การประชุมแถลงผลประกอบการไตรมาส 2 ปีงบประมาณ 2026 ของ Similarweb (SMWB): รายได้จาก AI แตะ 13%, ปรับเพิ่มแนวโน้มผลประกอบการ

ประเด็นสำคัญ

• รายได้ประจำไตรมาส 2 ปีงบการเงิน 2026 เพิ่มขึ้น 9% เมื่อเทียบรายปี สู่ระดับ 77.2 ล้านดอลลาร์ ซึ่งสูงกว่าช่วงประมาณการของบริษัท
• รายได้ที่เกี่ยวข้องกับ AI คิดเป็น 13% ของรายได้ทั้งหมด เพิ่มขึ้นจาก 11% ณ สิ้นไตรมาส 4 ปีงบการเงิน 2025
• ซิเมิลาร์เว็บ (Similarweb) มีกำไรจากการดำเนินงานตามมาตรฐาน GAAP เป็นบวกได้เป็นครั้งแรก ขณะที่กำไรจากการดำเนินงานแบบ Non-GAAP เพิ่มขึ้นสู่ระดับ 6.5 ล้านดอลลาร์ คิดเป็นอัตรากำไร 8%
• บริษัทได้ลงนามในสัญญาองค์กรแบบหลายปีจำนวน 3 ฉบับ ซึ่งมีมูลค่าสัญญารวมมากกว่า 60 ล้านดอลลาร์ แต่ละฉบับมีมูลค่าในระดับ 8 หลัก และมีเพียงฉบับเดียวเท่านั้นที่เกี่ยวข้องกับการฝึกฝนโมเดลภาษาขนาดใหญ่ (LLM)
• อัตราการรักษาฐานรายได้สุทธิ (NRR) ปรับตัวดีขึ้นสู่ระดับ 100% สำหรับลูกค้าทั้งหมด และอยู่ที่ 107% สำหรับกลุ่มลูกค้าที่สร้างรายได้ประจำต่อปี (ARR) มากกว่า 100,000 ดอลลาร์
• ผู้บริหารได้ปรับเพิ่มประมาณการสำหรับตลอดปีงบการเงิน 2026 โดยคาดว่าจะมีรายได้อยู่ที่ 340 ล้านดอลลาร์ถึง 380 ล้านดอลลาร์ และกำไรจากการดำเนินงานแบบ Non-GAAP อยู่ที่ 24 ล้านดอลลาร์ถึง 26 ล้านดอลลาร์

ข้อมูลทางการเงินที่สำคัญ

ผลการดำเนินงานทางธุรกิจและฝ่ายปฏิบัติการ

ซิเมิลาร์เว็บระบุว่าการปรับเปลี่ยนทิศทางไปสู่กลุ่มองค์กรขนาดใหญ่ขึ้นและสัญญาที่เน้นการใช้ข้อมูลกำลังได้รับแรงส่งที่ดีขึ้น ลูกค้าที่สร้างรายได้ประจำต่อปี (ARR) มากกว่า 100,000 ดอลลาร์ คิดเป็น 69% ของ ARR ทั้งหมด เพิ่มขึ้นจาก 63% ในช่วงเดียวกันของปีก่อน ขณะที่สัดส่วนของ ARR ที่อยู่ภายใต้สัญญาข้อตกลงแบบหลายปีเพิ่มขึ้นเป็น 66% จาก 57%

บริษัทปิดท้ายไตรมาสด้วยจำนวนลูกค้าที่สร้าง ARR มากกว่า 100,000 ดอลลาร์รวม 473 ราย เพิ่มขึ้น 9% เมื่อเทียบรายปี โดยมูลค่าบัญชีเฉลี่ยสำหรับลูกค้ากลุ่มนี้เพิ่มขึ้น 18% สู่ระดับ 438,000 ดอลลาร์

ลูกค้าที่สร้าง ARR มากกว่า 25,000 ดอลลาร์ มีจำนวนรวม 1,815 ราย เทียบกับ 1,809 รายในช่วงเดียวกันของปีก่อน โดยมูลค่าบัญชีเฉลี่ยของลูกค้ากลุ่มนี้เพิ่มขึ้น 19% สู่ระดับ 149,000 ดอลลาร์ ทั้งนี้ ฝ่ายบริหารระบุว่าการเติบโตของจำนวนลูกค้าที่มีอยู่อย่างจำกัดนั้น ส่วนหนึ่งเป็นผลจากการให้ความสำคัญกับการขยายธุรกิจภายในกลุ่มลูกค้ารายใหญ่ที่มีอยู่เดิม มากกว่าการไขว่คว้าโอกาสจากลูกค้ารายย่อย (SMB) ขาเข้าที่มีขนาดเล็กกว่า

