Symbotic Inc (SYM) 2026 Q3 法說會逐字稿

Good day, and thank you for standing by. Welcome to Symbotic's Third Quarter Financial Results Conference Call. (Operator Instructions)

大家好，感謝您稍候。歡迎參加 Symbotic 2026 會計年度第三季財務業績電話會議。(接線員指示)

Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Charlie Anderson, Vice President of Investor Relations. Please go ahead.

請注意，今天的會議將被錄音。現在我想把會議交給今天的第一位講者：投資人關係副總裁 Charlie Anderson。請開始。

Hello. Welcome to Symbotic's Third Quarter of Fiscal Year 2026 Financial Results Webcast. I'm Charlie Anderson, Symbotic's Vice President of Investor Relations. Some of the statements that we make today regarding our business operations and financial performance may be considered forward-looking. Such statements are based on current expectations and assumptions that are subject to a number of risks and uncertainties.

您好。歡迎收看 Symbotic 2026 會計年度第三季財務業績網路直播。我是 Charlie Anderson，Symbotic 投資人關係副總裁。我們今天就業務營運與財務表現所做的部分陳述，可能被視為前瞻性陳述。此類陳述係基於目前的預期與假設，並受多項風險與不確定性影響。

Actual results could differ materially. Please refer to our Form 10-K, including the risk factors. We undertake no obligation to update any forward-looking statements. In addition, during this call, we will present both GAAP and non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in today's earnings press release, which is distributed and available to the public through our Investor Relations website located at ir.symbotic.com.

實際結果可能出現重大差異。請參閱我們的 Form 10-K（包括風險因素）。我們不承擔更新任何前瞻性陳述的義務。此外，在本次電話會議中，我們將同時呈現 GAAP 與非 GAAP 財務衡量指標。GAAP 與非 GAAP 指標之調節表已收錄於今日的財報新聞稿中，該新聞稿已發布並可透過我們的投資人關係網站 ir.symbotic.com 向大眾取得。

On today's call, we are joined by Rick Cohen, Symbotic's Founder, Chairman and Chief Executive Officer; and Izzy Martins, Symbotic's Chief Financial Officer. These executives will discuss our third quarter of fiscal year 2026 results and our outlook, followed by Q&A. With that, I'll turn it over to Rick to begin. Rick?

今天的電話會議由 Symbotic 創辦人、董事長兼執行長 Rick Cohen，以及財務長 Izzy Martins 與會。兩位主管將討論我們 2026 會計年度第三季的業績與展望，隨後進行問答。接下來我把時間交給 Rick 開場。Rick？

Thank you, Charlie. Good afternoon and thank you for joining us to review our most recent results and business updates. We delivered strong third quarter results, highlighted by continued revenue growth and expanding margins, leading to continued GAAP profitability and adjusted EBITDA that more than doubled year-over-year.

謝謝你，Charlie。各位下午好，感謝各位加入我們，一同回顧最新的業績與業務更新。我們交出強勁的第三季成績，亮點包括營收持續成長與利潤率擴張，帶動 GAAP 持續獲利，且調整後 EBITDA 年增超過一倍。

Thanks to another strong quarter, we remain well on track to achieve the objectives we laid out at the start of the year. As a reminder, our first objective was to leverage our growing product portfolio and capabilities to broaden our opportunities with customers.

憑藉又一個強勁季度，我們仍然穩健地朝年初所提出的目標邁進。提醒一下，我們的第一個目標是運用不斷擴大的產品組合與能力，拓展與客戶合作的機會。

We are clearly seeing this play out as our BreakPack product to handle individual items or each's has now begun deployment at half of Walmart's regional distribution centers. In addition, we recently began installation of our first SymMicro system for e-commerce fulfillment at the back of a Walmart store, a significant step forward towards unlocking this exciting new category of our business.

我們清楚看到這一點正在發生：我們用於處理單件商品（each）的 BreakPack 產品，現已開始在 Walmart 一半的區域配送中心部署。此外，我們近期也開始在 Walmart 門市後場安裝第一套用於電商履約的 SymMicro 系統，這是朝著解鎖這個令人振奮的新業務類別邁出的重要一步。

We are also continuing to drive additional value for our customers that have existing operational systems by providing higher levels of performance through software to further optimize their supply chains. A recent example is using our software to more intelligently layer pallets and dynamically optimize freight delivery, specifically for seasonal events like back-to-school.

我們也持續透過軟體提供更高水準的效能，進一步優化供應鏈，為已導入營運系統的客戶創造更多價值。近期的例子是運用我們的軟體更智慧地進行棧板分層堆疊，並動態最佳化貨運配送，特別是針對返校季等季節性活動。

By doing so, we believe our customers can realize shorter delivery times and faster restocking during these critical periods. We believe customers are increasingly recognizing the impact our systems can have. And as a result, we are seeing additional opportunities to broaden the scope of our work with both existing and prospective customers.

透過這些作法，我們相信客戶能在這些關鍵期間縮短配送時間並加快補貨速度。我們認為客戶正日益認知到我們系統所能帶來的影響。因此，我們看到更多機會，得以擴大與既有客戶及潛在客戶的合作範疇。

For example, in the third quarter, we signed an agreement with Southern Glazer's Wine and Spirits for a second site after the success of their first facility. Southern Glazer's is a leading total beverage distributor serving 47 U.S. markets in Canada. As we drive additional value to customers, it is allowing us to realize the second objective we laid out at the beginning of the year, which was to enhance our margins and profitability.

例如，在第三季，我們與 Southern Glazer's Wine and Spirits 簽署第二個據點的協議，係在其第一座設施成功之後所促成。Southern Glazer's 是領先的全品類飲料經銷商，服務美國 47 個市場以及加拿大。隨著我們為客戶創造更多價值，也讓我們得以實現年初提出的第二個目標：提升利潤率與獲利能力。

Our forecast for the year implies full year adjusted EBITDA that is more than double that of last fiscal year. This continues to be a key focus area for us, and we see clear levers to continue enhancing our profitability driven by value creation for our customers and further operational efficiencies.

