管理層發言
Good day, everyone. My name is Lenius, and I will be your conference operator today. At this time, I would like to welcome you to HubSpot's Second Quarter 2026 Earnings Call. At this time, I would like to turn the call over to Vice President, Investor Relations, Geoff Koegler. Please go ahead.
Thanks, operator. Good afternoon, and welcome to HubSpot's Second Quarter 2026 Earnings Conference Call. Today, we'll be discussing the results announced in the press release we issued this afternoon. With me on the call this afternoon is Yamini Rangan, our Chief Executive Officer; Dharmesh Shah, our Co-Founder and CTO; and Kate Bueker, our Chief Financial Officer. Before we start, I'd like to draw your attention to the safe harbor statement included in today's press release. During this call, we'll make forward-looking statements within the meaning of the federal securities laws that are subject to risks and uncertainties, including statements regarding our financial guidance for the third fiscal quarter and full year 2026, future financial performance, business outlook and strategy. These statements reflect our views only as of today and, except as required by law, we undertake no obligation to update or revise them.
Please refer to the cautionary language in today's press release, our Form 10-Q and our other SEC filings for a discussion of the risks and uncertainties that could cause actual results to differ materially from expectations. During the course of today's call, we'll refer to certain non-GAAP financial measures as defined by Regulation G. Reconciliations to the most directly comparable GAAP measures can be found in today's press release. Now it's my pleasure to turn the call over to HubSpot's Chief Executive Officer, Yamini Rangan. Yamini?
Thank you, Geoff, and welcome, everyone. I'll start with our Q2 results and what drove them. Then I'll walk through what we learned in the first half, the deliberate choices we are making as a company in response and how we are accelerating our AI transformation across the company. Let's dive in. Q2 revenue grew 17.5% year-over-year in constant currency. We delivered 3 points of non-GAAP operating margin expansion year-over-year, bringing our operating margin to 20.3%. Our total customer count reached over 306,000 globally with 7,000 net additions in the quarter. I'm pleased to announce that our Board of Directors have authorized an additional share repurchase program of up to $1 billion, a clear signal of the confidence we have in our business and the growth opportunity ahead. We are still in the early stages of a massive shift with AI, and this quarter reflected that reality. Let me be direct about what happened.
April got off to a slow start and the quarter we expected did not fully materialize. Two factors drove the headwinds. The first was deliberate. As we discussed last quarter, customers adopting AI want proof of value before they commit and predictability in what it costs. So in April, we leaned into those trends and made changes across product, pricing and go-to-market. On the product side, we introduced trials so customers can turn on agents and AEO in their own environment with their data for their workflows. The goal is simple: let customers experience real outcomes before they buy. We knew this would create some near-term headwinds by extending the buying process, but we believe it's the right long-term trade-off. Lowering the barrier to adoption and building customer confidence and outcomes will ultimately accelerate AI adoption. So far, this approach has worked best with upmarket customers, where our teams and partners provide the support needed to make these trials successful.
Our focus is to scale that experience so every customer can adopt AI with the same confidence. On pricing, predictability has become a defining theme in AI adoption. Businesses have been hit with unpredictable token costs. They want pricing that is transparent and tied to value. In response, we introduced outcome-based pricing for several of our HubSpot agents, lowered entry price points and are providing customers clear visibility and control over usage and spend, including the ability to set thresholds that fit their budget. The second factor was a shift in the demand environment in Q2 with increased budget sensitivity. Businesses want greater confidence that their investments will position them for the AI platform shift. As a result, purchase decisions are facing greater scrutiny, buying committees are larger and more deals require C-suite and Board approval, leading to longer sales cycles.
For existing customers, unpredictable AI costs across the broader landscape are impacting budgets, leading to budget optimization and downgrade pressure. In response, we've evolved our execution playbook by engaging the C-suite earlier, accelerating time to value with our partner ecosystem, and demonstrating our pace of innovation that future-proofs their investments. These trends reinforce our conviction in accelerating our strategic pivot to deliver predictable pricing and clear value tied to AI. Now despite the headwinds in Q2, we continue to see real momentum in AI adoption, upmarket wins and multi-hub growth. On AI, customer adoption is accelerating. Data Agent has over 16,000 customers activated, up 80% quarter-over-quarter. Prospecting Agent has almost 17,000 customers activated, up 28%, and Customer Agent reached over 10,000 customers. We launched HubSpot AEO in April, both within Marketing Hub and as a stand-alone product.
