蝶网科技 (BFLY) 2026财年第三季度业绩电话会：嵌入式增长推动业绩指引上调

核心要点

• 2026财年第三季度营收同比增长52%至3260万美元，较公司指引上限高出9%，这主要由 Butterfly Embedded 业务带动。
• 在与 Midjourney 的合作以及与 Merge Labs 和 Bridge Neurotech 签署新协议的推动下，Butterfly Embedded 营收大增超1000%，达到1110万美元。
• 核心 POCUS 营收增长5%至2150万美元。本季度核心业务与 Embedded 业务共达成了5笔总合同价值达七位数和10笔达六位数的交易。
• 调整后毛利率从64%提升至73%，调整后 EBITDA 亏损收窄85%至120万美元。
• 管理层将2026财年全年营收指引上调至1.23亿至1.26亿美元，并将调整后 EBITDA 亏损指引改善至1200万至1500万美元。
• 现金及现金等价物总计1.22亿美元。管理层重申了利用现有资金在2028年前实现调整后 EBITDA 和现金流转正的目标。

关键财务数据

业务与运营表现

Embedded 业务驱动增长

Butterfly Embedded 是最主要的增长引擎。目前该业务的营收包括开发付款、年度授权费、设备及配件、工作说明书（SOW）开发收入以及芯片销售额。

管理层表示，随着合作伙伴产品投放市场，该模式还可拓展软件许可、经常性芯片采购和商业许可收入。然而，收入确认将因开发里程碑和项目时间节点而异。

公司在本季度新增了三家脑机接口合作伙伴，包括 Merge Labs 和 Bridge Neurotech，第三家合作伙伴暂未披露。此外，Mendaera 推出了一款采用 Butterfly 技术实现实时穿刺引导的手持机器人系统，不过管理层预计该产品可能需要一到两年时间才能形成显著的营收贡献。

Apollo 芯片研发推进顺利。管理层表示，该芯片预计将提供20倍的算力，并支持 AI 模型直接在设备端运行。Butterfly 还计划向 Embedded 联合开发伙伴限量发放可穿戴设备原型机以供研究。

企业级 POCUS 业务快速拓展

美国医疗系统业务营收增长15%。管理层表示，自 Compass AI 推出以来，其潜在业务管线已扩大了六倍，并正逐步转化为多年期软件和硬件合同。

本季度共签署了5笔总合同价值达七位数和10笔达六位数的协议。管理层澄清称，这15笔协议涵盖核心业务与 Embedded 业务，并不完全是 Compass AI 合同。

Compass AI 旨在连接多家厂商的超声设备、整合扫描数据、支持 DICOM 集成并优化报销工作流程。Butterfly 的策略是先行部署软件，随后随着时间的推移增加硬件投放。

Butterfly 还获得了 FedRAMP 认证，使其能够争取更大规模的美国联邦医疗业务机会。管理层认为这是2027财年的潜在催化剂，而非2026财年第四季度的实质性驱动因素。

重塑国际市场战略

由于多笔交易推迟至第四季度以及部分合作伙伴去库存，国际业务营收下降5%。Butterfly 计划在国际市场推广其美国企业级销售模式，将更多精力转向机构客户，而非主要依赖服务个体临床医生的分销商。

管理层预计国际业务表现将在2026财年第四季度及2027财年得到改善。所需支出已包含在当前指引中，并被描述为向更高回报机会进行的针对性资源重分配，而非重大新增支出计划。

产品与医疗服务路线图

Butterfly 预计将于2027年初推出采用 P5.1 芯片的新一代产品。管理层表示原型机已在测试中，组织谐波技术将提升图像质量，从而支持拓展至更高要求的临床场景。

居家及社区医疗业务在第三季度完成了项目审批。扫描预计将于11月开始，初始收入预计将在2026财年第四季度产生。管理层预计该业务将在2027财年实现增长，并可能在2028财年形成显著规模。

根据管理层公布的路线图，公司还准备在2028年进入规模达80亿美元的推车式超声市场。

管理层业绩指引

营收指引区间下限上调了400万美元，上限上调了300万美元。调整后 EBITDA 亏损指引下限改善了800万美元，上限改善了700万美元。

管理层表示，该展望包含了对核心 POCUS、Embedded、AI、半导体研发及国际扩张的持续投资。公司维持了到2028年实现调整后 EBITDA 和现金流转正的目标。

风险与关注事项

• Embedded 业务的营收由里程碑驱动，季度间可能持续波动。此外，2026财年第四季度的对比基准是2025财年第四季度——即获得实质性 Embedded 收入的首个季度。
• 国际市场表现依旧疲软，交易推迟和分销商库存挤压对营收造成了影响。
• 管理层将继续关注中东冲突、某些市场的关税以及与 AI 相关的供应链短缺。目前的影响被归类为轻微，并已计入指引中。
• Butterfly 提交了新的 Form S-3 搁置注册声明，并设立了最高达1.5亿美元的按市价（ATM）增发股权计划。管理层表示目前没有迫切的现金需求，并打算在股权稀释方面保持克制。
• 居家及社区医疗业务正在进入初始商业化阶段，而多项 Embedded 应用仍取决于合作伙伴的开发和商业化进度。

分析师问答环节亮点

• 企业级业务营收节奏：管理层表示，企业部署通常从 Compass AI 软件和少量探头开始。随着部署推进，硬件订单将逐步增加，而软件则根据多年期合同提供每月经常性收入。
• 2027财年 Embedded 展望：公司预计 2027财年 Embedded 营收将维持近期发展轨迹，但未提供正式的年度指引。管理层表示，业绩指引将主要基于已签署协议，而非假设未签署管线的转化。
• 国际市场投资：Butterfly 计划为国际合作伙伴提供更多企业级基础设施、营销能力和技术支持。管理层预计，随着营收表现的改善，该项投资将实现收支平衡。
• 可穿戴设备机遇：Butterfly 正在向探索心脏、乳腺和腹部成像等应用场景的合作伙伴提供研究用可穿戴设备。管理层将自动波束控制和 AI 支持的居家监测描述为潜在的应用场景。
• 工程研发能力：由于 Butterfly 加大了对 Embedded 开发的投资，研发费用同比增长了17%。管理层预计，在维持2028年盈利目标的同时，人员配置和开发资源将随已签署合作协议的增加而增长。

财报电话会议完整文本

---

完整财报电话会议逐字稿

管理层陈述

Operator

Hello, everyone. Thank you for joining us, and welcome to Butterfly Network, Inc. Q3 2026 Earnings Call. [Operator Instructions]

I will now hand the conference over to John Doherty, Chief Financial Officer. John, please go ahead.

John Doherty

Good morning, and thanks to all of you for joining our call today. Earlier, Butterfly released financial results for the third quarter ended September 30, 2026. We also provided a business update. The release, which includes a reconciliation of management's use of non-GAAP financial measures compared to the most applicable GAAP measures is currently available on the Investors section of the company's website at ir.butterflynetwork.com. Additionally, we also filed a new universal shelf registration statement on Form S-3, replacing the shelf that expired in August 2026.

I, John Doherty, Chief Financial Officer of Butterfly, along with Joseph DeVivo, Butterfly's Chairman and Chief Executive Officer, will host the call this morning.

