Data Portal is an important new addition to our stream governance suite that brings our vision for discoverable and reusable data streams to the forefront of Conclude Cloud. The excitement around Flink at Current was palpable. Flink sessions were among the highest rated and most attended sessions at the entire conference, highlighting the hunger Kafka users have for Flink. That’s why we’re so pleased with the launch of Flink public preview in Confluent Cloud. Since the announcement, we’ve seen incredible uptake with hundreds of customers opting into the preview and trying out our cloud-native and serverless Flink offering. The addition of Flink strengthens our position as the only complete data streaming product. We bring Flink together with the connectors that capture streaming data, the stream itself in Kafka and the governance capabilities to manage streaming data across an organization.
Each of these capabilities strengthens the other, and the combination comprise what we believe will be the most important data platform in a modern company. And finally, we introduced data streaming for AI, a set of new partnerships that span vector databases, CSPs and SIs. It also includes new product capabilities, including the Confluent AI assistant that will turn natural language inputs into helpful suggestions and code. We expect this initiative to address the demand we’re seeing across our customers who are building innovative new AI applications. A great example of this is Notion. Notion is an AI-powered connected workspace where modern teams can create and share documents, take notes, manage projects and organize knowledge all in one place.
Given Notion’s strong growth, a new data streaming platform was needed to meet its rapidly expanding needs. However, due to the lean engineering team, they required a fully managed service to focus on product development rather than managing Kafka. They began using Confluent Cloud for real-time data flows for internal analytics pipelines and to supply data into data lakes. After realizing the value it provided, Notion expanded the usage of Confluent Cloud to enhance product features, including search automation and Notion AI. Now Notion’s customers can automate tasks in real time, such as adding summaries, extracting key points and consolidating action items in meeting notes. This is only the beginning as Notion continues to innovate and actively explore how data streaming can power new AI applications.
Before turning things over to Rohan, I wanted to reiterate a couple of key points. We’re incredibly excited about the tailwinds to the business, Flink, data governance, AI and the rest of the data streaming platform components add to our business. And while it may cause some short-term headwinds, I firmly believe our accelerated transformation to a fully consumption-oriented business will put us in an incredibly strong position to drive the monetization of these new offerings. I’ve never been more confident in our ability to be the leader in the emerging $60 billion data streaming market. With that, I’ll turn things over to Rohan.
Rohan Sivaram: Thanks, Jay. Q3 demonstrates our ability to drive efficient growth in a challenging macroeconomic environment. Key highlights include robust subscription revenue growth, coupled with proactive management of rate and pace of investments which drove record higher gross margin, 20-plus point of operating margin improvements and our first positive non-GAAP EPS quarter. I’d like to take a moment to thank our employees and partners for their contributions in delivering these key milestones for the Company. Turning to Q3 results. Total revenue grew 32% to $200.2 million, exceeding our guidance. Subscription revenue grew faster, up 36% to $189.3 million. Within subscription, Confluent Platform revenue growth reaccelerated to 19%, ending the quarter at $97.7 million or 49% of total revenue and was driven by continued strength in regulated industries as companies in those industries are still early in their journey of moving to the cloud.
Confluent Cloud revenue grew 61% to $91.6 million, slightly below our guidance of $92.2 million and ended the quarter at 46% of revenue compared to 38% of revenue a year ago and 44% last quarter. As Jay called out earlier, cloud revenue performance in the quarter was modestly impacted by two large customers: One large online gaming customer who moved workloads back to their own data center from the public cloud as part of our broad internal cloud strategy review; and one of our largest customers who ramped slower than expected as they are currently in the process of being acquired. Additionally, we saw lower-than-expected consumption in a few other large U.S. digital native customers. While these select customers did have an impact on Q3, especially in the last few weeks of the quarter, we expect to see a greater impact throughout Q4 and next year.
I’ll cover more on this in a second. Turning to the geographical mix of revenue. Revenue from the U.S. grew 25% to $119.4 million. Revenue from outside the U.S. grew 43% to $80.8 million. We continue to be pleased with the relative balance of our business around the world. Moving on to rest of the income statement, I’ll be referring to non-GAAP results unless otherwise stated. Total gross margins reached a record high of 76.4%, up 540 basis points. This translates to 80.1% subscription gross margin, up 320 basis points, despite a continued revenue mix shift to cloud. Gross margin outperformance was driven by continued improvement in the unit economics and in product optimizations of our cloud business and our R&D investments in Kora, the cloud-native engine of our Confluent Cloud offering.
Turning to profitability and cash flow. Operating margin improved 22 percentage points to negative 5.5%, representing our fifth consecutive quarter of more than 10 points in improvement. Q3 operating margin outperformance was driven by disciplined spending and ROI-focused investment philosophy across every function of the Company. And we are pleased to achieve our first positive net income per share of $0.02 for Q3, using 303.9 million basic and 347 million diluted weighted average shares outstanding. Fully diluted share count under the treasury stock method was approximately 355.4 million. Free cash flow margin improved 24 percentage points to negative 6.5%. We ended the third quarter with $1.87 billion in cash, cash equivalents and marketable securities.
Turning now to other business metrics. In Q3, we added 80 net new customers, 41 customers with $100,000 or more in ARR and 8 customers with $1 million or more in ARR, bringing our total customer count to approximately 4,910, up 16%. Our $100,000-plus customer count to 1,185, up 25% and our $1 million-plus customer count to 155, up 38%. NRR in the quarter was healthy, just under 130%. NRR for hybrid and cloud were both comfortably above 130% with cloud NRR continuing to be the highest. Gross retention rate remained strong and was above 90%. Consistent with Jay’s comments about the drivers behind the shift to consumption, RPO was $824.1 million, up 24%. Current RPO, estimated to be 65% of RPO was $535.1 million, up 31%. RPO growth was impacted by a decline in both average deal sizes and average contract duration, a continuation of the trends we had called out last quarter.
In the current macro environment, we believe customers have less appetite for larger upfront commitments and prefer consumption against smaller commits, reflecting a consumption-first trend in how our customers derive value from Confluent. Now, turning to our Q4 outlook. We have factored in our guidance the impact of the following: First, we expect to have a full quarter impact from the two large customers mentioned earlier, whose consumption is impacted by their company-specific events around the shift of workloads back to on-prem and an acquisition. We estimate that these two customers account for roughly 50% of the expected consumption shortfall in Q4. Second, the macro uncertainty, including the ongoing geopolitical tensions will likely persist, impacting new use case deployments into production and driving lower-than-expected consumption.
