ZD 25.26: Picking a Data Platform Today
My guide for finding a modern data platform to work with, mid-2025 edition.
Data platforms have changed a lot. Here is an educated guide to some of the options around in mid-2025. Don’t miss the musical coda, a live-show punk rock anthem that rips.
This week, I am trying a different layout — The Distilled Spirit is still here and still just as awesome, it is now just below the fold. Let me know if you love it or hate it in the comments. Happy fourth🎉!
Choosing a Data Platform in 2025
I wrote and presented on Model Context Protocol (MCP) in the last week. As I crafted the content, I realized that the data platforms you use to power AI have been often misunderstood. They tend to be opaque to most people. At heart, I’m a database geek. I thought I would share a bit about the ones I am using today to hopefully help people find their next tool.
There are a ton of data platforms out there. I wrote about the ones I actively use today. If you think I missed one let me know in the comments.
The four platforms I use enough to speak about today are:
Google Sheets is probably familiar to most — it is Google’s spreadsheet in the sky in case you missed the memo. It is also capable of integrating with Google Forms and other automation platforms and has some pretty sweet AI integration. I am not talking about Google App Sheet which is very new and honestly looks very limited.
Notion is an amazing tool. I use it to manage this publication among other tasks. It is also a data platform capable of talking to other tools and is worthy of this conversation.
Airtable is another tool I spend a lot of quality time using. It hits a sweet spot for the data problems I usually tackle, and has some great platform features that make it a strong all-around contender.
Supabase is definitely not an end user tool — it is a feature-rich developer platform. It has enough integration that you can look at it in the same light as these other tools in the automation world.
Mini applications are having a renaissance. Make yours today!
The Big Picture
In case you want the quick version, check out the handy reference chart below comparing all of the platforms across nearly a dozen criteria. Read on for the explanations of the ratings.
The Breakdown
Here is the breakdown of the criteria and why I rated the platforms that way, broken down by section of the graph.
End User Accessibility
How accessible is the tool to an average end user? Do you need specialized training or skills to leverage the platform?
Google Sheets (😍😍😍) feels like the office apps everyone knows and loves. Clear winner here.
Notion (😍😍) is a strong second place with the very wiki-like, type and paste anything feel.
Airtable (👍👍) is not too hard for people to understand — it feels enough like a spreadsheet not to be scary at least.
Supabase (❌❌❌) is not for end users in general.
Traditional Spreadsheet
How much does this tool work like a spreadsheet. Can users add fields move columns do math on the fly and write inline functions?
Google Sheets (😍😍😍) is a spreadsheet program.
Notion (😩) can make tables and you can do math in a few circumstances. It will also let you put tables in pages. But it does not do spreadsheet things beyond superficial levels.
Airtable (😩) is a bit deceptive. It looks a lot like a spreadsheet, and many online tutorials compare it to a spreadsheet. But it is really a database tool not a spreadsheet and you often figure that out the hard way.
Supabase (❌❌❌) is robust enough that you could probably program a spreadsheet to use it as a back end. But it is not a spreadsheet application.
Relational Database
Does the tool do all the relational database things you would expect out of a modern RDBMS platform?
Google Sheets (❌) is a spreadsheet and while it has table functionality it does not do relationships in a meaningful way.
Notion (👍) tables have improved a lot, and you can leverage a passable relational model. Works in the context of Notion at least.
Airtable (😍) does relational database very well while not getting too far into relational algebra. The recently added backlinks for self-referencing tables cleans up a major weakness.
Supabase (😍😍😍) is a full-featured relational database and much more; it is just Postgres in the sky with a pretty slick interface and app platform added.
Automation Capabilities
Does the platform have native automation capabilities? How mature are they?
Google Sheets (❌) has macros natively, which I am not going to count.
Notion (😩) has some improving capabilities, at least on the paid tiers, but they are not quite on par with other platforms.
Airtable (😍) has excellent, accessible automation built into the platform and it is the best choice without getting into writing code.
Supabase (👍👍) has robust triggers and full cron access should you need to go that deep.
Development Platform
How functional is the platform for development? Can you deploy applications on it? Can it integrate with other tools?
Google Sheets (❌) there is not a development platform.
Notion (❌) has not managed to grow a development platform as of yet.
Airtable (👍) passes muster as a scripting platform, and they are doing a decent job of building enough distribution tooling that you can see it as a development platform if you squint enough.
Supabase (😍) has a robust development platform attached to the service, enough to build full-featured applications with enough vibe coding.
Reporting Platform
Does the tool have a native reporting platform? How can it put data in front of end users?
Google Sheets (👍) has robust graphing features and integrates nicely with google slides and google docs.
Notion (😍) is a reporting platform extraordinaire. The ability to juxtapose data and content is handy, as is the ability to share publicly.
Airtable (👍) can do some solid reporting. It falls pretty flat when it comes to print without bringing in third party tools.
Supabase (❌) does not have one but that is not Supabase’s job either.
Application Platform
Does the tool have a native application platform? a way for end users to create and update records and generate reports?
Google Sheets (👍) is its own application platform without needing outside tools. Limited in terms of distribution models, but extremely powerful in its own right.
Notion (👍) has grown enough forms and UI integration to make this work, at least for people using notion directly.
Airtable (🤩) has built a really robust application layer on top of the bases. You can deliver end users applications that do not need much explanation or context. You can do pretty effective line of business apps on the tool.
Supabase (👍) requires development or at least vibe coding, but it can host apps using its edge functions and CDN.
Automation Platform Integration
How well does the platform integrate with automation platforms like Make and Zapier?
Google Sheets (🤩) is very widely supported on all modern automation platforms. It is often the example application in demonstrations.
