The 60th Edition of the IT Press Tour has just concluded successfully in Silicon Valley. Led by Philippe Nicolas, the founder of Coldago Research and a veteran storage industry analyst with over 30 years of experience, the event brought together a distinguished media team from the US and Europe. Volumez was one of 9 IT companies that briefed the team the last week of January 2025.
Key Details:
- Topic: Maximize GPU Yield and Automate AI/ML Data Pipelines with Volumez
- Location: Samsung Research America building in Mountain View
- Date: January 29, 2025
IT Press Tour 2025 Silicon Valley: Dedicated articles
The IT Press Tour continues to be a crucial event for technology companies to showcase their latest innovations in IT infrastructure, cloud, networking, data management, storage, big data, analytics, and AI/ML. This edition follows the tour’s tradition of bringing together industry experts and media professionals to explore cutting-edge technological developments.
Let’s hear from some of the media professionals and what they said about the GenAI landscape and how Volumez is tackling some serious tech challenges.
“Despite massive cloud investments, underutilized GPUs, inefficient resource allocation, and unpredictable cloud costs are draining budgets. Infrastructure teams scramble to patch inefficiencies while executives question AI’s actual return on investment (ROI).”
“AI’s problem isn’t a lack of computing power – it’s how poorly that power is managed. Most artificial intelligence (AI)/machine learning (ML) infrastructure is built on guesswork and overprovisioning, leading to wasted resources, ballooning costs, and underutilized GPUs.”
“Half-utilized GPUs mean companies are effectively paying double for AI processing. This inefficiency is mainly due to slow data pipelines that don’t feed GPUs fast enough – not a lack of GPUs.”
“Instead of spending time fine-tuning AI models, data scientists are often stuck troubleshooting data infrastructure bottlenecks. DIaaS removes infrastructure management from the hands of AI teams, allowing them to focus on what matters: delivering results.”
“DIaaS isn’t just about making AI faster – it’s about making AI viable. For companies frustrated by wasted cloud spend, underperforming AI workloads, and overburdened data science teams, rethinking infrastructure efficiency is no longer optional – it’s essential.”
“The future of AI won’t be built on more GPUs and storage. It will be built on more innovative, dynamically optimized infrastructure, and DIaaS is leading the way.”
- Podcast 3178: Is AI Infrastructure Broken? How DIaaS is Changing the Game, hosted by Neil Hughes, The Tech Blog Writer and featuring John Blumenthal and Dianne Gonzalez, February 12, 2025.
“AI and GenAI workloads require cloud-conscious, precisely coordinated systems that carefully consider IaaS restrictions and coordinate compute, storage and storage resources to meet AI/GenAI’s high-performance requirements.”
“AI/GenAI requires efficient infrastructure and efficient operation, where every GPU, every storage and every human resource is used optimally to deliver peak performance without waste. This results in shorter training times, lower costs and the scalability required to achieve the maximum yield from state-of-the-art models.”
“Data scientists need tools to meet the requirements of the ML pipeline without the complexity of the infrastructure. This allows scientists to focus on innovation and experiments while ensuring optimum infrastructure performance.”
“Let’s talk about something that’s been bugging data scientists and AI engineers for a while now – the headache of managing infrastructure for AI workloads. You know that feeling when your expensive GPUs are sitting idle while data shuffles around? Yeah, that’s exactly what Volumez is tackling, and they’re doing it in a pretty clever way.”
“Volumez shared an interesting perspective: what if we stopped thinking about AI infrastructure as separate pieces and started looking at it as one interconnected system that needs perfect balance? It’s like conducting an orchestra – if one section is off-tempo, the whole performance suffers.”
“Remember when getting 1TB/sec throughput seemed impossible in the cloud? Well, Volumez just shattered that ceiling in the MLPerf Storage 1.0 benchmark. But here’s the really cool part – they did it using standard Linux data paths. No proprietary controllers, no special sauce in the data path, just intelligent configuration of existing cloud resources.
The Numbers That Made Everyone’s Jaws Drop:
- 1.14 TB/sec throughput
- 9.9M IOPS
- 92% GPU utilization
- And get this – they did it at costs 27%-479% lower than competing solutions.”
“One of the most refreshing things about Volumez’s approach is how they’re thinking about the user experience. Dr. Eli David, their advisor and a veteran AI researcher, put it perfectly: “I don’t care about storage. I don’t care about this. I just care about GPU.”
“Another compelling aspect of Volumez’s platform is how it simplifies infrastructure management for data scientists. Through integration with tools like PyTorch, data scientists can specify their infrastructure requirements directly from their notebooks without having to coordinate with MLOps teams.”
“Again, Dr. Eli David highlighted the significance: “For many state-of-the-art models that I’m training, I’m not getting 100% GPU utilization… 50% utilization just means I’m paying double what I should for my GPUs.”
“Remember the recent DeepSeek announcement that got everyone excited about more efficient models? Dr. David made an interesting point – as models become more compute-efficient, the bottleneck will shift even more toward data infrastructure. It’s like widening the highway only to find out your on-ramps can’t handle the traffic.”
