Snowflake vs Databricks: Navigating the Impact of Recent Acquisitions on Your Business Strategy
Recently, I’ve found myself in the company of various C-Suite executives, including CEOs, CHROs, and CIOs. The talk of the town? The tech world’s recent acquisitions — Snowflake’s acquisition of Neeva and Databricks’ acquisition of MosaicML. It’s causing quite a stir, and these leaders are now
evaluating their allegiances to these platforms, wondering whether they should stay the course or make a switch. The question isn’t idle chatter. It’s a strategic consideration that could shape the future of their respective companies.
As these discussions unfolded, it became clear that understanding these acquisitions and what they mean for businesses was crucial. So, I took the liberty of discussing these developments with the wider community of data and tech enthusiasts on LinkedIn. From these conversations,
I’ve derived a simple “rule of thumb” to guide these C-Suite executives in their decision-making.
Snowflake, with its acquisition of Neeva, is now a powerhouse in running inference at a large scale. The significance? In the world of machine learning, 80–90% of costs are typically tied to inference—making predictions using machine learning models. If your organization heavily relies on making predictions from
models, such as recommendation systems or predictive analytics, you might want to consider sticking with or moving to Snowflake. They’re stepping up their game.
Meanwhile, Databricks’ acquisition of MosaicML has streamlined the training of large language models (LLMs). If your organization has a keen interest in creating and training its own models, Databricks, armed with its new capabilities, might be the better choice for you.
But let’s dive a little deeper.
Use Cases:
Snowflake is focusing on indexing enterprise data (think of it as
organizing a vast amount of data in a searchable way) and facilitating quick
and cost-effective inference, particularly for search-related applications.
Databricks is enabling businesses to create their own large language models from their data, likely for applications like search or chatbots.
Rule of Thumb: If your business has large amounts of data that you want to make more searchable or usable, Snowflake might be the right choice. If you’re more interested in developing custom language understanding applications (like chatbots), look into Databricks.
Quality of Data:
Both companies may be underestimating the quality of enterprise data,
which is often unstructured and disorganized, particularly for machine learning applications.
Rule of Thumb: Before investing heavily in machine learning solutions, ensure your data is in a state that’s usable for machine learning.
If it’s too structured, you may not need large language models. If it’s very
领英推荐
unstructured, consider implementing an automated ETL (Extract, Transform, Load) process to organize your data first.
Big Cloud Providers:
These acquisitions show that Snowflake and Databricks are
positioning themselves as important centers of data, even beyond the scope of the big cloud providers like AWS, Google Cloud, and Microsoft Azure.
Rule of Thumb: Consider your needs beyond just data storage. These companies offer unique capabilities that might provide value beyond what the big cloud providers offer.
Lastly, let’s talk about the winners and losers in this scenario. According to our discussions, the real winners here are enterprises who will have more control over their data and models, without having to send their data to potentially unregulated environments. The potential losers? Companies that provide middleware for large language models. Snowflake and Databricks are expected to develop their own tools in this area, so if you’re in this space, it’s time to think about differentiating your value proposition.
These rules of thumb are not one-size-fits-all solutions, but they offer guidance.
It’s essential to consider your organization’s specific context and needs when making strategic decisions. Let these recent acquisitions and the ensuing discussion inform your choice, but ultimately, the best decision will depend on your organization’s unique requirements and goals.
Comparative Analysis: Snowflake vs Databricks — A Guide forStrategic Decision-Making
Please note that this table is a simplification of the broader discussion and the decision to choose Snowflake or Databricks should be based on a thorough analysis of your specific business needs and context.
Sources:
Snowflake’s acquisition of Neeva:
Databricks’ acquisition of MosaicML: 6. TechCrunch: Databricks picks up MosaicML, an OpenAI
competitor, for $1.3B .