Core Insights - Nvidia is significantly underestimated despite its market capitalization exceeding $4 trillion, indicating a larger ambition beyond just AI chips [5] - The introduction of DGX Cloud in 2023 was a strategic move to rent high-end computing power, but by 2025, its role shifted to internal infrastructure rather than a primary product for enterprise markets [6][11] - The new platform, Lepton, launched in 2025, serves as a marketplace for GPU leasing and scheduling, redirecting focus from direct competition with major cloud providers to a more collaborative approach [11][23] Summary by Sections DGX Cloud and Market Dynamics - DGX Cloud initially gained traction, achieving an annualized revenue of $2 billion by the end of 2024, but faced challenges as GPU supply improved and competitors like AWS and Azure reduced prices by up to 45% [8][9] - The direct customer model of DGX Cloud created channel conflicts with major buyers like Amazon and Microsoft, leading to concerns about long-term partnerships and potential shifts towards self-developed chips by these companies [9][11] Transition to Lepton - Nvidia's strategy evolved to focus on Lepton, which connects AI developers with a network of GPU cloud providers without directly competing with them, thus mitigating risks associated with channel conflicts [11][23] - Lepton acts as a "marketplace" for computing power, allowing users to submit requests that are matched with available resources across various cloud providers, enhancing flexibility and efficiency [21][23] Nvidia's Ecosystem Strategy - Nvidia has been investing in cloud service providers like CoreWeave and Lambda, creating a cycle of selling chips, renting back computing power, and ensuring a stable cash flow for partners while maintaining control over critical resources [13][15][19] - The approach of leveraging partnerships while maintaining a stronghold on the GPU market positions Nvidia as a central player in the AI ecosystem, akin to how Apple controls the mobile internet through the App Store [27][29] Future Outlook - The shift from hardware to a focus on "computing power" and "platform" indicates Nvidia's ambition to dominate the AI landscape, ensuring that regardless of where the computing occurs, Nvidia's technology remains integral [30]

Core Viewpoint - Nvidia's ambition to dominate the cloud computing space remains strong despite the withdrawal of its DGX Cloud service, as it shifts focus to a new platform called Lepton, which aims to connect AI developers with GPU cloud service providers [5][10][12]. Summary by Sections DGX Cloud and Its Transition - Nvidia's DGX Cloud was launched in 2023, offering high-end GPU instances for a monthly fee of $36,999, initially gaining traction in the market [5][7]. - By the end of 2024, Nvidia reported $2 billion in annualized revenue from software and services, including DGX Cloud [7]. - However, by mid-2024, the competitive landscape changed as major cloud providers like Amazon and Microsoft reduced prices for their GPU offerings, diminishing DGX Cloud's competitive edge [7][9]. - Nvidia decided to repurpose DGX Cloud for internal use rather than as a primary product for enterprise markets, indicating a strategic retreat [9][10]. Introduction of Lepton - In May 2025, Nvidia introduced Lepton, a platform designed to manage and distribute GPU resources without directly renting out its own GPUs [10][12]. - Lepton acts as a marketplace for computational power, directing user demands to appropriate cloud service providers, including AWS and Azure [19][20]. - This shift allows Nvidia to avoid direct competition with its major customers while still maintaining control over the ecosystem [22][23]. Strategic Partnerships and Investments - Nvidia has been investing in cloud service providers like CoreWeave and Lambda, creating a symbiotic relationship where it sells GPUs and then rents back computational power [15][16]. - This strategy allows Nvidia to secure immediate revenue from chip sales while ensuring access to necessary computational resources for its own development needs [16][17]. - Nvidia's venture capital arm, Nventures, invests in various AI startups, further embedding itself in the AI ecosystem and ensuring future demand for its chips [17]. Future Outlook and Market Position - Lepton is positioned to become a central hub for AI computational needs, similar to how Apple's App Store operates in the mobile internet space [25][26]. - By not owning a cloud service but controlling the computational resource marketplace, Nvidia aims to maintain its relevance and profitability in the evolving AI landscape [26]. - The company's transition from hardware to a focus on computational power and platform services reflects its broader ambitions in the AI era [26].

