As I complete my first year at Databricks, I thought it would be fun to summarize the bricks we're laying for ISVs and data providers to become your most strategic partner.
Over the last year, I've met with hundreds of you to learn what you think we're doing well, what we need to improve, and what you'd need to see for us to become your most strategic partner. Most of you said we're the "horse to bet on", but you wanted 3 stronger "bricks" in our wall: 1) better GTM support, 2) more prescriptive technical guidance, and 3) the ability to transact on our Marketplace.
At our Partner Kickoff in February, I announced how we've laid the first two bricks: a new partner tiering program offering increasing GTM support the more business we drive together, and a Partner Well-Architected Framework (or PWAF) with clear technical guidance for integrating with Databricks (read about all the goodies David Porter and his team have added to PWAF since here).
This week at DAIS, we're laying the third: you can now access our customers’ pre-committed spend to speed up your deals on our Marketplace.
But we didn’t stop there. You can now also list and share Databricks Apps and Genie Agents, which we believe opens the door for innovative new commercial models, like “pay-per-question” (keep reading if that sounds interesting).
To reiterate the vision we laid out in Feb, we aspire to offer customers any tool, model, dataset, or agent they need, and help you grow your business along the way.
We think these bricks give us a shot at fulfilling that vision, and we hope you're as excited as we are to add "another brick in the wall" with your brand on it. Together, brick by brick, we can deliver customers the most comprehensive data and AI capabilities on earth.
As always, let me know what we need to do better to keep building together (stephen.orban@databricks.com).
One wall. Many bricks.
More ways for partners to build, sell, share, and monetize with Databricks.
Anecdotally, most of you told us your Databricks partnership is growing faster than your other partnerships, and you're looking for ways to capitalize on our joint momentum.
By the numbers, more than 20,000 customers - including over 70% of the Fortune 500 - rely on Databricks to detect fraud, discover new drugs, optimize their supply chain, personalize customer experiences, and more. This diverse set of customers and use cases has propelled our annual run rate more than 65% YoY past $5B.
I realize many of the conversations we've had over the last year may have a selection bias toward Marketplaces, since many of you worked with me on AWS's or Google Cloud's Marketplace over the last decade. But, it's impossible to ignore how many of you asked us for new ways to reach our customers, access our committed backlog, and incentives for our sales teams to land your solutions.
So, without further ado, we're excited to announce a plethora of new Marketplace capabilities and new sharing/delivery models to drive commercial innovation, particularly for those with proprietary data sets.
Starting this week, customers with a Universal Commit can submit invoices for eligible partner solutions on Databricks Marketplace. Once we validate that the transaction runs on and/or shares data to Databricks, we'll decrement the customer's commitment obligation accordingly and compensate our sales team for helping you with their account.
Customers get the solutions they need on Databricks, you get access to pre-committed budget to accelerate deals, and our sellers are all motivated to lean in. Everyone wins.
Later this year, we'll launch transactability, where customers will be able to use their pre-paid spend and/or our billing systems to buy eligible partner solutions, and we'll handle the remittance minus a small (industry-standard) fee to cover our transaction costs.
Solutions must run on or share data to Databricks to be eligible, and partners with eligible solutions include AccuWeather, Acxiom, Anomalo, Anvilogic, Arctic Wolf, AtScale, Cognition, Datavant, Definitive Healthcare, Domo, Dun & Bradstreet, Engine, Epsilon, FactSet, Fivetran, Health Catalyst, Intent, Komodo, Kythera, LakeFusion, Monte Carlo, mParticle, Omni Analytics, OneTrust, Qlik, Replit, Sigma Computing, ThoughtSpot, TrendAI, ZoomInfo.