我們對全年之預測意味著全年調整後 EBITDA 將超過上一會計年度的兩倍。這仍是我們的關鍵聚焦領域；我們也看到明確的槓桿可持續提升獲利能力，動能來自為客戶創造價值以及進一步的營運效率提升。

The final objective we laid out was to continue to invest in our innovation engine to expand our capabilities and support future growth. The analogy I often use here is that our automation system is like an operating system, and we add apps to enhance its functionality for customers.

我們提出的最後一個目標，是持續投資於創新引擎，以擴展能力並支援未來成長。我常用的比喻是：我們的自動化系統就像作業系統，而我們透過新增應用程式（apps）來強化其對客戶的功能。

For us, this is playing out both organically and inorganically. Organically, we are making several functionality upgrades to our SymBots to enhance the performance of our system. For example, we deployed over 1,000 larger bots into our operational system this calendar year to handle a wider variety of SKUs.

對我們而言，這同時透過內生成長與外延成長在推進。在內生方面，我們正對 SymBots 進行多項功能升級，以提升系統效能。例如，今年曆年我們已在營運系統中部署超過 1,000 台更大型的機器人，以處理更廣泛種類的 SKU。

With this new bot, we've also built new modularized software development tools to give us enhanced flexibility to create different bots for different tasks and payloads with our SymMicro bot being a perfect example. We're also in the process of rolling out LiDAR, enhanced camera systems, Nyobolt advanced batteries and other updates, all with the aim of driving enhanced efficiency and performance for our systems.

隨著這款新機器人，我們也建立了新的模組化軟體開發工具，使我們能更具彈性地針對不同任務與載重打造不同機器人，而 SymMicro 機器人就是最佳例子。我們也正在推動 LiDAR、強化相機系統、Nyobolt 先進電池及其他更新，目標皆在提升我們系統的效率與效能。

Inorganically, we've made two tuck-in technology acquisitions that expand our capabilities. Fox Robotics for dock automation and most recently, ARMS Innovations for warehouse operations optimization. With ARMS, we have an opportunity to expand the reach of our software beyond our automation system to the entire warehouse operation, optimizing the movement of both equipment and people.

在外延方面，我們完成了兩項補強型（tuck-in）技術併購，以擴展能力。包括用於碼頭自動化的 Fox Robotics，以及最近用於倉儲營運最佳化的 ARMS Innovations。透過 ARMS，我們有機會將軟體的觸及範圍從自動化系統延伸至整個倉庫營運，最佳化設備與人員的移動。

In summary, we are focused on meeting our objectives and in turn, creating ravingly happy customers and expanding shareholder value. We also continue to have a solid balance sheet and backlog. As always, I want to thank our team for all their hard work, along with our customers and our investors, for their continued support.

總結來說，我們專注於達成目標，進而打造讓客戶極度滿意的體驗，並提升股東價值。我們也持續維持穩健的資產負債表與在手訂單（backlog）。一如既往，我要感謝我們的團隊辛勤付出，也感謝客戶與投資人持續的支持。

I'll now turn it over to Izzy, who will discuss our financial results and outlook. Izzy?

接下來我把時間交給 Izzy，由他說明我們的財務結果與展望。Izzy？

Thanks, Rick. Fiscal third quarter revenue reached $721 million, near the high end of our forecasted range and was up 22% year-over-year and up 7% quarter-over-quarter. We also improved GAAP profitability with $55 million in net income. Adjusted EBITDA of $95 million was above our forecasted range due to expanding margins and operational efficiencies.

謝謝你，Rick。會計年度第三季營收達 7.21 億美元，接近我們預測區間的上緣，年增 22%，季增 7%。我們也提升 GAAP 獲利能力，淨利為 5,500 萬美元。調整後 EBITDA 為 9,500 萬美元，因利潤率擴張與營運效率提升而高於我們的預測區間。

Our revenue growth was driven by the continued expansion in the number of systems in deployment and the growth of operational systems that generate recurring revenue. We started 11 new system deployments in the third quarter, including the new Southern Glazer's site highlighted by Rick, bringing us to a total of 77 systems in deployment at the end of the quarter.

營收成長主要來自部署中的系統數量持續擴張，以及可帶來經常性收入的營運中系統持續成長。第三季我們啟動了 11 個新系統部署，其中包括 Rick 提到的 Southern Glazer's 新據點，使得季末部署中的系統總數達到 77 套。

This expansion in the number of deployments drove systems revenue growth of 20% year-over-year and 6% sequentially to $671 million. We also had four systems go operational during the quarter, bringing us to a total of 56 operational systems.

部署數量的擴張帶動系統營收年增 20%、較前一季成長 6%，達 6.71 億美元。本季也有 4 套系統轉為正式營運，使營運中系統總數達到 56 套。

As our base of operational systems continues to expand, software revenue grew 57% year-over-year to $13 million and operations services revenue of $37 million grew 49% year-over-year, both in the fiscal third quarter. Turning to margins in the fiscal third quarter.

隨著營運中系統基礎持續擴大，軟體營收年增 57% 至 1,300 萬美元；營運服務營收為 3,700 萬美元，年增 49%；兩者皆發生於會計年度第三季。接著談談會計年度第三季的利潤率。

Gross margin expanded both sequentially and year-over-year due to strong project execution, cost discipline, benefits from scale and revenue mix. Operating expenses on a GAAP basis were $128 million in the fiscal third quarter. Combined adjusted R&D and SG&A expenses totaled $85 million, with SG&A down sequentially due to operational efficiencies.