During today's call, we will be making certain forward-looking statements. These statements may include, among other things, expectations with respect to financial results, future performance, development and commercialization of products and services, potential regulatory approvals, revenue attributable to Home and Community Care, as well as Embedded partnerships through revenue share, ship purchases or otherwise, and the size and potential growth of current or future markets for our products and services. These forward-looking statements are based on current information, assumptions and expectations that are subject to change and involve a number of known and unknown risks, uncertainties and other factors that may cause actual results to differ materially from those contained in the forward-looking statements. These and other risks are described in our filings made with the Securities and Exchange Commission. You are cautioned not to place undue reliance on these forward-looking statements, and the company disclaims any obligation to update such statements.

As a reminder, this call is being webcast live and recorded. To access the webcast, please visit the Events section of our investor website. A replay of the event will also be available on this page following the call.

I would now like to turn the call over to Joe.

Joseph DeVivo

Thank you, John, and welcome, everybody, to our third quarter conference call in 2026. I remain extremely pleased with our execution across the business. We delivered $32.6 million in revenue, beating the top end of our guidance by 9%. We delivered the highest gross margin in Butterfly's history, and we're raising our full year guidance on the top and bottom lines. What I'm most proud of is how revenue is being built. It's durable and it compounds.

In the U.S., our core business is increasingly anchored in multiyear enterprise contracts with a pipeline that has grown sixfold since introducing Compass AI. Internationally, we've developed a plan to take what's working in the U.S. and put that playbook to work in global markets.

In Embedded, every new partnership brings development and licensing revenue today, but more importantly, the potential for revenue share, if and when they commercialize, means it builds on itself. And as Home and Community Care begins generating revenue, that's upside. Each of these wins can expand our addressable markets.

So for me, I see our opportunity growing every day. I become more and more excited by Butterfly's potential for significant growth over the next few years. Starting with our POCUS business, I'm very proud of our domestic team. We delivered 15% revenue growth in our U.S. HHS business this quarter. Our enterprise strategy is taking effect, and we're talking to new enterprise accounts on a regular basis. We are seeing a rapidly maturing POCUS landscape.

Health systems are listening to clinicians who want to perform ultrasound, and they're providing more and more CME education for them. It's also becoming more recognized that a health system POCUS program and data management system like Compass AI is essential.

Our Compass AI pipeline is converting, and we are now closing software and hardware enterprise deals each quarter. The third quarter included 5 7-figure TCV deals and then 10 6-figure TCV deals, a clear demonstration that this pipeline is materializing, being mostly software, that revenue also compounds over time. These are also multiyear contracts, and that's exactly the kind of revenue that builds durability into our business. We expect this to keep accelerating.

So continuing in the U.S., our partnerships with the Edward Via College of Osteopathic Medicine, known as VCOM, has kicked off. Through the new Lighthouse program, VCOM is committed to providing a Butterfly probe to every one of its students. That's about 670 new students per year across 4 schools and then upper classmen are receiving devices this year also.

I attended one of their White Coat ceremonies where I stood on stage with the dean. He handed each student their stethoscope and I handed them their new Butterfly. The look of joy in their faces is a memory I will not soon forget. Graduates will take their pros with them into residency now, which is new. And because every device is networked through Compass AI, we'll work with VCOM to understand post-graduate activity and use it to inform best practices going forward.

To wrap up the U.S., we achieved FedRAMP authorization in the quarter as anticipated, which enables us to do business with the federal government at scale. Our government pipeline is now building organically, and we are now being proactively included in RFIs from government agencies on much larger efforts. This is entirely new for Butterfly. FedRAMP could be a real catalyst for Butterfly's POCUS business by creating access to the largest purchaser of healthcare in the world. More to come on that front.

So turning to international POCUS. After deal timing shifts last quarter and continued softness this quarter, we took a closer look at our international commercial strategy and identified clear opportunities to improve our performance. Historically, we've relied heavily on distributors focused on individual clinicians. We now see real opportunity to pursue larger institutional and enterprise customers, much like we've successfully done domestically where our market share is more than double that of any international region.

For competitive reasons, I won't share the specifics, but we are preparing to put our U.S. playbook to work internationally. We expect improved performance in the fourth quarter and into 2027. We believe this is an important step towards building a stronger, more durable and scalable international business.

Moving to our partnership ecosystems. We signed 5 new Garden partners this quarter, spanning emergency medicine, global health, telemonitoring and veterinary applications. We should now begin to see a steady stream of partners coming to market. Every Guardant app can expand what the probes already do in the field and become an enabler of adoption.

Just this week, DeepEcho launched its app into the Butterfly Garden. It brings Butterfly users AI-powered fetal imaging analysis and automated measurements designed to expand access to critical maternal healthcare. DeepEcho marks our first international Guardant application to go live. Its initial commercial focus is underserved international markets where access to prenatal imaging is limited, and it is also available in the U.S. for educational purposes now. We congratulate our friends at DeepEcho on this accomplishment.

On Embedded, I want to take a moment to reiterate how this revenue builds because it's one of the most important things to understand about Butterfly. Today, the majority of our Embedded partners pay us for development work, licenses and chips while they are doing technology development. Assuming products reach market, we then receive software licensing, ongoing chip purchases and in some cases, commercial licensing. So each new partnership isn't a onetime event. It adds a layer of immediate and long-term revenue that compounds as partners commercialize.

Embedded revenue won't always move in a straight line quarter-to-quarter, especially at this early phase. But what matters is that the underlying base keeps growing with every partner we add. This quarter, we hosted an event in the Bay Area that brought together the founders, researchers and DCs at the forefront of brain computer interface innovation.

Ultrasound is one of the few modalities that can interface with the brain noninvasively, both sensing neural activity and stimulating neurons. And Butterfly's Ultrasound-on-Chip is emerging as a key enabler.

While early BCI work is focused on restoring function for people with injuries or disabilities, we're seeing growing interest in approaches that would also enhance human capability in entirely new ways. Some of these applications will take years of clinical study, while others may have shorter path to market. But each new partnership expands our reach into what we believe could be a significant new market for our technology.

We added 3 new partners over the quarter, all in the BCI space; 2, we've already announced: Merge Labs and Bridge Neurotech. Both incorporate founders from our partner, Forest Neurotech, recently renamed Arbor Neuroscience, which continues to incubate new ideas from some of the brightest minds at Caltech. Both contributed to Embedded revenue in the quarter. And our third partner is still Stealth.

Our pipeline continues to grow and mature with compelling ideas from start-ups to very large companies. We've also strengthened our commercial efforts with new sales executives dedicated to specific categories of Embedded. So the future is exciting and partner innovation is already becoming a reality today. That's the moment where long-term revenue begins.

Mandera, a pioneer in vascular robotics and one of the first announced Embedded partners, has now launched its product with Butterfly technology, a handheld robotic system for real-time needle guidance. We're proud of our partnership and the enhanced capability it provides physicians in critical use cases.

We continue to make strong progress also with our Midjourney partnership and remain on track to achieve previously communicated milestones. Development of Apollo is also progressing on schedule. Apollo is a critical technology for Midjourney and several of our other Embedded partners. We believe it could be transformational, representing the first convergence of semiconductor-based ultrasound and physical AI. It's a platform investment serving not just us, but many of our partners.

Our R&D team has also advanced our wearable form factor, which has been one of the most requested capabilities from current and prospective partners. We plan to release a limited batch of wearables to Embedded co-development partners for research and development. We see it as a potential enabler of future Embedded opportunities.

