Hội nghị công bố kết quả kinh doanh Q3 2026 của AvePoint (AVPT): Tăng trưởng ARR tăng tốc nhờ nhu cầu về AI Trust

Điểm nhấn chính

• Doanh thu quý 3 tăng 19% so với cùng kỳ năm trước lên 131,1 triệu USD, tương đương mức tăng 20% theo tỷ giá không đổi, nhờ doanh thu SaaS tăng 26% lên 105,5 triệu USD.
• Tổng ARR đạt 495,1 triệu USD, tăng 27% theo báo cáo và tăng 25% sau khi điều chỉnh theo tỷ giá. ARR ròng tăng thêm đạt mức kỷ lục 30 triệu USD, tăng 34% so với cùng kỳ năm trước.
• Tỷ lệ giữ chân khách hàng gộp sau khi điều chỉnh theo tỷ giá đã đạt mục tiêu 90% của công ty, trong khi tỷ lệ giữ chân khách hàng ròng cải thiện lên 111%.
• Số lượng khách hàng sử dụng AgentPulse tăng hơn 50% so với quý 2 và số lượng tác tử được quản lý tăng hơn gấp đôi. Ban lãnh đạo nhận định độ tin cậy của AI là cơ hội thị trường lớn nhất trong lịch sử của AvePoint.
• Biên lợi nhuận hoạt động GAAP quý 3 mở rộng hơn 130 điểm cơ bản lên 8,7%. Dòng tiền từ hoạt động kinh doanh vượt 31 triệu USD trong quý và đạt tổng cộng 71,6 triệu USD trong 9 tháng đầu năm.
• AvePoint đã nâng mức trung vị của triển vọng ARR cả năm thêm 1 triệu USD và hiện kỳ vọng ARR năm 2026 sẽ đạt từ 524,8 triệu USD đến 528,8 triệu USD.

Dữ liệu tài chính quan trọng

AvePoint khép lại quý 3 với Quy tắc 44 (Rule of 44) tính trên cơ sở 12 tháng gần nhất, được tính bằng tốc độ tăng trưởng doanh thu cộng với biên dòng tiền tự do. Biên lợi nhuận hoạt động GAAP trong 12 tháng gần nhất của công ty đạt 10%.

Hoạt động kinh doanh và vận hành

Quản trị tác tử AI là chủ đề trọng tâm trong cuộc họp báo cáo kết quả kinh doanh quý 3/2026 của AvePoint. AgentPulse cung cấp khả năng phát hiện tập trung, quản trị và kiểm soát vòng đời trên Microsoft Copilot Studio, Google, Salesforce, Amazon và Claude của Anthropic. Công ty cũng đã ra mắt tính năng sao lưu và khôi phục cho các tác tử Microsoft Copilot Studio, đồng thời giới thiệu Kinetic Classification để phân loại độ nhạy cảm của dữ liệu và nhãn lưu trữ.

Từ quý 2 sang quý 3, lượng khách hàng sử dụng AgentPulse đã tăng hơn 50%, trong khi số lượng tác tử được quản lý tăng hơn gấp đôi. Tỷ lệ ARR mới gắn liền với các gói Control (bao gồm cả AgentPulse) đã tăng khoảng 65%. Ban lãnh đạo cho biết phần lớn việc áp dụng AgentPulse hiện nay vẫn dựa trên gói giải pháp kết hợp do gói SKU độc lập mới chỉ ra mắt vào tháng 7; tác động đáng kể hơn đến việc chốt hợp đồng và doanh thu dự kiến sẽ diễn ra vào năm tới.

Nhu cầu từ khối doanh nghiệp tiếp tục tăng mạnh. AvePoint kết thúc quý với 976 khách hàng mang lại hơn 100.000 USD ARR, tăng 28% so với cùng kỳ năm trước. Công ty đã có thêm kỷ lục 65 khách hàng gia nhập nhóm này, trong khi các nhóm khách hàng tạo ra hơn 250.000 USD, 500.000 USD và 1 triệu USD ARR đều tăng trưởng hơn 30%.

Hoạt động của các kênh đối tác và nhà cung cấp dịch vụ được quản lý (MSP) cũng tiếp tục đóng vai trò quan trọng. ARR từ MSP tăng 40%, và 72% mức ARR tăng thêm trong quý 3 đến từ các đối tác kênh. Kênh đối tác hiện chiếm 60% tổng ARR, so với mức 56% của một năm trước đó.

Tăng trưởng ARR theo khu vực diễn ra trên diện rộng sau khi điều chỉnh theo tỷ giá: Bắc Mỹ tăng tốc lên 23%, trong khi EMEA và APAC đều tăng 27%. Doanh thu SaaS tăng 24% ở Bắc Mỹ và tăng 28% theo tỷ giá không đổi ở cả EMEA và APAC.

Ban lãnh đạo cho biết khu vực công của Mỹ tiếp tục phục hồi trong năm 2026 và đóng góp tích cực trong quý 3, nhưng vẫn chưa hoàn toàn trở lại mức hiệu quả trước đây. Công ty kỳ vọng sẽ tiếp tục mở rộng hơn nữa vào năm 2027.

Kế hoạch hướng dẫn của ban lãnh đạo

Ban lãnh đạo cho biết triển vọng biên lợi nhuận hoạt động quý 4 thể hiện sự cải thiện so với quý 3, nhưng vẫn thấp hơn mức của cùng kỳ năm trước do AvePoint tiếp tục đầu tư vào các năng lực bảo đảm độ tin cậy của AI.

Công ty tái khẳng định mục tiêu đạt 1 tỷ USD ARR vào cuối năm 2029. Ban lãnh đạo cũng duy trì cam kết đối với mục tiêu biên lợi nhuận hoạt động 27,5% vào năm 2029, đồng thời lưu ý rằng tiến trình này có thể không diễn ra theo đường thẳng.

