As KubeCon + CloudNativeCon 2024 approaches, VMblog sat down with Kevin Cochrane, CMO, Vultr, the largest privately-held cloud computing platform serving 1.5 million customers across 185 countries. With a growing focus on AI-native applications and enterprise-level solutions, Vultr is positioning itself as the leading alternative to traditional hyperscalers. The company's recent expansion of its Serverless Inference platform and Vultr Kubernetes Engine (VKE) demonstrates its commitment to helping organizations deploy and scale AI models globally while simplifying Kubernetes management.
In this exclusive Q&A, we explore how Vultr is addressing the challenges of AI model deployment, offering cost-effective alternatives to OpenAI, and leveraging its extensive network of 32 cloud data centers to deliver high-performance computing resources at the edge.
VMblog: If you were giving a KubeCon attendee a quick
overview of the company, what would you say?
How would you describe the company?
Kevin Cochrane: Vultr is the largest privately-held cloud computing
platform, offering unmatched usability, performance, pricing, and global reach. With
1.5 million customers in 185 countries, we stand out as the leading alternative
hyperscaler, catering to enterprise-level businesses in sectors including
healthcare, finance, telecom, retail, media, entertainment, and manufacturing.
Today, we provide a range of services, including Cloud
Compute, Cloud GPU, Bare Metal, Managed Kubernetes, Managed Databases, Cloud
Storage, and Networking solutions, enabling customers to achieve global reach
and high performance while simplifying deployment and scaling of cloud-native
and AI-native applications worldwide, all at a reduced cost.
VMblog: How can attendees of the event find you? What do you have planned at your booth this
year? What type of things will attendees
be able to do at your booth?
Cochrane: We welcome attendees to stop by
the Vultr booth (P26) to learn more about how to leverage our solutions to
deploy and scale AI models globally. The team will be hosting demos and can
answer any questions attendees may have about getting started with Vultr.
VMblog: Can you double click on your company's
technologies? And talk about the types
of problems you solve for a KubeCon + CloudNativeCon attendee.
Cochrane: As mentioned, Vultr has a range
of services ranging from cloud compute and cloud GPU to bare metal, managed
Kubernetes and more. Most recently, we announced a few updates to our
Serverless Inference platform, aimed at helping enterprises and digital startups
alike thrive in the age of agentic AI.
We expect agentic AI to be the
next big frontier in AI, as AI agents are poised to completely transform
business. But to unlock their full potential, organizations need flexible,
scalable, high-performance computing resources at the edge, closer to the end
user. Serverless Inference is the only alternative to hyperscalers, offering
the freedom to scale custom models with a user's data sources without lock-in
or compromising IP, security, privacy, or data sovereignty.
The expansion of our platform
introduces powerful new capabilities to empower businesses to autoscale models
and leverage Turnkey Retrieval-Augmented Generation (RAG) in real time, to
deliver performant model inference at the edge - using Meta Llama 3 or
proprietary models. Turnkey RAG also eliminates the need to send data to
publicly trained models, reducing the risk of data misuse while leveraging the
power of AI for custom, actionable insights. Meanwhile, with Vultr's
OpenAI-compatible API, businesses can integrate AI into their operations at a
significantly lower cost per token compared to OpenAI's offerings, making it an
attractive option for organizations looking to implement agentic AI.
VMblog: While thinking about your company's
solutions, can you give readers a few examples of how your offerings are
unique? What are your differentiators? What sets you apart from the competition?
Cochrane: There
are a few things that set us apart from the competition. The first is our
global reach. Vultr is the only independent cloud vendor that competes with the
hyperscalers, across six continents. In fact, we have over 32 cloud data center
locations worldwide, providing frictionless provisioning of public cloud,
storage, and single-tenant bare metal.