So shifting to Home and Community Care. We finally moved from program building to execution. We completed all program approvals in the third quarter and are now delivering the implementation phase we committed to last quarter. Scanning starts in November, and we expect initial revenue in this fourth quarter.

We hope to expand to other markets through 2027 and are also active in discussions with other large potential partners about this application while exploring new service opportunities enabled by emerging use cases and advances in AI. I believe this revenue will grow throughout 2027 and could become meaningful in 2028.

So with that, I'll turn it over to John. John?

John Doherty

Thanks, Joe. Butterfly continued its strong and focused execution in the third quarter of 2026, with an increase in revenue driven by strong growth in our Embedded business. Core POCUS was also solid with performance reflecting our increased and expanding focus on the enterprise market. Gross margin increased, and we saw continued improvement in operating performance with further reallocation of resources towards higher ROI opportunities and markets.

Building on the operational highlights Joe outlined, let me translate these into the financial highlights in the quarter. We had strong revenue attainment, 9% above the top end of our guidance, the highest gross margin for Butterfly in its history; adjusted EBITDA that was above consensus and our guidance range with improvement driven by our revenue performance, higher gross margin and continued financial discipline while investing for growth, and growth in probe unit sales of 5%.

With that, let me move on to our results. Third quarter revenue was $32.6 million, an increase of 52% year-over-year. Our growth was primarily driven by Butterfly Embedded. In addition, in core, we had strong performance across every segment in the U.S., including U.S. health systems, medical education, out-of-hospital, federal, global health and e-com. This helped to drive strong year-over-year unit volume growth along with Compass AI sales gaining traction with key wins in major institutions.

Breaking things down between the U.S. and international channels, during the third quarter, U.S. revenue was $27.5 million, which was 71% higher year-over-year, driven by revenue from Embedded as well as solid demand in the core business with unit sales up 9%.

As Joe mentioned, international was down year-over-year. Total international revenue decreased by 5% year-over-year to $5.1 million in the quarter. Once again, we had a few deals expected to close in the previous quarters, move out into the fourth quarter, and we have some partners working through inventory.

As we execute our U.S. playbook globally, we expect improved performance in the fourth quarter and as we move into 2027. The launch in the Brazil market is gaining traction. We have a new distributor in Taiwan, and we are continuing to expand our global footprint and commitment to key markets.

Let's turn to our revenue split between our core POCUS business and Butterfly Embedded. As a reminder, the core business includes probe sales and related software, Compass AI, other services and Home and Community Care, which launched in the fourth quarter. Embedded revenue currently includes onetime NRE payments, annual license fees, devices and accessories, revenue from SOW-driven development work and chip sales to Embedded partners.

With that, core revenue for the third quarter was $21.5 million, an increase of 5% versus the third quarter of 2025. This increase was driven primarily by growth in volume in most of our U.S. segments, as I mentioned. Of note, and as Joe touched on earlier, the core result includes 5 7-figure TCV deals as well as 10 6-figure TCV deals demonstrating our continued traction as we move deeper into the enterprise market.

Sales of the iQ3 were up 8% in the quarter year-over-year, while sales of iQ+ were down 9% in the quarter year-over-year. ASP was up 1% year-over-year, driven by the higher mix of iQ3, offset somewhat by a Labor Day subscription holiday promotion we ran in September.

Butterfly Embedded revenue was $11.1 million, an increase of over 1,000% versus the third quarter of 2025. This increase was primarily driven by the MidJourney partnership and the new partnership agreements with Merge Labs and Bridge Neurotech.

Moving on to gross profit. Gross profit was also a record at $23.8 million in the third quarter of 2026, an increase of $27.6 million as compared to the prior year gross loss of $3.8 million. You will recall, we took a write-down of $17.5 million in the third quarter of 2025. Normalizing for this, gross profit was up $10.1 million or 74% year-over-year.

Adjusted gross profit margin percentage increased to 73% from 64% in the prior year period, a 913 basis point increase. Gross margin percentage was positively impacted by the increased mix and a higher margin of Butterfly Embedded revenue as well as the tariff refund under the International Emergency Economic Powers Act.

Moving to EBITDA and cash. For the third quarter of 2026, adjusted EBITDA loss was $1.2 million compared with a loss of $8.1 million for the same period in 2025, an improvement of 85%. The improvement in adjusted EBITDA loss in the third quarter was driven by contribution from higher-margin revenue and continued financial discipline.

Our cash and cash equivalent balance, excluding restricted cash, at the end of the third quarter was $122 million, and the use of cash in the quarter was $2.4 million. This compares to a use of cash of $3.9 million in the prior year quarter.

I want to speak further to our capital structure for a moment and the financial announcements made this morning. With our shelf registration expiring in August, we filed a new universal shelf registration statement on Form S-3 today. This filing allows us the flexibility to issue up to $150 million in various types of securities over the next 3 years.

As part of this shelf registration, we have also established an at-the-market, or ATM, equity program for up to $150 million. I want to emphasize a few critical points regarding this ATM program. This is good corporate housekeeping. We have no immediate need for cash and are still on course to be adjusted EBITDA and cash flow positive by 2028 as our results this quarter reinforce. This is purely a proactive measure to provide us with an efficient low-cost tool to raise capital directly or in response to inbound interest if we need to at some point over the next 3 years.

In terms of use of proceeds, any capital raised under this program will be deployed for general corporate purposes focusing on high ROI opportunities specifically related to where we can accelerate our strategic road map. We are committed to dilution discipline. We are highly sensitive to shareholder dilution. The ATM structure gives us complete control over the timing, price and volume of any potential share issuance, ensuring that we only issue equity when it creates long-term value for our existing shareholders.

With a strong cash position and this enhanced financial flexibility, we are well positioned to execute on a sustained long-term growth plan and to continue to invest in our business in areas where we see significant opportunities for additional growth and disruption, which includes scaling our focus business through global market and market segment expansion, continued penetration of Compass AI as a core operating system for health systems, app releases in Butterfly Garden and 2027 new product launches, continued AI and semiconductor innovation with the development of our fourth-generation chip, broadening and deepening Butterfly Embedded partnerships, and moving Home and Community Care to the commercial phase.

Before turning to guidance, I want to update you on the general global macroeconomic environment relative to Butterfly. We continue to monitor the conflict in the Middle East and the ripple effects on the global economy as well as pressure from tariffs in certain markets and AI-driven supply chain shortages. We have mentioned in the past that there are some impacts to our business. However, they continue to be minor, and we continue to manage through it and make the appropriate adjustments. Our third quarter 2026 results are indicative of this, and our full year 2026 guidance includes any expected impacts.

I would now like to turn to our outlook for the balance of the year. Our full year 2026 guidance reflects continued strength across the business, and we feel good about where we are headed into the fourth quarter. For the full year 2026, we are increasing our guidance for both revenue and adjusted EBITDA as well as narrowing the range.

We expect revenue to be between $123 million and $126 million, an improvement of $4 million at the bottom and $3 million at the top of the range from our previous guidance and an increase of approximately 26% to 29% over 2025. This implies $31 million to $34 million in revenue for the fourth quarter.

One item that I would like to call out and that Joe mentioned in his remarks is that Embedded revenue does not recognize evenly quarter-to-quarter. It's tied to partner program timing and development milestones and the movement of pipeline into contractual revenue. I also want to remind you that the fourth quarter of 2025 was the first quarter of material Embedded revenue.