Rủi ro và các yếu tố cần theo dõi

• Biên lợi nhuận gộp giảm từ 75,1% xuống 72,8% do biên lợi nhuận gộp mảng dịch vụ suy giảm, mặc dù biên lợi nhuận gộp mảng phần mềm vẫn ổn định ở mức 83%.
• Chi phí hoạt động tăng từ 53% lên 56% doanh thu khi AvePoint đẩy mạnh đầu tư vào bán hàng, tiếp thị và các năng lực bảo đảm độ tin cậy của AI.
• Các sản phẩm chuyển đổi dữ liệu tiếp tục tạo ra lực cản 2 điểm phần trăm đối với tỷ lệ giữ chân khách hàng gộp do tỷ lệ giữ chân của nhóm này vốn thấp hơn về mặt cấu trúc.
• Ban lãnh đạo ước tính hiện chỉ có 1% các doanh nghiệp hàng đầu đang triển khai các tác tử trong những quy trình làm việc trọng yếu; hầu hết các tổ chức vẫn chỉ đang sử dụng AI cho các tác vụ cơ bản.
• Hơn 90% lượng ARR hiện tại của AvePoint vẫn gắn liền với hệ sinh thái Microsoft, dù ban lãnh đạo nhận thấy thêm nhiều cơ hội mới trong lĩnh vực đa đám mây và AI.
• AgentPulse hiện được cấp phép theo số lượng nhân viên thay vì mức độ sử dụng tác tử, do đó sự gia tăng số lượng tác tử được quản lý không trực tiếp chuyển hóa thành doanh thu dựa trên mức tiêu thụ.

Điểm nhấn phiên hỏi đáp với chuyên gia phân tích

Khả năng thương mại hóa AgentPulse: Ban lãnh đạo cho biết gói SKU AgentPulse độc lập đang phát triển một tệp khách hàng tiềm năng vững chắc và có thể đóng vai trò là điểm tiếp cận mới vào các tài khoản khách hàng doanh nghiệp. Tuy nhiên, phần lớn lượng tiếp nhận trong quý 3 đến từ các gói Control, trong khi tác động doanh thu lớn hơn từ phiên bản độc lập dự kiến sẽ diễn ra vào năm tới.

Vị thế cạnh tranh: AvePoint tạo sự khác biệt trong cách tiếp cận so với các nhà cung cấp an ninh mạng truyền thống bằng việc nhấn mạnh vào quản trị lấy dữ liệu làm trung tâm, giám sát hoạt động của tác tử, khả năng khôi phục và kiểm soát cộng tác nội bộ. Ban lãnh đạo cho biết đồ thị quản trị của công ty kết nối hành vi của tác tử và quyền truy cập dữ liệu với các biện pháp kiểm soát, bao gồm các khung tiêu chuẩn của Đạo luật AI của EU, NIST và ISO.

Động lực mở rộng: Tăng trưởng từ các khách hàng hiện hữu chủ yếu được thúc đẩy bởi việc bán thêm sản phẩm, tính năng và bộ giải pháp thay vì mở rộng số lượng người dùng (seat). Ban lãnh đạo nhận định mức đóng góp từ khách hàng mới và sự mở rộng trong các tài khoản hiện hữu là tương đối cân bằng.

Tiềm năng giữ chân khách hàng: Sau khi đạt tỷ lệ giữ chân khách hàng gộp 90%, ban lãnh đạo cho biết vẫn có khả năng tiếp tục cải thiện và dự báo có thể tăng thêm một vài điểm phần trăm trong vài năm tới khi các khoản đầu tư vào bộ phận hỗ trợ thành công của khách hàng tiếp tục được duy trì.

Phân bổ vốn: AvePoint đã chi khoảng 31 triệu USD để mua lại 2,3 triệu cổ phiếu trong quý 3. Tính từ đầu năm đến nay, công ty đã mua lại 12,6 triệu cổ phiếu với giá trị khoảng 142 triệu USD, qua đó giảm ròng 5 triệu cổ phiếu đang lưu hành. Công ty còn khoảng 89 triệu USD theo hạn mức ủy quyền mua lại và cũng đã đảm bảo một hạn mức tín dụng tuần hoàn mới trị giá 100 triệu USD.

Toàn văn biên bản cuộc họp báo cáo kết quả kinh doanh

Toàn văn cuộc gọi công bố kết quả kinh doanh

Phần trình bày của ban lãnh đạo

Operator

Good day and welcome to the AvePoint, Inc. Third Quarter 2026 Earnings Call. [Operator Instructions] Please note, this event is being recorded.

I would now like to turn the conference over to Jamie Arestia, Vice President, Investor Relations. Please go ahead.

Unknown Speaker

Thank you, Operator. Good afternoon and welcome to AvePoint's third quarter 2026 earnings call. With me on the call this afternoon is Dr. T.J. Jiang, Chief Executive Officer, and Jim Kelly, Chief Financial Officer. After preliminary remarks, we will open the call for a question and answer session.

Please note that this call will include forward-looking statements that involve risks and uncertainties that could cause actual results to differ materially from management's current expectations. We encourage you to review the Safe Harbor statements contained in our press release for a more complete description. All material in the webcast is the sole property and copyright of AvePoint with all rights reserved.

Please note this presentation describes certain non-GAAP measures, including non-GAAP gross profit, non-GAAP gross margin, non-GAAP operating income, and non-GAAP operating margin, which are not measures prepared in accordance with U.S. GAAP. The non-GAAP measures are included in this presentation as we believe they provide investors with a meaningful understanding of how management evaluates the company's operating performance. These non-GAAP measures should not be considered in isolation from, as substitutes for, or superior to, financial measures prepared in accordance with U.S. GAAP.

A reconciliation of these measures to the most directly comparable GAAP financial measures is available in our third quarter 2026 earnings press release, as well as our updated investor presentation and financial tables, all of which are available on our Investor Relations website. With that, let me turn the call over to TJ.

Tianyi Jiang

Thank you, Jamie. And thank you to everyone joining us on the call today. Q3 was another strong quarter for AvePoint, driven by $30 million of net new ARR that surpassed last quarter's record.

