What is H2O.ai?

Can the blockchain community teach AI developers something? Better yet, is it possible for the AI industry to shift towards an open-source culture?

Judging by the mission of H2O.ai, the answer to both of these questions is a resounding yes.

Upon visiting their landing page, the easiest thing to notice is the company’s vision, which states that they aim to “democratize intelligence for everyone with our award winning “AI to do AI” data science platforms.” After reading this however, it’s logical to wonder what it means.

On a surface level, their key products are clearly data science platforms, but beyond that, things get a bit more complicated. For starters, their product line includes three open-source platforms and two enterprise platforms, despite the fact that their overall vision seems to be to make AI easily accessible for everyone.

Despite this, it is easier to understand what might make H2O.ai unique if everything is looked at through the lens of only two of these products, which judging by their vision, appear to be the most important. Their platform called H2O seems to lead their open source efforts while their platform called Driverless AI seems to spearhead their larger-scale, institutional efforts.

To understand the company’s strategy and differentiation, it can be helpful to begin with its’ partnerships. Since H2O’s partners include heavy hitters like IBM, Nvidia, Microsoft, Amazon Web Services, and Google Cloud, it’s not a stretch to say that their platforms may perform better than those which are offered by their competitors. Still, none of this includes hard data on the actual performance of the H2O suite of products.

Perhaps, as with many early stage projects in emergent industries, an even better way to understand where their differentiation lies is to look at where they started.

In other words, what’s their vision and where are they with it now?

Reportedly, H2O began as a company called 0xdata, which was running an early-stage of their H2O platform as its’ open-source wing. Apparently, they garnered so much interest in this particular part of their operations that in 2014, they re-branded themselves as H2O entirely.

At this point, they had already also gained companies like Cisco, eBay, Nielsen, and Paypal as clients.

If you’re wondering why, the answer is simple. H2O’s main platform was and is seen as extremely interoperable due to the fact that it was written in the Java, Python, and R programming languages all at once. With this, we can see at least one of its’ defining features.

The H2O platform and by extension, Driverless AI, its’ institutional-grade equivalent, continue to attract powerful investors and clients because they likely, easily integrate with just about any legacy framework. Lending credence to the idea that these platforms continue to drive significant growth for the company are the facts that in the last year, it has expanded is’ offices globally.

As of now, this means that you can find H2O.ai in Prague, London, Australia, Brazil, in China, in addition to California, which is where their efforts began. On top of this, just two months after this expansion started, a partnership came to light related to their Driverless AI platform, which can serve to illustrate its’ utility.

First, according to the original press release, Driverless AI is really a plug-and-play platform for all things Machine Learning that companies might need. In other words, with this product, H2O aims to make it simple for any company to integrate ML algorithms into their operations, with little to no specialized knowledge on the subject. Next, the partnership itself is important for one key reason.

Given that their chosen partner was Tech Mahindra, which is a multi-billion dollar company that recently introduced netOps.ai, H2O’s aiming to scale in a major way. The foundation for this assumption boils down to what netOps.ai is built to do.

Essentially, it is a framework of technologies that can reportedly speed up and streamline the roll-out of 5G-based networks. While it is not entirely clear whether this particular capability relates to H2O.ai’s partnership with Tech Mahindra, the idea remains that the latter company appears to be known for scaling technologies well. Backing this up is the fact that H2O representatives have stated the partnership’s key goal as improving the scalability of their platforms.

Before we end our discussion here, it’s important to circle back to the questions that I posed above. If anyone can lead the way toward the AI industry becoming more reliant on open-source projects, it’s H2O.ai, at least until a newcomer comes along who does the same thing in a more efficient fashion. They leveraged an open-source platform to aggressively grow into what appears to be an analytical empire of sorts. Given that their website lists their current number of institutional clients at 18,000, and the data science teams that trust their platforms as much higher, it’s easy to wonder what height they will reach next.