リングセントラル（RNG）2026年第3四半期決算説明会：AI ARRは4億ドル超、フリーキャッシュフロー見通しを上方修正

主な要点

• 2026年第3四半期のサブスクリプション売上高は前年同期比6%増の約6億5,300万ドルとなり、総売上高は5.9%増の約6億7,600万ドルとなりました。
• Non-GAAP営業利益率はほぼ200ベーシスポイント上昇し、過去最高の24.7%に達しました。GAAP営業利益率は320ベーシスポイント改善し、過去最高の8.0%となりました。
• フリーキャッシュフローは前年同期比21%増の1億5,700万ドルとなりました。RingCentralは通期のフリーキャッシュフロー見通しを6億2,000万〜6億3,000万ドルに引き上げました。
• 少なくとも1つのネイティブな有料RingCentral AI製品を使用している顧客は、総ARRの約15%（4億ドル超）を占め、前年同期比で倍増しました。この顧客グループはARPUが高く、純継続率は100%を超えています。
• AI Receptionist（AIR）の顧客数は22,000社を超え、前年同期比でほぼ4倍となりました。ACEは60%増の7,000社近くに達し、CEBの有料顧客数は前四半期比でほぼ50%増加して14,000社を超えました。
• 総負債は四半期終了後に同社の公表スケジュールを前倒しして10億ドルを下回りました。第3四半期末時点の純レバレッジは1.3倍でした。

主な財務実績

特定の国内繰延税金資産に対する5億4,500万ドルの評価性引当金の取り崩しは、Non-GAAP業績には影響しませんでした。経営陣は、同社が引き続き繰延税金資産を利用するため、短期的に現金支払税額に重大な変化が生じることは見込んでいないと述べました。

事業および業績

AIの導入が引き続き主な追加成長ドライバーとなりました。RingCentral AI顧客（少なくとも1つのネイティブな有料AI製品を使用している顧客と定義）は、総ARRの約15%（4億ドル超）を占めました。経営陣は、このARR基盤が前年同期比で倍増し、意味のある高いARPUとより強力な継続率をもたらしていると述べました。

AIRの顧客数は22,000社を超え、前年同期比でほぼ4倍となりました。同社のAI会話インテリジェンス製品であるACEは60%増の7,000社近くに達しました。RingCentralの軽量カスタマーエンゲージメントソリューションであるCEBは、有料顧客数が14,000社を超え、前四半期比でほぼ50%増加しました。CEB顧客のほぼ3分の2が、少なくとも1つのAI製品も購入しました。

経営陣は、AIRの従量課金制価格設定により、顧客の消費量増加に伴って追加の収益化が可能になると述べました。同製品は中小企業の間で最も強い初期の牽引力を獲得し、導入は中堅企業や大企業顧客へも広がり始めています。AirProはより複雑で複数ステップのワークフローを対象としており、完全なコンテキストを維持したまま人間のオペレーターに対応を引き継ぐことができます。

RingCentralは、AI導入の最も強力なバーティカル市場の機会としてヘルスケアと金融サービスを挙げました。同社はまた、16,000社を超えるチャネルパートナーおよびリセラーのネットワークを通じてさらなる潜在性を期待しています。当四半期中、同社はBT BusinessとのAIパートナーシップを拡大し、ニュージーランドのDigital Islandおよびイタリアの新たなパートナーを追加しました。

同社は年間約400億分の通話時間、140億件の通話、40億件のテキストメッセージを処理しています。経営陣は、この通信量、キャリアグレードのネットワーク、および統合された音声、コンタクトセンター、AIプラットフォームが、顧客対応全体により多くのAIエージェントを展開するための基盤になるとみています。

業績予想（ガイダンス）

2026年度について、RingCentralは売上高、利益率、フリーキャッシュフローの見通しを引き上げました。

2026年第4四半期について、経営陣は以下を予想しています。

経営陣は、GAAP営業利益率20%という中期目標を改めて表明しました。また、Non-GAAP収益性の向上と過去の株式報酬の付与による影響の減少に支えられ、今後3〜4年間でSBCを売上高の3%〜4%に抑えることも目標としています。

リスクと注目点

• 経営陣は、急速な導入拡大にもかかわらず、全体のAI浸透率およびRingCentralネットワーク上でAIが応答する通話の割合は、依然として非常に小さいと説明しました。
• AIは売上成長への寄与度を高めていますが、サブスクリプション事業全体に比べると依然として小規模にとどまっています。
• チャネルパートナーやグローバルサービスプロバイダーを通じたAIの導入は、依然として初期段階にあり、直接導入よりも一般的に進展が遅くなっています。
• RingCentralは最先端のAIモデルを使い続けています。経営陣は、オープンソースモデルや推論効率の向上による潜在的なコストメリットを見込んでいますが、AIの売上総利益率を維持できるかどうかは、価格設定、モデル選定、インフラの最適化にかかっています。
• 同社のコンタクトセンターにおけるパートナーシップは今後数年間継続する一方で、RingCentralは中堅企業および大企業顧客向けの独自製品の開発を継続しています。