Total ARR growth was 25% when adjusted for FX, the second straight quarter that this metric has accelerated. At the same time, our ongoing commitment to profitable growth enabled strong cash generation, giving us the flexibility to invest with conviction. And this quarter, we did exactly that, focusing on our go-to-market motion for AgentPulse and our AI trust capabilities. Because the opportunity in front of us is the largest in our company's history. Simply put, AvePoint is the AI trust company. It's this market opportunity that I want to focus on today and why AvePoint is positioned to win it. Let's jump in.

Our customers are no longer asking whether to deploy AI agents. This is already taking place at every customer and prospect we meet. Instead, they face a growing trust gap, the distance between what their AI can reach, do, and decide, and what anyone actually approved. They are asking whether they can trust these agents, a question confirmed by Gartner's findings that only 13% of organizations believe that they have the right agent governance framework in place today. And it's not surprising that they are investing heavily to answer that question. IDC found that enterprises are now allocating nearly 20% of their planned AI spending for governing and securing AI and agents. Our own State of AI Research, which surveyed 750 enterprise leaders across industries and regions, also aligns with these findings. Nearly 9 in 10 organizations experienced at least one AI agent-related security breach in the past 12 months. As we tell enterprise leaders, your AI will find everything you forgot you had.

And our promise to them is simple. We move them from unmanaged AI to measurable trust. These statements are resonating across the market, and it's clear that the demand is there for a complete agent management platform solution, a market that is ours to win.

Why do we believe this? There are 3 main reasons, and our Q3 wins show each of them in action. First, we are platform neutral by design and completely independent of what we govern. Enterprises today rely on multiple platforms and need 1 layer of proof that answers directly to them. Our AgentPulse solution provides centralized agent discovery, governance, and lifecycle control across Microsoft Copilot Studio, Google, Salesforce, Amazon, and now Anthropic's Claude in 1 registry. While native tools govern their respective ecosystems, AvePoint governs all of them.

This platform-neutral posture is winning us large enterprise customers. One of Asia's largest e-commerce companies became a new AvePoint customer this quarter, choosing our Control Plus bundle to secure both Google Workspace and Microsoft 365 across approximately 40,000 users. They replaced their incumbent data security vendor because we could accurately discover sensitive personal information across both clouds, including native local language detection. The deal also shows that our Google momentum extends well into the large enterprise, and it reflects the broader strength we are seeing across Asia Pacific, where SaaS revenue growth accelerated to 28% in constant currency this quarter.

Second, trust starts at the data layer, which has been our focus since founding AvePoint 25 years ago. As we have said many times, an agent is only as trustworthy as the data it can reach. The first step to AI trust is making data AI-ready by finding and fixing oversharing, labeling content, retiring stale or duplicate data, and curating grounding sources. We have spent 25 years managing our customers' most sensitive data, and agents are the next surface area for that same muscle.

One of the leading private research universities in the United States is a good example of the need for AI-ready data ahead of a broader rollout of Copilot and Gemini. They needed to find overshared and sensitive content across Microsoft 365 and Google Workspace with a small IT team supporting 60,000 users across many decentralized departments. They chose AvePoint over a competing data security vendor for our unified governance across both clouds, along with our managed services to add capacity where their team was stretched.

Third, we cover the full agent lifecycle and keep the business resilient. We ensure agents are discovered, owned, scoped, attested, renewed, and retired. And critically, we restore what an agent changed when something goes wrong. One of the largest natural gas infrastructure companies in the United States shows why that matters. A customer since 2023, they discovered roughly 3,000 agents had already been created across their environment with no visibility into who owned them, what data they could reach, or how to manage their lifecycle. They quickly addressed these challenges by upgrading to our Control Complete package in Q3, giving their security, compliance, and IT teams 1 place to govern agents in Microsoft 365 and Power Platform for approximately 7,000 users.

As more organizations face rapid agent sprawl throughout their businesses, we have seen an acceleration in the adoption of AgentPulse. From Q2 to Q3, the number of AgentPulse customers grew by more than 50% and the total number of agents that AgentPulse is managing more than doubled. And during the same time, the percentage of our new ARR tied to control bundles, each of which includes AgentPulse, grew approximately 65%. And with many organizations just beginning to address this growing and evolving risk, we expect the healthy demand trends seen thus far to only continue.

We're also pleased to share that Gartner included AvePoint in its first-ever Magic Quadrant for AI Agent Management Platforms. A year ago, this category barely existed. Today, it is becoming central to how enterprises govern AI agents, and we're proud that AvePoint is part of it.

We also kept up the pace of innovation in Q3. We became the first to market with backup and recovery for Microsoft Copilot Studio agents. We also launched Kinetic Classification, which uses AI to classify data by sensitivity and retention and keeps refining that classification over time so the data agents rely on is labeled correctly. And we expanded governance and protection across Salesforce, ServiceNow, AWS, and Azure.

We also recently announced the AvePoint Trust Platform, extending our platform to the full AI estate, not only to the data and agents, but to the AI models and tools which enterprises are deploying faster than they can govern. The AvePoint Trust Platform is built around 4 disciplines: Observe, Protect, Govern, and Optimize on a governance graph that joins global regulatory intelligence directly to each customer's business context. And the platform is built on the belief that AI is only trusted if you can prove it runs on governed data with least privilege access and can be recovered when it fails. Combined with expanded support for Anthropic's Claude and MCP-connected agents, these enhancements reinforce AvePoint as the AI trust company, and ongoing innovation will only deepen our competitive moat going forward.

These are clearly dynamic and fluid times, but the fundamental needs of our customers have not changed. They require integrated platform solutions for today's toughest data management challenges. And as they accelerate their AI deployments, they need to see every agent, govern it across every platform, and recover when something goes wrong so they can scale AI with confidence. While the short-term winners are those making AI possible, we believe that the long-term AI winners will be those that ultimately make it trustworthy with 1 control plane across the full AI estate. We are well positioned to lead this category as we continue on the path to our $1 billion ARR target by the end of 2029. Thank you again for joining us today. I will now turn it over to Jim.

