Inside Databricks’ $1B Deal for Next-Gen Data Platform  ## Metadata - Author: [[Hard Fork AI]] - Full Title: Inside Databricks’ $1B Deal for Next-Gen Data Platform - Category: #podcasts - URL: https://share.snipd.com/episode/fbb64008-540d-4c69-8e23-f01c62af2b49 ## Document Notes ## Highlights - **Use AI Box for Multi-Model Access** - Try AI Box to access top AI models and generate audio, text, and images in one platform for $20/month. - Use features like audio comparisons and mid-chat model switching to optimize your creative workflows. Transcript: Jaeden Schafer Our very first product is called the AI box playground. And I would love for you to try it out. Essentially, we allow you to access all of the top AI models, audio, text, image, all inside of one platform for $20 a month. So you don't need subscriptions to all of the different platforms. And in addition to that, we have way more features and some really cool tools. Some of my favorites are the fact that you can generate audio with it. So 11 labs is integrated if you need an audio for something, whether you're creating, you know, a voiceover or whatever, and you essentially can go and compare side by side the audio Tracked from different tools. So if you generate, you know, if you get OpenAI to generate something and you get 11 labs to generate something, you can get them side by side. You can listen to a comparison of what each of them sound like. And one of the cool things is inside of the platform, we have the ability for you to switch the AI model mid chat. So I might be chatting. I might've used OpenAI to generate a piece of audio. I could switch it over to 11 labs and I could go and check out all of the different voices in 11 labs. I can regenerate the same thing with a whole bunch of different voices, stick them all side by side, listen to all of them and compare, which I like best. So a ton of really cool features. One of my favorite is the side-by comparison of different model outputs. You do that with audio, you can do that with image and you can do it with text. You get a response from chat GP. You're like, I don't really like it. You can get Grok, Google Gemini, and Anthropics Cloud to all generate responses to the same question. And you can look at which one you think is best for that particular task. ([Time 0:00:27](https://share.snipd.com/snip/39930eab-5d9f-4bd1-9b5d-b7b1373e2784)) - **Databricks' Strategic $1B Acquisition** - Databricks acquired Neon, a startup building an open-source alternative to AWS Aurora Postgres, for $1 billion. - This complements their data intelligence services with a serverless relational database suited for AI workloads. Transcript: Jaeden Schafer All right. So one of the big, I think, shockers in this whole story is that we, of course, have a huge outcome, right? This is a neon. It's selling for a billion dollars. That's a massive price tag. Essentially, it is a startup that's building an open source alternative to AWS's Aurora Postgres. And so Databricks said that they're just pretty much excited to buy this because it lets them combine that startup's serverless relational database management system with their Own data intelligence services. Okay. Sounds like a lot, but pretty much what it's doing is it's letting customers deploy AI agents more efficiently to spin up databases. Okay, then the price and the sale of this is crazy. I'll go into like some of the AI stuff that this database company is doing because I actually do think it's impressive. And do not get me wrong, I'm going to like put a big caveat on the financial side here really quick. But don't get me wrong, this is a very fascinating company. I think it's doing some really interesting things. Like it's getting sold for a billion dollars. Obviously, that's amazing. It's doing a great job. But the ([Time 0:02:05](https://share.snipd.com/snip/b364d8da-1ba2-47ce-b696-d7b57941df59)) - **Financing Enables Databricks Growth** - Neon raised $129 million but is sold for under 10x that amount, reflecting interesting valuation dynamics. - Databricks, with $19 billion raised and a $62 billion valuation, can afford such large acquisitions to expand rapidly. Transcript: Jaeden Schafer Like it's getting sold for a billion dollars. Obviously, that's amazing. It's doing a great job. But the crazy thing is that they actually have raised combined cash up until this point, a ton of money. They've raised $129 million. And that's come from massive investors. So we have Microsoft's venture arm, which is M12, General Catalyst, Menlo Ventures, Notable Capital, huge people, right? And they most recently had raised a ton in financing. So I will break down all of that. But before we get into that, I want