Since then, 32% of Marketing Hub Pro+ customers have activated AEO and nearly 16,000 customers have activated a stand-alone AEO trial in Q2. More importantly, customers are seeing real business outcomes. Sesame HR, a 400-person HR software platform, could only respond to 70% of incoming support tickets before deploying Customer Agent. Now they cover all tickets received with 60% fully resolved without human escalation. After their initial trial, they purchased more than 1 million credits and are expanding into other agents. RevenueWell, a 250-person dental software company, combined HubSpot's buyer intent with Prospecting Agent to identify high-intent buyers and reach them with personalized outreach. Meetings booked increased from 8% to 28% and conversion rates climbed to nearly 10%, well above industry benchmarks. They have since grown their usage to more than 350,000 credits across six HubSpot agents.
Upmarket momentum continued to be strong: deals over $120,000 ARR grew 38% year-over-year, reflecting continued demand from larger customers for a unified AI-powered customer platform. And multi-hub momentum continued: 64% of new Pro+ customers landed with multiple hubs, up three points year-over-year. Taken together, these results reinforce our conviction that we are making the right strategic choices. Now let me step back from the quarter and talk about the bigger opportunity. The last 20 years were about helping teams do more work with software. The next 20 will be about helping them achieve better outcomes with AI. That's the opportunity we're building for. And that's why we are evolving every part of HubSpot from our products to our go-to-market to how we operate as a company. Our AI strategy is simple: drive growth for scaling companies. That means delivering real outcomes across the full customer journey, and that is exactly what our HubSpot agents do.
Customers don't want chaotic agent sprawl. They want a controlled, cohesive way to build demand, win deals and delight their customers. In Q2, we expanded our agent portfolio and added both breadth and depth in agent capabilities. We also added two new products: AEO, which shows marketers how their brand appears in AI search engines and tells them what to do about it, and Revenue Hub, which brings quoting, contracts, billing and payments into one place. Beyond HubSpot agents, we want to make building easy. AI is democratizing the ability to build workflows, agents and automation. We want every go-to-market builder on our platform to take advantage of that. In July, we launched Agent Builder and Agent Hub. Agent Builder lets anyone build custom agents, agentic workflows or mix and match both. What makes it powerful is that it connects deeply to HubSpot CRM and context, but it can also automate anything outside HubSpot.
Custom agents can take a rich set of actions from sending a WhatsApp message to invoking an LLM to calling a custom API, and Agent Hub gives customers one place to manage all their agents: HubSpot-built, customer-built and partner-built. Over 2,700 customers have already activated in beta and partners are leading the way here. SmartBug Media, a HubSpot Elite solutions partner, built a library of 20 custom agents on HubSpot spanning sales research, content creation and proposal compliance. We have expanded partner incentives to encourage this motion and expect a significant portion of customers and partners to build on top of our platform, given the trusted context we deliver. The biggest adoption jump this quarter was in Breeze Assistant. Breeze is no longer just a chat interface. It is how customers interact with the entire HubSpot platform. They use it to build reports, create automations, invoke agents and generate artifacts, all without writing a single line of code.
More than half of our Pro+ customers are using it, and weekly active usage has doubled since the start of the year. What makes it powerful is what's behind it: every action and every artifact is grounded in real CRM data, growth context and HubSpot's governance model. Customers get AI that actually knows their business. So how will you know whether our AI strategy is working? We measure it through four lenses: reach, depth, quality and growth. First, reach: our customers adopting AI. Today, more than 55% of our Pro+ customers use our agents or Breeze Assistant, up by double-digit percentage points since the start of the year. HubSpot agent adoption among our Pro+ customers has grown from high single digits to mid-teens this year. Second, depth: our customers using more over time. The total number of monthly agentic actions across our customer base has increased more than 3x this year. Customers are not just experimenting with AI; they are making it part of their daily workflows.