That said, interest in Embedded remains strong and our pipeline is growing. We're being thoughtful about how this is ultimately reflected in our guidance. We expect our adjusted EBITDA loss to be between $12 million and $15 million, an improvement of $8 million at the bottom and $7 million at the top of the range, from our previous guidance. This implies an adjusted EBITDA loss between $3 million and $6 million for the fourth quarter.

Our guidance for adjusted EBITDA for the full year includes continued investment in key areas to support innovation and revenue growth in our core business and our emerging Embedded business for 2026 and beyond.

In summary, we had another strong quarter. We had 52% revenue growth for the quarter and improving adjusted EBITDA. We beat both our revenue and adjusted EBITDA guidance. We are very well positioned to close 2026 out strong. And as the raise to our full year guidance reflects, we look forward to continued strong growth this year and beyond.

We remain focused on gaining share through deeper penetration of existing customers and adding new customers in our core focus business and expanding our Ultrasound-on-Chip partnerships. The business continues to get stronger across core POCUS and Embedded. This is happening while we continue our intense focus on driving operating efficiency across the business and return on investment. I continue to be excited about what is ahead for the company for the rest of the year and beyond.

Now let me hand it back to Joe for some closing comments.

Joseph DeVivo

Thanks, John. So before I look ahead, I'd like to comment that my intention for the ATM is purely opportunistic and strategic. I remain committed to get cash flow breakeven with our current capital in 2028, as John stated. I want to leave you with a few thoughts on where Butterfly is going.

So first, Butterfly's diversified strategy is working. Our core POCUS business continues to grow, driven by enterprise adoption, our expanding Garden ecosystem and leadership in medical education.

And now we're preparing for what we believe will be another major inflection point for our business, our next-generation product in early 2027 with our P5.1 chip, the final chip in the Poseidon family. This new chip will introduce tissue harmonics. What's exciting is that we're already scanning with our prototypes internally and our KOLs are very excited with the image quality that we're seeing.

With iQ3, we reached image quality parity with our largest handheld competitors. With our next-generation device, we expect to move beyond parity and set an entirely new standard with the best image quality in handheld ultrasound. This is the power of semiconductor technology. We are marching with the inevitability of Moore's Law, and we're continuing to push the boundaries of what's possible.

And better image quality doesn't just mean a better product. It opens doors. We expect it to strengthen our ability to compete in demanding clinical environments like intensive care and cardiology, where image quality is critical. It is essential as we enter the cart market in 2028. And when combined with AI applications from partners like HeartFocus and DeepEcho, it could also expand our opportunity in cardiac care and women's health.

This is where the pieces of our strategy really start coming together: best-in-class digital imaging, enterprise infrastructure through Compass AI, education and training through ScanLab and our medical school partnerships, a growing ecosystem of AI applications through Butterfly Garden, and new care delivery models through Home and Community Care.

These aren't independent initiatives. They tie together. They're reinforcing one another, creating a flywheel that becomes more powerful with every new product, partner and customer. We are building that momentum while the competitive landscape shifts.

Some smaller competitors are struggling to find their way and larger incumbents are actually pulling back from markets they've erroneously underinvested in. Butterfly is doing the opposite. We're leaning in by imagining, investing and creating a new future.

So that brings me back to our first cart form factor. As we look towards 2028, we're preparing to enter the $8 billion cart-based ultrasound market. We're not simply building another cart. We're bringing our semiconductor platform, digital infrastructure, AI ecosystem and connected workflows into a market dominated by a conventional technology for decades, which frankly has become unimaginative. And we see an opportunity to fundamentally rethink how ultrasound is delivered, managed and used across healthcare.

Then there's Apollo. And Apollo represents our next major leap forward. Apollo will usher in extraordinary imaging and then physical AI. Midjourney and other Embedded partners are asking for these capabilities with applications that extend well beyond the handheld pro: wearable devices, implantable chips, extraordinary machines, new ways for ultrasound to interact and power an entirely new age of human interaction. This is the bigger Butterfly story. We started with a mission to democratize medical imaging. And today, we're building a semiconductor-based ultrasound platform that can serve traditional healthcare, enable new models of care, and power entirely new industries.

Our core business is growing. We continue to compete at a high level, and our technology road map is delivering. And with each advancement, our addressable market gets bigger. We've spent more than a decade building the foundation for this moment, and now we're just beginning to see the scale of what that investment is unlocking. We are excited about what's ahead.

With that, operator, let's open it up for questions.

Operator

[Operator Instructions] Your first question comes from the line of Josh Jennings with TD Cowen.

分析师问答

Joshua Jennings

Great to see the momentum continuing across the board. I wanted to just start with a follow-up just on your end of the -- your last comments, Joe, just on the innovation pipeline and the progress there. Two things that stuck out that were -- seemed a little bit new to me. I may be a little bit behind. But maybe you could just build out on this physical AI channel and how Ultrasound-on-Chip and then semiconductor technology can enable that. And then also just the limited release of wearable research devices and how you see that wearable opportunity evolving with future Embedded partners?

Joseph DeVivo

Great. So thank you for that, Josh. So the -- it's really, really exciting when we think about what the continued development of semiconductors can bring. We've been fixated on improving image quality, and that has been a function of data processing speeds and the quality of the MEMS capabilities and voltage generation. And I think this next generation, it is extraordinary. We've seen what P5.1 can do, and we are thrilled for what that means for 2027. Our Apollo chip will have 20x the processing power. So we'll go from -- it will be 20x the processing speeds for data.

And what that means is that today, when we interact with AI, we do a query on our phones or on our computers. And then those queries are run actually in data centers in the Cloud. And then that information is sent back to our devices that we use. And in a hospital setting, that is relatively suboptimal when you're in front of a patient delivering care and especially with hospital networks and connectivity, et cetera. And so the ability to actually bring AI models into the device itself because today, when a Butterfly captures an image, it's actively communicating to the Cloud. And it's using some of the processing power to be able to deliver some of the extraordinary capabilities that we're delivering today.

And so now having a semiconductor that can literally take a full AI model and process it on the chip means that it can be encapsulated for immediate feedback for the clinicians. So they don't have to be tethered and process. A lot of times, even advanced cardiac images have to be pushed to a Cloud, processed and wait several minutes to come back, which is suboptimal. So there are many different use cases where that processing and that demand and the immediacy is essential, and we'll be bringing that to healthcare.

What was the wearables? So we've had a wearable form factor in prototype form for several years. But the key for a wearable is not that you can wear it. The key to the wearable is really to 0 in on the use case. And to do the final mile type research where that wearable blends into what is -- what it could be used for? Is it for breast imaging? Is it for cardiac imaging? Is it for some general abdominal imaging? And those use cases, one of the incredible things about Embedded is our core business is a major investment. We are doing market development in point-of-care ultrasound. We have sales and marketing teams. We have R&D teams. We have implementation teams, full organization to build one market.

What Embedded is giving us is the ability to have a shot on goal and to have a revenue opportunity into many different markets that we don't have to develop ourselves. We developed a core platform and then those who are in these different specific subspecialties can invest to develop that business. Our shareholders win because we take the same technology and now leverage it over multiple markets without having to make the type of investment that today we are making to grow the POCUS business. And so we are now feeding our partners with these wearables and the wearables can be anything from an implantable or something that can be a wearable for a patient in the home. And when we really invest in cash for the future, we're the only company in the world on a single device that can move the beam on a mobile device.