“Volumez, the startup that makes storage access for cloud-based virtual machines faster and cheaper, adapts its solution to online AI processing.”
“Let’s specify that Volume arrives these days in a specially revamped version 2.0 to take account of data loading as part of an AI model drive. Compared to the acceleration of data loading for VMs or application containers that it originally proposed, the parallelization of accesses to power GPUs would differ slightly, says the startup.”
“In practice, data scientists write a simple Python code that tells the credits to connect to Volumez, the location where the source data is located, and the work to be done on which cloud in PyTorch. It is also possible to use an infrastructure-as-code console such as Terraform or directly through the Volumez API. Volumez then calculates the necessary resources on the indicated cloud and launch the deployment order.”
“The rise of artificial intelligence and machine learning applications is putting increasing pressure on cloud infrastructures. Faced with this situation, the American-Israeli start-up Volumez, founded in 2020 by Jonathan Amit, proposes an alternative solution to traditional storage architectures. With the launch of its Data Infrastructure as a Service (DIaaS) platform, the company aims to address resource balancing challenges by maximizing the use of GPUs and automating data flow management.”
“80% of artificial intelligence workflows today devoted to the diagnosis and troubleshooting of entry and exit operations, to the detriment of calculations themselves. This highlights a major structural problem, the imbalance between storage, network, and computing capacity, which leads to under-use of resources and an explosion of costs.”
“One of the main advantages of the DIaaS platform is its ability to dynamically adapt infrastructure to the real needs of businesses, thus avoiding the costs of excessive allocation of cloud resources. A comparison by Volumez between its solution and AWS io2 Block Express shows substantial savings. In particular, optimising resources reduces overall expenditure by 75%, with a 92% reduction in computing power costs and 70% of storage costs.”
“Volumez has updated its cloud-delivered block storage provisioning service for containerized applications to GenAI with its Data Infrastructure-as-a-Service (DIaaS) product.”
“The company says its technology can maximize GPU utilization and automate AI and machine learning (ML) pipelines. Existing pipelines to deliver data and set up AI/ML infrastructure are sabotaged by storage inefficiencies, underutilized GPUs, over-provisioned resources, unbalanced system performance, increased cost, complex management, and poorly integrated tooling, which drains the AI pipeline team’s bandwidth and delays projects.”
“The context here is that public clouds such as AWS and Azure rent out compute and storage resources and are not incentivized to optimize how customers efficiently and evenly consume these resources. Volumez makes its money by operating in the zone between the public clouds’ optimization of their own resources for their own benefit, and customers’ need to optimize the efficient use of these resources in terms of performance, cost, and simple operations for their own benefit.”
Volumez named to the 2025 Gems list as a result of their vision, product development, recent announcements, and potential market disruption.
“As enterprises rapidly increase their AI investments, with global spending expected to reach $500B by 2027, according to IDC, many organizations are discovering that traditional infrastructure approaches create significant bottlenecks in their AI/ML pipelines. Volumez, a Silicon Valley startup, is tackling this challenge with a novel “Data Infrastructure as a Service” (DIaaS) platform that promises to maximize GPU utilization while dramatically reducing costs.”
“Volumez executives explained that the fundamental problem lies in the inherent imbalance of current cloud infrastructure configurations. This imbalance manifests in several ways: I/O bottlenecks, storage inefficiencies, underutilized GPUs, complex management requirements, and data scientists spending excessive time on infrastructure instead of model development.”
“Rather than acting as a traditional storage company, Volumez provides a SaaS platform that configures and optimizes cloud infrastructure components. The platform’s key innovation is its “cloud awareness” – a deep understanding of cloud provider capabilities and constraints that enable it to create perfectly balanced systems for specific workloads.”
“A key benefit of the Volumez approach is simplifying infrastructure management for data scientists. Integrating with tools like PyTorch allows data scientists to specify their infrastructure requirements directly from their notebooks without having to coordinate with ML ops teams.”
“The company’s focus on automation and simplification, combined with its impressive performance metrics, suggests it could play an essential role in helping organizations overcome the infrastructure challenges that currently limit AI adoption and effectiveness.”
“Volumez, a young storage software company, has made clear progress and illustrated earlier today during the 60th IT Press Tour its Data Infrastructure as a Service with some real applications to AI oriented services but before we jumped into this in a future post, I wish to spend time and (re)introduce their storage engine.”
Conclusion
As you can see, Volumez has come a long way from its single use case of cloud-native block storage for data-intensive workloads like databases. Volumez’ Data Infrastructure as a Service (DIaaS) for AI/ML is a platform that composes cloud resources and intelligently orchestrates them to build a balanced data infrastructure tailored to the AI/ML data pipeline. Volumez makes AI/ML data services highly performant, scalable, and efficient across any cloud, keeping GPU utilization near 100%, resulting in faster time to model deployment with new economics.
Volumez is available on the AWS Marketplace, Microsoft Azure Marketplace, and Oracle Cloud Marketplace. Please contact us for a demo.