Third, quality: are our agents delivering real outcomes? Customer Agent now resolves 72% of support tickets without human escalation. Prospecting Agent is generating response rates on par with human-written outreach, leading to more meetings booked and more deals closed. Finally, growth: is AI driving growth of our business? Total credit consumption grew in Q2 despite the pricing changes we made in April. Credit usage is now evenly distributed across Data Agent, Prospecting Agent, Customer Agent and Buyer Intent. That is an important signal. Customers aren't adopting a single AI use case. They're using HubSpot agents across the entire customer journey. Transforming our product and pricing is only part of the story. We're also transforming how HubSpot operates. And that internal transformation is driving leverage. We have reorganized into smaller, more focused teams operating in six-week sprints.
This lets us move at the speed of AI, experiment continuously and quickly scale what works. AI is now embedded in how we build, how we sell and how we run the business. We're doing more with less and getting faster as we go, and it is showing up in the numbers. Operating margins are expanding in 2026 even as we invest aggressively in AI innovation. In addition, we expect to deliver two to three points of operating margin expansion in 2027, a meaningful step-up that reflects the operating leverage we are building as an AI-first company. I want to close with what I keep coming back to. We are in the middle of a real transition to AI, and we are making deliberate choices to lead in it. While some of these choices create near-term headwinds, they will help us drive long-term compounding growth. Our core fundamentals are solid. Our AI momentum is real and accelerating. And the adoption indicators are moving in the right direction.
We have a clear strategy, a focused team and high conviction in where we are headed. With that, I'll hand it over to our CFO, Kate Bueker, to walk you through our financial and operating results. Kate?
Thanks, Yamini. Before diving into Q2 results, I want to reiterate the business trends that Yamini shared in her remarks. April got off to a slow start, and the quarter that we anticipated did not fully materialize. We faced two headwinds this quarter. First, the deliberate changes we made across product, pricing and go-to-market were a headwind. We believe these are the right long-term decisions for HubSpot to win in the AI era. Second, we saw a shift in the demand environment in Q2 with increased budget sensitivity. These demand trends reinforce our conviction in accelerating our strategic choices to better assist customers through this transition. With that, let's turn to our second quarter 2026 financial results. Q2 revenue grew 20% year-over-year as reported and 17% in constant currency. Q2 subscription revenue grew 20% year-over-year, while services and other revenue increased by 8%, both on an as-reported basis.
Domestic revenue grew 17% year-over-year in Q2. International revenue growth was 23% as reported and 18% in constant currency, representing 49% of total revenue. We added 7,000 net new customers in Q2, bringing our total customer count to over 306,000, growing 14% year-over-year. This was below our 9,000 to 10,000 expectation and driven primarily by weaker conversion rates and increased buyer hesitancy. Average subscription revenue per customer was $11,800 in Q2, up four points year-over-year as reported and two points in constant currency. Customer dollar retention remained healthy in the high 80s, while net revenue retention was 102%, down one point year-over-year as continued seat and credit expansion was offset by other net upgrade headwinds as a result of customer budget optimization. Q2 calculated billings were $930 million, growing 14% year-over-year as reported and 17% in constant currency.
Non-GAAP operating margin was 20%, up three points compared to the year-ago period. This expansion reflects our continued disciplined approach to head count spend, partially offset by AI costs. GAAP operating margin was 5% in Q2 compared to a negative operating margin of 3% in the year-ago period. This eight-point expansion reflects our non-GAAP operating income expansion and a more than four-point reduction in stock-based compensation expense as a percentage of revenue. Non-GAAP net income was $165 million, and non-GAAP net income per diluted share was $3.26, up 40% and 49% year-over-year, respectively. GAAP net income was $43 million in Q2, and GAAP net income per diluted share was $0.86. In the second quarter, the company generated $168 million of free cash flow or 18% of revenue. Our cash and marketable securities totaled $1.4 billion at the end of June. During the quarter, we bought back more than $500 million of stock under our current $1 billion share repurchase program.