Notion (👍👍) has solid support, especially on the database side, though working with the content can be a bit tricky due to the nature of the beast.
Airtable (🤩) is another automation platform darling — it needed them before it had it’s own automations.
Supabase (👍👍) has good integrations if you have the know-how — it can be accessed as Postgres if there is no Supabase option.
AI Integration
How integrated is AI into the tool? Does it have any AI capabilities that are accessible to end users?
Google Sheets (👍) the Gemini ai() function is pretty rad and lets you do lots of cool things in line.
Notion (🤩) has amazing AI integration inside the product from both an assistance and a retrieval standpoint. They have done a great job integrating the technology into the tool.
Airtable (👍) has decent AI functionality within your base. They have made major updates, see this video for a rundown on the features. At the time of writing I have not had a chance to use any of them, this might be up for a re-rating soon.
Supabase (👍) does not have much AI directly but it has tools that power a lot of your AI integrations. It can host your knowledge and act as your vector store. Edge Functions can handle embeddings.
MCP Support
Does the platform provide any native MCP tools? How good are those tools?
Google Sheets (❌): No native MCP client. It should be said that most AI providers work with Google Drive so you can easily share your sheets with your chatbot of choice.
Notion (🤩): Notion has official MCP support in the cloud and I’ve been impressed with the implementation. This is where the others should aspire to be.
Airtable (❌): No native MCP client supported by the vendor. There are some local ones if you need access.
Supabase (👍): Supabase has official MCP support but it is really a developer tool, like the rest of the stack. The developer tool is excellent.
Pricing
How expensive is the platform to use?
Google Sheets (🤩) has a great free level and the paid versions of Google Workspace are not that pricey. It is a very accessible platform.
Notion (🤩) has a pretty generous free tier for single users, and it is not expensive in the enterprise software sense.
Airtable (😩) has a lot of bells and whistles but those bells and whistles come at a price, especially when you want to share access to the tool.
Supabase (👍) has a great free tier and does not get that expensive. You can self-host too.
Choose Your Weapon Wisely
Each of these data platforms brings something different to the table. If all you really wanted was a spreadsheet then Google Sheets is a win. Notion is an amazing wiki with embedded database functionality. Airtable is a fully featured database app. And Supabase is a back-end to many of your AI-workflows in other apps.
No single platform is perfect—each one excels in a different lane. My rule of thumb: pick the simplest tool that solves today’s problem, then graduate as complexity grows. Start small, experiment, and iterate. The next, better AI is just around the corner. Get ready to be a context engineer today.
Musical Coda
The Distilled Spirit
Living in Exponential Times
We live in amazing times, here are a few discussions of how fast technology is advancing.
🆕 50 Things That Did Not Exist 5 Years Ago ()
It has been an astounding half decade of rapid advances. Here are some things that only came to be in the last half decade of exponential progress.
📊 10 Charts that Capture How the World Is Changing ()
10 charts exploring AI, protein intake, Gen Z politics and more. Always worth a look.
👽 The Dawn of the Posthuman Age ()
For the first time in written history, humanity will shrink. Technology and low birth rates are going to change the species.
🔍 The End of the Anonymous Internet ()
Texas’ supreme court victory means the end of internet privacy as we understand it.
Zohran Wins!
Zohran Mamdani won the New York City mayoral primary in a win that is spilling loads of digital ink. Here are some good reads.
👀 It Is the Attention Economy Stupid ()
Zohran understood this lesson implicitly and knew how to leverage the attention economy.
😃 The Fun Factor ()
Being unfiltered and fun might actually help you win elections.
📔 A City We Can Afford ()
How economics drove Zohran’s success in New York.
📈 How Zohran Won in 4 Charts ()
Where Zohran did and did not succeed within New York’s electorate.
Model Context Protocol Marches On
MCP is amazing, see last week’s discussion for some details. Now that it is gaining mindshare, here are a few choice articles on the subject to continue the discussion.
🗣 Talk to Your MCP With 11.ai (Elevenlabs)
Voice AI leader Elevenlabs built a MCP host you can have a conversation with. Ever wanted to talk to your Notion content?
🔌 A MCP Connection Too Far ()
MCP’s network effect is amazing, but when you can go from your database all the way to your web enabled toaster it might be a step too far.
💸 MCP Value Creation ()
The failure of the API gold rush has lessons for MCP builders.
How To Do Things With AI
Builders are getting seat time with AI agents and living to write about it.
📖 Getting the Most out of Deep Research ()
Deep Research is amazing and worth a long look. It is tricky to get a handle on though, but see this article for guidance on how to get the most out of it.
6️⃣ Six Builders Who Will Thrive in The Era of AI ()
Six archetypes of people who will thrive in the AI era.
🤷♀️ Why Just Asking AI to Code It Does Not Work ()
Some thoughts about the serious limitations of vibe coding.
How to Do Things Without AI
Ways you can improve your writing and thoughts this week.
📝 7 Writing Lessons I Learned the Hard Way ()
Really good ideas on ways to improve your writing process.
🏃♀️ Resistance Train Your Mind for the Age of AI ()
Ways to keep thinking even when AI wants to do it for you.
🙃 The Power of Inversion ()
Inversion, when properly used, is a very effective problem solving technique.
The Look
Defense R&D is seed corn for technology.
Did you enjoy reading this post? Hit the ♥ button above or below because it helps more people discover great Substacks like this one and it helps train your algorithm to get you more posts you like. Please share here or in your networks to help us grow!
Inversion is also a great way to use AI. Instead of having it answer your questions have it help you figure out the right questions to answer. Use it to challenge the answers you already have!