สัญญารายใหญ่ 3 ฉบับที่ลงนามในระหว่างไตรมาส 2 มีมูลค่าสัญญารวมกันมากกว่า 60 ล้านดอลลาร์ โดยหนึ่งในนั้นเป็นการทำสัญญากับบริษัทเทคโนโลยีชั้นนำระดับโลกที่ใช้ข้อมูลของซิเมิลาร์เว็บสำหรับการฝึกฝนโมเดลภาษาขนาดใหญ่ (LLM) ทั้งก่อนและหลังการฝึกฝน (Pretraining และ Post-training) ส่งผลให้บริษัทดังกล่าวกลายเป็นลูกค้ารายที่สามของบริษัทที่มี ARR มากกว่า 10 ล้านดอลลาร์ ส่วนอีกสองฉบับที่เหลือเป็นองค์กรที่นำข้อมูลของซิเมิลาร์เว็บไปใช้กับการประยุกต์ใช้งานในวงกว้างขึ้น แทนที่จะนำไปใช้ฝึกฝน LLM นอกจากนี้ ซิเมิลาร์เว็บบริษัท ยังได้ลงนามในสัญญารายใหญ่อีกฉบับหนึ่งในเดือนกรกฎาคม

ฝ่ายบริหารอธิบายว่า AI เป็นตัวขับเคลื่อนที่ทำให้เกิดการใช้ข้อมูลเพิ่มสูงขึ้น โดยองค์กรต่าง ๆ สามารถนำข้อมูลดิจิทัลของซิเมิลาร์เว็บไปปรับใช้ในทีมงาน แอปพลิเคชัน และกรณีการใช้งานอัตโนมัติที่หลากหลายมากขึ้น ซึ่งช่วยเพิ่มศักยภาพในการขยายการดำเนินงานกับลูกค้ารายเดิม สำหรับโครงสร้างราคาไม่ได้มีการเปลี่ยนแปลงอย่างมีนัยสำคัญ แต่การเติบโตนั้นมาจากสิทธิ์การเข้าถึงชุดข้อมูลที่เพิ่มขึ้นและปริมาณการใช้งานที่สูงขึ้น

กลยุทธ์ด้าน AI ของบริษัทประกอบด้วยการจัดหาข้อมูลให้กับ LLM และเอเจนต์ AI การพัฒนาผลิตภัณฑ์แบบ AI-native และการจัดจำหน่ายข้อมูลผ่านแพลตฟอร์ม AI ของบุคคลที่สาม ทั้งนี้ Similarweb AI Studio ช่วยให้ผู้ใช้สามารถถามคำถามทางธุรกิจด้วยภาษาธรรมชาติได้ และรองรับโมเดลการสร้างรายได้ตามปริมาณการใช้งาน นอกจากนี้ บริษัทยังอ้างถึงการจัดจำหน่ายผ่าน MCP บน Claude และ ChatGPT ควบคู่ไปกับการขยายความสัมพันธ์กับ Perplexity และ Manus

ประมาณการจากฝ่ายบริหาร

ฝ่ายบริหารกล่าวว่าค่ากลางของประมาณการรายได้ตลอดปีคิดเป็นการเติบโตประมาณ 12% เมื่อเทียบรายปี โดยคาดว่าการเติบโตของรายได้จะเร่งตัวขึ้นในช่วงครึ่งปีหลัง ซึ่งได้รับแรงหนุนจากการเติบโตของ ARR ในไตรมาส 2 และแรงส่งทางธุรกิจที่แข็งแกร่งยิ่งขึ้น

นอกจากนี้ บริษัทยังคาดว่าจะมีกระแสเงินสดอิสระปรับนอร์มาไลซ์เป็นบวกในแต่ละไตรมาสหลังจากนี้ ส่วนค่าใช้จ่ายทางการเงินแบบ Non-GAAP และค่าใช้จ่ายภาษี คาดว่าจะยังคงอยู่ที่ระดับใกล้เคียงกับไตรมาส 2 ในช่วงที่เหลือของปี

