trends and cost-effectiveness in AI adoption.

The dialogue focuses on identifying key milestones for economic growth, emphasizing the importance of large energy companies partnering with the company to monetize energy assets, the increasing use of inference AI in data centers, and the profit per megawatt hour as a critical operational KPI. It also touches on strategic flexibility in business decisions.

The dialogue emphasizes the strategic decision to integrate AI inference operations with energy asset control, aiming for a fully vertically integrated business model. The approach avoids public disclosure to prevent competitor emulation, focusing instead on gradual market awareness through consistent announcements, aligning with a long-term vision set a year ago.

Discussion shifts to sovereign and foreign government initiatives, focusing on competitive edges in the load management sector, questioning what grants a leading position against rivals.

By partnering directly with sovereign entities like ADP and IHC in UAE, the company establishes a reputation for grid balancing and efficiency, leading to welcomed opportunities in countries such as France, UK, Kenya, and Saudi Arabia, emphasizing a commitment to enhancing grid stability and integrating new loads like AI data centers.

The dialogue underscores the necessity of growing hash rate to maintain a significant presence in the Bitcoin network, emphasizing strategic energy sourcing and partnerships. It highlights plans to double flair gas capacity, leverage low-cost wind and solar energy, and allocate resources between Bitcoin mining and AI, aiming to support network security and diversity while fostering long-term success through governmental and industrial relationships.

A mix of advanced systems, particularly those with unique load balancing capabilities, are being integrated into fleets to meet growing utility demands. Additionally, investments in AI-related businesses, such as securing large language models and developing ultra-high-speed cluster interconnect switch technology, are highlighted as strategic acquisitions for future solutions.

The dialogue discusses how publicly traded Bitcoin mining companies have not fully realized their market value, emphasizing the importance of integrating energy generation and AI for future growth and investor recognition.

The call concludes by thanking participants, encouraging unresolved questions to be directed to the investor relations team, and wishing everyone a pleasant day.