Databricks swapped the old partner labels this year; Registered / Select / Elite / Global Elite are gone, replaced by Bronze, Silver, Gold, and Platinum. Most write-ups on this are aimed at buyers choosing a vendor. I wanted to flip it for the people who actually are the partner: the engineers and architects whose certifications and delivery work are what these tiers are quietly made of.
Here's the part that matters if you're the one doing the work:
Your certifications belong to you, but they roll up into the firm's tier. Databricks separates accreditations (can be self-assessed) from certifications (proctored exams). The certifications are the ones that carry weight toward a partner's standing. So when your company pushes for the team to get certified before a tier review, that's the mechanism. Your proctored cert is a line item in the org's Gold or Platinum case.
A quick practitioner-oriented read of each tier:
- Bronze - enrollment only. Access to training, portal, brand assets. No delivery history implied. A firm can hold it with zero live implementations, which is worth knowing if you're evaluating a codebase someone else built.
- Silver - first earned tier. Real customer activity + certified people. If you're at a growing practice, this is usually the first badge your work directly moves.
- Gold - requires at least one Brickbuilder specialization plus customer-success evidence. Specializations are the interesting bit here: they're validated on actual technical delivery in a product or industry area, not just headcount. This is where your domain depth (Unity Catalog, streaming, ML, a vertical) becomes an organizational asset.
- Platinum - invite-only, multiple specializations, strategic scale. Big-program governance and geographic breadth. Note: a Platinum firm can still staff a small engagement with junior people, the tier says nothing about who's on your project.
Why this is useful even if you never pick a partner:
If you're inheriting a partner-built platform, the tier tells you almost nothing about code quality. What actually predicts it is whether the named people who built it held current, relevant certifications and had done that architecture before. A Gold specialist with senior engineers on the work routinely produces cleaner, lower-risk delivery than a Platinum generalist who rotated in juniors. The badge is org-level and aggregated; the code is built by specific humans.
If you're at a partner and thinking about tier progression, the practical levers are: proctored certifications on the delivery team, a Brickbuilder specialization that matches where you actually have depth, and documented customer outcomes. Training completion alone doesn't move it.
For deeper insights on the full structure, the C&SI scorecard logic, how specializations differ from tier, and a layered way to verify capability at the team level rather than the logo level, I came across this extended version, if it's useful: Databricks Partner Tiers Explained (Bronze, Silver, Gold and Platinum).
For those of you working at partners: has the 2026 relabel changed how your org tracks certs internally? And for anyone who's inherited a partner-built Databricks environment, did the partner's tier match what you actually found in the code?
Announcements