ปัจจัยเสี่ยงและประเด็นที่ต้องจับตา

• ค่าเงินเชเคิลอิสราเอลที่แข็งค่าขึ้นยังคงเป็นปัจจัยกดดันต่อความสามารถในการทำกำไร เนื่องจากพนักงานประมาณครึ่งหนึ่งของซิเมิลาร์เว็บประจำอยู่ในอิสราเอล อย่างไรก็ตาม ฝ่ายบริหารระบุว่าวินัยในการควบคุมต้นทุนช่วยชดเชยแรงกดดันนี้ได้ในไตรมาส 2
• โอกาสทางธุรกิจขนาดใหญ่ด้าน AI รายที่สองที่เคยกล่าวถึงในไตรมาสก่อน ๆ ยังคงอยู่ในท่อส่งดีล (Pipeline) ซึ่งฝ่ายบริหารกล่าวว่าอาจสามารถปิดดีลได้ภายในสิ้นปี แต่กรอบเวลายังคงมีความไม่แน่นอน
• ฝ่ายบริหารยอมรับว่าจำเป็นต้องดำเนินการเพิ่มเติมเพื่อขยายความสัมพันธ์กับลูกค้าองค์กร และพัฒนาโครงสร้างพื้นฐานข้อมูลดิจิทัลอย่างต่อเนื่องสำหรับการใช้งานด้าน AI
• จำนวนผู้เข้าชมเว็บไซต์ลดลงเนื่องจากพฤติกรรมของผู้ใช้งานที่เปลี่ยนไป แม้ว่าฝ่ายบริหารจะไม่ได้ระบุถึงแรงกดดันที่ผิดปกติในกลุ่มลูกค้ารายย่อยก็ตาม

ประเด็นสำคัญจากการตอบคำถามนักวิเคราะห์

นักวิเคราะห์ให้ความสนใจในประเด็นที่ว่า AI จะสามารถขยายการใช้ข้อมูลภายในองค์กรขนาดใหญ่ได้อย่างมีนัยสำคัญหรือไม่ โดยฝ่ายบริหารกล่าวว่าองค์กรระดับก้าวหน้าสามารถประมวลผลข้อมูลได้มากขึ้นและสร้างผลตอบแทนได้สูงขึ้นจากข้อมูลนั้นโดยใช้ AI ซึ่งจะช่วยสนับสนุนสัญญารับจ้างที่มีขนาดใหญ่ขึ้นและครอบคลุมมากขึ้นของซิเมิลาร์เว็บ

ในประเด็นเรื่องการรักษาฐานลูกค้านั้น ฝ่ายบริหารอธิบายว่า NRR ที่รายงานนั้นคำนวณจากค่าเฉลี่ยของสี่ไตรมาสก่อนหน้า ดังนั้น แนวโน้มการรักษาฐานลูกค้ารวม (Gross retention) และการขยายธุรกิจที่แข็งแกร่งเมื่อเร็ว ๆ นี้ จึงช่วยสนับสนุนความคาดหวังว่า NRR จะยังคงปรับตัวดีขึ้นเรื่อย ๆ

ฝ่ายบริหารกล่าวว่าศักยภาพของฝ่ายขายในปัจจุบันสามารถรองรับโอกาสในกลุ่มลูกค้าองค์กรได้ พนักงานฝ่ายการค้าระดับสูงและผู้จัดการฝ่ายดูแลลูกค้า (Account manager) ได้ปรับเปลี่ยนบทบาทมาเน้นการขายสินค้าเกี่ยวเนื่อง (Cross-selling) และการขายสินค้าที่มูลค่าสูงขึ้น (Upselling) ให้แก่ฐานลูกค้ารายใหญ่ระดับโลกที่มีอยู่เดิมของซิเมิลาร์เว็บมากขึ้น

สำหรับโครงสร้างของสัญญานั้น มีเพียงหนึ่งในสามสัญญารายใหญ่ของไตรมาส 2 เท่านั้นที่เกี่ยวข้องกับการฝึกฝนโมเดล LLM ทั้งก่อนและหลังการฝึกฝน โดยฝ่ายบริหารระบุว่า อุปสงค์กำลังเริ่มปรากฏขึ้นในกลุ่มแอปพลิเคชัน OEM แบรนด์ต่าง ๆ และนักลงทุนที่ต้องการบูรณาการข้อมูลของซิเมิลาร์เว็บในระดับขนาดใหญ่ (at scale)

ถอดความการแถลงผลประกอบการฉบับเต็ม

บทถอดเสียงฉบับเต็มของการประชุมทางโทรศัพท์ผลประกอบการ

คำชี้แจงจากฝ่ายบริหาร

Operator

Hello, and welcome, everyone, joining today's Similarweb Q2 Fiscal 2026 Earnings Call. [Operator Instructions] Please note this call is being recorded. [Operator Instructions] It is now my pleasure to turn the meeting over to Rami Myerson, Vice President, Investor Relations. Please go ahead.

Rami Myerson

Thank you, operator. Welcome, everyone, to our second quarter 2026 earnings conference call. Joining me today are our CEO and Co-Founder, Or Offer; our Chief Financial Officer, Ran Vered; and Maoz Lakovski, our Chief Business Officer. This morning, we released our results for the second quarter and published an investor presentation with a strategic overview of the business as well as a summary presentation of second quarter results on our Investor Relations website at ir.similarweb.com.