Our Board of Directors has authorized an additional share repurchase program of up to $1 billion, reflecting the confidence we have in our business and the growth opportunity ahead. Our continued strong balance sheet and free cash flow provide us with the flexibility to return capital to shareholders while maintaining our focus on investing in organic innovation and opportunistic M&A. Before we dive into guidance, let me share our current thinking on the second half and our expectations on the near-term performance of our KPIs. Our Q3 and full year guidance reflects our expectation that the headwinds we saw in Q2 will persist throughout the remainder of the year. We now expect quarterly net additions to be approximately 5,000 to 6,000, and ASPRC growth to be in the low to mid-single digits in constant currency. For the full year of 2026, we expect net revenue retention to be roughly flat year-over-year.
We expect customer dollar retention will remain strong and stable in the high 80s, and we continue to expect net upgrade rates to be pressured in the second half as a result of customer budget optimization. We now expect net new ARR growth to be below constant currency revenue growth for the fiscal year of 2026. With that, let's dive into our guidance for the third quarter and full year of 2026. For the third quarter, total-as-reported revenue is expected to be in the range of $924 million to $925 million, up 14% year-over-year on an as-reported basis and 15% in constant currency. Non-GAAP operating income is expected to be between $187 million and $188 million, representing a 20% margin. Non-GAAP diluted net income per share is expected to be between $3.25 and $3.27. This assumes 49.3 million fully diluted shares outstanding. And for the full year of 2026, total-as-reported revenue is now expected to be in the range of $3.678 billion to $3.686 billion, up 18% year-over-year on an as-reported basis and 16% in constant currency.
We continue to expect non-GAAP operating income to be in the range of $762 million to $766 million, representing a 21% margin. Non-GAAP diluted net income per share is now expected to be between $13.23 and $13.31. This assumes 50 million fully diluted shares outstanding. Before we turn to some modeling notes, I'd like to provide additional color on our margin expansion trajectory. As we transform how HubSpot operates, our internal transformation is driving leverage. Our guidance contemplates two points of non-GAAP operating margin leverage this year. Looking ahead, we expect to deliver two to three points of incremental non-GAAP operating margin leverage in 2027. In addition, we remain committed to driving stock-based compensation as a percentage of revenue down year-over-year. We'll share more detail on this at Analyst Day next month. As you adjust your models, please keep in mind the following.
We continue to expect CapEx as a percentage of revenue to be 5% to 6% for the full year of 2026 and continue to expect free cash flow to be about $750 million. Before we open the call for questions, I want to invite you to join us at our Annual Analyst Day at Unbound taking place on September 17 in Boston. We look forward to seeing you there. With that, I will turn the call back over to the operator for questions.
分析師問答
Your first question comes from the line of Samad Samana with Jefferies.
Yamini, I appreciate all the details you gave us. In the press release you referenced the deliberate choices, and we understand the pricing model change and the sales training. There was also the terms of service change during that period and the subsequent rollback. As I'm processing Kate's guidance and the information you provided, were there additional choices this quarter that disrupted results and influenced the outlook beyond what we've already discussed? Looking ahead, are there more aggressive moves the company needs to make in this fast-moving AI environment? I appreciate the progress you've made, but is there more work to do and more decisive action required?
Samad, thank you for that question. The simplest way to describe what we observed in the quarter is that as customers are navigating both a platform shift and budget pressure they're cautious. I mentioned two specific headwinds, but let me be very specific and unpack what those meant. When I look at the first half, Q1 was solid. But as we moved through Q1, it became very clear that customers are evaluating and buying AI differently. So in April, we leaned in and made a set of deliberate changes across product, pricing and go-to-market that you referenced. Specifically, customers want proof of value before they buy. This is a different buying motion than traditional SaaS where they bought features, and so they want to see the agents work within their data, within their workflows and their environment. We shifted to providing trials for several of our HubSpot agents and AEO. We knew evaluations would slow deal cycles, but would provide higher confidence for customers as they adopt AI.
That would allow us to surface as many use cases as early as possible because once they buy the first agent and the second agent and begin to see clear outcomes, they are much more progressive about adopting the third, fourth, fifth agent. That is exactly why we did that. The second thing we did is reflect on customers wanting predictable AI costs. That's an industry-wide concern. So we lowered the price for many of our key agents. We introduced outcome-based pricing so we could tie the agent value to the value they're getting and we gave customers much better visibility and spend control. They can set thresholds for what they want to spend. These decisions create a near-term headwind, and the decision we made is we'd rather remove friction and build customer confidence at the beginning of the AI journey and then help them drive much more adoption as they continue. We believe those are the right trade-offs to becoming the long-term winner.