アナリスト質疑応答のハイライト

アナリストは、サブスクリプション成長の改善が中核であるUCaaSの安定性を反映しているのか、それともAIによる寄与の拡大を反映しているのかに焦点を当てました。経営陣は、この傾向を持続的なリカーリングレベニューの中核、健全なプラットフォーム利用、堅調な顧客需要、および有料AI製品からの貢献増加によるものと説明しました。

AIの収益性に関して、経営陣はAI製品の売上総利益率が概ね会社平均と同水準であると述べました。価格設定は顧客価値と使用状況に基づいて設計されており、コストはモデル選定、推論効率、および共有プラットフォームインフラを通じて管理されています。

資本配分に関して、RingCentralは製品および技術への年間2億5,000万ドル超の支出により、自社の製品ロードマップを内部資金で賄うことができると述べました。経営陣は買収に対してオープンな姿勢を維持していますが、現在の計画実行が非連続的な成長（M&A）に依存していないことを強調しました。同社は第3四半期に8.5%のハイイールド社債のうち1億1,000万ドル以上を償還し、約1,800万ドルの自社株買いを実施し、1株あたり0.125ドルの四半期配当を支払いました。自社株買いの授権枠には3億ドル以上が残っています。

経営陣は、今後のAI拡大は、顧客セグメント全体におけるより広範な導入、ヘルスケアおよび金融サービスにおける浸透の深化、チャネル販売の追加、ならびにAIRおよびAirProによって処理されるより複雑なワークフローからもたらされると述べました。

決算説明会文字起こし全文

決算説明会の完全なトランスクリプト

経営陣による説明

Operator

Good day, and welcome to the RingCentral Third Quarter 2026 Earnings Conference Call. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Steven Horwitz, VP of Investor Relations. Please go ahead.

Steven Horwitz

Thank you. Good afternoon, and welcome to RingCentral's Third Quarter 2026 Conference Call. Joining me today are Vlad Shmunis, Founder, Chairman and CEO; Kira Makagon, President and COO; and Vaibhav Agarwal, CFO.

Our remarks today include forward-looking statements regarding the company's business operations, financial performance and outlook. These statements are subject to risks and uncertainties, some of which are beyond our control and are not guarantees of future performance. Actual results may differ materially from our forward-looking statements, and we undertake no obligation to update these statements after this call. If the call is replayed after today, the information presented may not contain current or accurate information. For a complete discussion of the risks and uncertainties related to our business, please refer to the information contained in our filings with the Securities and Exchange Commission as well as today's earnings release. Unless otherwise indicated, all measures that follow are non-GAAP with year-over-year comparisons. A reconciliation of all GAAP to non-GAAP results is provided with our earnings release and in the slide presentation which you can find at the Financial Results section at ir.ringcentral.com.

With that, I'll turn the call over to Vlad.

Vladimir Shmunis

Good afternoon, and thank you for joining us. We delivered another strong quarter in Q3 with subscription and total revenue above the high end of our guidance. We also achieved record GAAP and non-GAAP operating margins. Free cash flow generation remained strong, which enabled us to reach an important milestone of reducing gross debt to under $1 billion as of today, which is ahead of schedule. These results reflect RingCentral leadership in agentic voice AI, healthy demand and our continued focus on expanding operating margins through disciplined execution. We're building on this momentum by investing in innovation that strengthens our market position and brings AI and human agents together on a complete integrated customer engagement platform.

Our strong momentum with agentic voice AI has been recognized by Business Insider, which named RingCentral to its America's Most Innovative Businesses list. Innovation has always been a driving force behind RingCentral's success. We pioneered the shift from legacy communication systems to the cloud. The mega trend of AI represents an even larger opportunity and the powerful tailwind for RingCentral. We're leading this next transformation with agentic voice AI, making customer interactions more efficient, intelligent and actionable.

Our AI results provide strong proof points for our building success. AI Receptionist or AIR now has more than 22,000 customers up nearly 4x year-over-year. A, our AI conversation expert is approaching 7,000 customers, up 60% year-over-year. And our newest lightweight contact center offering, CEB or customer engagement business solution has reached more than 14,000 paying customers. Growing nearly 50% quarter-over-quarter.

Nearly 2/3 of CEB customers have also purchased 1 or more of our AI products. Importantly, RingCentral AI, which we define as customers using at least 1 native paid AI product now represent approximately 15% of total ARR or over $400 million, doubling year-over-year. This cohort of customers also has meaningfully higher ARPU and net retention. These trends demonstrate how AI is expanding wallet share, strengthening retention and creating a more durable foundation for profitable growth. The momentum we're seeing is rooted in our well-established position as a leader in consumer to business communications.

Voice and SMS remains central to how consumers engage with their business providers, such as doctors, bankers, insurance agents and the host of other service providers. RingCentral is a leader in consumer to business communications with the platform carrying approximately 40 billion minutes, 14 billion calls and 4 billion text messages annually. Importantly, our customers can utilize the bridge context in their own conversations to generate actionable insights, automate tasks and drive resolution through the use of our AI products. Our carrier-grade network, global scale, security and decades of operating expertise in invoice create an opportunity for us to become a critical infrastructure layer for business communications and is a significant barrier to entry.