And conceptually, we can put a wearable on a patient's chest in the home. We can actually move the beam in an automated fashion. AI can determine if there's any anomalies and then that image can be sent to a subspecialty clinic for diagnosis, taking out even having to have a caregiver to acquire the image. And these are the things that we've been dreaming about for years. And now amplifying our partners with wearable is a way for us to accomplish those dreams.

Joshua Jennings

And just for my follow-up. You mentioned that you're going to launch the U.S. playbook globally. I was hoping you could just build out on that. Help us think about the steps and, John, any incremental spend and when this initiative could bear fruit. Maybe just some more details on using the U.S. playbook that's been so successful in building out internationally.

Joseph DeVivo

Thanks, Josh. So we're really seeing -- we're hitting our stride now on the enterprise side in the United States. It is becoming aware now finally to administrators, health systems, department heads, that having a centralized data management tool is not only necessary for compliance, but it has a very good ROI because we're capturing scans and capturing reimbursement. And so we built a team and an infrastructure and a marketing plan that allows us to go and access those health systems. And it has to do with our platform architecture. It has to do with our infrastructure.

We have then basically outsourced our international business to partners, and we haven't empowered them with that type of infrastructure and that type of selling capability. And there's been some large deals that we could have won, but we didn't because we didn't set ourselves up like we have in the U.S. So it's more about making technology available. It's more about now investing, and it's also more about taking the marketing capabilities that we've built in the U.S. and bringing it around the world.

And you make -- when you're a small company and you're trying to make investment decisions, there are things that hit the cutting floor, and we haven't invested in international as much as we have in the U.S., and that's going to change. But I think we can do it through operating. I think we can do it by leveraging the capabilities. I don't think it's a massive new spend, but it's one that -- it's just become stark.

We had actually a distributor meeting in September in Amsterdam, and I had the opportunity to look eye to eye with all of our distributors and talk about our plans. And it became evident to me that we need to bring the type of energy to the U.S. We need to support our international team more. We need to put more resources, and we need to take the technology solutions that we've been so successful with in the States to the rest of the world.

It's an evolution thing. It's something that I'd love to be able to spend maybe $0.5 billion in OpEx a year and just do everything, but you have to make your decisions. And now with the market maturing the way it is, Josh, I think it's time for us to be able to get to that next level. Yes.

John Doherty

And Josh, just on investment, I mean, this is really more about driving operating efficiency and doing what we do in the U.S. globally. It's an investment that pays for itself. As I mentioned in the prepared remarks, we're looking to do -- allocate capital to higher ROI opportunities, and this is absolutely kind of one of those. The investment will pay for itself given what we expect to be able to drive globally and the improvement in our overall revenue results there. And everything that it requires is included in the guidance that we gave.

Operator

Your next question comes from the line of Bill Plovanic.

William Plovanic

I wanted to start off just on the enterprise. You signed 15 deals in the quarter. And I'm curious -- 5 superscaled. And I'm curious as to how should we think about the forward revenue contribution timing? Does this drive incremental unit sales as well? Just how should we think about it as we go into Q4 and then even into '27 and maybe frame it as over the first 6 or 12 months, kind of what that kind of cadence looks like from the deal with our patients?

Joseph DeVivo

So I'll start, and then I'll ask John to fill in. So the way our enterprise deals work today is we solve the problem initially of deploying software. And deploying software allows them to connect all their point-of-care ultrasound devices, allows them to compile all their patient data from the POCUS scans, allows them to bring the DICOM images into their DICOM servers and then allows them to have the data ready for their revenue cycle management platforms for reimbursement.

And so what we normally do is we put the software in. We normally start a program up with a smaller amount of probes. And then over time, as we are rolling the software out through the enterprise, we see sales of probes kind of occur on a weekly, if not monthly, basis and whether it's one by one or whether it's in chunks. And so it's a way of building up the enterprise.

There's no big bangs where someone buys 1,000 probes day 1. It's all about the evolution of the software, building of the systems. But I can tell you, there's a top 5 customer who started off with a couple of hundred probes and had put their enterprise platform in, and they're probably at 700 or 800 probes today. And so that -- over 1.5 years. So that's an evolution of what it takes.

Now we have to implement all of these -- and so we have implementation time lines. And then once the program start, then it's -- these are monthly type SaaS revenue for the software. So it will take time for it to get into the pipeline, but also the ones that we've closed before are also in the pipeline.

So John, any other comments on how we would phase that?

John Doherty

Yes. Just for crystal clarity here. Of the 15 deals that we mentioned, they include Compass deals. They also include non-Compass deals. So for example, the Merge deal, the Bridge deal are also deals that would be included in that count. On the core business side of things, I look at it primarily with the Compass deals and some of these deals being larger deals, they are also multiyear deals as kind of firming up the foundation of our revenue over time through -- in the core business. So that's kind of -- it's like a layering effect. And like I said, over a 3-year window of time, but it's a mix of deals and they contribute a bit differently in each of the Embedded side as well as the core side, but ultimately, just improving the overall foundation by which we're growing revenue also.

William Plovanic

Okay. And then just following up on guidance. It -- if you look at the Q4, the high end of the guide is kind of where consensus sits now. I think the general question is just, obviously, it's the lumpiness of the software sales component. But is this conservatism? Is this -- do you have expectations that maybe with the new 5.1 coming out in early '27 that that could have some impact on the product side? I mean, how should we think about the guidance relative to where consensus sat prior to the report?

Joseph DeVivo

Thank you, Bill. So first of all, next quarter, we'll be able to guide for 2027, and we'll give a view as to how 2027 looks. We've been guiding this guidance the entire year. We gave our annual guidance in the first quarter. We've updated it, I think, in the second quarter, and we've now increased our guidance in the third quarter. And so our annual guidance really hasn't changed. It's -- I think you're looking at what the quarter-to-quarter growth rate is. We just increased our annual guidance, the top and bottom. We narrowed the range. And so our annual guidance grew.

Now when we look at the individual quarter, we look at -- we guide what we know we can deliver and then we look to exceed that. We have a very large pipeline in Embedded. And I know everyone is asking when is the next Midjourney, when is the next pop. It will happen. We're working on a ton of stuff. But we are -- I think when you look at the year-over-year growth percentage, we're anniversarying the Midjourney deal in the fourth quarter.

And so you'll see the numbers are as we've guided all throughout the year. We've been very consistent in what we said we were going to deliver. And I think if we have the next big deal and next thing happens, we will update our guidance accordingly. But we don't get ahead of ourselves. And we don't guide to things that we don't have in our hand. And the ability to grow this business, the amount of opportunities we're looking to close across all the different areas have never been better.

John, do you want to add anything to that?

John Doherty

Yes. No, just to say, and as we said upfront, we're incredibly positive on the business as we close out 2026 and head into 2027. Just a couple of numbers. In the fourth quarter of '25, Embedded was $7.3 million. Everybody will recall that's when we announced the Midjourney deal. So it was kind of the first step in our journey up the Embedded curve of revenue growth, and we do continue to expect that to move up to the right.

The other point is we've talked about what we expected Embedded to be each and every quarter this year. In the last couple of quarters, frankly, we're trending well above what we had anticipated through the signing of the 2 new deals this quarter as well as the work we're doing against the Midjourney contract. We continue to expect that trend to continue. However, we just want to be thoughtful as we close out the year.