Unknown Speaker

Thanks, TJ, and good afternoon, everyone. Thanks for joining us today. Throughout 2026, we have consistently demonstrated the strength and durability of AvePoint's financial model, delivering strong recurring revenue growth, expanding GAAP profitability, and meaningful cash generation, while continuing to invest in the opportunities ahead. Our third quarter results represent another step forward in that journey, reinforcing both the growing demand for our platform and our ability to execute consistently across the business. As we look toward the end of the year and beyond, we remain confident in our ability to sustain attractive growth, generate strong cash flow, and drive long-term shareholder value.

Q3 had a number of highlights, including ARR growth of 25% after adjusting for FX, the second straight quarter that the growth of this key metric has accelerated; record net new ARR of $30 million, representing year-over-year growth of 34%, the 14th straight quarter that this metric has been double digits; FX-adjusted gross retention rates reaching 90%, which to remind you, has been our target for several years; and FX-adjusted net retention rates improving to 111%; the fourth straight quarter of GAAP operating margin expansion as Q3 grew more than 130 basis points year over year, and GAAP operating margins are now at 10% on a trailing 12-month basis; and finally, another strong cash flow quarter as we generated more than $31 million of operating cash flow in Q3 and have generated more than $71 million year-to-date.

So with that, let's dive a bit deeper into the quarter. Total revenues in Q3 were $131.1 million, representing 20% growth year over year on a constant currency basis and above the high end of our guidance. On a reported basis, total revenue growth was 19%. Q3 SaaS revenues were $105.5 million, representing 80% of total Q3 revenues and growing 26% year-over-year on both a reported and constant currency basis. Term license and support revenue was $8.9 million, continuing its expected decline and representing 7% of Q3 revenues compared to 11% a year ago. Lastly, services revenues grew 21% year over year to $16.7 million and represented 13% of Q3 revenues, in line with a year ago.

Turning to our revenue performance on a regional basis. In North America, total revenues grew 20% year over year, driven by SaaS revenue growth of 24%. In EMEA, total revenue growth was 21% year over year, driven by SaaS revenue growth of 26%. And in APAC, total revenues grew 18% year over year, driven by SaaS revenue growth of 27%. On a constant currency basis, EMEA SaaS revenue growth accelerated to 28%, while total revenue growth was 22% year over year. And for APAC, SaaS revenue growth accelerated to 28%, while total revenue growth accelerated to 18%.

Switching to ARR, which we believe is the most important metric for investors, we saw strong performance from all 3 regions in Q3 as North America ARR growth accelerated to 23%, EMEA ARR growth was 27%, and APAC ARR growth accelerated to 27% on an FX-adjusted basis. Taken together, we ended Q3 with total ARR of $495.1 million. This represents 27% year-over-year growth and 25% after adjusting for FX, which was an acceleration from Q2. As a result, Q3 net new ARR was a record $30 million, representing growth of 34% year over year. I'd also like to point out that our North America business reached $200 million in ARR in Q3 and congratulate the team on another strong quarter of execution. On a reported basis, North America ARR growth was 23%, EMEA ARR growth was 32%, and APAC ARR growth was 27%.

The broad-based demand we see also applies across our customer segments, each of which grew roughly in line with our total ARR growth. And I would call out a few additional highlights. First, the year-over-year growth in the number of both our enterprise and mid-market customers was the highest we have ever achieved. Second, we ended the third quarter with 976 customers with ARR of over $100,000. This represents 28% growth year over year, which is both an acceleration from Q2 and the highest growth for this metric in 4 years. The 65 net additions to this cohort were also the highest we have ever delivered. And lastly, our larger customer cohorts of greater than $250,000, greater than $500,000, and greater than $1 million of ARR each grew comfortably above 30%.

Taken together, these results continue to show the durable and accelerating enterprise demand for our ability to solve their most critical data management challenges. And similar to last quarter, the strong demand from our enterprise segment was matched by our smaller customers. The year-over-year growth in the number of our SMB end customers was the highest in 8 quarters. Our continued success here is attributable to the rapid growth of our MSP practice, where ARR grew 40% year-over-year in Q3, and to our strategy of driving more business through channel partners. This led to 72% of our Q3 incremental ARR coming through the channel. And as a result, 60% of our total ARR today comes through the channel compared to 56% a year ago. As we continue catering to companies of all sizes, our channel investments will remain a top priority in supporting our commitment to profitable growth.

Turning now to our customer retention rates, which are adjusted for FX. As I mentioned, our Q3 gross retention rate was 90%, and our Q3 net retention rate was 111%, both of which were an improvement from Q2. To remind you, 90%-plus has been our GRR target for several years, and with our actions demonstrating ongoing progression over this time, we are pleased to achieve the goal this quarter. These actions include investments in our customer success teams, especially in North America, where our improved efficiency and greater utilization of telemetry and analytics has driven a meaningful improvement in the retention rates of the accounts where we leverage a pooled customer success strategy. And lastly, we are pleased to achieve this goal while still facing the 2-point headwind from our migration products, given their naturally lower retention rates. On a reported basis, Q3 GRR was 90% and Q3 NRR was 112%, both of which were improvements over Q2.

Turning back to the income statement, Q3 gross profit was $95.4 million, representing a gross margin of 72.8%. This compares to 75.1% a year ago, and while the year-over-year decline is again the result of lower services gross margins, the 83% gross margins on our software products is in line with both the prior quarter and prior year. Q3 operating expenses totaled $73.4 million, or 56% of revenues. This compares to 53% of revenues a year ago, reflecting our increased investments across the business in 2026. As a result, Q3 non-GAAP operating income was $22 million, which represents an operating margin of 16.8% and at the high end of guidance. Importantly, we continue to focus on GAAP profitability through our ongoing management of stock-based compensation, which was 8% of Q3 revenues compared to 10% a year ago. As a result, Q3 GAAP operating margins were 8.7% and expanded more than 130 basis points year over year and are 10% on a trailing 12-month basis. For the Rule of 40, which is the sum of revenue growth and free cash flow margin, we finished Q3 at the Rule of 44 on a trailing 12-month basis.