We are seeing encouraging signals in terms of AI adoption; July continued that adoption. Specifically, Prospecting Agent is now at 17,000 customers, Customer Agent at 18,000 customers growing significantly, Data Agent similarly at 16,000, and buyer intent adoption is strong. Customers are getting more comfortable, and it's balanced across multiple agents in terms of what they're adopting. Those are all the deliberate changes we made, and we feel good because, yes, we take a short-term step back, but we are doing that so we can take multiple steps forward. The second trend is that in Q2 we saw a shift in the demand environment with increased budget sensitivity. For prospects we're seeing greater scrutiny, larger buying committees, and more deals requiring C-suite or Board approval, which lengthens deal cycles. We are seeing a lot of large opportunities in the pipeline, larger than before, and they are closing, but often a month or a few weeks later.
With that change, we are engaging with customers earlier, providing clarity on outcomes, rolling out plays to support flexibility in spend, and working closely with partners to give confidence and deliver outcomes. Those are the two specific trends we saw and how we are responding within the quarter. The factors I talked about resulted in slower customer acquisition and net new ARR headwinds, and that is what we are contemplating in guidance for the rest of the year, but we're confident we're making the right trade-offs to position the business for the much larger AI opportunity in front of us, and that is automation.
Our next question comes from Rishi Jaluria with RBC.
Look, I appreciate all the detail. I think a lot of us are willing to be patient and long-term focused. But I'd love to understand in terms of what you're seeing from a buying behavior. I totally understand some of the business model shifts and a lot of the noise out there. You talked a little bit at the beginning about how a lot of companies are struggling to manage their token bills and we obviously saw the trend of token maxing basically be DOA. So my question is, has that kind of token burn impacted budgets and how companies both new and existing are thinking about deploying HubSpot? And maybe related to that, is there any confusion or noise in the market from adjacent vendors that are relatively new to marketing themselves as more AI-native that's leading to that higher level of scrutiny you talked about? Any color there would be helpful.
Yes, Rishi, thank you. There are two parts to your question. One is what we're seeing in the spending environment and whether that's because of token maxing. The second is around competitive positioning. On spending, in Q2 we saw a more cautious environment. There will be a sorting phase, and there will be clarity in terms of not token maxing but value-maxing, which is why we are leaning into delivering outcomes. For new customers, we're seeing larger buying committees and more C-suite and board involvement. That wasn't as common a few quarters ago. We are seeing more pipeline and larger deals, but because of the scrutiny they take longer. For example, replacing an incumbent for a 500-plus seat deal recently took two more weeks than normal to go through CEO and board approval. Customers navigating a platform shift want clarity on outcomes and predictability in cost. Our execution now meets them where they are: engaging C-suite earlier, making value clear, accelerating migrations and implementations with partners to deliver fast outcomes.
On competition, the market remains competitive and has always been so. It's not winner-take-all. Where we continue to win is because we are easy to use and provide a platform. Customers do not want agent sprawl—ten agents from ten vendors increases unpredictability of cost and reduces the ability to see outcomes in one place. They look to us to have agents in a single place with CRM context, governance and workflows. Our overall market position is solid; we see that in upmarket win rates and the strength of conversations, but deals are taking longer because of the environment.
Your next question comes from the line of Brian Peterson with Raymond James.
I appreciate all the detail. Yamini, I'd love to understand the demand environment maybe from a customer size perspective. I know you're talking about CEO involvement and bigger committees. I think we normally associate that with enterprise. I think you're saying that maybe that is a stronger area of the market. I just want to make sure I understand that. How can we think about large customers versus SMB adoption through Q2?
Brian, thanks. Let me unpack what we are seeing upmarket and downmarket. In the downmarket, we saw a more cautious buying environment, leading to higher scrutiny in decision-making. Our priority is to make it super easy for downmarket businesses to adopt hubs, adopt agents and deliver measurable outcomes. The actions we took in Q2—trials and predictable pricing—will help downmarket adoption. Upmarket is where we saw the larger buying committees and more approvals needed. The biggest change upmarket is the number of people involved in the buying process, which elongates evaluation time. That said, upmarket continues to be strong: deals over $120,000 ARR grew 38%, indicating increasing interest. When upmarket customers talk to us, they do so because we can consolidate their platform, reduce TCO and provide clarity on the AI roadmap. So the clear distinction is the number of people involved in the process, which lengthens cycles. Downmarket faces higher scrutiny, where trials and cost predictability are helping. We are leaning in to meet customers where they are; these are the right long-term decisions to position HubSpot to win in the AI era.
