A：Nevis Group highlighted that the company had a very strong year in 2025, successfully scaling capacity and delivering it to customers with speed and reliability. They also made great progress in developing their hyperscale AI cloud. The company is now one of the world's leading and most reliable AI cloud compute providers, having attracted a diverse client base and an amazing team. Notable achievements include exceeding financial targets, significantly outperforming capacity plans, robust demand, sold out capacity in Q3 and Q4, and an increase in the average contract duration of new cloud customers by 50%.

A：The company is accelerating its capacity plans to meet the robust demand environment and is on track to deliver more than 3 GW of capacity in 2026, up from 2 GW previously. They announced nine new data centers globally to support this growth. The company remains confident in its plans for the year ahead, reiterating its annualized run rate revenue target of $7 to $9 billion by the end of 2026. They believe they are well-positioned in the early stages of a significant industrial and technological revolution, becoming the AI cloud provider of choice.

A：The core AI cloud business saw impressive results in Q4 with revenue growth of 830% year over year and 63% quarter over quarter, driven by high utilization, strong pricing, and execution. The company sold out capacity once again in Q4, demonstrating that demand significantly exceeded available capacity. Despite strong growth, operating leverage and spending discipline enabled the company to achieve progress on the bottom line. Specifically, group adjusted EBITDA expanded from 19% in Q3 to 24%. The company ended the year with $3.7 billion in cash and cash equivalents, having generated $834 million in operating cash flow in Q4, which was primarily from term payments from long-term agreements.

A：For the full year starting in Q4, the expectation is to achieve between 3 and $3.4 billion in revenue.

A：After the deployment of all branches, the anticipated adjusted EBITDA margin is approximately 40%.

A：The 2026 EBIT forecast is at a loss. The commitment is to maintain a medium-term EBIT target of 20% to 30%, with the potential to go higher starting in Q1 2026.

A：The two guardrails for capital alternatives are: 1) Focus on raising debt relative to business needs and be prudent with the cost of capital, and 2) Mindful approach to shareholder dilution if choosing to issue equity.

A：Neil Doshi, Vice President of Investor Relations, is moving into a new strategy function role. Gili Naoichi is the new Vice President of Investor Relations, coming from deep research and strategic finance experience at Goldman Sachs, UBS, and Mindera.

A：The evidence includes changes in daily habits across private and corporate lives, the integration of AI into various industry workloads and capabilities, and direct signals observed in the business with large clients discussing expanded capacity and AI business growth. Specific examples include AI startups and enterprises ordering significantly more GPUs, indicating increased demand and growth.

A：The two parts of the capital expenditure question are what the optimal capital expenditure for their own business in 2026 should be and how this capital expenditure will be financed.

A：The company will finance the capital expenditure in 2026 primarily from its cash flow, including cash on hand and cash generated from its co-business. A significant amount received from long-term contracts will finance around 60% or more of all CapEx needs in 2026. For the remaining amount, the company plans to use tools to move towards a more optimal capital structure, maintaining a healthy disciplined balance sheet.

A：Given the company's full vertical integration, its CapEx is mainly allocated to ensure secure power for its business, which is crucial for future hyper-growth. The company is focusing on securing power at a relatively low cost and has made significant strides in this area.

A：The company's data centers have been one of the best investments, enabling rapid deployment of GPUs in a short time frame when there is high demand. The company has secured demand from cloud clients and AI labs, and plans to continue capitalizing on these investments in the future.

A：The rationale behind the 16 to $20 billion CapEx forecast for 2026 is that it is based on the ability and demand from 2026 and 2027, as well as the interest received from various participants in their platform. It will support meeting their revenue targets while laying the foundation for hyper growth in 2027.

A：The company does not have concrete plans to use the take-or-pay program at the moment but considers it another valuable tool in their toolbox that could be useful for the business and shareholders in the near future.

A：The company expects to finance its growth and investments by deploying capital in a prudent, balanced, and disciplined manner. The potential for future capital injections is seen as a strength that contributes to their growth plans while maintaining a disciplined balance sheet.