Turning to the balance sheet and cash flow statement, we ended the quarter with $418.3 million in cash, cash equivalents, and short-term investments. And for the first 9 months of the year, operating cash flow was $71.6 million, or a 19% margin, while free cash flow was $67.7 million, or an 18% margin. This compares to operating cash flow of $55.6 million and free cash flow of $52.6 million in the same period a year ago, representing year-over-year growth of 29% for both metrics.

Turning now to share repurchases, an important pillar of our capital allocation philosophy. In Q3, we spent approximately $31 million to repurchase 2.3 million shares, and combined with our first half actions, we have spent approximately $142 million to repurchase 12.6 million shares this year. Compared to the first 9 months of last year, we have repurchased more than 7 times as many shares at approximately 70% of the cost per share. These repurchases have more than offset the dilution from our employee equity programs, reducing shares outstanding by a net 5 million year-to-date. We have approximately $89 million left in our share repurchase program, and we will continue to be opportunistic in evaluating additional buybacks.

Before I turn to guidance, I'd like to provide 2 additional financial updates. The first is that we recently entered into an agreement with JPMorgan for a new $100 million revolving credit facility, replacing our existing $30 million credit facility at what we believe are very favorable terms. And while we have not needed to leverage these facilities to fund our business, this new upsized facility, alongside our strong cash position, gives us the confidence to continue executing on the 3 key pillars of our capital allocation strategy: investing in the business, strategic M&A, and returning capital to shareholders through share repurchases. And second, we were pleased to be added to the S&P SmallCap 600 Index earlier this week, which is another important step in our evolution as a public company and reflects the continued financial maturity of our business. This is especially true given the index's requirements tied to GAAP profitability, where we have emphasized for many quarters as a differentiator between AvePoint and many other software companies of all sizes.

Turning now to our guidance, where we are again raising our expectations for full-year ARR by $1 million at the midpoint. Additionally, our updated full-year outlook for revenue and non-GAAP operating income includes the full Q3 outperformance relative to guidance, and our outlook for all metrics includes the current expected impact from the fluctuation in foreign exchange rates. As a result, for the fourth quarter, we expect total revenues of $138.4 million to $140.4 million, or growth of 22% at the midpoint, both on a reported and constant currency basis. We expect non-GAAP operating income of $24.7 million to $25.7 million, representing an operating margin of approximately 18% at the midpoint, a step up from Q3, but below the prior year as we continue to invest behind our AI trust capabilities. For the full year, we now expect total ARR of $524.8 million to $528.8 million, reflecting the $1 million raise at the midpoint I just mentioned, which is growth of 26%, both on a reported and FX-adjusted basis. We now expect total revenues of $511.3 million to $513.3 million, or growth of 22% at the midpoint. This includes the Q3 beat of $1.9 million. On a constant currency basis, we continue to expect revenue growth of 21% at the midpoint, although this reflects a modest acceleration from our prior guidance. And lastly, we now expect full-year non-GAAP operating income of $87.5 million to $88.5 million.

We have again included a slide in our investor presentation that provides a walk from our prior full-year guidance in August to today's updated outlook.

In summary, Q3 showed accelerating demand, with FX-adjusted ARR growth accelerating for the second consecutive quarter, record net new ARR and additions to our $100,000 customer cohort, and the achievement of our 90% gross retention target. Importantly, we delivered these while continuing to invest with discipline and return capital to shareholders. And as we look to close out a strong year, we remain focused on profitable growth and on our path to $1 billion of ARR by 2029. Thanks for joining us today. And with that, we would be happy to take your questions. Operator?

Operator

[Operator Instructions] The first question comes from Joseph Gallo with Jefferies. Please go ahead.

Okay, we'll move to the next question. The next question comes from Shrenik Kothari with Baird. Please go ahead.

Phần hỏi đáp

Shrenik Kothari

You have been highlighting improving U.S. public sector traction following a difficult period last year with the second half expected to benefit more. It was interesting to hear the wins around multi-cloud, data governance, AI readiness, all of that. It's all really encouraging. Just wanted to understand how much of the recovery and the strength you saw reflects some of these sort of procurements finally converting versus more broadening of demand for the resilience and the AI governance. And as you look into '27, how should we think about the public sector returning to their historical growth contribution and the larger opportunity emerging from these new use cases? And I had a quick follow-up.

Unknown Speaker

Sure, thanks, Shrenik. So I think you're right. We talked about it last quarter and I would describe it similarly that we talked about our public sector in recovery mode this year. And I think we saw that for the first 2 quarters and we saw it again here in the third quarter where we saw really good performance compared to, particularly compared to last year. But I would say there's still room to go. I don't think we're back to 100%. And I think that's both encouraging because we had really nice contribution in Q3, but I think that gives us hope and really optimism for really improved performance in 2027. So I would still have called this year recovery. We're going to see expansion in '27. So we're looking forward to even more contribution in '27.

Shrenik Kothari

Great, very helpful. And then on TJ, there's definitely clear substantial interest in agent governance and AgentPulse numbers that you disclose growing 50% sequentially and agents under management doubling or more than doubling. I wanted to understand just this distinction between customers trying to discover these unmanaged agents as illustrated by your example around the energy customer versus those deploying them at scale into business-critical workflows. Like, where are we right now? What are you seeing in respect to that transition happening most rapidly? And is spending being unlocked? Are starting to get unlocked by this kind of more broadening of access or it's still compliance, runway costs, just wanted to understand what are the trends here.

Tianyi Jiang

Yes, thanks for that. We see a lot more agents running within enterprise. And again, it could just be smaller workloads. It doesn't necessarily mean the critical workflows that you're referring to. We think that companies actually doing the critical workflows with agents are still, you know, top 1%, if you will. So majority of the organizations are still very much leveraging AI to do basic tasks. And of course, we are there to help them discover and uncover agents and making sure that we can control the risks, lay down the guardrails, and at the same time, we also do this cost analysis for enterprise customers as well.