Our ability to connect consumers with their service providers at scale well positions us to deploy AI agents across all stages of our consumer business interaction. Whether it be human to human, human to AI or human AI and bad to human, otherwise known as a human in the loop. For a little more color, AI agents can handle routine inquiries, automated workflows and resolve issues autonomously. When judgment, empathy or specialized expertise is required, a human can step in.

In health care, for example, AI can help manage routine interactions while bringing in a qualified professional when a patient needs medical advice. Another strong competitive moat for us is our distribution reach. Alongside our direct sales force, we have a global ecosystem of more than 16,000 channel partners and resellers including some of the leading global service providers such as AT&T, British Telecom, Charter Communications, Optus, Telis and Vodafone, amongst others. Our GSP and channel partners are increasingly incorporating our AI portfolio into their product lines. Their established customer relationships and local expertise create a distribution advantage that is difficult to replicate.

Let me close with 3 key points. First, RingCentral has a deep and defensible mode in a large and expanding market. Our global voice infrastructure, billions of calls, text messages and minutes that we process every year, and integrated customer engagement platform provides a distinct foundation for strengthening our leadership through Agentic voice AI. Our partner ecosystem gives us the reach to scale these capabilities globally.

Second, our investment in agentic voice strategy are delivering tangible results. Growing paid adoption, broader use of our AI portfolio and stronger customer economics demonstrate how AI is expanding the value of our platform and accelerating our business.

Third, our large and growing TAM, combined with disciplined execution and operating model leverage is resulting in a strong and improving financial profile. We are growing revenue, expanding profitability and generating substantial free cash flow and free cash flow per share with our gross debt now below our stated $1 billion target, we now have the balance sheet strength to continue investing in innovation while also delivering sustained long-term shareholder value via share buybacks and dividends.

I want to thank our employees for their innovation and execution, our customers and partners for their trust and our shareholders for their continued support. I'm excited about the opportunity ahead and confident in our ability to deliver. With that, I will turn it over to Kira.

Kira Makagon

Thank you, Vlad, and good afternoon, everyone. As Vlad highlighted, RingCentral brings together a complete and integrated customer engagement platform with agentic voice AI and humans working together. Our leadership in voice puts us at the heart of consumer-to-business conversations and gives RingCentral AI the foundation to turn those conversations into better outcomes. We are building on that foundation to meet customers wherever they work, including the elements they use every day. This quarter, we introduced plug-ins for open AI, ChatGPT and MCP connectors for [indiscernible]. We were also selected as a cloud for small business launch partner. RingCentral provides the conversation context and the AI models provide reasoning. It's great to see how our innovation is helping our customers turn conversations into action as part of their everyday workflows.

In less than a month, we are already seeing early traction with thousands of active customers. [indiscernible] in humans and AI to work together is core to our strategy. and where we are uniquely positioned because we bring voice, contact center and AI into 1 platform, so AI can either handle the entire task or hand off to humans as needed. For example, an industrial parts manufacture and supplier replace its fragmented copes with AI CEB, our customer engagement business solution. Bringing advanced copus SMS, live agent tracking and virtual cobot onto 1 place.

Customers now spend less time on hold get issues resolved in fewer handoffs and report higher satisfaction, which means lower cost to serve and stronger retention. They are also leveraging ACE, our AI conversation expert for rough score caloric and coaching. Plus their deploying AIR are AI exceptions for handling inquiries with context aware routing. We see this in action with customers across all segments, including 1 million TCV deals, such as our rapidly growing national business operator with millions of members.

Already using [indiscernible], this customer added ACE to identify key coal drivers than AIR to handle a high volume of recurring increase. This illustrates our AI flywheel. ACE uncovers opportunities and AIR automate interactions, giving cause immediate answers while freeing staff to build a number of relationships, lead clock tours and support membership roles.

As they recognize value, they extended their investment with us by 4x. This past quarter, we also saw strong adoption of AirPro, AirPro extends the value of our agentic voice AI, delivering was first omnichannel AI agents that autonomously complete multistep workflows and bring in human agents when judgment and embassy are needed. As an example, 1 of the largest technology equipment and procurement providers for senior living in the United States already uses in CX and ACE. This quarter, they added Airport to help them take their customer experience to the next level. Airpro understand intent and automates more complex workflows. When a customer holds Airpro verifies the color's identity measures the caller to their record and identifies the reason for the call. That takes action. When needed, Airflo hands off to human with full context. And then follows up. In just the first month, this customer achieved an over 95% accuracy rate in end-to-end workflow completion.

We're also delighted to see our existing customer destination Pet, a nationwide pet care provider extend their use case of our AI. They started with ACE to analyze cold transcripts and identify opportunities to improve service and recently expanded to to automate workflows around scheduling appointments funding confirmations and updating records for customers across locations. This supports destination pet goal of capturing every opportunity, improving customer experience, increasing revenue and lower labor costs. When our solutions work together as they do for these customers, it creates a virtuous loop or items conversation context into action. That's the value of the essentials and agentic voice AI. It's changing the game for businesses, and that is something we're really excited about.

Speaking of changing the game, today, we announced the expansion of our decade-long partnership with Golden State Warriors to be named the official agentic voice AI platform for fund engagement. Together, the 2 organizations will work to elevate plan experiences from ticket employees to gain the equations and ongoing engagement. We continue to expand our Agentic voice AI solutions to more businesses around the world through a network of global service providers.