A：The company is accelerating the build-out and deployment of capacity in 2026 and expects continued acceleration in 2027. They are well on track to achieve the 800 to 1 GW goal around year-end. The company's investments in various sites, both smaller and larger projects, are expected to contribute to reaching the 3 GW of contracted power by year-end. New projects are expected to ramp up in 2027, contributing to the growth in contracted power.

A：The New Jersey Data Center site is progressing well. The first phase was delivered to Microsoft on time, and remaining commitments are on track to be fulfilled. The company's partners have secured the necessary components and are working hard to get everything operational. Safety margins and buffer times have been built into the project to ensure execution, and there is a high confidence in the timeline.

A：The difference between ARR and revenue guidance reflects the deployment schedule of the company's capacity throughout the year, with the majority of this capacity being installed in the second half of the year. The revenue guide simply reflects the ramp up in capacity coming online.

A：The company's continued success with AI natives and early progress with ISVs and enterprises set the foundation for capturing the market. They remain opportunistic with large deals while balancing the opportunity with their long-term strategic positioning and the expansion of their AI cloud.

A：The first tranche to Microsoft was delivered according to the plan in November with ongoing scaling, and the remaining capacity will be delivered throughout 2026 and more, with more than half landing in the second half of the year. The company delivered both contracted tranches to Meta on time, and these are now fully in service. Financial recognition for the tranches delivered to Meta is expected to be 12 months and roughly 11 months for the second tranche. The Microsoft deal is expected to ramp revenue over the course of the year, with full year revenue starting in 2027.

A：The upside in December 2025 ARR was driven by solid execution, strong pricing, and utilization. The company continued to make progress in adding new logos and expanding with existing customers, experiencing strong pricing across all families of GPUs and full utilization with the sale of all available capacity. As for Q1, demand remains extremely robust with continued pipeline growth, positive deal trends, and progress with their vertical strategy.

A：The company is managing data center equipment shortages by having a portfolio of sites to avoid dependency on any specific single data center project. They are ahead of their contracted power and have developed 9 contracted sites. They have also secured the majority of their long-lead items around their own sites to ensure capacity deployment beyond 2026. For Microsoft and Meta deals, they were able to secure necessary components last year for the full scope of those contracts before any price increase.

A：The company evaluates ownership versus leasing based on the scale and specific needs of projects, with a focus on bringing most large projects in-house to maximize control and efficiency. In the long term, they expect this preference for ownership to continue, but still use leases and partnerships to support growth and fill gaps before fully developing their own infrastructure.

A：The main reasons for preferring ownership over leasing data centers include better overall control over projects due to expertise in building and operating one's own sites, the ability to achieve greater efficiency and scale by tailoring designs to specific technical requirements, and the time it takes to bring large projects to fruition.

A：The most common software tools used by customers include those from the AI cloud. 100% of AI cloud customers are utilizing the AI cloud software. New products such as Token Factory and Aether releases have been launched to expand into enterprise, and the acquisition of Tavai has extended the platform capabilities with agentic search for AI developers. There is a growing demand for embedded storage across various verticals, leading to the creation of vertical-specific solutions. The company is in the early stages of monetizing its software and services to make the platform sticky, which allows for higher pricing per GPU hour. They are exploring other monetization models including consumption-based ones like per token pricing with Token Factory.

A：Confidence in the 40% EBITDA margin target comes from the group's achievement of a 7% adjusted EBITDA margin in Q4 and the strong demand for their AI cloud business. The company is investing appropriately to serve this demand, and the core AI cloud business has a significantly higher margin. As the AI business scales, most revenue and margin will continue to be driven by the core AI business, which supports the guidance of a 40% adjusted EBITDA margin.

A：The strategic rationale behind the acquisition of Tavii is to fit into the company's strategy of becoming a platform for AI developers from startups and enterprises to build their AI applications and agents. Tavii has served many Fortune 500 customers and has a great adoption and developer community. The acquisition is expected to enhance customer engagement, increase lifetime value, and strengthen the company's position as a full-stack AI cloud provider.

A：The company is building one of the largest platforms for AI developers by allocating capital between scale and functionality product dimensions. To build scale, they are investing in data centers and GPUs. To enhance the product, they allocate capital to acquisitions for more talent and to develop faster. The company is also focusing on internal organic development by their own developers, but recognizes the need for strategic partnerships and acquisitions to complement their offerings and provide a great developer experience.