March 2023 Summaries

Cloud GPUs are pivotal in democratizing access to artificial intelligence (AI) and machine learning (ML), enabling companies worldwide to harness the necessary computing power to innovate and compete. Historically, access to such resources was limited to well-funded corporations, creating a significant digital divide. The emergence of AI models like ChatGPT, which require substantial GPU power to operate, highlights this divide and the necessity for scalable cloud solutions. As cloud computing has rapidly advanced, it has allowed businesses to accelerate their digital transformation, with cloud GPUs offering a scalable, cost-effective means to access the computational resources needed for AI technologies. This shift towards cloud-based AI solutions aims to level the playing field, making advanced AI capabilities accessible to businesses of all sizes and fostering a new wave of digital disruption and innovation across industries.

Vultr, the world's largest privately-held cloud computing company, has announced the availability of the NVIDIA HGX H100 Tensor Core GPU, enhancing its cloud infrastructure with cutting-edge accelerated computing capabilities to support AI and machine learning initiatives. Partnering with Domino Data Lab and Anaconda, Vultr aims to streamline and expedite the work of data scientists and MLOps engineers by providing a comprehensive and integrated platform that simplifies infrastructure configuration and enhances collaboration. This strategic alliance allows users to quickly access high-performance computing resources, enabling faster model iterations and deployment while maintaining affordability and accessibility. Vultr's expanded GPU offerings, alongside its flexible cloud services, position it as a leading provider for businesses focused on AI-driven innovation and machine learning projects. With this partnership, Vultr seeks to empower data science teams by offering near-instantaneous access to necessary tools and environments, thereby facilitating breakthrough innovations and driving the AI/ML lifecycle forward.

An incident of unauthorized access to Vultr's marketing data occurred at a third-party vendor, ActiveCampaign, which Vultr used for email marketing until September 2022. This breach exposed 188,228 marketing records, with most containing only email addresses, while others included first name, last name, and country. Importantly, Vultr's core systems, customer subscriptions, and assets were never at risk, and all sensitive account details remained uncompromised. Vultr transitioned to a new service provider in October 2022, ensuring current marketing communications are unaffected. Customers are advised to stay vigilant against potential phishing attempts and verify any suspicious communications. Vultr has taken steps to purge remaining customer data from ActiveCampaign and is conducting a thorough review of security policies to safeguard customer data.

Vultr is introducing the NVIDIA HGX H100 platform to its cloud services, aiming to enhance the capabilities of virtual machines and bare metal servers for artificial intelligence and high-performance computing workloads. This platform, featuring NVIDIA H100 Tensor Core GPUs built on the Hopper architecture, offers significant performance improvements with fourth-generation Tensor Cores and NVLink interconnect technology, making it capable of delivering up to 1.5 exaflops of AI performance. This advancement is particularly beneficial for handling large and complex models in industries like healthcare, finance, and manufacturing. Vultr's offering allows customers to access this advanced technology on a pay-as-you-go basis, providing a flexible and cost-effective solution without the need for significant upfront hardware investments. Additionally, the platform supports simultaneous workloads, making it suitable for machine learning, deep learning, and data analytics applications, and can be further optimized with NVIDIA AI Enterprise software.

Vultr is excited to participate as a Platinum sponsor at the upcoming NVIDIA GTC, a global conference dedicated to AI and the metaverse, taking place online from March 20-23. The company aims to highlight its high-performance cloud computing solutions that are scalable and cost-effective, particularly emphasizing the importance of GPUs for complex workloads in AI, machine learning, data analytics, and scientific simulations. Vultr's CEO, J.J. Kardwell, will deliver a session on "Cloud Economics: Why Your GPU Strategy Needs To Embrace the Cloud," discussing the advantages of adopting cloud-based GPU strategies over traditional data center models. The conference offers a platform to learn from industry leaders and explore advancements in AI and accelerated computing, with the added benefit of being free to attend online.

In the digital age, businesses are increasingly focusing on reducing operational spending, particularly in cloud computing, which has traditionally been a significant cost center. The trend of moving workloads from major hyperscalers like AWS and Google Cloud to independent clouds is gaining traction as companies seek to optimize costs without sacrificing performance or capabilities. Independent clouds offer a versatile and cost-effective alternative, coming in various forms, such as niche, region-specific, industry-specific, or broad-based global solutions like Vultr. This shift is driven by the rise of co-location providers, infrastructure-as-code, and new cloud orchestration platforms that allow businesses to tailor their cloud infrastructure to meet specific compliance and ROI requirements. The movement reflects a broader change in mindset towards optimizing cloud spend by workload and region, ultimately providing businesses with greater freedom, choice, and flexibility in their infrastructure decisions.