Certain statements made on the call today constitute forward-looking statements, which reflect management's best judgment based on the currently available information. These statements involve risks and uncertainties that may cause actual results to differ from our expectations. Please refer to our earnings release and our most recent Annual Report on Form 20-F filed with the SEC on March 2, 2026, for more information on the risk factors that could cause actual results to differ from our forward-looking statements.

Additionally, certain non-GAAP financial measures will be discussed on the call today. Reconciliations to the most directly comparable GAAP financial measures are available in the earnings release and the earnings presentation. We will begin with Or and Ran's highlights of the quarter, and then we will open up the call to questions from sell-side analysts.

With that, I'll turn the call over to Or. Or, please go ahead.

Or Offer

Thank you, Rami, and welcome, everyone, today. I'm extremely proud of how the Similarweb team delivered in the second quarter. Similarweb is at inflection point. Our core business is getting stronger while AI is opening a significant new growth opportunity. Larger deals, longer commitments, improving retention and expanding profitability are validating the strength of our business.

At the same time, strong demand from leading AI companies demonstrate that our proprietary digital data has become critical infrastructure for the AI ecosystem. Over the last several quarters, we have been very focused on 3 things: strengthening our data moat, deepening our relationship with the largest enterprise in the world; and positioning Similarweb to capture the enormous opportunity created by AI. In Q2, we started to see those pieces come together. We delivered the strongest quarter in Similarweb history for net new ARR.

We had one of the strongest quarter ever for gross retention. We achieved positive GAAP operating profit for the first time ever, NRR improved to 100% across all customers and 107% for customers above $100,000 of ARR. We exceeded our expectation for both revenue and operating profit and are raising our full year guidance for the second time this year.

AI-related revenue reached 13% of revenue in the second quarter, up from 11% at the end of the fourth quarter of 2025, and we will continue to expand it moving forward. And perhaps most importantly, we signed 3 very large multiyear enterprise contracts representing more than $60 million of accumulated contract value. So when I look at this quarter, I see more than just a strong set of numbers.

I see evidence that the strategy we've been executing is working. Let me walk you through why I believe this is so important. And let's start with the financial performance. Revenue grew 9% year-over-year to $77.2 million above the top end of our guidance. We also delivered non-GAAP operating profit above our expectation and for the first time in our history, positive GAAP operating profit.

This is a direct result of the operational disciplines we have been implementing across the company. We have been very focused on improving sales productivity, sharpening our go-to-market execution and becoming more efficient across the organization. At the same time, we continue to invest in the opportunities where we believe we can generate the highest long-term returns.

We generated $8.7 million of normalized free cash flow in the quarter, representing an 11% free cash flow margin. Moving into the second half of 2026 and beyond, we remain committed to expanding our margin and leveraging the operational efficiencies that come with our increased scale. So we are seeing the combination we have been working on: growth, improving retention, profitability and strong cash generation.

Our customer metrics also continue to improve. NRR increased to 100% across all customers and to 107% for customers above $100,000 of ARR. This is particularly encouraging because as we discussed in the last several quarters, we have been focused on improving the expansion motion with our existing enterprise customers.

We are seeing better gross retention, stronger customer engagement and increasing demand for additional Similarweb data and products. And we believe there is more room for -- improve from here. But the part of the quarter that I'm most excited about is what we are seeing with the largest enterprise. Our strategy to move upmarket is working. During Q2, we signed 3 very large multiyear contracts with a cumulative value of more than $60 million.

Those are strategically important contracts. They are also very different from the way Similarweb historically sold. For many years, our model was primarily a land and expand model. We would land the customer with a relatively small initial contract and then expand over time across additional products, teams and geographies. That model remains important, but something new is happening.

The largest companies in the world are increasingly looking at Similarweb not simply as a software application, but as strategic source of digital data. And when they do that, the size and scope of relationship change dramatically. Those contracts require a significant amount of work across sales, R&D, data scientists, finance and legal. And I want to recognize the entire team because those deals are truly company-wide efforts.

The strategic ALO go-to-market team that we created at the end of 2025, specifically focused on AI, LLM and OEM opportunities has been instrumental in this success. The team is building relationships with some of the most sophisticated companies in the world and is creating a pipeline that is materially larger than what we have historically seen.

In fact, in July, we signed a fourth large contract and we continue to see a very strong pipeline of additional opportunities like never before. And let me explain to you why this is different. Historically, large 7-figure contracts were relatively rare for Similarweb. As recently as 2025, we were still seeing large 7-figure opportunities infrequently a handful of times a year at most.