To Joe's point, we upped guidance in the second quarter. We upped guidance in the third quarter based on the beat that we had in both the second and third quarter, and we're being prudent about what we're putting out there for the fourth quarter. But we're -- nothing's changed. We're well on track. And we're excited about being able to do what we usually do, which is put out revenue results at the JPMorgan conference in the second week of January. And at that time, we'll likely also talk about 2027, but we're really excited about where the company is and where the company is going.

Operator

Your next question comes from the line of David Saxon with Needham.

David Saxon

So I'll maybe just follow up to kind of how to think about fourth quarter and '27. So I mean, we've been getting a lot of questions on how to think about Embedded revenue in '27, particularly since there's been a lot of milestones recognized in '26 that won't repeat. I think you've historically talked about kind of $7 million to $8 million a quarter for Embedded. Last 2 quarters, we're obviously above that. So any framework or guardrails you'd point to as we kind of look out to '27 for Embedded, just given the milestones in '26 and kind of how you're thinking about the portfolio?

Joseph DeVivo

You want to start, John?

John Doherty

Yes, sure. So to my comments based on the last question, we are trending above what we had said. We do expect to kind of live in the space that we're in now. We also expect to carry that on into 2027. Of course, we're not prepared to give full year guidance on that. But our trajectory is we expect it to be consistent.

In addition, as Joe has said in the past and as well as I have, we're not -- we're folks that will tell you what we do when we do it. So as and when we get additional partners to Embedded, we'll make sure that we announce that, and we'll be very clear about it as we have in the past. And as we said, we remain incredibly confident about what's happening with Embedded.

In addition, and as we've talked about and we'll talk more about in early 2027, is the core business is also moving in a really strong direction with what we're doing with international, with what we're doing in preparation for our new product launches next year, with the launch of Home this quarter, things are looking very, very positive for us.

Joseph DeVivo

And the only other thing I'd say on Embedded is we have the contracts signed that we have and the work that we are doing, and it's reflected in a line of sight as we see things, and we'll communicate that for '27. But we also have a very large pipeline. And now we have a commercial team that is actively engaging new opportunities. And the more and more we get comfortable with the ability to close those opportunities and more it repeats them, the more we'll be willing to take some more risk in the guidance looking forward.

But we've -- John and I, first of all, we're very conservative. We've hit everything we've told you guys. And we want to make sure that we support our investors and we support the way that we put our vision forward. But we don't get ahead of ourselves. And so the tailwinds in this business are real and they're real across the board. And I wouldn't look too much into -- I wouldn't get too spooked with some guidance that's basically been there since the beginning of the year.

John Doherty

And has improved each quarter.

Joseph DeVivo

And has improved each quarter.

David Saxon

Okay. I appreciate the color there. And then in the script, you talked about some competitive dynamics, Healthineers moving out of general ultrasound and then Philips recall too. How are you thinking about Butterfly's positioning to catch some of that demand, particularly with the kind of next-gen probe and IQ spacing around the corner? And I don't know if you have any data that you could kind of quantify how you think about that competitive opportunity or how meaningful it could be?

Joseph DeVivo

Thank you, David. My -- I don't want to make a comment about a specific company. That's really not my objective. I do think that when it comes to semiconductor-based ultrasound, when it comes to point-of-care ultrasound, when it comes to the next wave, I think what you see in the larger companies is this desire to continually move upstream and to move into more and more niche capabilities.

And look, if I'm a patient, I want those capabilities. I want you to be able to see my disease even better and better in these different care settings. So I don't regret a strategy. But I do think there is the part of the market or the individual user and the expanding out into homes and empowering individuals in the main part of the market. I think that's an area that big ultrasound is underinvested in. I think they've looked down their nose at this. I think they've looked down their nose at Butterfly. And that builds a chip on our shoulder.

We do have a chip on our shoulder, and we are proving the world, just like every other digital company that's disrupted analog, we are going to do it, and we actually are doing it. And I think what's important is we have 600 or 700 patents now. It's -- we're going to launch our fourth-generation chip and our first generation cost -- it took $300 million to get the first chip out the door. Our competitive moat is big. Our competitive moat is real and our competitive moat is long.

And so I think if you really sit back and look at this segment, and I think of people being honest with you, they say, wow, that thing has really taken off and what are we going to do about it? Now I don't think we factor into any of those other companies' decisions. I think they're bigger with a bunch of other things that they're concerned about. But from our perspective, we are making big investments in the future of this industry, and we've been building a moat, whether it's been quietly or whether or not. But I think that is real. And it is hard to be able to just now decide today. It would be 5 years if someone decided today to replicate what we're doing.

So I think what's so exciting is not only is our technology getting better, but the health systems are waking up to this now. And this is becoming more present, and we're seeing more activity. And so competition will definitely get harder, and it will definitely -- and that's -- any market that matures segments, and you'll see more and more people getting in and more and more competition. But specifically in how we address this problem, it's a very, very unique situation where we've been so far ahead for so long. And I think we're about to reap the rewards of that.

Operator

Your next question comes from the line of Daniel Markowitz with Evercore ISI.

Daniel Markowitz

Number 1, I wanted to ask on Embedded. Firstly, are you able to quantify the 3Q revs for Midjourney? And then I recognize Midjourney might be somewhat lumpy going forward, but I wanted to talk about the overall Embedded business ability to grow in '27 versus '26. Can you -- I guess, in framing that question, can you talk about how big of a factor the Midjourney lumpiness will be versus the ramp that you'll be able to see in some of your other Embedded partners and whether that should be kind of moving in the right direction?

John Doherty

Daniel, we're not going to specifically call out Midjourney revenue. We'll obviously continue to talk about Embedded revenue. If things are material over time and we have to file, we'll put things out there. But yes, I mean, Midjourney is foundational for us in Embedded, but we're not going to talk specifically to it. The guidance that I've given in the past around Embedded as a whole still kind of stands. We've trended a little better than that. Some of that is driven by Midjourney, but also driven by some of the new partnerships that we've signed this quarter. Merge was a big contributor in the quarter.

And we do expect -- and we've talked about this in the past, and I want to remind folks too, we said if Midjourney current course and speed is successful what they're doing, there's nothing we've seen so far in working with them that proves anything that they're not going to be. In 2030, our revenue will be at least $550 million. And that's what we've talked about. We've also talked about how that would be effectively only taking into account 1/4 of what we would expect from Midjourney.

Now we called it out at the time because it was such a large partnership. They had the announcement in mid-June. We thought it was important and prudent that we were open and transparent with you guys, the analyst community, as well as with our investors. Current course and speed, if we just achieved the midrange of our revenue for this year, that would be $124.5 million. Certainly, Joe and I, as we've said, we do what we say, and we feel we're going to do better than that. But even at that level, that's a 45% compound annual growth rate for revenue for the business as a whole.

And as I said, we feel that's somewhat conservative. But on a quarter-to-quarter basis, we'll announce the deals that we signed. We'll obviously announce revenue each quarter. We'll continue to break out Embedded, but we're not going to talk specifically to any individual customer. We did it for the onetime because of kind of the journey that we're on with them, and I'll leave it at that.

Joseph DeVivo

I think the only thing I'd add is if we sign something of a magnitude of Midjourney, it will be material. There will be an 8-K on it. We're in this really kind of uncomfortable place where we're a public company and our investors want to have a very clear model and roadmap for the future. And our partners want everything, but they don't want us to talk about this. Midjourney hated the fact we had to put the 8-K out. And all these companies, unless there's a press release that's associated on the technology, whatever is in the press release is what they want to talk about.