由於專案執行強勁、成本紀律、規模效益以及營收組合的帶動，毛利率無論按季或按年皆有所擴張。以 GAAP 基礎計算，本財年第三季的營業費用為 1.28 億美元。合併後的調整研發（R&D）與銷售、一般及行政（SG&A）費用合計 8,500 萬美元，其中 SG&A 因營運效率提升而較上季下降。

Net income for the fiscal third quarter was $55 million, an improvement from a net loss of $21 million in the third quarter of fiscal year 2025. This included an unrealized noncash gain on the fair value of our strategic investments of $19 million in the quarter, which was primarily driven by an increase in the value of our investment in Nyobolt, our next-generation battery supplier.

本財年第三季淨利為 5,500 萬美元，較 2025 財年第三季 2,100 萬美元的淨損有所改善。其中包含本季對策略性投資公允價值的 1,900 萬美元未實現非現金收益，主要來自我們對下一代電池供應商 Nyobolt 的投資價值上升。

GAAP net income improved both year-over-year and sequentially, reflecting this impact as well as expanding margins and operating leverage. As Rick highlighted, adjusted EBITDA of $95 million was more than double the $45 million in the third quarter of fiscal year 2025. Our backlog of $22.5 billion remains strong.

GAAP 淨利按年與按季皆改善，反映上述影響，以及利潤率擴張與營運槓桿提升。如 Rick 所強調，調整後 EBITDA 為 9,500 萬美元，較 2025 財年第三季的 4,500 萬美元高出一倍以上。我們 225 億美元的在手訂單（backlog）仍然強勁。

The slight decrease from last quarter primarily reflects revenue recognized in the quarter, offset by final pricing adjustments on projects started in the quarter and the addition of the new Southern Glazer's site. We finished the quarter with cash and cash equivalents of $1.7 billion, down from $2 billion last quarter due primarily to timing of cash receipts related to project starts, along with the timing of cash usage related to project activity.

較上季的小幅下降主要反映本季認列的營收，部分被本季啟動專案的最終定價調整以及新增 Southern Glazer's 站點所抵銷。本季末現金及約當現金為 17 億美元，較上季的 20 億美元下降，主要由於與專案啟動相關的現金收款時點，以及與專案活動相關的現金使用時點所致。

Now turning to the outlook. For the fourth quarter of fiscal 2026, we expect revenue between $760 million and $780 million and adjusted EBITDA between $100 million and $105 million. With that, we now welcome your questions. Operator, please begin the Q&A.

接下來談展望。對於 2026 財年第四季，我們預期營收介於 7.60 億至 7.80 億美元之間，調整後 EBITDA 介於 1.00 億至 1.05 億美元之間。接下來我們歡迎各位提問。接線生，請開始問答環節。

(Operator Instructions)

(接線生指示)

Andy Kaplowitz, Citigroup.

Andy Kaplowitz，花旗集團（Citigroup）。

So Rick, I know you said that you've now installed a SymMicro prototype into a Walmart store. So maybe you can give more color into where you are in that development process. I think you've said previously, you could see conversion on the $5 billion in Walmart backlog before this -- the end of the calendar year. Is that still the right time frame?

Rick，我知道你提到你們已經在一家 Walmart 門市安裝了 SymMicro 原型機。也許你可以更詳細說明目前開發流程進展到哪一步。我記得你之前說過，在今年曆年年底之前，你可能會看到 Walmart 50 億美元在手訂單的轉換。這個時間表仍然正確嗎？

And then as SymMicro has evolved, how have you thought about the ultimate opportunity even beyond the initial $5 billion? I think, for instance, you've been working on solving perishables with the smaller system. So maybe just an update would be helpful.

另外，隨著 SymMicro 的演進，你們如何看待超越最初 50 億美元的最終機會？例如，我想你們一直在用較小的系統來解決生鮮（perishables）問題。所以如果能提供一些最新進展會很有幫助。

I think you've covered the whole waterfront there. SymMicro, we are installing, it will take I don't know, about six months into our first Walmart -- the new version of our system into the first Walmart store. We're running 19 of the old versions, but we've been working with Walmart to develop this. And so that will be come to life about six months from now. That should trigger expanded.

我想你在問題裡幾乎都涵蓋了。SymMicro，我們正在安裝；我不確定，大概需要六個月左右，才能把我們系統的新版本在第一家 Walmart 門市完成導入。我們目前在運行 19 套舊版本，但我們一直與 Walmart 合作開發這個新版本。因此大約六個月後就會正式上線運作。這應該會觸發擴大。

We expect that will work very well. We have a second site that will follow shortly after that. And then that should trigger a bunch more sites once Walmart actually sees the system working. Your second question on perishables. We have a lot of interest in perishables.

我們預期它會運作得非常好。第二個站點會在那之後不久跟進。接著，一旦 Walmart 實際看到系統運作，就應該會觸發更多站點。你第二個問題關於生鮮。我們在生鮮方面有很多興趣。

It seems like something clipped in the rest of the world. And so I think the realization that with the new structure that you can save so much money on the construction cost of these perishable buildings, and they're so expensive to start with. We've had a lot of interest. So we would expect within the next 6 months, I guess, I would say, to begin building our first prototypes and testing stuff.

看起來在世界其他地方有些內容被截掉了。因此我認為大家逐漸意識到，透過新的結構，可以在這些生鮮建築的施工成本上節省很多，而這些建築本來就非常昂貴。我們收到很多關注。所以我們預期在接下來大約 6 個月內，應該會開始建造第一批原型並進行測試。

Very helpful. And then, Izzy, maybe just revenue is beginning to accelerate now in Q4 as per your guidance. Given the new storage structure, it seems like it's allowing you to accelerate deployments. Ultimately, do we continue to see continued acceleration in FY27 in revenue, at least how do we think about that if you don't want to give specific guidance?