Core Viewpoint - Nvidia's DGX Cloud has shifted from being a competitive AI cloud service for enterprises to primarily serving as internal infrastructure, with most of its computing power now dedicated to Nvidia's own research rather than customer-facing services [4][6]. Summary by Sections DGX Cloud's Rise and Fall - DGX Cloud was launched in 2023 with a pricing model of $36,999 per month for each H100 instance. Initially, this pricing was justified due to GPU shortages, but as supply improved, the value of this "scarcity alternative" has significantly decreased. AWS has reduced rental prices for H100 and A100 GPUs by up to 45%, making DGX Cloud less attractive to customers [6]. Strategic Shift to Lepton - Nvidia has redirected its focus to the Lepton GPU rental market, which operates differently from DGX Cloud. Lepton acts as a "traffic coordinator," routing computing demands to partners like AWS and Azure, rather than directly renting GPUs. This strategy positions Nvidia not as a direct competitor in cloud computing but as an aggregator within the AI cloud economy, allowing it to maintain influence over the GPU market without owning cloud infrastructure [8]. Impact on Developers and the Industry - For developers, transitioning from DGX Cloud to Lepton means accessing GPU computing power at more competitive prices through existing cloud services like AWS or Azure. For Nvidia, this shift reduces conflicts with channel partners and enhances its control over global GPU workload distribution [10].

Core Viewpoint - The importance of AI and NVIDIA's role as a foundational infrastructure provider for AI was emphasized by CEO Jensen Huang during his keynote at Computex 2025, highlighting the future necessity of ubiquitous AI similar to the internet and electricity [1]. Group 1: AI Development and Infrastructure - Huang discussed the evolution of AI, introducing concepts like Agentic AI, which possesses reasoning and perception capabilities, allowing it to understand, think, and act [5][6]. - The introduction of Physical AI, which understands the real world and its physical laws, is seen as crucial for the robotics revolution [8]. - NVIDIA's new Grace Blackwell system, which has entered full production, is designed to enhance AI capabilities, with the GB300 version offering 1.5 times the inference performance and doubled network connectivity compared to its predecessor [9][10]. Group 2: Performance and Technological Advancements - The Grace Blackwell GB300 system achieves 40 PFLOPS, equating to the performance of the 2018 Sierra supercomputer, showcasing a 4000-fold performance increase over six years [9]. - NVIDIA's AI computing power is projected to increase by approximately 1 million times every decade, supported by new manufacturing processes in collaboration with TSMC [9]. - The introduction of NVLink Fusion aims to build AI infrastructure that can scale to millions of GPUs, integrating with various cloud service providers [11][13]. Group 3: Robotics and AI Integration - Huang highlighted the need for robots to learn in virtual environments that adhere to physical laws, addressing the challenges of data strategy in robotics [24]. - The GR00T-Dreams system generates synthetic data to train AI models, enhancing the efficiency of robot training through simulated tasks [25]. - NVIDIA's humanoid robot foundational model, Isaac GR00T N1.5, has been updated to improve its adaptability in material handling and manufacturing tasks [28][29]. Group 4: Personal AI Computing - The DGX Spark personal AI computer is set to launch soon, allowing individuals to own a supercomputer, with pricing determined by companies [18]. - The DGX Station, capable of running large models with 1 trillion parameters, is also being introduced, showcasing NVIDIA's commitment to personal AI computing [18]. Group 5: Future Directions in Computing - NVIDIA is developing quantum-classical computing platforms, predicting that future supercomputers will integrate GPU, QPU, and CPU technologies [22]. - Huang emphasized the need for storage systems to evolve, integrating GPU computing nodes to handle unstructured data more effectively [22].