A：The company is building its AI native hire canler across four dimensions: first is quite in scale, second is product and functionality, third is customers and demand, and finally, capital and execution across all of these dimensions.

A：The company has built more than 2 GW of power capacity and is targeting more than 3 GW by the end of the year. They have already contracted more than 3.5 GW, and are now targeting at least 4 GW of contracted power for the year.

A：The significance of the company owning the entire stack is that it makes their platform most efficient. Currently, owned and contracted capacity accounts for more than 75% of total capacity, and they are building towards full-stack ownership, which means they offer cloud services spanning the AI lifecycle.

A：The company has made acquisitions of the C.M., Vile Agen, and Carfino. These acquisitions bring industry-leading engineers and research capabilities, with a focus on AI and inference optimization solutions. The acquisitions also extend their platform reach to agentic search and bring expertise in developing what this new class of developers need.

A：The company's cloud platform has seen strong demand with powerlink generation growing 3.35 times over the fourth quarter. The full-stack platform allows the company to capture and service a diverse range of hundreds of customers, and they are experiencing broadening demand across industries.

A：The updated 2026 CapEx guidance indicates that the company is investing in 2027 capacity that will come online during the next year. This investment is expected to contribute positively to revenue in the first half of 2027, where they already have customer commitments in place, including a contract with Meta.

A：The recent capital raised of more than $6 billion, with over $4 billion coming from convertible notes and $2 billion from Nvidia equity investments, is significant for the company's strong financial position. It provides a strong foundation and enables the company to tap the market at scale, invest in capacity, and support future growth.

A：The company's upcoming plans and priorities include continuing to execute on expanding capacity, building a cloud platform, expanding their customer reach, and financing growth intelligently.

A：The group revenue grew by 681% year on year to $399 million, with a particularly strong performance in Q4, selling out capacity due to demand exceeding available supply.

A：The AI business saw revenue growth of 141% to $390 million, with an 82% quarter over quarter increase. Group Adjusted EBITDA increased to $130 million from $30 million in the previous quarter and a loss of $54 million a year ago, indicating strong top-line growth and improved profitability.

A：The increase in Group Adjusted EBITDA and margin was driven by strong revenue growth. The AI business Adjusted EBITDA margin expanded to 45%, from 24% in the previous quarter, primarily due to strong revenue growth.

A：NaaS, alongside its consolidated investments in Simple Tens, is still in the air stage and requires substantial operating investments as it scales. The group's financial status is affected by these investments and their future consolidation, which is part of the strategy to find sustainable financial partners for these businesses.

A：The company has taken several steps to strengthen its financial position, including a private offering of convertible senior notes raising $4.3 billion at attractive premiums and coupons, an equity investment from Nvidia, and an increase in customer prepayments, resulting in a sharp increase in cash flow and overall capital.

A：The company is building for a future infrastructure expansion driven by strong market demand, with customer commitments already in place. It plans to invest in capacity and continue to leverage a diversified range of funding sources, including asset financing, corporate level debt, and its At the Market program, to ensure it can access a wide range of potential funding sources.

A：The company has reiterated its full year 2026 guidance with a revenue run rate of 7 to 9 billion, group revenue between 3 and 3.4 billion, and a group adjusted EBITDA margin of around 40%.

A：The company's growth profile and EBITDA margin progression will be influenced by utilization, pricing, and capacity. The timeline for employing new capacity and its impact on top and bottom-line results is particularly important.

A：The company expects a nonlinear quarterly adjusted EBITDA margin progression, with a back-end weighted nature to capacity being brought online. Adjusted EBITDA margins are expected to return to Q1 levels in Q3 before moving even higher in Q4.

A：The stronger GPU pricing is reflected in the core AI business with a sell-out capacity across all chip types at higher prices. The impact on older, shorter-term contracts that could benefit from this pricing dynamic is not quantified in the transcript, but the text indicates a very dynamic market with strong pricing across both old and new GPU generations, extended contract durations, increasing average contract values, and more significant prepayments.

A：The projected capacity growth is reflected in the CapEx increase as the company expects to have much more capacity in the first half of 2027 than in the current year. This expansion necessitates higher CapEx investment starting from now on.

A：The impact of component inflation on the 2026 program was quite material, amounting to low single-digit percentages of total spend. This was mainly because the company secured a lot of 2026 capacity back in 2025 at previous price levels before inflation increased.

A：The adjusted EBITDA margin is projected to progress as follows: it nearly doubled from Q4 to reach 45% in Q1. The company expects some quarter-to-quarter variability throughout the year, with margins in Q2 expected to be lower due to timing dynamics of investments. Margins are expected to return to Q1 levels in Q3 and increase further in Q4, culminating in a full-year group margin around 40% as guided.