非常有幫助。另外，Izzy，依照你們的指引，第四季營收現在開始加速。考量新的儲存結構，似乎讓你們能加快部署。最終來看，我們是否會在 2027 財年看到營收持續加速——至少如果你不想給出具體指引，你會如何看待這件事？

Yes. I think, as Rick said, you covered it in your question. Obviously, as we unveiled the next-generation storage structure, we were expecting that inflection point. I think we're just starting that out. I think the sequential improvement quarter-over-quarter, including our guide is, call it, steady.

是的。我想如 Rick 所說，你在問題裡已經點到了。顯然，當我們揭示下一代儲存結構時，我們就預期會出現那個拐點。我認為我們才剛開始進入那個階段。我認為按季的改善（包含我們的指引）可以說是穩健的。

I think though the real inflection point of the next-generation storage structure will really happen in the second half of next year as we proceed with the installation of that.

不過我認為，下一代儲存結構真正的拐點，會在明年下半年、隨著我們推進安裝時才會真正發生。

Matt Summerville, D.A. Davidson.

Matt Summerville，D.A. Davidson。

A couple of questions. Can you maybe provide an update on where you are with customer acquisition for Exol and maybe update where you're at with site launches? And then I'm also curious as to what initial inbound interest is with respect to that ARMS acquisition you referenced earlier.

有幾個問題。你們能否更新一下 Exol 的客戶開發進度，以及站點啟用的最新狀況？另外我也好奇，對於你們先前提到的 ARMS 收購案，目前初步的外部主動詢問（inbound interest）情況如何？

So on Exol, our Atlanta site has gone live. We are receiving product there. Customer was asked not to be named yet, but that site is now live and receiving product. And then our Lathrop site, which the customer is C&S, that Symbotic system is now complete.

關於 Exol，我們亞特蘭大站點已經上線。我們正在那裡接收產品。客戶目前要求暫不具名，但該站點已經上線並開始收貨。接著是我們的 Lathrop 站點，客戶是 C&S，那套 Symbotic 系統現在已經完成。

And so that site will go live within the next 60 or 90 days, and that will be a nice revenue-producing site because right now, we have the system in there, but we're not getting any revenue as the cases go through Symbotic. So we're feeling good about these sites are coming online. It's been a journey to get these sites filled up, but the reality is we have to get the buildings built and show people.

因此該站點將在未來 60 到 90 天內上線，這會是一個不錯的營收貢獻站點，因為目前系統已經在那裡，但當箱件（cases）通過 Symbotic 時，我們還沒有取得任何營收。所以我們對這些站點陸續上線感到樂觀。要把這些站點填滿是一段旅程，但現實是我們必須先把建築蓋好並展示給大家看。

So -- and we have a lot of incoming. We have five buildings. And so the fact that we have 5 buildings, we're able to talk to bigger customers as well as smallest customers. But it's a process. But the answer to your first question is we are live in Atlanta and receiving product. And in Lathrop, California, we will start filling out about 100% of what we plan there within the next 60 to 90 days.

所以——我們也有很多主動洽詢。我們有五棟建築。因此有 5 棟建築這件事，讓我們能與大型客戶以及較小型客戶洽談。但這是一個過程。不過回答你第一個問題：我們在亞特蘭大已經上線並接收產品。而在加州 Lathrop，我們將在未來 60 到 90 天內開始把那裡的規劃填充到大約 100%。

Give him the insights on ARMS.

給他 ARMS 的洞見。

And then on ARMS. ARMS we're doing the integration of ARMS software with the operating system from Symbotic. We have a site -- our first site that we're doing the integration will be the testing, and then we'll be able to show people how that will work. And we think that will be a very nice revenue -- software revenue business for us because we think it creates great value in improving the efficiencies of the maintenance whole system and process.

至於 ARMS。ARMS 方面，我們正在把 ARMS 軟體與 Symbotic 的作業系統進行整合。我們有一個站點——我們進行整合的第一個站點將用於測試，之後我們就能向客戶展示其運作方式。我們認為這會成為一個非常不錯的營收——軟體營收業務，因為我們相信它能透過提升維護整體系統與流程的效率而創造很大價值。

As a follow-up --

作為後續追問——

Yes, go ahead.

是的，請繼續。

Go ahead.

請繼續。

No, that will actually be one of the best examples of integrating AI with a software system because that system will actually be able to tell an operator what's wrong, where the inventory is, which operator should go fix it. So that's going to be a very sweet little business for us.

不，那其實會是將 AI 與軟體系統整合的最佳案例之一，因為那個系統實際上能告訴操作人員哪裡出了問題、庫存在哪裡、以及應該由哪位操作人員去修復。所以那會是我們一門非常不錯的小生意。

Appreciate that color. Curious if Southern Glazer's is using that next-gen storage structure and maybe remind us what the site opportunity may ultimately look like with that customer.

感謝提供這些補充說明。我想了解 Southern Glazer's 是否在使用那種新一代的儲存結構，也請你提醒我們，與該客戶合作的據點機會最終可能會是什麼樣子。

So with Southern Glazer's specifically?

所以是針對 Southern Glazer's 具體來說？

Yes.

是的。

Southern Glazer's is not -- their second site is not using the newest structure in part because the way that the liquor industry works, the cases are more standardized. And so I think the third and fourth sites probably will, but the second site was already started in design. And these are heavy -- these are liquor heavy bottles and the case sizes are pretty standard.

Southern Glazer's 不是——他們的第二個據點沒有使用最新的結構，部分原因是酒類產業的運作方式使得箱裝規格更標準化。因此我認為第三與第四個據點可能會使用，但第二個據點在設計上已經啟動了。而且這些很重——這些是以酒類重瓶為主，箱子的尺寸相當標準。

So it's not -- the new structure is beneficial to them, but in more varied block sizes, it's even more beneficial. But the real answer is they would have used it, but we already started with the old structure when we designed it, and it's just too far down the road.