Operator

Thank you. The next question comes from Joe Gallo with Jefferies. Please go ahead.

Joseph Gallo

Hey guys, thanks for the question. And I'm sorry about earlier. My call dropped. TJ, you talked a lot on the call about, you know, being a neutral bodyguard and that's what's needed in agentic security. I know the pitch versus the mega platforms, but when you consider pure play cyber vendors or other cyber resilience vendors, they have a similar pitch in terms of kind of being that neutral bodyguard. So can you just kind of go through a little bit more what the differentiation is there? Is it the depth on Microsoft or what are the 1 or 2 selling points you have against the more pure play cyber vendors in terms of agentic security?

Tianyi Jiang

Yes, that's a great question. So if you look at the Gartner Magic Quadrant around agentic management platform, we are 1 of the only 13 vendors listed there, given our rich history around specific unstructured data analysis and also our governance graph that actually tracks various signals and looking at how actually these agents, what system they're touching, what data they're touching. So our grounding is very different than the traditional cyber players. The cyber players are very focused around periphery, very focused on identity management. We are very much focused on data because we come from data content management space. So data access and also what do these agents do in terms of, you know, modifications, changes, and if those cause damages, when we do the remediation. So recovery as well as, of course, cost monitoring. So we think that our approach is very different. It's not a peripheral or aggressive external-facing cyber posture, but more internal collaboration, productivity type of governance. So it's a different perspective. It is a very big space.

Joseph Gallo

That's really helpful. Thank you for that. And maybe just as a follow-up, Jim, you talked about capital allocation. You've obviously leaned into share buybacks, you know, in the past year. But when you think through investing in your internal business, where's the higher ROI? Is it building out the go-to-market and pedal to the metal on that side? Or is it building out the product portfolio with M&A? And is there anything that you need on the agentic side that we should kind of expect you to kind of build out? Thank you.

Unknown Speaker

Well, thanks, Joe. You know, I think the beauty of where we sit today is we do have the flexibility to do more than 1 of the things you mentioned. You know, with our strong balance sheet in terms of our cash position, in terms of our free cash flow generation that we're creating every quarter, we do have a lot of flexibility that we can actually execute on all 3 of our capital allocation strategies. So the first, as you mentioned, investing in the business, I think we've demonstrated that. We talked about it last quarter about some of the investments we're making in our sales and marketing efforts and the go-to-market strategies that we're employing. So I think we want to invest there, and we're doing that. In addition to that, we still have enough capital, obviously, to look at M&A opportunities to provide incremental benefits to the platform that we have and where we can shore it up and where we can add incremental value to it. So we're doing that as well. And then third, we're obviously, as you mentioned, significantly repurchasing our shares and being very opportunistic when we do that. So we've really done all 3. And, you know, I'm happy that we're in a position and business is executing really well that we can do that not only today, but tomorrow. Thank you very much.

Operator

Thank you. The next question comes from Adam Holets with Bank of Montreal. Please go ahead.

Tianyi Jiang

Hi, thank you for taking the question and congrats on the good results. So I want to ask a question similar to Shrenik on AgentPulse, but I believe last quarter you mentioned you guys offered 2 different flavors of it, you know, 1 for visibility and 1 for more actionable items. I'm curious which 1 is resonating more with your customers? And then secondly, for that more advanced offering, how far down does it go to the path of execution? Thank you.

Tianyi Jiang

Great question. So AgentPulse allows the customer to discover agents and then bring them under management, and when there's outages, do remediation. And also importantly, we also support this concept of democratization of agent management, so keeping the human in the loop to bring about good governance posture changes in the enterprise. I think what you're referring to specifically is that previously we had included AgentPulse as part of the Control bundle, and in July we actually made AgentPulse a singular SKU that customers can just deploy as the first solution point. So that's going very well and we think that from a material perspective, it will go into next year. So that's where we will see the good growth impact. But overall, I think from a functionality perspective, it is our overall governance and management posture around AI.

Operator

Thank you. The next question comes from Todd Weller with Stephens. Please go ahead.

Unknown Speaker

Thanks for taking the question. TJ, on the resilience side of the business, can you give us an update on kind of the trends you're seeing there, growth trajectory, drivers, and then how much is being driven there from new customer acquisition versus expansion, whether that would be seats or upselling suites?

Tianyi Jiang

So yes, on the resiliency side, it's continued to be a very strong grower for us. Along with Control and Modernization, it's really we don't see massive seat expansion. So when we license, we do license based on the entirety of the enterprise employee count. So when we do see expansion from existing customers, it's really more features and SKUs that we upsell. So it's not so much from the seat expansion itself. But then new customers, it's 50-50. We not only grow from our existing customers from upsell and cross-sell of new products, but we also have new customers from new logos that contribute to our new ARR growth.

Unknown Speaker

And are you seeing any AI tailwinds on the resilience side of the house at this point?

Tianyi Jiang

Oh, absolutely. We think that part of the reason that resilience continues to be a very important aspect of our platform is that AI tailwind from potential risks and damages that processes operating at machine speed will do. So fundamentally, our history has been in the governance and management and protection of collaboration across human actors. Of course, now it's a step-up into agency, the agentic AI era where these things are operating at machine speed. Same guardrail, same framework. It just elevates up into the AI trust layer that we talk about, the AI Trust Platform.

Operator

Thank you. The next question comes from Rudy Kessinger with D.A. Davidson. Please go ahead.

Benjamin Smith

Hey, this is Ben Smith on for Rudy. Thanks for taking the question. And nice to see GRR hit the 90% target there. Where can we go from here? Is it possible that we see further improvement into the low 90s over the next year or so? And then I'd follow up as well.