This quarter, we expanded AI partnership with BT business. This expansion enables BT to customers to seamlessly layer native intelligent automation over their existing telephony setups. Together, we're empowering the [indiscernible] organizations to capture lost revenue and drive operational efficiency. Additionally, we added 2 new partners to our GSP portfolio Digital Island, our first partner in New Zealand will sell, implement and support our agentic voice AI.

In Italy, [indiscernible], our first partner there will bring Index with the full suite of RCAI agents onto its network. These partnerships are key to our overall go-to-market strategy and international reach. In closing, our customer wins show our vision in action. [indiscernible] and AI working together onto 1 platform to turn conversations into better outcomes. Growing AI adoption and expanding global reach, we're bringing that opportunity to more businesses around the world.

With that, I will turn it over to Vaibhav.

Vaibhav Agarwal

Thank you, Kira, and good afternoon, everyone. Q3 was a solid quarter with key metrics above the high end of our guidance, reflecting continued execution against the financial framework and commitments we have outlined. Our core business remains resilient with growing contributions from our AI portfolio. Just as importantly, we are expanding margins and free cash flow, reducing SPC and making meaningful progress towards our target of 20% GAAP operating margin and steady state SBC of 3% to 4% of revenue. We also reached an important milestone reducing gross debt to below $1 billion as of now, significantly strengthening our balance sheet while continuing to return capital to shareholders through share repurchases and dividends. This combination of growth, improving profitability and disciplined capital allocation is translating into stronger free cash flow per share growth.

With that context, let me turn to our third quarter performance. Starting with growth. Subscription revenue was approximately $653 million, up 6% year-over-year, while total revenue was approximately $676 million, up 5.9%. Underlying customer trends remain healthy, with steady new customer additions and monthly net retention above 99% and underscoring the mission-critical nature of our platform.

Notably, our RCAI portfolio continues to gain traction, which reinforces the role of AI, not only as an incremental growth driver but also as a meaningful contributor to strong customer economics. We believe these trends, combined with our established competitive position provide a solid foundation for durable growth over time. Turning to profitability. We delivered another quarter of record GAAP and non-GAAP operating margins alongside further reductions in SPC. Subscription gross margin remained strong at above 80%. Non-GAAP operating margin reached 24.7%, up nearly 200 basis points and SBC as a percent of revenue declined 50 basis points, reflecting our continued discipline around equity compensation. The combination of stronger non-GAAP margin and lower SBC drove a record GAAP operating margin of 8%, an improvement of 320 basis points. GAAP EPS was $6.50 per share, which included a noncash benefit of $545 million from the release of a valuation allowance against certain domestic deferred tax assets. This adjustment was driven by an improvement in GAAP profitability.

The valuation allowance release had no impact on non-GAAP results and we do not expect material changes to our near-term cash taxes as we continue to utilize deferred tax assets. Excluding this benefit, diluted EPS was $0.28 and an improvement of 47% year-over-year. Non-GAAP EPS increased about 18% to $1.33. Our margin expansion trend is driven by multiple structural drivers such as strong gross margins, operating leverage, disciplined cost management broader internal adoption of AI and a focus on our highest return investments. Together with SPC reduction, these drivers reinforce our confidence in long-term sustainability and durability of margin expansion while investing in growth.

Turning to cash flow. We generated $157 million in free cash flow in Q3, up 21% year-over-year. Our recurring revenue model, strong gross margins and improving operating efficiency continue to translate into substantial cash generation. Reflecting this performance, we are raising our full year free cash flow outlook to $625 million at the midpoint representing nearly 24% of revenues. This represents over $7 of free cash flow per share, up about 23% year-over-year. We believe the durability of our recurring revenue base, AI tailwinds, structural margin expansion, disciplined capital allocation and continued reduction in provide a strong foundation for sustained free cash flow generation over time.

Turning to capital allocation. In Q3, we prioritized repayment of our highest cost debt the 8.5% high-yield notes that became callable reducing gross debt by just over $110 million and bringing net leverage to 1.3x. In fact, as of today, our gross debt is now below $1 billion. As a reminder, there are no debt maturities until 2030, and we have $305 million of undrawn credit capacity. We also continue to return capital to shareholders, spending approximately $18 million on share repurchases during the quarter and paying a dividend of $0.125 per share. Diluted share count declined about 5% year-over-year to approximately 87.7 million shares. We have over $300 million remaining under our share repurchase authorization. Our substantial cash generation and our stronger balance sheet gives us the flexibility on capital allocation while maintaining a consistent focus on long-term value creation.

With that, let me turn to guidance. For fiscal 2026, we are raising subscription revenue to be 2.563 billion to 2.569 billion representing growth of 5.6% to 5.9% raising total revenue to $2.650 billion to $2.657 billion. representing growth of 5.4% to 5.6%. GAAP operating margin of 9.2% to 9.6% and expanding by 460 basis points, raising non-GAAP operating margin to 24% to 24.2% in raising free cash flow to $620 million to $630 million, up 18%. SPC in the range of approximately $240 million to $245 million, representing 9% of revenues, fully diluted share count of approximately 87.5 million shares, raising non-GAAP EPS to be between $5.07 to $0.52, up about 17%. This results in free cash flow per share of $7.09 to $7.20 per share for the year, up 23%.