That has changed. The number and size of opportunities we're seeing from Large Enterprise has increased significantly. And I believe there is a very important reason for that. AI is changing the economics of data. Let me explain you what I mean. Historically, the value an enterprise could extract from Similarweb was limited in part by the number of people who could actually work with the data.

You had an analyst that analysts needed to understand our data. They needed to know how to query it. They needed to analyze it. They needed to connect different data sets. And then they needed to turn those insights into recommendation for the business. That is powerful, but it's still constrained by human capacity.

AI fundamentally changed this equation. Now you can take Similarweb digital data and make it available to AI systems that can analyze enormous amount of information across thousands of questions and use case at a speed at scale that humans simply cannot match. The result is that the ROI from the same underlying data can increase dramatically.

And this is the part of the story that I think is still underappreciated. AI doesn't make our data less valuable, it make our data much more valuable. Because the better the AI becomes its reasoning, the more valuable high-quality, comprehensive and trusted the data become. This is why I believe the opportunity for Similarweb goes far beyond simply selling data for LLM training. But still, the LLM opportunity is big and growing.

One of the largest contracts we signed this quarter is with a leading big tech company for data used to train its large language model. Following this expansion, this customer become our third customer with more than $10 million of ARR engagement. It is remarkable to see how Similarweb data become a fundamental source of digital intelligence for top large language model in the world.

But what excites me even more is that the other large contracts are not only about LLM training, they demonstrate that enterprise can use Similarweb data at scale for many different AI-driven use case and applications. And that is much bigger opportunity because if our data can be used across multiple AI use case and applications inside a Large Enterprise, the potential consumption of our data increased dramatically.

One customer can have multiple teams, multiple teams can have multiple use case and each use case can consume more data. That creates powerful expansion opportunity. The more use case we unlock, the more valuable our data becomes. And more valuable the data become, the more Similarweb can expand within the organization. This is a very different model from selling seats of software.

We are increasingly monetizing access to data and the consumption of the data, and we believe AI will accelerate this transition. Similarweb has become an enterprise data company. You can already see this transformation in our business. The portion of our business driven by customers generating over $100,000 in ARR has reached almost 70%, a significant rise from 63% a year ago.

Furthermore, the share of our revenue tied to a multiyear commitment expand to 66%, up from 57% a year ago. And we see this momentum continue. Those metrics underscore a fundamental transition. Similarweb has become more deeply integrated in the world's largest and most enduring enterprise organization. We are increasingly an enterprise-focused business, providing digital data to some of the largest, most sophisticated companies in the world.

And our goal is to become even more deeply embedded in those organizations, not just with one product, not just with one team, but across multiple teams, multiple products and multiple use cases. This is the expansion opportunity in front of us. Let me now turn into our AI strategy.

Over the last several quarters, we have talked about our AI strategy through 3 pillars: powering AI system with our data, building an AI-native product ourselves and expanding distribution throughout the AI ecosystem. We are seeing strong progress across all 3. First, we are powering LLM and AI Agent. We continue to see strong demand from leading AI companies for our digital data for both pretraining and post-training use case, but we are also seeing increased demand from AI Agents.

Agents need trusted, structured and comprehensive information about the digital world in order to perform their task efficiently. Our data is built for both human and Agent, and that is becoming increasingly important. Second, we are building our own AI-native products. Gen AI intelligence is shaping up as a huge opportunity. It's lucrative, fast-growing market that is top of mind for every CMO and executive at Large Enterprise right now.

We're already seeing an early sign we can win here and become a leader at this category. Our solution helps brands understand how they can show up across Generative AI platforms. We think it's an entirely new category, and our data give us a real edge. And earlier this year, we launched Similarweb AI Studio. The response has been extremely strong.

AI Studio changed the way people interact with Similarweb. Instead of needing to know exactly which report to open or which data set to use, user can simply ask a business question in natural language and receive an actionable answer with insights and recommendation. This dramatically expands who can use Similarweb. And importantly it creates a new consumption-based monetization model.

We believe this is the direction the industry is moving. Third, we are expanding distribution. Increasingly, research and decision-making are happening inside AI platforms. So we want Similarweb to be there. Our data available through MCP on Claude and ChatGPT. During the quarter, we expanded our relationship with Perplexity to bring Similarweb digital data directly into its AI-native workflow, and we expand our partnership with Manus following the successful launch of Similarweb data on the platform.

Those partnerships are more than integration. They are new distribution channels for Similarweb. They allow us to reach users who we couldn't not reach through our traditional go-to-market motion. They expand our TAM, and they reinforce our position as a critical data layer for AI-driven research and decision-making.