And so we want to be able to inform you, but we also have to protect kind of the roadmap and the privacy of our partners. So it's really a tight rope that at times is uncomfortable. But the thing I would say is we have a big pipeline. We have a lot of conversations. There are many big deals out there. And there's a lot of -- and there's a ton of base hits that pretty quickly add up to big deals. I think it's going to be a beautiful contributor to '27. But John and I are not in the business of guiding to something that within a period, we don't have in our fingertips and can't deliver. So you're going to see everything that we say is going to have a conservative side to it.

Daniel Markowitz

Super helpful. And I definitely appreciate it. It's a tricky one, but that's a really helpful answer. The second thing, hopefully an easier question on core POCUS. It's encouraging to see the acceleration from low single-digit growth last quarter to mid-single-digit growth in the third quarter. And it seems like there are more catalysts for acceleration in 4Q and beyond between Home and Community, Compass AI contracts. Is it fair to say we should continue to expect acceleration in 4Q and '27 on the back of these tailwinds?

Joseph DeVivo

Yes. I don't think there's any reason to be looking back. And one of the things I mentioned in the prepared remarks was FedRAMP. FedRAMP is the hardest security claims to get in the world. That means that we can operate inside the federal government systems. And we're a Cloud-based vendor. So we're not some brick device that we claim that we're FedRAMP. Our stuff is networked. That means we have to ride their rails. And so for us to be able to get FedRAMP and ride inside the government is nuts. And we now are having conversations that are meaningful.

And actually, I didn't realize it would happen this quickly, but I don't think there's many companies that get across that FedRAMP threshold. And once it happens, now there's things -- and I think the government's biggest purchaser of healthcare. And from what I'm seeing, I think that is going to be not a fourth quarter. It's a little too soon, although we have seen some really nice base hits. But I think there's an opportunity in '27 for it to really catalyze a whole new opportunity for us. And I'm very appreciative of our IT security leadership who have led this for us.

Operator

Your next question comes from the line of Steven Lichtman with William Blair.

Steven Lichtman

Congrats on the progress. Just first question, thinking about the Home Care opportunity with that kicking off here in earnest. You talked about the healthcare cost savings as an important aspect to that story. When do you think we'll start to see enough data from the first partner to start showing those savings? And as you're talking with other potential customers, are the discussions focused on CHF or moving to other indications as well?

Joseph DeVivo

Yes. I mean I wish we can do an Investor Day just on answering that question because it is really, really exciting because remember, our device has many different use cases. And the initial use case we're doing with our partner is for congestive heart failure and to reduce the revolving door into hospitals, which every patient that you can eliminate an admission or readmission saves them $10,000 to $12,000 for each patient. And so putting a protocol in place that allows us to manage those patients and keep them healthier and better managed than having a virtual cardiologist with them and having them getting pulmonary scans for pulmonary wetness really is a major deal. But the same partner also now wants to do bladder scanning, wants to do scanning for sepsis, wants to do cardiac echoes, and we're telling them, no, let's just focus on this first. The opportunity --

Oh boy. We're in a building that is testing its fire system. So I apologize, I don't know how to --

John Doherty

We'll just keep talking through it.

Joseph DeVivo

Yes, we'll just talk through it. That is unbelievable. That's great. All right. Thank you for that guys. All right. So we know that our fire system is being tested right now, so we can put that into our note.

Now there are other types of clinical opportunities. And there's -- right now, this opportunity is in nursing homes to keep a revolving door, but there's opportunities where millions of patients are being managed in their home and they're being managed by home healthcare nurses that come in and see them. And there are use cases that we are working towards with other large organizations too.

So as far as the data is concerned, that's another really delicate source because our partner is not one to want their data to be marketed out in the marketplace. So we'll find an elegant way to communicate the progress without violating their trust. So sorry about that announcement in the background. We're in New York City in our offices and we can't control that. So thank you.

Steven Lichtman

Yes. We've been there. Just really quickly, one depth on Mendaera and any initial feedback you've heard from the field. Obviously, I understand of the Embedded opportunities at the largest, but it's sort of the first to hit here commercially. Any initial feedback that you've heard from them relative to Focalist?

Joseph DeVivo

Yes. Well, actually they were in our offices yesterday, and we had a great update. And we're very excited for their launch and what will occur with them in 2027. They're a start-up. So they have a -- it's not like there's all this existing revenue that just pops over, but they're out there selling and they're in account by account by account.

And it is just such a wonderful multiplier for us because on one side, they're bringing ultrasound into some interventional cases, doing some more complex cases that would be hard to do with the human hand. And it's leveraging our core platform and it's a part of that flywheel because as they go into health systems doing more advanced procedures, leveraging on a Butterfly imaging stack and then they're doing these advanced procedures. It's -- there's a lot of synergies in our sales process. But it will be a year or 2 before that becomes meaningful revenue for us, but they're launching now and the revenue is coming in, and we're very excited.

Operator

The next question comes from the line of Chase Knickerbocker with Craig-Hallum.

Unknown Analyst

This is Nick on for Chase. So you guys touched on this a bit in the previous question. But just as you guys build out more of a proactive outbound sales effort for Embedded, how is that changing the size and mix of the pipeline? And what could this mean for the pace of new partner adds in 2027? And then also on top of that, just any verticals outside of healthcare where you guys are seeing strong interest?

Joseph DeVivo

Yes. So thank you for that. So the base is definitely broadening. And the question really is, as we broaden the base, I think there's going to be a lot more base hits. So there's going to be a lot more opportunities to get people on and then to start seeing some revenue. And I think they're all going to stack. And then we hope to pull some diamonds out of that stack and create more Midjourneys.

As we -- the thing right now is so much of our earlier conversations have been with more start-up new tech generation. And some of our conversations are now being held with much larger entities and some of them are large entities with mature products. And so there is the possibility if we sign a deal with a large entity with a mature product that that revenue happens very quickly because they're already standing on hundreds of millions or billions of dollars of product revenue of things that they're already selling. So us getting into that environment all of a sudden would be like you're not in the R&D of a new market, you're converting or you're upgrading their technology to new capability.

So there are so many different things, and it's early and it's very hard to predict. And that's why John and I are taking a conservative posture that we basically guide what is signed. We're not going to extrapolate. We're not going to guess because each of these are unique deals and each of these are unique situations. But one deal can completely change the complexion of the company.

And I think -- I have a friend of mine, the CEO of Biotech. I just checked their revenue today. They're about $5 million in revenue as a biotech and their market cap is about $3 billion. Because people look at and say, okay, if they can do -- if this molecule can do what this can, then this opportunity is massive.

We're signing deals like that every day. And Butterfly should not be looked at as a revenue multiple today. It should be looked at as the magnitude of opportunity that we are capturing for our future and for investors. And so it becomes very difficult to predict each one of these. But as I said in my prepared remarks, we are layering in opportunity on top of opportunity on top of opportunity, and it expands our TAM and it expands the leverage. And the beautiful part is that it's leveraging our core tech.

We just had a laboratory 2 days ago with a very large potential partner who came up with a very specific use case and said, can you do this? And our team spent a month, looked at exactly what they were trying to do, and we knocked it out of the park in the lab yesterday. So -- over 2 days ago. So it is a very, very exciting time. And I'd be looking at what is Butterfly able to deliver over the next several years then worry about how this phases on a quarter-by-quarter basis.