所以並不是——新結構對他們是有幫助的，但在更為多樣的區塊尺寸下，它會更有利。但真正的答案是：他們本來會用，只是我們在設計時已經採用舊結構並開始推進了，現在進度太後面了。

And then for the potential there, as Rick mentioned in his prepared remarks, we're starting the second one. And as you know, they serve in 47 U.S. markets, including in Canada.

至於那邊的潛力，如 Rick 在事先準備的發言中提到的，我們正在啟動第二個。而且如你所知，他們服務美國 47 個市場，包含加拿大。

Joe Giordano, TD Cowen

Joe Giordano，TD Cowen

Just a couple of clarifications. The Atlanta site for Exol, is that like a one customer site. I know you mentioned the customer doesn't want to be named. Is that customer planning on taking the whole --

我有幾點需要釐清。Exol 的亞特蘭大據點，那是像單一客戶的據點嗎？我知道你提到客戶不希望被點名。那位客戶是否計畫承租全部的--

No, that's a multi-customer site. We're just receiving the first customer, and we haven't determined how much space they're going to need, but they're building up pretty quickly. That will be a multisite.

不，那是多客戶據點。我們只是先接收第一位客戶，我們還沒確定他們需要多少空間，但他們擴張得很快。那將會是多據點。

And then on the micro fulfillment, I'm just curious as you said you're going to deliver -- you're building it out 6 months and then you'll do a second. What's the mechanism in the contract? I thought the contract was kind of once they accept it, it automatically triggers the $5 billion and the 400-store order. What is required to have that hit?

另外關於微型履約（micro fulfillment），我想確認一下：你說你們會交付——你們會建置 6 個月，然後會做第二套。合約裡的觸發機制是什麼？我原本以為合約是他們一旦驗收，就會自動觸發 50 億美元以及 400 家門市的訂單。需要什麼條件才會讓那個生效？

The way we've done things with Walmart in partnership is we build a prototype, we build them so that they work, but we also know that where you can tell from the prototype. We're building it into a store. I'm not sure I'm supposed to announce the store, but it will become obvious pretty soon.

我們與 Walmart 的合作方式是：我們先建一個原型，確保它能運作，但我們也知道從原型就能看出——我們會把它建到一間門市裡。我不確定我是否應該公布是哪一家門市，但很快就會很明顯。

But we'll build it into the store and then we'll -- we overbuild it to make sure that it works. And then we redesign it to make sure that we've got the cost out, in this case, make it smaller, make it more efficient. Walmart may add items; they may delete items. And so -- and when we do the second version, that's usually what triggers, okay, we want 400 of these.

我們會把它建到門市裡，然後——我們會先做得更高規格一些，以確保它能運作。接著我們會重新設計，確保把成本降下來；在這個案例中，就是做得更小、更有效率。Walmart 可能會新增品項；也可能會刪減品項。所以——當我們做第二版時，通常就會觸發：好，我們要 400 套。

Got it. Okay.

了解。好的。

The most important thing with these sites is there's the coordination of the hardware, but most of the time, what's happened with these micro fulfilment sites is that the software hasn't been flexible enough and the software and the automation haven't been coordinated enough. So we're going to overbuild this, but we probably won't build 400 of the version we're building now. But I think the one after this, we will.

這些據點最重要的是硬體的協調，但多數時候，這些微型履約據點遇到的問題是軟體不夠彈性，而且軟體與自動化之間的協調不夠。所以我們會先做得更高規格一些，但我們可能不會用現在這個版本去建 400 套。但我認為下一個版本會。

Great. Izzy, just how should we think about the pacing of system adds maybe for next quarter and into the near future?

很好。Izzy，對於下個季度以及近期未來，系統新增的節奏我們應該怎麼看？

Yes. So as you noticed, right, we had a great three quarters in a row. I had originally mentioned a couple of quarters ago, maybe the fourth would be a little light. But actually, now as I'm seeing the trajectory, I think the fourth quarter will be in line with the third, maybe just a little short of the third. So great, great expectations where we've been in the last three quarters and where we're going to land for the year.

是的。如你所注意到的，我們已經連續三個季度表現很強。我在幾個季度前原本提到，第四季可能會稍微偏淡。但現在看這個走勢，我認為第四季會與第三季大致一致，可能只會略低於第三季。所以，對於過去三季的表現，以及我們今年最終落點，我們有非常、非常高的期待。

Ken Newman, KeyBanc Capital Markets.

Ken Newman，KeyBanc Capital Markets。

Maybe for my first question, Izzy, maybe you can help us to think about -- I'll ask the new storage system or the revenue question on systems a little bit differently. As you think about the new storage system now being fully implemented, how should we think about the cadence of segment gross margins on that improvement, just given that you do expect that to maybe ramp, it sounds like maybe later in the back half of next year.

我的第一個問題，Izzy，也許你可以幫我們換個角度思考——我想用不同方式問一下新儲存系統或系統收入的問題。當你們的新儲存系統現在已全面導入，你認為我們應該如何看待該改善對分部毛利率的提升節奏？考量到你確實預期它可能會逐步爬升，聽起來可能要到明年下半年較後段。

But just trying to think about the opportunity for gross margin improvement there and the cadence of that in coming quarters.

我只是想理解毛利率改善的機會，以及未來幾季的節奏。

Okay. So let me unpack your question a little bit. First with -- just let me repeat what Rick was saying on the micro fulfillment. So we're starting now the first prototype. We expect to get into, after that, or maybe in the middle of that, getting the second prototype.