So now let me walk you through what I believe is happening. First, our core business is getting healthier. Gross retention is improving, NRR is inflected, sales productivity is improving, and we are seeing better expansion across enterprise customer. Second, our enterprise strategy is working.

We are seeing larger contracts, longer commitments, more multiproduct relationship and increasing demand from the world's largest companies for digital data. And third, AI is dramatically expanding the opportunity for our data. It's created new customers, new use case, new distribution channels and new ways to monetize consumption.

Those 3 things reinforce each other, and this is why I believe Q2 represents an important inflection point for Similarweb. And as I have to say before, AI is the engine, but data is the fuel. Regarding the CEO search, we are making good progress, and we are interviewing very strong candidates.

And with that, I will hand it over to Ran, our CFO.

Ran Vered

Thank you, Or. It is great feeling to deliver a strong set of results and raise guidance for the year. I'll provide highlights of our financial performance and guidance for the third quarter and the full year of 2026. Turning to our quarterly results. We generated $77.2 million of revenue in Q2, a 9% increase year-over-year and above the guidance range we provided.

Revenue growth was driven by good performance across the book of business, including new sales and upsells as well as growth in AI-related revenues that reached 13% of revenues in the second quarter, up from 11% at the end of the fourth quarter of 2025. I would like to remind you that the second quarter of 2025 provided a tough comparison for this quarter.

We expect revenue growth to accelerate in the second half of 2026, supported by the growth in ARR in the second quarter and the accelerated momentum in our business as Or discussed. Non-GAAP operating profit for the quarter was $6.5 million, reflecting an 8% margin compared to $2.4 million in the second quarter of 2025.

Non-GAAP operating profit was also above our guidance range, thanks to top line growth and disciplined cost control that more than mitigated FX headwinds. We continue our efforts to offset the headwinds to profit presented by the strengthening of the Israeli shekel versus the U.S. dollar. As a reminder, approximately half of our employees are based in Israel. Non-GAAP finance expenses was $108,000.

Non-GAAP tax expense was $1.2 million in the quarter compared to $86,000 and $1.2 million in the second quarter of 2025. To help with your modeling, we expect these items to remain approximately at this level on a quarterly basis for the rest of the year. Non-GAAP diluted earnings per share was $0.06 compared to $0.01 in Q2 2025.

ARR contracted under multiyear contracts continue to expand to 66% of ARR from 57% last year. We believe that this metric is very important and demonstrate the durability of our revenues and the importance of our data to our customers. Good cash generation and strong balance sheet are critical for our business at any stage.

We generated $8.7 million of normalized free cash flow and 11% free cash flow margin despite the FX headwinds. We believe we will generate positive normalized free cash flow on a quarterly basis going forward. We ended the quarter with approximately $73.9 million of cash and cash equivalents and no debt. We also have an available line of credit of $75 million. Our remaining performance obligations, RPO, totaled $345 million at the end of Q2, up 26% year-over-year. We expect to recognize approximately 66% of total RPO as revenue over the next 12 months.

The growth in RPO provides us with confidence in our full year guidance. We are also proud that our deferred revenue increased to a total of $141 million, a 21% increase year-over-year. In Q2, overall NRR was 100% across all customers and 107% for customers with over $100,000 of ARR. We are proud of the improvement in NRR in the quarter, which came through earlier than expected.

We expect further improvement in NRR over 2026. At the end of the second quarter, we had 1,815 customers with ARR above $25,000 compared to 1,809 in 2025. This sequential trend reflects our decision to prioritize go-to-market resources and focus on large-scale opportunities within our existing customer base over smaller inbound SMB deals.

Consistent with this focus, the average account value for this cohort grew 19% year-over-year to $149,000. The number of customers with ARR over $100,000 increased to 473 at the end of Q2, up 9% compared to 2025. Average account value for this customer cohort increased 18% to $438,000 compared to 2025.

We believe that accounts generating more than $25,000 and $100,000 of ARR that account for 90% and 69% of ARR, respectively, demonstrate that Similarweb is an enterprise-focused data company. Moving to guidance; for the full year of 2026, we are raising our revenue guidance range and expect total revenue in the range of $340 million to $380 million, representing approximately 12% year-over-year growth at the midpoint of the range.

In Q3 2026, we expect total revenue in the range of $80.5 million to $82.5 million, representing [ 17.5% ] year-over-year growth at the midpoint, accelerating versus Q2 2026. For the full year, we are raising our guidance for non-GAAP operating profit to be between $24 million and $26 million. Non-GAAP operating profit for the third quarter of 2026 is expected to be in the range of $7.5 million to $9.5 million.