Your next question comes from the line of Raimo Lenschow with Barclays.
This is Eamon Coughlin on for Raimo. It's great to see that credit consumption continues to grow healthily despite April pricing changes. And I recognize that it's very early, but is there any way to think about how much of total credit consumption today is coming from paid usage versus bundled or trial usage?
Thank you. The way we look at credit consumption starts with breadth of AI usage — how many customers are using AI features — then depth of usage, and then the number of outcomes they drive which leads to credit consumption including both included and additional consumption. We saw positive leading indicators: reach, depth and quality. Reach: more than 55% of our Pro+ customers are using Breeze Assistant and agents. Agent usage moved from high single digits to mid-teens this year and accelerated into the high teens in July. Breeze Assistant usage jumped significantly; it's how customers interact with HubSpot, take agent actions, build reports and artifacts. Depth: agentic actions taken grew 3x from the beginning of this year, showing it's not just experimental. Quality: Customer Agent resolves 72% of support tickets without human escalation and Prospecting Agent is delivering results comparable to human outreach. Those are leading indicators for credit consumption. Despite reduced pricing, credit consumption increased in Q2 and that trend continued into July. We're very early in the adoption cycle, especially for SMBs; as we give them confidence in predictability of cost and quality of output, consumption will grow.
Your next question comes from the line of Terry Tillman with Truist.
Giancarlo on for Terry here. Just on the product roadmap, you had mentioned that there's a lot of progress moving forward. How do you actually train the sales force to sell those new products?
That's a good question. Stepping back on product strategy: we start with HubSpot agents, leveraging deep domain expertise in marketing, sales and service to deliver outcomes. Agent Builder is a major part of the strategy: it enables customers and partners to build workflows, agents and automation on top of HubSpot because we provide context, CRM data, permissions and governance that make it practical. In terms of training, our teams use HubSpot every day and understand the capabilities firsthand. We're training them to lead with outcomes and the value of the full agentic platform, and to explain AI pricing and credits. Customers want to understand how this works economically, and that's a new sales motion. The bigger change is how customers buy: AI adoption requires customers to see outcomes in their environment and understand ongoing economics before committing. Our sellers are adjusting to this buy motion and helping customers navigate the change.
Your next question comes from the line of Gabriela Borges with Goldman Sachs.
Yamini, all of your commentary makes sense. My question for you and Kate: I know you said you're anticipating the environment to stay the same in the second half. When does it get better? I would imagine there's a period where customers are exploring and the products are becoming more mature. Do you envision getting to a more steady-state environment? Or is this the new normal, like what we were talking about post-COVID, where we kept waiting for things to get better and it just turned out this is the new baseline? What's your hypothesis for the timeline for customers to get more comfortable and for the products to mature so the buying process becomes more comfortable for everyone?
Great question. We're in a moment where customers are figuring out how to adopt AI and how to budget for it. The changes we made are intended to help customers through this environment: trials, solution architects, partners to validate outcomes in customers' environments, and predictable cost models including outcome-based pricing and thresholds for spend. Historically, transitions like this start with experimentation and then move to broader adoption as customers see peers getting value. For SMBs, adoption is in early stages; as we give them predictability and clear outcomes, comfort will increase. We are assuming Q2 trends persist into the second half, and that's reflected in guidance, but we're doing everything possible to provide comfort for customers as they navigate this transition.
Your next question comes from the line of Alex Zukin with Wolfe Research.
This is Ivan here for Alex. Maybe one question on the product side since HubSpot was always known as the company with beloved products. This one is about the agents. What do Customer Agent, Prospecting Agent, AEO and these other agentic products need to get right over the next two quarters for the H2 numbers to land? Where are these products still maturing today? Our sense from the field is that most AI products are still very early. From your perspective, what closes the gap? And when do these products start impacting revenue and offsetting other weaknesses?