For Q4 '26, we expect Subscription revenue of approximately $653 million to $659 million, total revenue of approximately $673 million to $680 million. GAAP operating margin of 13.3% to 14.7%, up 740 basis points, non-GAAP operating margin of 25% to 25.5% and up 250 basis points, non-GAAP EPS of $1.32 to $1.37, up 14% and fully diluted share count of approximately 89 million shares.

As we look ahead, our execution is delivering tangible results with a durable core business growing momentum across our AI portfolio and continued improvement in profitability and cash generation. Together with a stronger balance sheet and disciplined capital allocation. We believe these fundamentals position RingCentral to compound free cash flow per share and drive meaningful long-term shareholder value.

With that, let's open the call for questions.

Operator

[Operator Instructions] The first question comes from Elizabeth Porter with Morgan Stanley.

質疑応答

Elizabeth Elliott

Congrats on the quarter. Subscription revenue growth has been mostly improving for several quarters now, and the beat was more pronounced in Q3? So could you just help us unpack what drove the upside in the quarter? I'm curious if there were any specific drivers to call out whether it was any stronger AI-driven demand or potentially starting to see some easing of those enterprise renewal headwinds you guys briefly discussed.

Vaibhav Agarwal

Yes. Thanks, Elizabeth, for the call out. And look, the beef was driven by product execution across -- we had a solid quarter. We delivered results on top line and the bottom line. Fundamentally, we are executing discipline consistent results on all our commitments. And as you called out, we are showing the sustainable and a stabilized growth profile. And that's what's getting reflected in our full year guide, basically reflecting the confidence in the underlying growth drivers, particularly on the top line, our core business continues to be a scale recurring revenue business that is durable supported by strong retention metrics and steady customer demand usage on the platform continues to be healthy, and it just underscores some mission-critical nature of our solutions.

And on top of the core, we are seeing growing contributions from our AI-led products. As we called out in our prepared remarks, adoption is strong across the board. We are seeing strong customer engagement, customers who are using 1 plus paid AI products are now approaching 15% of ARR doubling year-over-year. And within these cohorts very importantly, we see strong customer economics in terms of increased ARPU and over 100% net retention rate.

So net-net, it reinforces both the monetization opportunity as well as it demonstrates the stickiness of the base. And very importantly, we are not only growing, but we are also very focused on the quality of the growth. So we delivered record GAAP and non-GAAP operating margins and a strong free cash flow profile. So net-net, we have a recurring model. We have a diversified customer base from small to enterprise customers -- our solutions are mission critical. We have a deep competitive mode test built on platform scale, the depth of our product and differentiated GTM. And these fundamentals are frankly contributing to the durability and the sustainability of growth and our ability to expand both margins and free cash flows.

Elizabeth Elliott

Great. And then just as a follow-up, I wanted to hit on capital allocation. It was great to see that the gross debt target was ahead of schedule, and you have free cash flow growing faster than revenue. And when we think about the sort of opportunities, how should we think about the case for increasing any sort of organic investment to go after AI and contact center relative to pursuing acquisitions or prioritizing further capital returns?

Vladimir Shmunis

Yes. Elizabeth. Here. Look, this question comes up fairly regularly. Look, fundamentally, we are a very tech-heavy organization. we view ourselves as an innovator at heart. I think as you all know, my background is quite technical. Many of the original contributors, including the CTO, still with us, which I think is not that common. So we feel we have a lot of talent now in-house. And our commitment and spend on product technology continues to be quite healthy over $0.25 billion per year. So we can do a lot ourselves.

Now having said that, we're always looking out opportunistically. People are calling us. We do have an M&A function in-house, and we are evaluating multiple opportunities always. And I think clear as a bar, we consider them and then we move forward. So I'm not going to tell you that we will never do a sizable acquisition. But I'm also saying that we're very confident in our ability to deliver on our road map using capacity we have at hand. So we'll just have to see what the future brings. But as you can see for now, we're able to deliver the results without having to grow outside.

Operator

The next question comes from Siti Panigrahi with Mizuho.

Sitikantha Panigrahi

Congrats on a good quarter. Just wanted to ask you about the new product ARR that you guys talked about $100 million run rate exiting $25 million. How fast that's growing in Q3? And how should we think about that exiting this year? And I have a follow-up.

Vaibhav Agarwal

Yes. Siti, thanks for the question. This is Vaibhav. So in terms of the ARR for new products, the metric we provided this quarter was customers adopting at least 1 paid AI product or as we call it RCAI. So that's approaching 15% or $400 million. and it's doubling year-over-year. And as I mentioned in my previous comment, that's -- the customer economics there are very healthy. ARPUs are higher, net retention rates are above 100%. And -- and the reason that we are focused on that metric, and we manage the business, frankly, based on that metric is it's a very -- it's the clearest proof point of AI adoption and penetration, both into the base as well as with new sales.

And what that metric is clearly showing is that there is healthy customer demand. customers are moving beyond experimentation. They are getting tangible ROI benefits from the use of AI, and it's leading to the base being more sticky. So over this metric has been continuously growing, and we provided an update on this every quarter. So our expectation is that as adoption continues to grow, -- and as customers adopt more AI products that this metric will continue to grow.