With that, Or and I are ready to take -- to answer your questions. Following Q&A, Or will share some closing remarks. Operator, please open the line for questions.

Operator

[Operator Instructions] And we'll take our first question from Raimo Lenschow with Barclays.

ช่วงถาม-ตอบ

Raimo Lenschow

Congratulations. That's an amazing evolution. Or, can you talk a little bit more about those customers that you signed and they are kind of showing up in the pipeline? Is this kind of using the similar rep data in a much broader sense than we've seen classically?

And in theory, if that's the case and we open it up for more end users, then in a way, the sky is the limit because like you should like every single customer could do a lot more with your data. So that we -- it sounds almost like we are at the beginning of a very great journey here. Can you talk a little bit more on that?

Or Offer

Yes. So first of all, yes, thank you for the question. And yes, you're right. The most advanced enterprise are now realized that with AI, they can get -- they can first crunch much more data and they can get much more better insight and recommendation and the ROI dramatically for the same data is now much higher, and they can consume much more data.

So we're seeing this trend. It's very exciting, and we're very happy about that. And we think that we continue to see this great success with onboarding more enterprise to use our data in that approach.

Raimo Lenschow

And then Ran, if you think about it, where are we sitting on this new approach of getting against the big enterprise, where are we on sales capacity, et cetera? Like can the organization kind of support the growth that potentially is coming your way?

Ran Vered

Right now, yes, we are set up correctly. We're starting with the engagements we currently have in our book of business. We already work with the biggest and best enterprise in the world, and we start rolling up and start increasing those engagements.

Operator

We'll take our next question from Ken Wong with Oppenheimer & Company.

Hoi-Fung Wong

Fantastic. I think you guys mentioned that NRR potentially could trend up still. I'm just looking at the levels of -- there's a big step-up. What gives you the confidence there? How much of that is mechanical? I recall you guys already were exiting at a pretty high rate after Q1 and how much of that is kind of underlying activity that you're seeing that gives you confidence that, that number could keep trending higher?

Or Offer

Yes. We have a very, very high confidence because the NRR we report to the market, as we said, is the average of the previous 4 quarters. And because we know that this quarter NRR was very, very high, we already know that the next quarter's NRR will continue to increase.

Ran Vered

And just to add on that, we see very strong -- Or mentioned it in the prepared remarks, we see very strong GRR trends. And the changes that we did last -- at the beginning of the year that the account management are more focused on expansion, this is a focus on GRR.

We already see the fruit of this change. So we see very strong GRR that continues to be at a very strong level. And on top of it, the account management that are focused on expansion provide us with the confidence that the NRR is going to increase.

Hoi-Fung Wong

Understood. And then, Or just on the really strong pipeline results, again, fantastic in terms of what you guys saw in Q2. When you look at the quarter, how much of that execution was some of the labors from last year, you guys kind of refreshed the go-to-market, you added capacity.

How much of that is, as you said, truly an underlying change in terms of some of the customer actions like where data is now kind of proliferating across organizations or is there still more of that to come?

Or Offer

I think, as I said in the earnings, a lot of it came just focusing the people on doing the right things that can produce the highest outcome. One of them that we changed and took some of our best people and just put them on the ALO team, we call it, the strategic sales team to build better, bigger relationship with the top enterprise and basically unlock this potential.

So just moving those priorities, including the account manager, focusing on expansion, all of those decision of changing strategy really are bearing the fruits. So these are the result of that.

Operator

We'll take our next question from [ Arun ] Bhatia with William Blair & Company.

Willow Miller

Hi, it's Willow Miller on for Arjun Bhatia. Nice quarter. A couple of quarters ago, you saw some variability in, I believe, 2 large AI deals. I believe you closed one last quarter. But can we get an update on the second? Is that still in the pipeline and how are you thinking about the timing of closing it?

Or Offer

Yes. Funny enough, this second one is still in the pipeline. And we still think that this is another nice big surprise that can come by the end of the year. Yes.

Operator

We'll take our next question from Patrick Walravens with Citizens.

Kincaid LaCorte

Great. This is Kincaid on for Patrick. Super excited to see this narrative playing out for you guys. Is it possible to get a little bit of color when I think about the $60 million as well as just the total spectrum of these new deals, you mentioned it's not just pretraining and post-training, but it also includes like correct me if I'm wrong, like Agents in production that want to understand how the Internet is structured and where they should go next. Can you give us a sense of like how that ARR is split? Is it 50% post-training, 25% pretraining and 25% in production? What's that breakdown?

Or Offer

I cannot go specifically to the full breakdown, but we had -- what we said, we had 3 big deals. Each one of them is above 8-figure engagement. Only one of them is for LLM training pre and post. So only one is pre and post. The other 2 are just big enterprise leveraging our data in different ways than LLM training that is much more exciting and very big opportunity.