好的。我先把你的問題拆解一下。首先，關於微型履約——我先重申 Rick 剛才說的。我們現在開始做第一個原型。我們預期在那之後，或可能在進行到一半時，就會著手第二個原型。

I really am not expecting just yet the micro fulfillment, call it, the store order that is mentioned in the contract probably until early 2028. So then when you think about margins, right, our whole journey of improving margins, right? This contract is more profitable from that perspective.

我目前並不預期微型履約——也就是合約中提到的門市訂單——會在 2028 年初之前發生。所以當你思考毛利率時，對吧，我們整體提升毛利率的旅程。從這個角度來看，這份合約的獲利性更高。

So you just have to think of it as we continue the mix, right? The first step, as I've been talking about is probably closer to the second half of next year. We get the inflection point of really having the installation of the next-gen system, which will improve margins.

所以你只要把它視為我們持續在做組合（mix）的調整。第一步，如我一直在談的，可能會更接近明年下半年。我們會迎來真正的拐點：開始大量安裝下一代系統，這將改善毛利率。

Then you also then end up adding in the back of stores and that being also a big part of the mix, which gets us to, in this journey, how our margins continue to improve. The one thing I will say about margins, we had a great quarter from a margin perspective.

接著你也會看到我們把系統加到門市後場，而那也會成為組合中的一大部分，進一步推動我們在這段旅程中毛利率持續改善。關於毛利率我想說的一點是：我們這一季在毛利率方面表現非常好。

As I said last quarter, I was expecting stable margins. The quarter was really, really strong. I think the fourth quarter will behave very similar to our exit trend in the second. So I hope that helps, [Ken.]

如我上季所說，我原本預期毛利率會持平。但這一季真的非常、非常強。我認為第四季的表現會非常接近我們第二季的期末趨勢（exit trend）。希望這能有所幫助，[Ken。]

Third.

第三。

Third.

第三。

Yes. That's very helpful. I appreciate that. Maybe for the follow-on here, Rick, it was interesting to see a couple of bolt-on deals this quarter. You did a bolt-on last quarter as well. As you look at the forward innovation pipeline, is there any color you can give on just other types of deals that you're looking to maybe help you drive faster deployments?

是的。這非常有幫助。謝謝。接著延伸問一下，Rick，這一季看到幾筆補強型（bolt-on）交易很有意思。你們上季也做了一筆補強型交易。當你看向未來的創新管線時，能否分享一些你們正在評估的其他類型交易，看看是否能幫助你們推動更快的部署？

And I would also be curious just if there's anything you can kind of talk about on what you're spending on AI development in terms of token spend versus the hardware spend on R&D?

我也想請教一下，關於你們在 AI 開發上的支出，能否談談在代幣（token）消耗相較於研發硬體支出方面的大致情況？

Yes. So -- we are looking at more bolt-ons. It's an interesting time. As you guys know, there's so much money chasing AI that a lot of the traditional automation companies are running into funding problems or -- and so we become a very good place for people to approach us as investors or acquirers. So we're -- that's why we've built up our balance sheet.

是的。所以——我們正在看更多的補強式併購（bolt-ons）。這是一個很有意思的時點。如各位所知，有大量資金追逐 AI，導致許多傳統自動化公司遇到募資問題——因此我們成為許多人來找我們、把我們視為投資人或收購方的理想對象。所以我們——這也是為什麼我們把資產負債表做得更強健。

We guessed right about that. We are right about that. I think we'll see continued opportunities there to acquire hardware. In the case of ARMS, it was a software. So -- and some companies we're looking at are a combination of interesting technology, both hardware and software in vision.

我們當時判斷得很準。我們現在也仍然認為如此。我想我們會持續看到在收購硬體方面的機會。以 ARMS 的案例來說，那是一個軟體。所以——我們正在看的某些公司，具備有趣的技術組合，在視覺領域同時涵蓋硬體與軟體。

The question you asked about AI is, so the way I would describe it is, I think we were doing AI five years ago before anybody called it AI. So we've been doing self-driving cars. We've been doing vision; we've been doing LiDAR. We generate, I think, 1 trillion bits of data every day at every site. Maybe it's 100 billion.

你問到 AI 的問題，我會這樣描述：我認為我們在五年前、在大家還沒把它稱作 AI 之前，就已經在做 AI 了。我們一直在做自動駕駛車。我們一直在做視覺；我們一直在做 LiDAR。我們每天在每個站點產生，我想，大約 1 兆位元（bits）的資料。也可能是 1,000 億。

It's an incredible amount of data at every site. And so we're looking for -- to economically store it in the cloud. And then we are writing our own AI agent. So yes, we're using some AI to audit code. And that's helpful.

每個站點的資料量都非常驚人。因此我們在尋找——以具經濟效益的方式把它儲存在雲端。然後我們正在撰寫自己的 AI 代理（agent）。所以是的，我們有用一些 AI 來稽核程式碼。這很有幫助。

But mostly, what we will do is we will develop our own AI agents that will actually be able to predict and tell us what's going to go wrong with our systems before they go wrong and then actually communicate to the robots, drive them out of the system, tell the maintenance people what's wrong with them and fix them.

但主要來說，我們會開發自己的 AI 代理，實際上能在系統出問題之前就預測並告訴我們哪裡會出狀況，接著直接與機器人溝通，把它們從系統中引導出去，告訴維修人員它們哪裡有問題，並把它們修好。

So that's not something that we're going to pay a lot of money for outside. That's something that we've been building here for a long time. And that's why I think most people consider us one of the leading companies in the world with physical AI. And I think there's a lot of misnomers about that, but we're actually doing it, and we've been doing it for a long time before we used to call it machine learning.