Sitikantha Panigrahi

That's correct. Just on the margin side, I think what's impressive is your subscription margin still holding well despite AI investment. Can you help us like how is that -- how should we think about the gross margin as you ramp on AI?

Vaibhav Agarwal

Yes. So look, overall gross margins, as you called out, are strong at 80% plus. And our gross margins on AI products are generally holding in line with the corporate average as well. And really, we are optimizing it with 2 levers there. One is pricing and the other 1 is cost. So on the pricing front, relative to AI, we are being very deliberate in our pricing strategy. The idea is to kind of charge customers for the value and the usage that they receive. And customers are willing to pay because they are getting tangible ROI benefits both from a revenue standpoint as well as a cost avoidance or cost savings standpoint. And that's -- the uplift is indicated in the metric that I just talked about.

In terms of the costs, we are optimizing against model selection, call it, inference efficiency and infrastructure. So in terms of inference, it's getting better with different models, the price per unit is coming down. In terms of model selection, we are currently using frontier models. And over time, we will benefit from the use of open source models. And a very important point is that these AI products are being delivered through the same platform and engine -- and therefore, we get very good incremental unit of economics. So I think overall, our expectation is that gross margins will continue to hold at around 80%.

Operator

The next question comes from Brian Peterson with Raymond James.

Brian Peterson

So I wanted to hit on the AI adoption a bit. I'm curious if there's any part of the customer segment, whether that's enterprise or U.S. or SMB that's really kind of embracing some of that portfolio? And where do you think that next unlock may come as we think about that expanding beyond the mid- to high teens number you just gave us.

Kira Makagon

This is Kira. So EV adoption across all segments in -- from enterprise to customers and for all our products, our ARs and. So Brian, this is really -- we're happy to see that and show the across the board. Move to the second part of the question.

Brian Peterson

Just as we think about where the incremental adoption is coming from, is there a certain part of a geography or customer base or maybe even thinking about some partner channels where you could see that leverage kind of expanding beyond the number you just gave us?

Vladimir Shmunis

Yes, I can say that out here. Yes, very good question. Look, so we've been on record with our what we call golden verticals, which is fundamentally health care and financial services. So as we continue, we are seeing more AI within those sectors that were stronger for us to begin with. And again, as we've been saying for quite some time now, our best traction is in consumer to business interactions. Many of them, most are by voice and text. And we are very strong in both of those channels, moving billions and tens of billions of minutes through our platform.

Unknown Executive

Okay. So that as far as verticals is concerned. On the partnership side, it's also very interesting because we are seeing increasing adoption of our AI portfolio. by the partner community and in particular, global service providers. And there were some releases relatively recently, for example, with British Telecom, some others. So when we say -- as a maybe a word of caution, when we say that it's still pretty early, very early with AI in general, for this type of an application. It is yet earlier through the channel and in particular, with GSPs, it just moved a little slower. But once they move, they move was very strong for us. So we're pretty optimistic that as we get our formula down, that we'll be able to propagate it through our channel, which, as you know, is large and unique.

Brian Peterson

And Vaibhav, maybe just a follow-up. I know the margins have been ramping up. Can you remind us just from a GAAP perspective, -- how are you thinking about goals like stock comp and targeting that? And how should we be thinking about kind of the non-GAAP and GAAP operating margins going forward?

Vaibhav Agarwal

Yes. Thank you, Brian. So look, we are very pleased with the work that we've done on both operating margins from a GAAP standpoint as well as a non-GAAP standpoint over the years. So we laid out a medium-term target of reaching 20% in GAAP operating margins and we remain confident in our ability to get there. And the proof point there is if you look at our guide for 2026, we are expanding GAAP operating margins by about, call it, 400 basis points or so. And about half of that is coming from improvements in non-GAAP operating margins and the other half is coming from BC. So call it, 200 basis points each. And if you were to project this out into the outer years, that's how we remain confident in our ability to reach the 20% target.

In terms of the non-GAAP operating margin, look, we have a lot of structural drivers behind that. We have a scale recurring revenue model that producing gross margins at 80%, which is industry-leading. So we inherently get a lot of leverage in our operating model, and that, coupled with the discipline in cost across the board that we have is resulting in material improvements in non-GAAP operating margins. And then on the SBC side, we are continuing our discipline in terms of managing grants, managing SBC. And what's flowing through the SBC model now is also the prior grants as well as the new brands and as the impact of the prior grants that were given at higher prices as that kind of tails off, that gives us the confidence to continue to reduce SBC over time. And that's where we've laid out a target of getting to 3% to 4%. And in the next 3 to 4 years. So net-net, it's a combination of expansion in non-GAAP operating profit as well as reductions in that allowed us to kind of make those targets, and we remain on track to achieve that.

Operator

The next question comes from Peter Levine with Evercore.

Peter Levine

Maybe the first 1 I have, if you think about, I guess, the trajectory of subscription revenue growth into '27, I know you haven't guided, but how should we think about, I think, distinguishing between, I think, the underlying performance of like the core UCaaS business versus how much of that is coming from like newer AI, some of your customer engagement -- customer engagement products. So I'm just trying to understand like how much growth is dependent on like upsell, cross-sell of AI versus stabilization in your core business?

