In this issue:
New Portfolio Company: Mythic
Market Perspective: MLOps
Solutions Update: Scality
Events: HPE Discover and Pathfinder With Antonio
Welcome to the July edition of Pathfinder Insights!
For those of us with school-age children, it is hard to believe that the summer is already in the back-half. Where did it go? It does feel like many people are taking a much needed vacation but we don’t see much of a break in venture market activity. We continue to see record fundraising in hot markets around data, AI/ML, and software. Congratulations to portfolio company DataRobot for its $300 million raise at a valuation of $6.3 billion!
In this issue, we officially welcome Mythic to the Pathfinder portfolio. Mythic produces a low power AI processor for the edge. Later in this newsletter we highlight several companies in this ecosystem in our recaps of Discover and Pathfinder with Antonio. Also, check out our abstract on the MLOps market, a capacity critical to helping customers realize value from their AI initiatives.
Don't hesitate to reach out with questions or comments. We look forward to your feedback. Enjoy the issue!
Portfolio Company Updates
New Portfolio Company: Mythic
In May, Pathfinder announced it had co-led Mythic's $70 Million Series C round of funding. Mythic is an analog silicon chip startup building efficient, low-power AI accelerators for edge-based inference workloads.
Executing useful AI workloads at the Edge requires low-power, high efficiency compute that analog silicon delivers. We expect Mythic to be a strong partner complementing HPE’s edge strategy. Read more
Market Perspective: MLOps
What is MLOps?
Machine Learning Operations – or MLOps for short – is an ecosystem of products and practices that enable machine learning models to be used in production reliably and at scale. These technologies and processes exist to help organizations use (rather than just build) Artificial Intelligence (AI) applications. MLOps aims to achieve the core principles of DevOps and apply those concepts to AI workloads: automated development, scalable deployment, continuous testing & continuous training, and infrastructure considerations.
The market for MLOps is expanding…fast; the global machine learning market is projected to grow from $7.3 billion to $30.6 billion in 2024 (a CAGR of 43%), and some more aggressive TAM estimates go as high as $126 billion by 2025¹ .
Why is MLOps necessary?
The path to enterprise ready, at-scale machine learning is riddled with complex and nuanced roadblocks that often prevent AI projects from being deployed and producing results. AI projects often get stuck without the right processes, data, or expertise to guide engineering teams through the lifecycle of training, deploying, and monitoring machine learning models. Moreover, once built and ready for production, models need to be scalable, sustainable, and trustworthy. Without these boxes checked, AI applications are not deployable.
The pandemic accelerated the business imperative for digital transformation, and AI applications based on ML models is a critical piece of that transformation. Successful implementation faces numerous obstacles and MLOps is ready to address these challenges.
The MLOps startup ecosystem has exploded with companies addressing customer roadblocks across the entire machine learning stack (see below). At Pathfinder, we are most excited about partnering with open core solutions and best-in-class tools that can provide significant value to multiple industry use cases. Pathfinder continues to explore investment opportunities across model serving, inference optimization, data prep, model monitoring/governance, etc.
¹Global Machine Learning Market Research Report" Market Research Future, February 2020
Solutions Update: Scality
New Scality ARTESCA launched!
Pathfinder portfolio company, Scality, has been a strong partner with HPE since 2014 with Scality RING. Now, HPE and Scality have announced a new addition to the joint portfolio, HPE Solutions for Scality ARTESCA, a lightweight, cloud-native object storage platform for Kubernetes environments.
Several trends have contributed to the new solutions developed by HPE and Scality:
(1) Application modernization using the cloud-native model to develop and deploy applications within a Kubernetes environment
(2) New application workloads, such as in-memory processing, analytics, and machine learning, are changing the requirements for storage performance and management to be able to search a massive amount of metadata
(3) Flash memory is becoming denser, and more cost effective, allowing storage of large data sets on flash to result in cost-effective high-performance storage with low latency and high throughput
These trends result in object storage increasingly becoming a primary storage option for many use cases. HPE supports Scality ARTESCA on a broad portfolio of HPE all-NVMe flash and hybrid flash data center and edge servers, as well as via HPE GreenLake with as-a-service cloud economics and a managed services experience on-premises.