所以這不是我們會花很多錢向外部購買的東西。這是我們在這裡長期建置的能力。也因此我認為多數人把我們視為全球在實體 AI（physical AI）方面的領先公司之一。我想外界對此有很多誤解，但我們確實在做，而且在它還被稱為機器學習之前，我們就做了很久。

They used to call it a whole bunch of other things. But now we're actually learning how to use AI, not just to generate reports, but actually to communicate directly with our robots and in some cases, fix them, in some cases, tell them what to do, in some cases, tell them where to go to the exit ramp and get fixed.

以前人們用很多不同的名稱來稱呼它。但現在我們正在學習如何使用 AI，不只是產出報告，而是能直接與我們的機器人溝通；在某些情況下修復它們；在某些情況下告訴它們該做什麼；在某些情況下告訴它們該往哪個出口匝道離開並去維修。

If you just -- if I could just clarify that last point, Rick. When I -- when you talk about scaling that infrastructure on the AI software side, does that require an incremental or scale up in tokens needed to operate that system? Or is that really just on the inference that you get to scale?

Rick，如果我可以釐清剛才最後那點。當你談到在 AI 軟體端擴大那套基礎設施時，這是否需要額外增加或擴大運作該系統所需的代幣（token）用量？還是說主要只是推論（inference）端的規模化？

Yes. So that's a great question. We're using some tokens, but there's a lot of open source AI. There's a lot of AI that -- and we're also looking at different forms of AI. There's some AI that we can actually not have to go to the cloud.

是的。這是個很好的問題。我們會用到一些代幣，但也有很多開源 AI。也有很多 AI——而且我們也在看不同形式的 AI。有些 AI 我們其實不必上雲端。

We can actually imbue that technology right into our bots because with the new NVIDIA chips, we have 4x as much storage, and we'll have more storage on our bots that we didn't have 2 years ago. So I don't think tokens, I don't think AI expense is going to be a major issue for us, and we're very focused on doing as much as we can internally ourselves.

我們可以把那項技術直接賦能到我們的機器人上，因為有了新的 NVIDIA 晶片，我們的儲存容量增加了 4 倍，而且我們的機器人端會有比兩年前更多的儲存空間。所以我不認為代幣、也不認為 AI 費用會成為我們的重大問題；我們非常專注於盡可能在內部自行完成。

So we don't -- one of the things that we've learned is that about 80% of the AI that maybe we looked at using last year was a lot of formatting. It was not actually using the data that we needed. So one of the things we're focused on is because we generate so much data, because we've always mined our own data, we're actually looking at what's the most efficient way to use our data that's cost effective.

所以我們不——我們學到的一件事是，去年我們可能考慮使用的 AI 裡，大約 80% 其實是在做格式整理。它並沒有真正使用到我們需要的資料。因此我們聚焦的一件事是：因為我們產生的資料量非常大，而且我們一直以來都在挖掘自有資料，我們正在尋找以具成本效益的方式、最高效率地使用我們資料的方法。

Mark Delaney, Goldman Sachs.

高盛（Goldman Sachs）的 Mark Delaney。

I think better margins was one of the key highlights from the quarter. I believe revenue was $11 million above the midpoint of your guidance, but EBITDA came in $12 million better. So can you share more on what led to the degree of margin improvement in 3Q and the upside relative to your expectation?

我認為更好的利潤率是本季的主要亮點之一。我相信營收比你們指引區間的中點高出 1,100 萬美元，但 EBITDA 則好 1,200 萬美元。所以你們能否分享更多，第三季利潤率改善幅度的原因，以及相對於你們原先預期的上行因素？

Sure. I'll take that. So just to unpack the margins, right? If you think about it just in the amount of revenue we had in the systems, those margins came in quite solid. Quarter-over-quarter, they actually came a little bit better than I was expecting originally.

當然。我來回答。先把利潤率拆解一下，對吧？如果你只看系統業務的營收規模，那些利潤率表現相當穩健。與前一季相比，實際上還比我原先預期的略好一些。

But those really come down to the project execution and the mix of business we had in the quarter. I think the other thing that came in nicely this quarter was the fact that op services, right, it continues to deliver profitability. Maybe it was a little bit better than I expected. But at the end of the day, I expect next quarter to be in line with this year -- with this quarter's revenue.

但這主要取決於專案執行，以及本季我們的業務組合。我想本季另一個表現不錯的部分是營運服務（op services），對吧，它持續帶來獲利能力。可能也比我預期的略好。但歸根結底，我預期下季的營收會與今年——與本季的營收水準一致。

And then last but not least, as you hit about on EBITDA margin, the operating leverage was really good because when you look at the non-GAAP OpEx year-over-year, it was only up 3%. So a combination of all those things, be it systems, operation services and really the scale that we're getting in our OpEx really allowed us to deliver a more profitable quarter.

最後但同樣重要的是，正如你提到的 EBITDA 利潤率，營運槓桿非常好，因為你看非 GAAP 的營業費用（OpEx）年增只有 3%。因此，系統、營運服務，以及我們在 OpEx 上取得的規模效益，這些因素的組合讓我們交出更高獲利的一季。

Very helpful. My other question was on cash flow. Izzy, you talked about timing as the reason that the free cash flow was a headwind in the quarter. Help us understand how to think about free cash flow for the upcoming quarter, should some of those timing issues persist or maybe the better EBITDA will drive improved cash flow?

非常有幫助。我另一個問題是關於現金流。Izzy，你提到時點是本季自由現金流（free cash flow）形成逆風的原因。請協助我們理解接下來一季該如何看待自由現金流：這些時點問題是否會持續，或是更好的 EBITDA 會帶動現金流改善？

Yes. I would certainly look at the free cash flow for the quarter just as a timing item and not even timing that I have to wait for the whole fourth quarter. Those are really payments that just came in a week later. So I would say if I had a week more in the quarter, you wouldn't see no blip in that.