to say Neon, right? They've raised $129 million. They're getting sold for less than 10x the cash that they have raised total, a billion dollars. ([Time 0:03:03](https://share.snipd.com/snip/e91da5c6-e528-42e8-b276-0369b32c80de)) - **Neon's Efficient Database Scaling** - Neon offers a managed cloud database that auto-scales compute and storage based on usage. - It enables rapid cloning and point-in-time recovery, reducing wasted cost from over-provisioning. Transcript: Jaeden Schafer Like, what is this actually capable of doing? So what I found interesting was the fact that this was founded back in 2021. So a lot of the companies at the time, peak Zerp era, went bankrupt, really struggled. But they essentially had a free usage-based plan that I think really helped them grow a ton. This was a managed cloud-based database platform. Essentially, it was going to let developers clone a database, and then they could look at all the changes that were made before it would actually go into production. So it was very useful, able to really quickly spin up clone databases, and then it automatically scales processors, memory, and storage according to whatever the usage was needed. So they had isolated database instances for testing and development, as well as point in time recovery. So this is impressive because it kind of is automatically determining how much storage it needed, how much memory needed, what type of processors it needed. And these are really expensive things. Like a lot of times what happens is you'll just say, hey, I want like, I'm going to buy like X amount of capability for my server, and you just get charged for the whole thing, even if you're Not using it. So this is kind of cool that they allowed you to scale it. And I think that's what a lot of people are excited about. ([Time 0:05:37](https://share.snipd.com/snip/649c1fe6-4322-4605-b6b7-c63942a1f734)) - **AI Agents Automate Database Deployment** - 80% of Neon databases are provisioned automatically by AI agents instead of humans. - This aligns perfectly with Databricks' focus on AI-native agent-driven applications. Transcript: Jaeden Schafer Databricks evidently was also very excited about acquiring this, because they said that this is, you know, ideally suited to workloads that are run by AI agents. This is what Databricks is pushing heavily into is AI right now. And so when they see kind of this technology, they're like, hey, this is pretty much exactly what we need to run AI agents because they operate much faster than human developers, which, You know, kind of need supervision and you have to control for errors. So because of all of this, Databricks essentially cited, which was kind of interesting, that 80% of the databases that were, quote, provisioned on Neon were created automatically By AI agents rather than humans. So four out of five databases that were spun up over on Neon were actually done by AI agents or AI running. ([Time 0:06:52](https://share.snipd.com/snip/5949f666-69d1-48ca-8ad9-1cffc03a2de1)) - **Databricks CEO on AI-Native Databases** - Databricks CEO Ali Ghodsi highlights the shift to AI-native, agent-driven apps reshaping database needs. - Neon’s serverless Postgres supports agentic speeds and pay-as-you-go economics with open-source advantages. Transcript: Jaeden Schafer And here's the quote that they specifically said. This is Ali Godsky, the co-founder and CEO of Databricks. They said, quote, the era of AI native agent driven applications is reshaping what a database must do. Neon provides it, proves it. Four out of five of every databases on their platform are spun up by code, not humans. By bringing Neon into Databricks, we're giving developers a serverless Postgres that can keep up with agentic speed, pay-as economics, and the openness of the Postgres community. I think at the end of the day, what's cool is that this is kind of like an open source project or open source tool or has those capabilities. ([Time 0:07:45](https://share.snipd.com/snip/f6f92194-c1b9-45e7-bb70-d62515663982)) - **Acquisition Strategy Boosts Databricks** - Databricks’ acquisition strategy helps it rapidly grow its footprint and consolidate its AI infrastructure position. - Buying successful companies like Neon and others strengthens its market competitiveness and innovation pipeline. Transcript: Jaeden Schafer And by buying companies that are obviously successful and doing quite well, Databricks seems to continue to grow its footprint. So I think the acquisition play is a great one for Databricks. You can see a clear path to success here. So I'm really excited and looking forward to see what they do. ([Time 0:08:49](https://share.snipd.com/snip/39d8d01b-cbcd-48ba-9df5-2905890a86d8))