Vladimir Shmunis

I think we've been sharing this before that it's roughly 50-50 from the base and the other half from new customers. And we think we will always also saying that as new customers, also about half of them take AI and the other ones. Do not yet take AI, but they may well be upgraded across all to AI over time. But for now, those as I mentioned.

Peter Levine

And then with the AI reception -- I'm sorry, go on, please.

Vaibhav Agarwal

Yes. Peter, just to add up on that, I think what Vlad was mentioning is the contribution from -- on bookings from the AI product. In terms of 2027, look, we are not providing formal guidance here today. But I think the underlying growth drivers would be similar to what I talked about before, which is the core business continues to be durable margins are high. It's a recurring revenue base and the retention metrics there are strong and AI will be a layer on the top, and that is growing rapidly today. And it's becoming an increasing contributor to growth. The thing is that it's growing off a smaller base relative to the size of the overall subscription business. But net-net, we are seeing strong adoption to Vlad's point in customer engagement across the platform. And we talked about the margin expansion, the progress there in free cash flow and we expect to continue to kind of do a lot of the same.

Peter Levine

That's a good segue. My other question is AI reception has gained a lot of traction. How do you think about the opportunity, I guess, then to expand monetization within that installed base? Like what are other avenues usage or use cases, pricing mechanisms that you think you could use with an AI Receptionist or even AIR?

Vaibhav Agarwal

Yes. So thanks for the call out on the momentum in AIR. So it now has over 22,000 customers. So the reason it's kind of seeing this momentum is the ROI that we hear from customers is real. Customers are able to address miss calls and they're capturing board leads, which is resulting in more revenue. So from a pricing standpoint, Peter, it's usage-based pricing. So the more customers use the product, the more they consume it, it kind of contributes to revenue -- it's currently selling, I think, in a big part in our small segment, small business segment. And over time, it's also kind of going up market to the mid-market and the enterprise customers. So that's 1 kind of kind of way to think of it.

In terms of the use cases, customers are looking for more and more AI agents. Air is a form of an AI agent that helps customers address missed calls -- and over time, the idea is for air to kind of handle more sophisticated use cases and workflows over time and any great into more kind of platforms and workflow to do more and more, if you will.

Kira Makagon

And as we just came as we mentioned in our prepared remarks, we have air product addresses more complex use cases. those more complex use cases usually have more workflow attached to them and take on sort of more turns in a conversation. So they're more autonomous. And the big thing for our AI, whether it's air, AirPro, any of our products, that we think -- as we think about, we think about human in the loop, we think about handling as much as possible by AI, but also making sure that for judgment sentiment or more complex questions and required, those are handled onto human B and then they work together and preserving that context.

Our customers are telling us is very important. We provided some of those scenarios in our prepared remarks of our customers are finding that it's especially valuable to be able to do the swarm handoffs context and if necessary, take it off from human then back to ever. You kind of have this virtuous continuous loop -- and those create additional use cases for us, and we see those as expansion opportunities as customers get more and more comfortable adopting AI for what initially we went to market with AI exception known the schools, which continues to be a big driver because business is doing the school, miss opportunities, but then it gets into handoffs was in the business and more involved use cases such as handling more complex billing increase in terms appointment settings that require dips into databases, knowledge bases.

So those are being now handled with Airpro and the use cases will continue as we see from our customers. So that will continue to go -- they're almost growing think of it as incremental 1 by 1 as customers are realizing value with our products.

Operator

[Operator Instructions] The next question comes from Ryan MacWilliams with Wells Fargo.

Ryan MacWilliams

Vlad, I thought your comments on the voice potential for ring was interesting. As AI agents become more of the inbound call volume for businesses, how do you think RingCentral Air benefits from businesses 1 thing like an AI line of defense from inbound calls from other AI agents? Like AirPro seems to get to sit here like let Air talk with a use agent or another voice the agent and the human calls, then you can flip it over to another heating the chat. Like how do you think like the soon to be increase in voice AI means for the demand for your proxy.

Vladimir Shmunis

Yes. So yes, if I understand the question. Look, so there are 2 fundamental growth drivers for RingCentral. One is more goals and the other 1 is more AI. So they tend to go together. That's the good news. So the more close through the system, the more AI, we have a chance to apply. And the larger the percentage of AI-assisted calls on the network, the better off we are because obviously, we're able to extract better value -- deliver better value, but also extract better value for intelligence and insights that AI brings. So they work together. I can tell you that our traffic as growth has been strong, but linear. Our AI growth has been explosive obviously.

But overall penetration of AI and AI answer calls on our network is still very, very small percentage wise. So we think that there is a very meaningful opportunity. to simply, if you will, in the very simplest terms, help our customers assign AI to more of their traffic. And this is very much front and center what our teams are working on as well as obviously increasing overall traffic, adding logos, new better engagement with the base in general.

Operator

The next question comes from James Fish with Piper Sandler.

James Fish

Just wanting to work off of the last couple of questions here. Really, what I'm trying to understand is what you guys view as the opportunity in terms of the ability to cross-sell CB, ACE and AIR in terms of where we could see that penetration go over the next 2, 3 years? And you kind of alluded to this in the 1 answer, but how does penetration differ between sort of major customer accounts versus the more SMB commercial customer base.