Great question. SMB adoption is still early. Our strategy starts with HubSpot agents across the customer journey to deliver outcomes: build demand, win deals and delight customers. Agents we have include AEO to show up in LLMs, Data Agent to build audiences or TAM, Prospecting Agent for outreach, smart deal progression for self-updating CRM, and Customer Agent. The adoption pattern: customers start with internal-facing agents like Data Agent and smart deal progression because they are comfortable with internal use cases. Agents interacting directly with customers — Prospecting Agent and Customer Agent — need additional validation of outcomes. We measure discovery, activation, the first 'aha' moment and ongoing usage for each agent and optimize those steps. That is how we control adoption and maturity. Agent Builder is the second part of our strategy, enabling customers and partners to extend HubSpot with custom agents and workflows tied to CRM context and governance; we already have 2,700 customers and 1,000 agents adopted in beta. From a product perspective, we are focused on driving that first moment of value, reducing friction, and delivering clear outcomes. This is a multi-step adoption process and we're optimizing each point of friction to drive broader adoption over time.
Your next question comes from the line of Jackson Ader with KeyBanc.
I have one around the budget scrutiny. Customers are scrutinizing budgets more — what are they opting to spend their money on instead of HubSpot at the moment? What gives you confidence this is temporary? How long might this temporary last?
Jackson, I don't have a crystal ball on exact timing. In the first half, we saw a lot of token maxing, which is unhealthy because it's not tied to value. There will be a sorting phase where organizations focus on value-maxing. We're leaning into clear outcomes, outcome-based pricing and predictability to give customers confidence. In early technology adoption there is experimentation; customers will settle on what drives clear impact and ROI. HubSpot will win because our platform connects directly to business outcomes and we can pick and choose among AI models to provide cost-efficient ways to deliver outcomes. I can't give an exact month or quarter, but we're being proactive in communicating value and predictability to help customers through the shift.
Your next question comes from the line of Siti Panigrahi with Mizuho.
Yamini, appreciate the color. Among customers who are upticking AI products, are they asking for more discounts on their traditional core hub? How are they funding that spending? And on NRR, what gives you confidence that NRR will be flat? Are you seeing churn among customers who are not upticking AI?
Siti, why don't I start. I'll give Yamini a chance to add more. On net revenue retention, think about its components. It starts with customer dollar retention — in Q2 that remained strong and stable in the high 80s. Seat and credit expansion, which has been a key driver of NRR, continued to benefit NRR this quarter. But the agent pricing reductions and the introduction of trials created a near-term headwind for credit expansion. Outside of seats and credits, we did see pressure in other upgrade motions; as customers come up for renewal we're seeing some downgrade pressure as they optimize overall spend. Given those trends, we took down our full-year view for net revenue retention to be basically flat for 2026. Year-to-date we are flat. NRR tends to be higher in Q3 and Q4 with UNBOUND and Q4 is a big quarter for new business, renewals and upmarket activity. That said, the actions we're taking — lowering friction to get started with AI use cases — position us well to drive stronger NRR over time as those customers expand.
Siti, on how adoption is funded: there are a couple of patterns. Some customers start with a trial to validate an agent outcome in their environment, and then they expand. For example, RentSpree started with Customer Agent, validated that 70% of conversations could be resolved, and then expanded with an additional 350,000 AI credits. Another pattern is customers swapping seats for credits to manage their budgets while driving AI adoption. We provide flexibility so they can control budget while adopting AI. Trials provide confidence that lead to expansion; flexibility in spend gives them control. Both motions are what we're leaning into.
Your final question for today comes from Keith Bachman with BMO. Keith, your line is open. You can ask your question.
Operator, we'll take the next question, please.
Your next question comes from Parker Lane with Stifel.
Kate, one for you on the two pieces that impacted the quarter here: the deliberate actions you took and then the shift in the demand environment. Was the demand environment fairly level and the pressures consistent throughout the quarter? Or did that become more pronounced as you exited the quarter?
I think there's not a distinct difference between pressure at the beginning of the quarter and pressure at the end of the quarter. And just to be clear, we did see it continue into July.
Thank you. This concludes today's call. You may now disconnect.