Ran Vered

Yes. Maybe just to add [indiscernible] the way we think about it, the various markets that we're seeing tailwinds and demand in different markets. One is the LLM that we were speaking about. The other one is OEMs, so build on top of our data. Then we have the brands and last we have the investors.

And we see great demand across the board in all of them for large deals to integrate our data at scale because of the ability to digest and ingest more data at scale with AI. So what we've seen in the past, which was deal every now and then, deal every quarter of our pipeline. So now it's becoming much more sustainable, much more robust. It helps us increase our TAM and increase our average deal value.

Kincaid LaCorte

Spectacular. And then if you have any update with specifics on your Gen AI intelligence product, I've always found that one fascinating.

Or Offer

Yes. There is -- the demand there is increasing. We do see a nice increase there. And I think there is now more strategic team start building up on the Gen AI data that we have that's becoming very interesting for brands to connect the full journeys of the users with what people are asking those LLMs and the outcomes. So everything around how the interaction with LLM change the consumer behavior. So this is -- we're seeing also a very big success on that.

Operator

[Operator Instructions] We'll take our next question from Scott Berg with Needham.

Scott Berg

Really nice quarter here. Or, you obviously spoke a lot about the new use cases that you're seeing in some of these large deals. They're not all LLM training exercises. But as you think about the new use cases, are you changing the pricing of the platform to evolve with some of these new use cases or are the pricing mechanisms the same?

Or Offer

The pricing is the same because most of those big deals are around data. So it's more about consumption and data access. It's less about the software. And so the price didn't change and just selling much more data and much -- many more different data sets because as I said, AI is changing the equation. Now companies can consume much more data and get much more ROI of data.

Scott Berg

Very helpful there. And then Ran, as we think about a couple of these deals here, these large deals, are there any onetime revenues to be cognizant of -- from these? I believe that maybe one of them has some onetime revenues. And if there is, what is the timing of those revenues look like for modeling purposes?

Ran Vered

Scott, thanks for the question. So those deals are ARR deals. And if there is a onetime element, it's really quite negligent and material. And usually, it's recognized on the initiation of the deal or a couple of months later at max. But those deals are pure ARR deals with, again, a small and immaterial amount of onetime, if at all.

Operator

We'll take our next question from Luke Horton with Northland Securities.

Lucas John Horton

Congrats again on the great quarter here. Obviously, the momentum is really building on the enterprise side. Just curious if you're -- if there's anything to call out on the kind of SMB or smaller customer cohort. Are they seeing any sort of macro pressures or I guess, how is growth looking on the smaller customer cohort side?

Or Offer

Nothing special there. I think overall, the traffic was going down, the visitor traffic to the website, the world change and there is less traffic coming from [ SaaS ] or like market broader. But overall, AOV is the same, [indiscernible] is the same. And I think a lot of more commercial people are moving now to drive more expansion as we can see this opportunity.

Lucas John Horton

Okay, great. And then just wondering about as far as adding net new customers, are you seeing any trends with initial customer contracts coming in at higher price points now? And kind of how much of the focus is on the cross-sell and upsell with existing customers versus kind of outbound net new customers?

Or Offer

So I think that most of our tenure senior commercial people that were doing outbound and enterprise new sell, we removed them more to focus on expansion. As I said before, we have an amazing book of business, and we already engaged with most of the best and biggest companies in the world. And now we just come and have a huge opportunity on cross-sell and upsell with this new change that's happening with AI.

Operator

At this time, there are no further questions in queue. I will now turn the meeting back to Or Offer for closing remarks.

Or Offer

So thank you for all the questions today. So now let me recap what we discussed today at the call. When we entered to 2026, we said this was going to be a transformation year for Similarweb. We said we were moving from building to scaling. We said our priorities were to strengthen our data mode, deepen our enterprise relationship and scale our AI-first solutions.

6 months into this year, we are seeing tangible evidence that this strategy is working. We delivered the strongest net new ARR quarter in our history. We achieved positive GAAP operating profit for the first time ever. We improved NRR to 100%. We signed multiple transformational enterprise contracts, and we continue to see expanding demand for our data from AI companies and Large Enterprise.

There is still a lot of work ahead of us. We need to continue improving execution. We need to continue to expand our enterprise relationship, and we need to continue building the best digital data infrastructure for the AI era. Thank you, everyone, on the call for your continued support. We're looking forward to speaking to you again over the coming weeks. Thank you all.

Operator

Thank you. This brings us to the end of today's meeting. We appreciate your time and participation. You may now disconnect.