Breaking down Insight Partners State of Enterprise Tech 2024 Report

The team at Insight Partners just released their State of Enterprise Tech report for 2024. There is a lot to consume in the 60+ slides, but we cherry picked the things that should be interesting to our audience - and frankly there is a lot of interesting stuff. 

I will leave the survey methodology stuff for you to consume, but suffice to say the sample size was large and the audience was very enterprise-y. They looked at four key areas:

We are going to focus disproportionately on the first two.  

Let’s start with Data & AI

Data & AI

Here is the top line takeaways from this section:

So what is interesting here? Well not the fact that nearly half the respondents were working on foundational models…if anything, the number we might have expected the number to be higher. What will be interesting will be to see how the % of open source increases in the coming quarters. Right now, it is effectively 80% proprietary and 20% open source. But the tack towards “open” AI models is fairly strong and there is no putting the genie back in the bottle - no matter what your philosophical position is. Open source offerings are on par with proprietary models according to objective sources. Open source works with other open source. The enterprise increasingly operates on open source. The 20% will only grow.The second point is more interesting from our perspective. Nearly half the respondents indicated they were modernizing their foundational data layer a top priority. There was some lumpiness in the relative positioning depending on size, but moving to a modern data stack was critical to their AI success. We have made this point on numerous occasions in recent quarters. If you want to succeed in AI, you better have good data and scalable data infrastructure. That means adopting the architecture of the modern data lake/lakehouse. This architecture uses parts of the data lake and data warehouse, uses object storage as the underlying storage substrate and leverages open table formats like Apache Iceberg (or Deltalake or Hudi). It is multi-engine and built for scale. If you pick the right object store (hint, hint) you can deliver incredible performance and do so at exascale. There is a real urgency in this replacement cycle. You simply can’t execute at scale with older technologies like Hadoop, appliances or POSIX. 

The other takeaway (although it is not particularly well backed up by the data) that Insight Partners makes is that “Cloud technology will be critical to successfully scaling GenAI.” The other way of saying this is “The cloud is an operating model and not a place.” We have hammered this point for years - the cloud operating model is software-defined, containerized, orchestrated, API driven (S3 as a case in point) and microservices oriented. Done correctly, that means the “cloud” runs anywhere, public, colo, private. This is noted by NYSE’s Chuck Atkins:

Part of the argument for the cloud as an operating model is that the private cloud offers superior economics. This is another point we have hit on time and time again, particularly as we enter the age of AI, where petascale becomes exascale pretty quickly. That demands foresight in the planning. Again we have that covered, in our groundbreaking work around ARM chips which are a fraction of the power footprint of traditional chips.

Infrastructure and Dev Ecosystem

We live in the devops world as much as we live in the core IT world. We have a cult-following with developers as evidenced by our 45K GitHub Stars, 31K Slack Channel members and countless endorsements on places like Gartner Peer Insights. 

AI doesn’t happen without this key constituency and Insight Partners rightly surveys the landscape. Here is the top line:

Once again, there are a few things that jump out at us. One is the use of the term “cloud technologies” again. It is clear that Insight and the survey respondents get the cloud in an operating model. The evidence here is that the concept of the hybrid-cloud or multi-cloud is on the minds of enterprise technology leaders. Almost every enterprise has more than one cloud - often 3,4 or 5. They need technologies that can span the public, private, edge and colo worlds, providing developers with industry standard approaches which in turn delivers consistency. Consistency is a powerful lever in the game of scale and developers love it, which is why we focus on creating “greatest common denominator” experiences. 

There is also an emphasis on container technologies. Those technologies demand software defined solutions. This also speaks to the modernization effort underway - away from things like storage appliances or non-cloud native technologies like SAN/NAS. 

There are additional findings around the implementation of CI/CD, infrastructure as code and serverless. We have written extensively about the CI/CD world and how it interacts with object storage and see it as an important change in thinking for organizations when you think about the concept of  hybrid-cloud or multi-cloud. With having multiple cloud providers and multiple environments, CI/CD is what brings harmony when deploying and managing your infrastructure. By laying a solid foundation to your infrastructure you can automate the most mundane tasks and focus on scaling your infrastructure to the next level.

The last area we will comment on is the rise of the edge. Edge compute doesn’t get enough love in our opinion and that might be because no one seems to be making a lot of dollars in the space. According to Insight Partners, Microsoft Azure seems to be keeping their success a bit of a secret.

We can’t really validate this one way or the other. What we can say is that, with a binary of less than 100 MB, MinIO can run anywhere, from rockets to cars to, obviously, the datacenter. The ability to have a consistent data platform from edge to core is what draws many enterprises to MinIO and why every defense/intelligence organization seems to be running MinIO at scale. 

It is an excellent report and echos much of what MinIO has talked about in the previous year or so. We encourage you to take it in and think about how it relates to your organization. We are always available to discuss how MinIO would fit into your architecture - from AI to archival. Drop us a note at hello@min.io and we can get started.