MongoDB CEO Dev Ittycheria talks AI hype and the database evolution as he crosses 10-year mark

Quite a bit has occurred since Dev Ittycheria took the reins at MongoDB, the $26 billion database firm he’s led as president and CEO since September 2014. Ittycheria has taken MongoDB to the cloud, steered it through an IPO, overseen its transition from open supply, launched a venture capital arm, and grown the shopper base from just a few hundred to one thing approaching 50,000.

“Once I joined the corporate, it wasn’t clear if folks would belief us to be a very mission-critical expertise,” Ittycheria instructed TechCrunch. “Once I joined, it was doing roughly $30 million in income; now we’re doing near $2 billion.”

It hasn’t all been peaches and cream, although. 5 months in the past, MongoDB was hit by a security breach, which, whereas comparatively contained, did momentarily threat its status in an business the place status is paramount.

Throw into the combo the whirlwind AI revolution that has engulfed nearly each business, and there was a lot to debate when TechCrunch sat down with Ittycheria at MongoDB’s new London workplace, which opened in Blackfriars final 12 months.

MongoDB’s London workplace.  Picture Credit: Paul Sawers / TechCrunch

Vector’s embrace

Databases have come a great distance since IBM and Oracle first popularized relational databases greater than half a century in the past. The web’s rise created demand for versatile, scalable, and cost-effective knowledge storage and processing, paving the best way for companies reminiscent of MongoDB to thrive.

Based in 2007 by a trio of veterans hailing from on-line adtech firm DoubleClick (which Google acquired for $3.1 billion), MongoDB was initially called 10Gen till a rebrand to the name of its flagship product six years later. It has since emerged as one of many preeminent NoSQL databases, serving to firms retailer and handle giant volumes of knowledge.

Previous to becoming a member of MongoDB, Ittycheria based and exited a server automation firm referred to as BladeLogic for $900 million in 2008, and went on to serve in numerous board member and investor roles (together with a 16-month stint at Greylock) earlier than joining MongoDB as president and CEO approaching for 10 years in the past now. Ittycheria changed Max Schireson, who stepped down for household causes after simply 18 months within the function.

Constructed on a document-oriented mannequin, MongoDB has grown off the again of the explosion in cell and internet purposes the place versatile, dynamic knowledge buildings are at play. The present synthetic intelligence wave is driving an analogous shift, with vector databases the hot new thing in town.

Like NoSQL, vector databases additionally concentrate on unstructured knowledge varieties (e.g., pictures, movies, social media posts), however are significantly nicely suited to giant language fashions (LLMs) and generative AI. That is because of the manner they retailer and course of knowledge within the type of vector embeddings, which convert knowledge into numerical representations that seize relationships between totally different knowledge factors by storing them spatially by relevance. This makes it simpler to retrieve semantically comparable knowledge and permits AI to higher perceive context and semantics inside conversations.

Whereas a slew of devoted vector database startups have emerged these previous few years, the incumbents have additionally began embracing vector, together with Elastic, Redis, OpenSearch, Cassandra, and Oracle. Cloud hyperscalers, together with Microsoft, Amazon, and Google have additionally ramped up help for vector search.

MongoDB, for its half, introduced vector search to its flagship database-as-a-service product Atlas final June, an indication that the corporate was making ready for the oncoming AI tsunami. This mimics different historic developments the place single-function databases emerge (such as time-series) with some utility as stand-alone options however that may even be higher built-in into a bigger multi-purpose database stack. That is exactly why MongoDB introduced support for time-series databases just a few years again, and why it’s doing the identical with vector.

“Numerous these firms are options masking as merchandise,” Ittycheria mentioned of the brand new wave of devoted vector merchandise. “We constructed that into the platform, and that’s the worth — quite utilizing some stand-alone vector database after which your OLTP [online transaction processing] database after which your search database, we will mix all three issues into one platform that makes the lifetime of a developer and architect a lot simpler.”

The concept is that database suppliers that undertake a multipronged method can mix all the info in a single place, making life simpler for builders to work with.

“There’s most likely like 17 various kinds of databases, and doubtless about 300 distributors,” Ittycheria mentioned. “There’s no buyer on this planet that desires to have 17 totally different databases. The complexity that creates, and the price of studying, supporting and managing these totally different applied sciences turns into overwhelming. It additionally inhibits innovation, as a result of it creates this tax of complexity.”

MongoDB’s Dev Ittycheria. Picture Credit: MongoDB

An excessive amount of hype

Regardless of the preparation, Ittycheria reckons there may be an excessive amount of hype round AI — for now, no less than.

“My life has not been remodeled by AI,” he mentioned. “Sure, perhaps I can write an electronic mail higher by means of all these assistants, however it’s not basically remodeled my life. Whereas the web has utterly remodeled my life.”

The idea is that regardless of the hullaballoo, it’ll take time for AI to seep into our on a regular basis lives — and when it does, it is going to be by means of purposes integrating AI, and companies constructing on it.

“I believe with the adoption of any new expertise, we see worth accrue on the backside layer first,” Ittycheria mentioned. “Clearly, Nvidia is making money hand over fist, and OpenAI has been essentially the most talked about firm since they launched ChatGPT. However the actual worth will come when folks construct purposes on high of these applied sciences. And that’s the enterprise we’re in — we’re within the enterprise of serving to folks construct purposes.”

For now, it’s all about “easy apps,” as Ittycheria places it. This contains chatbots for customer support, one thing that MongoDB itself is doing internally with CoachGTM, powered by MongoDB’s vector search, to deliver its gross sales and buyer groups immediate information about their merchandise. In some methods, we’re at the moment within the “calculator apps” stage that the iPhone discovered itself in practically 20 years in the past when the idea of the App Retailer hit the lots.