Many of our customers choose Sigma because of how well-integrated we are across Databricks' vast surface area. Over 300 customers choose Sigma on Databricks for best-in-class AI and Analytics. Being able to speed up our sales cycles on the Databricks Marketplace is a natural extension of our relationship, and we're excited to help Databricks shape and grow their offering.— Mike Palmer, CEO, Sigma Computing
We launched Databricks Apps in November of 2024, and it's quickly become one of our fastest-growing product lines by almost every metric. We’ve seen 5X growth since last year’s summit, with more than 5,000 accounts running Apps in production weekly. Our top 100 Apps customers have deployed on average over 250 production Apps, with use cases ranging from custom analytics dashboards, ops portals, job workflow managers, AI chat interfaces, data tagging workflows, custom model interfaces, and more.
And, a significant chunk of Apps adoption is coming from line-of-business users, with the data teams and app builders sharing these apps to democratize data and AI for all.
While customers love sharing these Apps across their enterprise, they've been unable to share these Apps with their customers... until now.
Starting this week, you can create Apps, list them on our Marketplace, and reach our 20,000 customers around the world. Customers can discover these Apps on the Marketplace, deploy these Apps in their own Workspaces, and give them access to their own data plus whatever data you've either shipped in the App or OpenShared with them. All while keeping your code and core IP closed from the customer.
Launch partners include Acxiom, Bright Data, Capital One, Datavant, Meta, Slalom, Stagwell, The Trade Desk, and more than twenty others across financial services, healthcare, advertising, marketing, and analytics.
Deploying Datavant’s tokenization technology as a native Databricks App eliminates the need for data movement. By building directly on Databricks, we allow clients to securely tokenize healthcare records right where their data already lives in a simple and intuitive manner. Combining our trusted tokenization with Unity Catalog’s governance means organizations can safely unlock clinical value in minutes instead of months.— Sara Livengood, SVP Product Management, Datavant Life Sciences
Many of you are familiar with Delta Sharing, the world’s first and most widely adopted completely open protocol for data sharing. Hundreds of partners like Adobe, Atlassian, Aveva, Circana, FactSet, LSEG, Stripe, The Trade Desk, and Walmart are using it to deliver data to their customers, many in part because they don’t want to be locked into the walled gardens offered by other sharing platforms.
While Delta Sharing pioneered open sharing for structured data, and has since expanded with notebooks, support for Iceberg, and sharing to any endpoint (including Snowflake), we think a new approach is required for agents to participate in cross-party collaboration and monetization.
So, last week, we announced OpenSharing, the industry’s first open-source, vendor-agnostic sharing protocol designed for the AI era.
OpenSharing enables secure, cross-cloud, and cross-platform sharing of both structured data in Delta or Iceberg formats and unstructured data, volumes, agent skills, AI models, semantics, and more. In addition to these new types of shares, we also shipped improvements to Databricks’ managed version of OpenSharing, including SecureConnect for cross-cloud storage to simplify network configuration and Global Distribution to proactively create/cache cross-region replication to reduce egress fees.
OpenSharing is aligned with our LSEG Everywhere strategy, and helps us deliver trusted, AI-ready financial data and AI assets wherever our customers work, regardless of which cloud, tools, or AI model they want to use it in. We're excited to be helping Databricks shape this offering to serve our joint customers worldwide.— Ron Lefferts, Divisional CEO of Data & Analytics at LSEG
While most enterprises are moving to the cloud, the majority of the world’s data still sits in on-premises data centers. Whether it’s to meet data sovereignty requirements, maintain low latency at the edge, or avoid transfer costs, customers with large on-premises storage footprints have been unable to take advantage of modern, cloud-hosted analytics and AI capabilities….. until today.
Last week we also announced the Open Lakehouse Storage Ecosystem, where world-leading storage vendors including Cohesity, Commvault, Everpure, HPE, MinIO, NetApp, Nutanix, Qumulo, Rubrik, and VAST Data have or will soon adopt OpenSharing to allow their customers to quickly and easily zero-copy share their exabytes of data stored on-premises to Databricks.