A：Beginning in Q2, capacity additions are expected to ramp up, with significant improvements in Q3 and Q4. Key sites like Pennsylvania are projected to reach full capacity by around 2030 or possibly the beginning of 2033. Notably, the project schedule anticipates adding substantial capacity yearly, reaching 1.2 GW by 2033.

A：There are no reported delays at the Finland New Jersey site. Regarding the Microsoft contract, the delivery schedule is extended into the second half of the year, with the first tranches already delivered and more volumes expected in Q3 and Q4. The delays do not impact the previously delivered commitments.

A：The recently announced Meta deal is a five-year contract for a total of $27 billion, consisting of a $12 billion commitment to dedicated compute capacity starting in early 2027 and an additional $15 billion of capacity, at the company's discretion, for either Meta or AI cloud customers. Meta is committed to buying any capacity in these clusters during the five-year contract. This deal allows for asset-based back financing at attractive terms and the potential to sell to AI cloud customers at higher market prices, providing higher margins, less risk, and more revenue visibility.

A：The acquisition of Agan and Clarify is expected to significantly enhance the AI platform capabilities and improve customer stickiness. The move is based on the rationale of integrating talented teams from these organizations to bolster the company's AI efforts.

A：The company's strategy is to own the compute stack by leveraging vertical integration, a deep supply chain, and hardware engineering advantage. They aim to build the full cloud solution and use software as an enabler. The company partners when the right path is identified and uses M&A selectively to accelerate their roadmap, bring in proven developer adoption, or add complementary capabilities. They evaluate every potential deal based on criteria such as deepening customer engagement, increasing lifetime value, targeting new customer categories, and strengthening their full-stack AI cloud position.

A：Software plays a critical role in the company's strategy as an enabler that unlocks new capabilities and opportunities for the company and their customers. It is not built primarily for generating separate revenue streams but to support the company's growth and customer engagement. The company views M&A as a way to bring in rare talent or proven adoption that would take longer to build organically. They partner where it is the right path and use M&A selectively to strengthen their software plays and capability complementary to what they are building.

A：The company finances additional CapEx by relying on a strong balance sheet, supported by operating cash flow mainly from customer payments. Approximately 90% of the projected CapEx for 2020 is already secured by cash and contractual commitments. For incremental capacity, the company will use additional financing, which may include secured debt financing against contracts with Microsoft and Meta and corporate-level equity financing. They have also established an at-the-market program for issuing up to 25 million plus shares and have maintained disciplined capital structure guidelines to ensure consistent cost, capital, and shareholder dilution controls.

A：The AI cloud business pipeline has grown 3.5 times quarter-over-quarter and includes qualified opportunities across core AI cloud and token factory products, as well as across key customer segments like AI natives, software vendors, and enterprises. The pipeline does not include strategic hyperscaler deals like the Meta deal. Conversion rates have been maintained with solid win rates, accelerated sales cycles, and increased average selling prices. Some notable wins include contracts with Vistria in healthcare and life sciences, Rota in core automation, and new customer wins for token factory. The company is focusing on enabling consultative customer conversations and scaling their go-to-market and success teams to help customers realize their plans, which in turn creates durable revenue for the company.

A：The primary inference product for the company is their factor, which is positioned as the fastest-growing segment in their stack. The company has a winning combination of capacity and scale with a strong software stack that supports both popular open-source models and specialized models. This allows the company to provide the best total cost of ownership and optimize the stack for their customers. The company is confident in the product-market fit and expects that their factor will be foundational for building at scale, with plans to provide the necessary tools and platforms to optimize workloads in an evolving landscape.

A：The company assesses customer concentration and risk by considering the size and diversity of their revenue base beyond large contracts with Meta and Microsoft. While these contracts are significant, the company's focus is on customer diversification and the rest of their revenue base. They have a revenue model that spans across various customer segments, which helps to mitigate concentration risk. The company aims to maintain a strong, diversified customer base that contributes to their overall revenue and business resilience.

A：The diversified core of the AI customer base includes a range from developers to enterprises, focusing on the AI cloud business which is experiencing strong traction across various products and customer segments.

A：The concrete deliverables include a strategic investment of $2 billion from Nvidia, capacity commitments extending to the end of 2030, and support for future chips in databases and CPU platforms. This collaboration does not affect the timing or scale of the company's Rub employment in the second half of 2026.

A：The company is sold out for Q1 and most of the year, with the majority of capacity for AI cloud customers already under contract or earmarked. However, they retain some capacity for self-service and manage allocations as demand evolves.

A：The company's approach to data center construction involves efficiency, transparent communication, and long-term partnerships with communities. They focus on technological waste heat reuse, clear engagement in communities, and see themselves as contributing not just through building but also through education and research collaborations.