The Rise of Machine Learning Operations Startups

In the United States, it’s getting tough for me to pay attention. It’s difficult to express if you aren’t from the US. In a nutshell, several flaws in police and judicial systems were exposed as the week drew to a close.

As a consequence, today’s Exchange newsletter will be less comprehensive than intended. Goodbye, dear ones and everyone else.

The DevOps market is thriving and well-funded. I had a conversation with Opslyft the other day, for example. The firm is straddling India and the United States, building a unified DevOps service that combines tools for post-deployment development of software.

It’s a great company, and I’ll probably spend more time covering it when it raises money. Recently, GitLab debuted as a pre-deployment DevOps solution to pick another example from memory.

There are a number of tools that DevOps organizations may use. To begin with, both large and small technology companies are creating DevOps solutions.

We’re seeing a surge in the MLOps (machine learning operations) market, which is similar to the more well-known MLops industry.

This week’s announcement by Comet that it has raised 3 million brought back memories of our previous coverage of a MLOps firm’s most recent Weights & Biases round.

We caught up with Sapphire Ventures’ Jai Das the other day to get more information for our AI fundraising story, and I brought up the prospect of AIOps becoming a third “Ops” category for us to keep an eye out for.

However, according to Das, “MLOps is essentially AIOps,” so we may mostly confine our thinking to the two primary categories.

However, because the lines between AI and machine learning are not always distinct—let’s not dig ourselves into a quarrel—it will be worth examining whether AI and ML may coexist in the same framework.

More on AI

We’ll continue the AI discussion today, with a touch more on the AI market. Anna has notes to get you started, expanding on our prior entry about worldwide trends in artificial intelligence investment.

She has ideas about where AI money is going right now and how changing definitions of what constitutes “AI” might expand the dollar reach for startup activity:

While the geographic disparity caught our attention, we expect dollars to be more evenly distributed as the definition and applications of AI broaden. For instance, the two newly minted Latin American AI unicorns in Q3 were NotCo, a food tech company, and Unico, a digital ID provider, while a major round also went to Mexican lending company Kueski, which we’d have called a fintech but is also AI-enabled. If that’s the new reality of AI, we wouldn’t be surprised to see more money flowing into startups leveraging it to tackle real-world problems anywhere in the world, including in Latin America, but also in Africa.

If you live in Canada, we have something coming that you’ll want to read next week; let’s finish up our AI work for the day with an answer from Point72 Ventures’ Sri Chandrasekar, which came in a little late for our previous AI article but is nevertheless worth reading.

When it comes to AI-focused startup economics, the investor had a few things to say:

In my view, most of the recent interest in AI has been driven by revenue growth of companies that are raising large rounds. But the reason behind that revenue growth is pretty simple: high demand for products and low labor participation. We’re seeing this across the Point72 Ventures deep tech portfolio. AI has the ability to augment humans and make them more productive, and in some cases, replace them in tasks that are highly suitable to automation — freeing them up to focus on more value-add, strategic activities. Historically, the friction to introduce this automation has been high, but when you can’t hire someone to handle a customer service request or to man a desk, automation suddenly becomes a lot more interesting.

Macro conditions are becoming increasingly important to consider for startups. From inflation hurting insurtech margins to the Great Resignation, which is propelling demand for AI software. Something to bear in mind.

There are a few more factors to consider.

  • We’re noting a recent PitchBook entry looking at the bigger startup scene in Utah, given Podium’s recent huge round. As you might guess, the figures are trending in the right direction.
  • Faire just raised a Series G, and it’s got some notable growth numbers to report. So what? It had some interesting growth statistics to share, in its own words. Faire is an “online wholesale marketplace,” according to the firm, and it’s growing quickly. The company claimed “3x” revenue increase and more than 1 billion in annual volume, which piqued our interest. If the private market isn’t trying to convert it into venture capital foie gras, the business would be an IPO candidate.
  • What else is there? Koan, the OKR startup that sold to Gtmhub, came to an end after failing to raise a Series A. We would have gone deeper into the topic on a less busy week. However, given our regular coverage of the OKR software industry throughout the years, I wanted to draw attention to it. (Koan’s CEO was kind enough to share some insights about his firm’s conclusion both publicly and privately with us so that we may continue discussing it next week, depending on availability.)

Braze, in addition to VideoStaff. Finally, there’s Braze. This week, New York-based software unicorn Braze went public, and The Exchange caught up with the firm’s leadership on its IPO day.

The company in question was subject to rather tight limitations on what it could say (not much) and what it could not (nearly everything).

Still, we received some feedback on the preparation process, with the firm stating that it began preparing for its IPO several years ago, but that it has only recently begun the actual process of going public.

We wanted to know why the business — which hasn’t had to raise funds since 2018 – hadn’t sought a direct listing. The traditional IPO may not be as inflexible as many people believe, according to Braze’s Bill Magnuson.

That’s worth considering as we get closer to 2021 when the final public debuts will take place. After raising 65 per share in an initial public offering, Braze is currently valued at 94.16 per share.