Secondly,
Vultr is the only composable/MACH Alliance-certified
global cloud vendor, enabling enterprise and innovator teams to scale their
digital AI infrastructure without traditional vendor lock-in. Last year, we
launched the Vultr Cloud Alliance, which includes a marketplace of plug-and-play
services from leading Infrastructure-as-a-Service (IaaS), Platform-as-a-Service
(PaaS), and Software-as-a-Service (SaaS) providers, to enable customers to
build agile cloud operations that can scale and evolve to meet their needs at
every stage. The Cloud Alliance gives customers a simple, intuitive control
panel that makes it easy to deploy infrastructure and add services from one
central portal. Meanwhile, composable enterprise-grade cloud infrastructure and
services, along with powerful API automation, allow developers to seamlessly
assemble and scale modern cloud operations on demand - regardless of location.
Lastly,
we are the only independent cloud vendor that enables teams to train their AI
models anywhere, but scale everywhere. As it becomes increasingly complex to
manage and deploy AI models, Vultr
Cloud Inference leverages our global
infrastructure network to accelerate the time-to-market of AI-driven features,
such as predictive and real-time decision-making while delivering a compelling
user experience across diverse regions. This in turn enables AI innovations to
have maximum impact by simplifying deployment and delivering low-latency
inference around the world through a platform designed for scalability,
efficiency, and global reach.
VMblog: Where does your company fit within the
container, cloud, Kubernetes ecosystem?
Cochrane: Vultr is the leading alternative
hyperscaler. As such, we are paving the way for AI-driven applications,
collaborating closely with our customers to address key challenges and
implement cutting-edge cloud infrastructure. Our solutions are designed to help
organizations efficiently scale their Kubernetes deployments, positioning them
for success in the ever-evolving AI landscape.
Kubernetes is complex, and we
believe that our customers should not have to spend their time managing
clusters. The Vultr Kubernetes Engine (VKE) is a fully-managed product offering that makes Kubernetes easy
to use. We manage the control plane, worker nodes and provide integrations with
other managed services such as Load Balancers, Block Storage, and DNS.
VMblog: With regard to containers and Kubernetes, is
there anything holding it back from a wider distribution? If so, what is it? And how do we overcome it?
Cochrane: Kubernetes
is becoming easier to use, thanks to cloud providers like Vultr, which simplify
the experience for developers by offering managed services that ensure 100%
uptime for Kubernetes clusters globally. However, a significant issue that
often goes unaddressed is the challenge of applying Kubernetes to new AI-native
applications and managing the scalability of AI inference models within these
clusters. There's a need to rethink the operational practices and guidelines
for hosting containerized applications on Kubernetes.
This
is where Vultr comes in, assisting customers in adapting to a new framework for
managing containerized inference models on Kubernetes. Until recently, there
has been limited progress in developing tools for an integrated pipeline of AI
models-covering training, tuning, inference, and global scalability within a
Kubernetes cluster. At Vultr, we are leading the way in this new era of
AI-native applications, collaborating with our customers to tackle these
challenges and establish a cloud infrastructure that enables organizations to
scale their Kubernetes deployments for AI advancements.
VMblog: Are
companies going all in for the cloud? Or
do you see a return back to on-premises?
Are there roadblocks in place keeping companies from going all
cloud? And if so, what are they, and how do they address that challenge?
Cochrane: In a new industry report commissioned by Vultr and conducted by S&P
Global Market Intelligence, The New Battleground: Unlocking the Power of AI
Maturity with Multi-Model AI, research found that in 2025, the AI
infrastructure stack will be hybrid cloud with 35% of inference taking place
on-prem and 38% in the cloud/multi-cloud. I think we can contribute this to
companies embracing cloud solutions in recent years, recognizing the
flexibility and scalability they offer. Rather than companies returning to
one-premises solutions, I foresee us entering an era of composable cloud
architectures, which will allow for organizations to mix and match various
cloud services to make their perfect configuration, while maintaining critical
on-premises components as needed.
Data security and compliance are top concerns that have hindered a
complete shift to the cloud, especially for industries handling sensitive
information. Additionally, legacy systems and integration complexities create
challenges for organizations, as many companies that already have substantial
investments in on-premises infrastructure may find the shift to cloud to be
daunting. To address these challenges, companies can adopt a phased approach to
cloud migration by assessing their existing workloads, prioritizing
applications and leveraging compatible cloud strategies to create an
environment that supports innovation while addressing security and compliance
hurdles.
##