Unknown Speaker

Thanks for pointing that out again so we can reiterate. Yes, I agree. We're really pleased with the attainment of that 90%. It's been our long-term target now for a couple of years to get to 90%-plus. So really pleased with the team's execution and being able to deliver that 90%. I don't think it's the end. I do think there's still room to go, room for improvement. Some of the things that we've been executing on the last couple of years, I still think there's a lot more we can be doing. So we do think there's improvement. We've invested a lot in technologies, as you know, we've talked about over the past couple of years. And we're going to continue to invest. You may have seen a recent press release that we put out where we added for the first time a Chief Customer Officer, again demonstrating our commitment to ensuring that our customers are receiving the absolute best from us and ensuring they get the full value of our platform and making sure they can take advantage of all we have to offer and solving as many of their challenges as possible. So we're going to continue to invest in this area. And I agree, I think we could definitely see a couple percentage point improvement here over the next couple of years.

Benjamin Smith

Great. Yes, it's nice to hear. And then I know we'll probably get an update at year-end, but is there any color that you can share on how the different suites did in Q3 and year-to-date so far?

Unknown Speaker

Yes, we're seeing really good growth really across all the suites. I mean, TJ just alluded to it. We think obviously that governance alone would be the headline, but I think the AI agentic world now is really focusing people's attention on not only what the agents are doing, but their ability to recover and the resilience that is needed. And so the conversation may start out talking about how are we governing agents, but it quickly goes into other aspects of what the customers actually need, which is really a holistic solution. And so that's been working really well. So we're seeing growth really across all 3 suites in the platform.

Operator

Thank you. The next question comes from Jason Ader with William Blair. Please go ahead.

Jason Ader

Yes, good afternoon. Thanks. So, TJ, your growth has been incredibly consistent. I'm talking about ARR kind of in that mid-20s range. What would it take for you guys to really accelerate that to, let's say, closer to the high 20s or even 30? Is it a specific SKU? Is it better cross-sell execution or something else that could get you there?

Tianyi Jiang

Thank you, Jason. Well, clearly we're also increasing some investments as we started talking about last quarter as well. The reason for that is we do see the market opportunity. We're leaning into that. So the launch of AI Trust Platform is another indication of how we're elevating our overall broader offering to lean into the market opportunity. You will hear more from us in the coming weeks and months as we refine that through Ignite and into next year. We also have an Investor Day coming up in March. Yes, effectively we are aiming to do exactly what you're referring to, to accelerate the growth here because the market opportunity is massive.

Jason Ader

Okay, and just to be clear, so you're saying that it'll likely be new AI products? Is that the right way to think about it?

Tianyi Jiang

So the bigger TAM is the AI trust opportunity that's layered on top of our data and infrastructure trust and governance. That's for sure.

Jason Ader

Got you, okay. And then 1 quick follow-up on the Microsoft relationship. And I'm not sure if you guys said this on the call. I might have missed it. But I think you've said that the $1 billion ARR target in 2029, that does not require a significant expansion into non-Microsoft workloads. Is that correct?

Tianyi Jiang

Yes, we always say that for our growth trajectory, we could do it both organically as well as even just focus on our current concentration. We'll be able to do that. Of course, we have more multi-cloud opportunities and now, of course, AI opportunities to lean into. So there's no shortage of growth opportunities here.

Jason Ader

And you still stand by the 30% of ARR long-term coming from non-Microsoft workloads?

Tianyi Jiang

Up to, yes. We still stand by that, yes.

Jason Ader

Okay, thank you.

Unknown Speaker

Yes, and I think as you were saying, Jason, that it could be up to 30%, but it's not a requirement for us to get to our $1 billion as TJ was alluding to, multiple vectors to get there. And again, we just have multiple paths to do that.

Operator

Thank you. The next question comes from Kirk Materne with Evercore ISI. Please go ahead.

S. Kirk Materne

Yes. Hi, TJ, I was curious to the earlier question on sort of competition with the cyber guys or the traditional cyber guys. I mean, you guys have always come at things from a data perspective, but as people sort of link data with security going forward, is that allowing you to tap into more traditional maybe IT security budgets? Do you see the conversations you have expanding perhaps more outside just the core IT folks into cyber? I was just kind of curious if the broader trends in terms of, you know, as agents come in and the importance of data is getting you into a little bit more of a cyber conversation than you've had previously.

Tianyi Jiang

Absolutely, that's actually a very intentional evolution. We historically talked to IT like you mentioned, but more and more we talk to security, talk to CISOs and information security officers because this is an evolving surface area. So what we do obviously is data-centric, but now extending to AI trust that very much has to do with security concerns. And of course, there are a lot of areas that we can expand into, including identity management. So it's a very exciting area. And we're very pleased to see so far the Gartner recognition, but also the industry uptake by both our customers and partners.

S. Kirk Materne

Great. And then just as a follow-up, you guys had some really nice momentum in your sort of bigger customers and some of the cohorts of your customers above $100,000. Do you feel like you're taking share there? Meaning, is the growth coming from just the activity level in the market, meaning, or do you feel like you're actually getting more people to kind of consolidate on your platform so it's not just sort of a market dynamic, you're taking share as well? Thanks.

Tianyi Jiang

There is definitely a consolidation dynamic here. Increasingly, we see customers prefer platform vendors. Also the trust of regional and international vendors when it comes to enterprise security in cloud, that's paramount. Even in last quarter alone, we've seen some vendors go through some challenges there. So there's platform, there's security credentials and credibility and track record. And also, now with AI really being mainstream, especially in autonomous agents, there's been a balancing of budgets, buckets that we can go after as well.

Operator

Thank you. The next question comes from Erik Suppiger with B. Riley. Please go ahead.

Erik Suppiger

Yes, thanks for taking the question. So it looks like you're still standing by your target for '29 for operating margins of 27.5%. Can you talk about the trajectory for getting there? And I think 2026 has been a year of investment. How are you thinking about 2027 in terms of planning how you are spending relative to '26?

Unknown Speaker

Yes, thanks, Erik, for the question. So, you know, obviously it's a little early to be providing guidance for '27, but I appreciate the question in terms of directionally. You're right, '26 was definitely an investment year, so we've definitely stepped back in terms of the operating income. Clearly we're still delivering very nice operating income, but below where we were last year. And that was intentional, an investment year. I do expect us still to be on the trajectory to get to those levels that we have projected for 2029. We've always said it wouldn't necessarily be a linear trajectory. And so, you know, I don't think I'm prepared yet to say exactly what '27 will be, but we do still feel committed to get to those levels by 2029.