Unknown Analyst

Got it. And that kind of leads into my next question is just as these existing Embedded partners progress and you continue to add new ones, how much capacity do you guys have within the current engineering team? And how should we think about any incremental operating expense going into next year to support Embedded growth?

Joseph DeVivo

Yes. I mean I think there is a variable component there. So as we sign new deals, there is a demand on development resources, and we will continue to grow it. Again, we won't get ahead of ourselves. We actually -- the first parts of these deals, we get a lot of leverage off of what we've already done. And then the incremental, we will bring in the best talent.

We've been very fortunate actually in 2026 so far. We brought on some extraordinary people, and we're so delighted and we do need to continue to do that as we sign these deals. But it's something that we will be able to just create more and more leverage. But yes, of course, we will have to continue to invest as well. Anything?

John Doherty

Yes. I mean our results reflect that. I mean we were up -- R&D was up 17% year-over-year in the quarter. Not that I want to sit there and call out an increase in operating expense, but effectively, as we've also said, our adjusted EBITDA was down, and we're trending in the absolute right direction. We're still well on track to be...

Joseph DeVivo

Adjusted EBITDA loss was down.

John Doherty

Yes, I'm sorry, adjusted EBITDA loss is down. We're still on track to be adjusted EBITDA positive and cash flow positive in 2028. And actually, as our results are being driven out, I would say that that's moving closer for us.

And in addition, we haven't really talked about it much on the call. We talked about in the prepared remarks. We did file a new shelf. I want to remind folks that's over a 3-year period of time that that would be active for us in addition in ATM. So should we need capital at some point in time, we're prepared for that. And it was just general -- good general corporate housekeeping. But current course and speed, we're headed in the right direction, and we are investing in the areas of the business that we need to, and our results are reflective of that.

Operator

The next question comes from the line of Ben Haynor with Lake Street Capital Markets.

Benjamin Haynor

Just following up on the Embedded sales force, how is that kind of structured? Are they going after individual areas, PCI, industrial, what have you? And then how do you make the decision whether or where to add folks to that group?

Joseph DeVivo

So thank you for the question. So first of all, we are trying to get through the pipeline. And so -- and we're -- and it's not just in the selling folks, but there are application specialists who are able to do labs and able to do demos, et cetera. So we're just -- first of all, we're getting through the pipeline and trying to get the pipeline closed because each one of these, it's not just a cookie-cutter sale, it's a partnership, and we have to learn about the partners and tailor ourselves to their needs.

But there are very large ultrasound markets. And there are very large players that we are proactively targeting. And we are looking at it from a market-by-market basis. And so it starts off by cultivating the pipeline and then learning from the pipeline to extrapolate it into different markets. And then now we're starting for the first time to -- we're in the process of being proactive, co-calling, having dialogue, going -- and we can be more proactive with existing incumbents by communicating to them and by going out and having that dialogue. Usually, the start-up companies are the inbounds to us and then the more existing ultrasound companies are the ones that we are outbound to.

But we're also starting to target certain markets where we think we have great applicability. I wish I can outline them for you. But if I outline them for you, then you'll be like, oh, so you're calling -- you're talking to this company or that company. And so I don't want to do that. But we are proactively going into the largest ultrasound markets that are not competitive with Butterfly and talking to them about what we think we can do to refresh and upgrade and take their current technology to the next level.

Now we are also, though, targeting non-healthcare industries. And we are -- I think I mentioned in the last quarter that we are talking to someone who is not a traditional healthcare company and it is our first market that we're having dialogue in and doing labs. And so we're going to do everything we can to build on that. But we've been building our infrastructure, building a team. It's one thing to talk to a customer. It's another thing to have the technical support in the lab, the ability to make modifications that they might want to see and show them. So we've been building that all together. And right now, I'm very pleased. We've been blessed with David Horsley and Steve Cashman, who have been leading this effort. And we are very excited about the new people that we brought on. So the more I can tell you, the more I will, but I can't kind of outline each of the markets for competitive reasons.

Operator

Your next question comes from the line of [ Aralum Konectgazi ] with Freedom Broker.

Ilya Zubkov

This is Ilyara. Congrats on a strong quarter and record gross margin. I also want to focus on Compass AI. This quarter, you closed 5, 7-figure and 10 6-figure enterprise deals, which suggests that health systems are starting to buy point-of-care ultrasound as a managed program. But historically, hospital pocket adoption has felt less on hardware and more on workflow, and that is exactly the layer Compass AI is designed to solve. My question is, how do these enterprise wins represent system-wide standardization on Butterfly? Or are they still mostly departmental deployments that can expand over time? And over the long term, what do you expect to come from recurring enterprise software? And how should we think about the contract structure behind these deals?

Joseph DeVivo

So that is a beautiful question, a very, very good question because when we built Compass AI, we didn't build it for Butterfly alone because we knew if we were going to put an enterprise software in place, you can't put enterprise software in for one product. And so what we're doing is solving a problem the hospital has. And the hospitals are buying GE POCUS carts. They're buying Sonacite POCUS carts. They're buying carts for Mindray and all these other companies that have -- the POCUS is divided into 2 different pieces of hardware, a handheld device and then a scaled-down version of a cart. And that's, again, why we are getting into the IQ station market and to get into that cart market to begin with.

But hospitals have POCUS carts all over the hospital, and they're not -- and probably 30% of the scans are getting recorded into their EMR. And so they're not compliant and they're losing revenue. So when we put Compass AI in place, we are solving a major problem that the hospital has, whether or not they're a Butterfly customer. When we solve that problem for them and we put our infrastructure in place, of course, Compass AI integrates with our hardware much better than anyone else because we can control both sides, and we have the ability to make our hardware more -- the software that sits inside of our hardware more capable.

So it starts off with solving a problem that is -- that the entire industry is creating by putting POCUS into hospitals. And then -- but once we are in, then we use that as a land to expand by building out the handheld program. So that's why our IQ station dovetails so beautifully in because now we will have a handheld, we'll have a cart and we'll have the enterprise software platform to be able to have a really smooth, elegant and incredibly modern way of managing these programs. But it's a very astute question because a lot of times, sometimes 80% of the scans that are coming in the Compass are from all of our competitive companies.

So our Butterfly's vision has been to promulgate and to move point-of-care ultrasound forward and Compass AI is dedicated to solving health system problems in point-of-care ultrasound and then, second, to be able to pull through our hardware. And that's why our IQ station is going to be such an important launch because we're going to go after existing business. We're going to go after the competitive footprint on point-of-care ultrasound and the integration of our software and hardware is going to be so tight that it's going to make us the choice for the future. But thank you for such a great question.

Operator

There are no further questions at this time. I will now turn the call back to Joe DeVivo, President and CEO, for closing remarks.

Joseph DeVivo

I know it's been a long call. We're going to have to -- we've been blessed with having 3 more analysts this quarter. We're so appreciative of the attention and support. So we're going to have to reduce our prepared remarks and have more conversation going forward because those questions were just fantastic.

I just want to thank everybody for their support. We are so excited about our future. So much is happening and I can't wait to provide more updates at the end of the year and of course, at the upcoming JPMorgan conference. So thank you all and have a wonderful holiday season.

Operator

This concludes today's call. Thank you for attending. You may now disconnect.