Scality RING achieves SEC 17a-4(f) and FINRA Certification
The U.S Securities and Exchange Commission (SEC) stipulates recordkeeping requirements for the Securities Broker Dealer industry regarding books and records retained on electronic storage media. Scality engaged Cohasset Associates, a highly respected consulting firm with more than 40 years of experience with the records management practices of companies regulated by the SEC, to assess the capabilities of Scality RING for the recording and non-rewriteable, non-erasable storage of electronic records as set forth in SEC Rule 17a-4(f).
After a thorough study, Cohasset concluded that Scality RING, when properly configured and utilized to retain time-based records, meets the five requirements of SEC Rule 17a-4(f) and FINRA Rule 4511(c) that relate to the recording and non-rewriteable, non-erasable storage of electronic records.
For additional information about solutions from HPE and Scality, view the new web page!
During last month’s Discover conference, HPE President and CEO Antonio Neri shared that “it is no longer about simply capturing data. It is about how fast we can extract value from it.” As HPE’s customers increase their usage of data to make business decisions, several of Pathfinder’s latest investments have focused on AI technologies that help customers turn their data into insights:
• DataRobot offers an enterprise AI platform that automates the process of building, deploying, and maintaining AI at scale.
• Flywheel automates information processing and machine learning pipelines and provides for secure collaboration in life sciences, clinical and academic research.
• Lightmatter is building optical processors for accelerating AI inference workloads in the datacenter with increased performance and dramatically lower power consumption.
• Mythic has produced an analog AI processor for edge-based inference workloads that overcomes technical constraints that traditional types of compute are unlikely to achieve without sacrificing performance.
Three of Pathfinder's portfolio companies sponsored this year's event: Cohesity, Dragos, and Scality. Click on the links to visit their virtual booths and learn more about about their solutions and how they work with HPE.
Pathfinder with Antonio
In mid-July, the Pathfinder team hosted another exciting Pathfinder With Antonio featuring four start-up companies.
BigID offers a data intelligence platform that enables organizations to know their enterprise data and take action for privacy, protection, and perspective. Customers deploy BigID to proactively discover, manage, protect, and get more value from their regulated, sensitive, and personal data across their data landscape.
OctoML is a machine learning acceleration platform that aims to accelerate model performance while enabling seamless deployment of models across any hardware platform, cloud provider, or edge device.
Accurics provides a platform that self-heals cloud native infrastructure by codifying security throughout the development lifecycle. It programmatically detects and resolves risks across Infrastructure as Code before infrastructure is provisioned and maintains the secure posture in runtime by programmatically mitigating risks from changes.
Pachyderm is an enterprise-grade, open source data science platform that makes explainable, repeatable, and scalable machine learning and artificial intelligence possible. The platform combines version control for data with the tools to build scalable end-to-end machine learning and artificial intelligence pipelines.
Pathfinder has been busy in the virtual video studio as well. For a refresher on how Pathfinder aligns with HPE’s strategy and how the program engages with startups we have some exciting new videos.
• First, listen to Hang Tan, Chief Strategy Officer for HPE, explain the Pathfinder program’s relationship to HPE’s strategy and the virtuous cycle of Insights, Investments and Solutions: https://youtu.be/6gvMsV64zio
• Next, from the pre-show to Discover's day 3 keynote watch Paul Glaser, VP and Head of Pathfinder, provide insights on disruptive innovation.
• For an overview of Pathfinder and how the program engages works with startups, watch a replay of Eric Chin, Solutions Manager of Pathfinder, and Michele Fowler, Marketing Manager of Pathfinder, discuss how Pathfinder helps harness external innovation.