Vladimir Shmunis

So I'll start with the second part of the question. CEB is fundamentally a smaller business product. because it is basically an informal contact center and larger enterprises tend to have formal contact center. -- they would be more right to buy our Rig CX product, which is a full up contact center, okay? Now how do we deliver it to them, each and every way we can. But fundamentally, certainly, our sales teams are well aware we have -- look, it's been growing really, really well for us. That means that we have a lot of satisfied customers and lots of reference cases. So our sales teams are able to leverage that. And of course, we are doubling down on product-led growth, PLG, and that is showing results as well. So we are also working on ways on specifically explaining to customers what CEB can do for them. And in as much as possible, we try to make them understand that RingCentral does not need to be sought off as a cost center, just a cost center, but we can actually help them generate revenues, capture more leads, do better leads processing CBD that and, in particular, the flavor of CEB that we sell with AI. That really is our most explosive new offering this year.

Operator

Our next question comes from Andrew King with Rosenblatt Securities.

Andrew King

Just wanted to touch base quickly on 2 of your partnerships nice obviously began reselling Ring in July. Any early pipeline wins to share to call out there. And then also, if you could just touch base on the ACO transition, if you're seeing any increased churn as you go through that transition, that would be great.

Vladimir Shmunis

Yes. So I think the second question is no. It was very orderly. Remember, those customers -- the experience does not change. It's just a small tweak in the logo in the upper left corner, and that's about it. Not about that is actually -- so there is no friction whatsoever. So I'm not aware of any churn related to that. As far as partnerships in general, yes, look, definitely there is nice. That's kind of the new. And it's early. And remember, they're very much kind in enterprise. So we don't expect many logos there just because they don't have that many logos basically.

But we are coping for friction for traction transition traction with some whales through that. Rest of the channel is performing, it's already in the up as you know, various folks are doing channel checks on the industry and us in particular. And centers pretty up at this point. I think people understand that voice is here to stay in business. And as long as voice stays in business, RingCentral stays in business because we are the voice leader.

Operator

The next question comes from Shebly Seyrafi with FBN Securities.

Shebly Seyrafi

So related to the [indiscernible] partnership, what's your plan for RingCentral contact center when the 1-year [ NICE ] extension ends. And it seems to me that 1 year is rather small. Why wouldn't you have like a 3-year extension or something like that?

Vladimir Shmunis

So first, I'm not sure where you got the 1 year from, I don't think that we've ever disclosed any specifics. But just as it may and they're not suggesting that is 1 year base. But in any case, it's an extension to the whatever term was there to begin with. So we still have very much multiple years left on this contract. But to fast forward a number of years, look number of things. This partnership could continue being extended as it has been for the last, I don't 12 years.

But also very importantly, our [indiscernible] product is getting stronger and stronger. We believe that we have a very robust offering already today with [indiscernible] for the mid-market. And moving forward, we think that will be quite viable and attractive for the larger enterprise as well, especially if you take into account the fact that the world is changing, and that these gigantic call centers that traditionally were utilized by large enterprises. I will and the situation there for sure. While smaller businesses very much price themselves and differentiate on their ability to answer customer increase with the human being. And that at least for the time being, basically regardless of how good AI gets because that is part of the services they provide, larger enterprises, it is a cost center. And I can do a very good job there, and we're seeing that as well in our own case as well as through our customers. So net-net is we're very confident that this base will be taken care of well taken care of and increasingly so with AI involved while still keeping human resolved.

Operator

The next question comes from Taylor McGinnis with UBS.

Taylor McGinnis

Sorry for any background noise. But -- if I look at the 4Q subscription revenue guide, it's a bit stronger than what we saw in 3Q and 2Q. So it seems like there was a little bit more of like an unlock or an uptick in growth that you guys are seeing -- so could you just unpack for us what's driving that? Is that more stability with the UCaaS business, but now we're starting to see the strength in AI revenue start to show through? -- any contribution from the NICE partnership. Maybe you could just unpack that as we think about that potential momentum going into next year.

Vaibhav Agarwal

Yes. Thanks, Taylor, for the question. So I think it's kind of most of what you said, look, our core business continues to be durable. Retention metrics there are strong. And it's a recurring revenue base for us, which -- and as Vlad said, the usage metrics on the platform are very healthy. which underscore the mission-critical platform. And then on top of the core, AI is continuing to add more contributions. It's growing off a smaller base, but it's already additive. And as it becomes a larger part of the mix, its contribution will grow to the overall number will grow over time.

So I think it's a combination of a durable core, growing contributions from our AI products, strong retention are kind of all the drivers for growth. And then in addition to that, I called out, we had growing and at the same time, we are also expanding margins and free cash flow in a very meaningful way. And we are being very disciplined in terms of how we are using that money and allocating that capital -- so all in all, the goal for us is to be a compounding free cash flow per share story. And -- our free cash flow per share that we've guided to this year is over $7 and growing faster than free cash flow at above 20%. So I think that's how I would characterize kind of the quarter and things kind of going forward.

Operator

That's all the time we have for questions today. This concludes the question-and-answer session and today's conference call. Thank you for attending today's presentation. You may now disconnect.