This ecosystem allows customers to use the same Databricks capabilities across cloud and on-premises storage, without having to move data or make copies. Because, why choose?
VAST was built to provide the performance, scale, and accessibility that the AI era demands. Our work with Databricks via OpenSharing delivers on that vision: the massive data estates organizations have spent years building on-premises don't need to move to the cloud to become AI-ready. This provides our mutual customers a zero-copy, high-throughput path to activate their data where it sits.— Jeff Denworth, Co-founder, VAST Data
Genie Agents is one of our newer, magical capabilities that allow customers to create natural-language chat interfaces informed by their business context on Databricks. For example, I use Genie Agents daily to answer “what partners are driving the most impact?”, “Are customers churning off one partner to another?”, and “What partners are driving anomalistic growth?” in seconds. These are questions that used to take days or weeks for an analyst to run, with follow-up questions taking longer, now right at my fingertips.
Starting this week, you can sprinkle that magic beyond your own account by sharing Genie Agents with your customers via OpenSharing, and you will soon be able to list them on our Marketplace for our 20k+ customers to discover and use.
Partners don’t just share data anymore.
They can share an AI experience over their data.
When sharing via OpenSharing, you control how many questions can be asked and/or rows can be accessed, and your customers have no schema to learn, no UI to build, and don't need access to your underlying tables (unless you allow that as well).
By making Kythera Labs' Genie Spaces available through OpenSharing, we're helping healthcare providers and life sciences organizations get answers from complex healthcare data more quickly and confidently. The real opportunity isn't just making data easier to access. It's shortening the distance between a question and a decision. Teams can explore patient journeys using natural language within Databricks while maintaining the governance, security, and control required for healthcare data.— Jeff McDonald, CEO, Kythera Labs
I've been in/around data licensing my whole career (11 years at Bloomberg and 3 at Dow Jones before 10+ years at AWS and Google). While it's too early to be sure, I think the combination of these capabilities have the potential to reduce sales cycle time, lower customer acquisition costs, and open new commercial models we haven’t even imagined yet.
For example, one of the biggest challenges with licensing data is the pre-sales "stalemate" between data providers and their prospects.
Customers want to know everything that's in a data set they're evaluating so they can match its relevance against their own data, but data providers don't want to give access until they know they'll be compensated for it. Questions like "does this data set cover the securities in my portfolio?" and "are my target customers in your audience?" can be really hard to answer without access to the entire data set, so there's often a prolonged back/forth of samples consisting of partial or stale data sets that extend the sales cycle and frustrate both sides.
Enter “pay-per-question”.
OpenSharing and Marketplace opens the door for a seemingly infinite number of technical and commercial possibilities. Data providers can now share any combination of data sets, notebooks, Genie Agents, and Apps, and, as we add transaction capabilities to our Marketplace, will be able to commercially innovate with "pay-per-question" and/or highly-targeted applications or natural-language interfaces that can seamlessly blend data provider and customer data together. All with the governance you have come to expect from Databricks behind it.
While AI continues to disrupt most segments, I think there's never been a better time to be sitting on proprietary data sets the models don't already have, and we're excited to reinvent the future of data licensing together.
We see tremendous opportunity in agentic AI to fundamentally change how organizations access and apply data. Our collaboration with Databricks reflects a commitment to more flexible, scalable consumption models - making it easier for clients to explore and evaluate our data in real time while accelerating innovation. At the same time, we maintain the highest standards of governance and protection for our proprietary assets.— Pat Dodd, CEO at Cotality
We heard you loud and clear. You wanted better GTM support, technical guidance, and Marketplace-accelerated deal cycles. We've delivered a new partner program with transparent gives/gets, an AI-ready Partner Well-Architected Framework, and many new Marketplace capabilities to help you close deals faster and innovate commercially, all in the past year.
But we're just getting started. Thank you for building with us, challenging us, and making our wall to the future stronger, together, brick by brick.
Keep building,
— Stephen
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。