“The true subtle [AI] apps shall be utilizing real-time knowledge, having the ability to make real-time selections on real-time occasions,” Ittycheria mentioned. “Possibly one thing’s occurring within the inventory market, perhaps it’s time to purchase or promote, or it’s time to hedge. I believe that’s the place we’ll begin seeing far more subtle apps, the place you may embed real-time knowledge together with all of the reasoning.”

The SaaS path

One of many greatest developments throughout Ittycheria’s tenure has been the transition from a self-deployed mannequin, the place clients host MongoDB themselves and the corporate sells them options and companies. With the launch of Atlas in 2016, MongoDB launched into the acquainted SaaS path the place firms cost for eradicating all of the complexities of self-hosting. On the time of its IPO the next 12 months, Atlas represented 2% of MongoDB’s income — as we speak that determine sits at practically 70%.

“It’s grown in a short time, and we’ve actually constructed that enterprise as a public firm,” Ittycheria mentioned. “What the recognition of Atlas confirmed was that persons are comfy consuming infrastructure as a service. What that enables them to do is delegate what they think about ‘non-strategic features,’ like provisioning, configuring and managing MongoDB. To allow them to give attention to constructing purposes which might be actually reworking their enterprise.”

One other main growth got here when, a 12 months after going public, MongoDB moved away from an open supply AGPL license to a source-available SSPL (server facet public license). In some methods, this was the bellwether of what was to come back, with numerous infrastructure firms occurring to desert their open supply credentials to stop the cloud giants (e.g., Amazon) from promoting their very own model of the service with out giving again.

“We really feel very blissful about it [the license change],” Ittycheria mentioned. “The fact is that whereas it was open supply, 99.9% of the event is finished by our personal folks — it’s not like communities contributing code. It’s not some easy, trivial utility — it’s very advanced code, and we have to rent senior, gifted individuals who value some huge cash. We didn’t suppose it was truthful for us to spend all this cash to construct this product, then somebody takes that free product, monetizes it, and never give us something again. It was fairly controversial in 2018, however wanting again, our enterprise has solely grown sooner.”

And grown it has. As with nearly each tech firm, MongoDB’s valuation soared through the pandemic, peaking at an all-time excessive of $39 billion in late 2021, earlier than plummeting south of $10 billion inside a 12 months — roughly the identical as its pre-pandemic determine.

Nonetheless, MongoDB’s shares have been in ascendency within the 18 months since, hitting $35 billion simply a few months in the past, earlier than dropping once more to round $26 billion as we speak — such is the risky nature of the inventory markets. However given the corporate’s comparatively modest $1.8 billion valuation at the end of its first day of buying and selling in 2017, MongoDB has carried out pretty nicely for public market traders.

Dev Ittycheria with MongoDB colleagues at its 2017 IPO. Picture Credit: MongoDB

4 months in the past, although, MongoDB revealed a knowledge breach that uncovered “some buyer account metadata and make contact with data” — it concerned a phishing assault by means of a third-party enterprise device (Ittycheria wouldn’t verify which). This brought on its shares to drop 3%, however within the months that adopted, MongoDB’s valuation surged again to a two-year excessive. This highlighted how little influence the breach had on affairs on the firm, actually in comparison with high-profile knowledge breaches at the likes of Equifax and Target, which hit the companies exhausting and compelled senior executive departures.

Whereas MongoDB’s cybersecurity incident was considerably smaller in scope, what stood out was how rapidly the entire thing went away — it was reported in a number of shops (including TechCrunch), however the story disappeared into the foggy ruins of time simply as rapidly because it arrived.

“A part of the reason being that we had been very clear,” Ittycheria mentioned. “The very last thing you need to do is cover data and seem such as you’re misrepresenting data. Now we have plenty of banks who put a variety of very delicate data in our knowledge platform; we’ve plenty of different firms which have a variety of delicate data. So for us, it’s actually about ensuring that our structure is strong and sound. And this actually pressured us to double down. I might by no means declare that we’re by no means gonna get hacked once more, however we’re doing every part in our energy to make sure that it doesn’t.”

Nothing ventured

It’s commonplace for the most important tech firms to launch their very own funding automobiles, as we’ve seen by means of the years with Alphabet (which has several funding offshoots), Microsoft, Amazon, and Salesforce all ingratiating themselves with the startup fraternity. However a more moderen wave of enterprise company enterprise companies have entered the fray, too, together with Slack, Workday, Twilio, Zoom, HubSpot, and Okta.

In 2022, it was MongoDB’s turn to launch such a fund, and within the two years since, MongoDB Ventures has invested in some eight firms.

“That is for us to construct deeper relationships — we work in an ecosystem that consists of enormous firms and in addition small firms,” Ittycheria mentioned. “The place we see a small firm that we predict might be fascinating to work with, we are saying, ‘Hey, we wish an opportunity to spend money on you,’ in order that additional worth’s created. We are also the beneficiaries of making a few of that worth.”

MongoDB solely has a handful of individuals in its company growth staff which might be largely targeted on the enterprise fund, and Ittycheria stresses that MongoDB takes a again seat with its investments. It additionally usually invests alongside different VCs, because it did with its inaugural funding in 2021 (predating the formal launch of its fund), when it quietly joined the likes of Perception Companions and Andreessen Horowitz in Apollo GraphQL’s $130 million Series D round. 

“We all the time take a minority place, we don’t take a board seat, and we don’t set the phrases,” Ittycheria mentioned. “However the motive startups are thinking about us is as a result of they need to leverage the MongoDB model. Now we have 1000’s of individuals within the subject, so that they [startups] can leverage our distribution channels.”