Erik Suppiger

And then real quick, just following up on Jason's question. Can we still assume that Microsoft is still 90%-plus of where you're generating your ARR today?

Unknown Speaker

Currently, yes. That ecosystem is where we generate most revenue today.

Operator

Thank you. The next question comes from Derek Wood with TD Cowen. Please go ahead.

James Wood

Great, thanks. Congrats from me. Great to hear the acceleration in AgentPulse. I mean, is most of this coming from the Microsoft ecosystem or are you starting to see more activity in Google? And just like any specific verticals that are proving to be early adopters? And lastly, I mean, as you position more as an agent management platform, are you running into any different types of competitors?

Tianyi Jiang

Yes, great question. So definitely not only the Microsoft space when it comes to AgentPulse. The prepared remarks actually cited examples in the Google space as well, as well as Anthropic. So we cover that. And in terms of competitive landscape, this is a very fast-moving and growing area. So there are different vendors going at it from different angles, as I mentioned earlier. Some come from an identity management perspective. Some come from external aggressive attack defense perspective, where we are looking at effectively internal governance and collaboration guardrails where our governance graph essentially allows privacy and cyber teams to actually adhere to EU AI Act, NIST, ISO, and other different good governance controls around AI. So it really runs the spectrum. From an industry perspective, we have grocery chains, we have analyst firms, we have pharma as well as FinServ. So, and also what's really interesting is that we see the sales from a population perspective, we proportionally also come from enterprise as well as medium to small businesses, also North America, EMEA, and Asia. So we actually see demand from all the segments and all the geos we operate in, and that's very encouraging to see.

James Wood

Right. Jim, follow-up for you. I mean, so we're hearing, I think, the number of agents managed by existing AgentPulse customers doubling. I think that's the second straight quarter where that's been doubling. Just wanted to ask how that translates into revenue. I mean, does revenue scale with agents managed? Are there bands that customers are in and ultimately burst out of? Or are there not really any volume limits right now?

Tianyi Jiang

Right now we actually license them just like how we license the rest of our subscription software by number of employees in companies because we see that at these initial stages, it's important to not have any friction and also give customers that predictability of subscription costs. From a budgeting perspective, we see a lot of our CFO approvers much rather the subscription modeling versus consumption. So we're doing right now still seat-based licensing.

Operator

Thank you. The next question comes from Joe Vruwink with Scotiabank. Please go ahead.

Unknown Speaker

TJ, you talked about the AgentPulse opportunity, some big platform wins, talked about, you know, winning in larger enterprises. And I think you answered this question just a minute ago, but I guess is AgentPulse helping you get a foot in the door with enterprise customers that maybe historically have been a little bit harder for AvePoint to reach? Any commentary there would be helpful.

Tianyi Jiang

Yes, I mean this is actually the exact reason why we made it a unique SKU by itself, because we do see the demand there and opportunity to become a new tip of spear for us to get into new accounts.

Unknown Speaker

Okay, that's helpful. And then 1 for Jim. Last year we saw a really nice jump in 4Q '25 net new ARR. So if seasonality holds, it would seem like there's possibly some upside in 4Q. But any thoughts about what's baked into the 4Q guide would be great to hear.

Unknown Speaker

Yes, so you know, obviously when we exited Q2 and we provided guidance for the second half of the year, we were committing at that point to $60 million of incremental ARR. We just delivered $30 million. And at the time, I had said that it seemed like it was going to be roughly 50-50 between Q3 and Q4. We delivered 30. We're raising our guidance by $1 million for the full year. So, you know, we feel good about that in terms of it's an acceleration, which reflects the momentum we're seeing. But again, we feel good about that number. It would be really good growth for Q4 as well. As you pointed out, Q4 was really strong last year. Q3 was a little stronger this year than last year because of the public sector, but we feel good about the number we just put out for Q4, and we're excited about closing the year strong and delivering another set of really good results.

Unknown Speaker

Very clear. Thank you.

Operator

Thank you. And our last question comes from Steven Brunner with Northland Capital Markets. Please go ahead.

Steven Brunner

Yes, TJ, you mentioned that AgentPulse customers grew more than 50% Q-over-Q. Would I be wrong to assume that that is AgentPulse customers with Control bundles attached, given that AgentPulse wasn't split out until July?

Tianyi Jiang

That is correct. So majority of the pickup is the bundle, and the unique SKUs only started in July, yeah.

Steven Brunner

Is there any possible indication you can give us on how well the standalone has done in 3, 4 months since it's been split out?

Tianyi Jiang

Yes, from a pipeline perspective, it's very good. As I mentioned earlier in the Q&A, the actual net closing and revenue impact, we expect more to be next year. And you will hear a lot more from us as part of our AI platform launch to expand these capabilities in the coming months and quarters.

Operator

Thank you. This concludes our question and answer session. I would like to turn the conference back over to TJ Jiang, CEO, for any closing remarks.

Tianyi Jiang

Thank you for joining us today. I'll close where I started. Q3 was another strong quarter. Record net new ARR, FX-adjusted ARR growth accelerating once again, and our 90% gross retention target achieved, all while we continue to invest with conviction. But what excites me most is what's in front of us. Our customers are no longer asking whether to deploy AI agents. They're asking whether they can trust them. With strengthening demand for AgentPulse and the AvePoint Trust Platform, extending governance across the full AI estate, we're moving our customers from unmanaged AI to measurable trust. This is the largest opportunity in our company's history. We're investing to capture it. AvePoint is the AI trust company, and our conviction in the path to our $1 billion ARR target has never been stronger. Thank you again for joining us today, and we look forward to speaking with you more this quarter.

Operator

This concludes today's conference. Thank you for participating. You may now disconnect.

This live transcript is auto-generated without human intervention or review.