是的。我會把本季的自由現金流視為純粹的時點因素，甚至不是那種必須等到整個第四季才會反映的時點。那些其實就是款項晚了一週才入帳。所以我會說，如果這一季多一週，你就不會看到那個小波動。

I think the better way to think about it is, to your question on fourth quarter, I would expect a positive free cash flow. And I think just in general, given our business, it's better to measure us over a longer period of time and that the free cash flow will be on an annual basis will be positive.

我認為更好的看法是，回到你對第四季的問題，我預期自由現金流會是正數。而且整體而言，考量我們的業務特性，用較長期間來衡量會更合適；自由現金流以年度基礎來看會是正數。

Quinn Fredrickson, Baird.

Quinn Fredrickson，Baird。

Now that you've owned Fox Robotics for a little bit, curious if there's any updates to their product that you've made or are contemplating that improves the integration with your system. And I think you've also mentioned some of their largest customers are not Symbotic customers. So any updates on discussions with any of their customers and whether they could be potential customers?

現在你們持有 Fox Robotics 一段時間了，想請教你們是否已對他們的產品做了任何更新，或正在考慮哪些更新，以改善與你們系統的整合。另外我記得你們也提到，他們一些最大的客戶並不是 Symbotic 的客戶。那麼，與他們客戶的討論有沒有任何最新進展？他們是否可能成為潛在客戶？

Yes. We've been very encouraged. All of the Fox customers are actually delighted that we bought the company. We're in talks with all of them. We've hired some new folks there. We've hired some new salespeople there. We're sitting down and doing a complete review with two of the larger customers and just talking about what they would like for next versions, what they would like for next steps.

是的。我們一直非常受到鼓舞。所有 Fox 的客戶其實都很高興我們收購了這家公司。我們正在與他們全部洽談。我們在那邊也聘請了一些新同仁。我們也聘請了一些新的業務人員。我們正與兩家較大的客戶坐下來做全面檢視，並討論他們希望下一版本有哪些功能、下一步想要什麼。

And so I think that's going to be a very nice business. We've been very encouraged. We've had no headwinds. We're actually -- I think the customers we're talking to are saying, we're really excited to own this company. And in some cases, they might want a Symbotic system.

因此我認為那會是一門很不錯的業務。我們一直非常受到鼓舞。我們沒有遇到任何逆風。實際上——我認為我們正在交流的客戶都在說，他們對我們擁有這家公司感到非常興奮。而在某些情況下，他們可能也會想要一套 Symbotic 系統。

And in some cases, they're actually really interested in the combination of Fox, the ARMS software, some of the other software we're looking at and actually helping them with a dock management system. So -- it just -- we just started. It's a very small company, but I think it's got a very big potential. And we've been excited about the reception that we've got from all of the Fox customers.

而在某些情況下，他們其實對 Fox、ARMS 軟體、以及我們正在評估的其他一些軟體的組合非常感興趣，並希望我們實際上協助他們打造一套碼頭管理系統。所以——這才剛開始。這是一家非常小的公司，但我認為它有非常大的潛力。我們也對所有 Fox 客戶給予的回饋感到振奮。

And for my follow-up, Izzy, you've been on a nice sequential EBITDA margin progression for the better part of two years. You did mention the fourth quarter guide kind of implies flattish EBITDA margin sequentially despite higher revenue. Could you just unpack maybe why margins wouldn't continue to improve with operating leverage?

我的追問是，Izzy，你們在過去將近兩年的時間裡，EBITDA 利潤率一直呈現不錯的逐季改善。你提到第四季的指引在營收更高的情況下，隱含 EBITDA 利潤率可能與前一季大致持平。你能否拆解一下，為什麼在營運槓桿之下，利潤率不會持續改善？

I think right now, I just want to make sure that we see it coming. So right now, based on our latest forecast, we do expect OpEx to just increase slightly, and that would be more on the SG&A side. So maybe it comes in better, but right now, my expectation is that the OpEx would be just a slight uptick.

我想目前我只是想確保我們看得到它正在發生。所以就目前、基於我們最新的預測，我們確實預期營業費用（OpEx）只會小幅增加，而增加主要會在銷售、一般及行政費用（SG&A）那一側。所以也許結果會更好，但就目前而言，我的預期是 OpEx 會小幅上升。

I think the other part, as I mentioned earlier, the gross margins where we landed on a non-GAAP basis of 25% this quarter. Right now, I'm going back to what I said I was expecting stabilization at the end of the second quarter. So if they come in closer to in line with the second quarter, those are the 2 main reasons why you would see that EBITDA margin would be flat.

我想另一部分，如我先前提到的，本季我們在非 GAAP 基礎下的毛利率落在 25%。目前我回到我之前說的：我原本預期在第二季末會趨於穩定。因此如果毛利率回到更接近第二季的水準，這就是你會看到 EBITDA 利潤率持平的兩個主要原因。

Guy Hardwick, Barclays.

Guy Hardwick，Barclays。

Just whether you could update us on the remaining performance obligations. I think the 10-Q says $22.5 billion and 15% realized over the next 12 months. It doesn't look like the changes were as significant this quarter than the previous quarter. Just whether there's anything unusual or just -- or is it just regular kind of contract plus ups as you begin deployments, particularly? I think there's another -- I think you said 11 starts.

想請你們更新一下剩餘履約義務（remaining performance obligations）。我記得 10-Q 上寫的是 225 億美元，且未來 12 個月內將實現 15%。看起來本季的變動不像前一季那麼顯著。想確認是否有任何不尋常之處，還是只是隨著你們開始部署、特別是在部署啟動時，屬於一般性的合約新增與加量？我想你們還提到有 11 個啟動。

That's correct. So the $22.5 billion and the banding of within the next 12 months of 15%, that's exactly what we put out there. I think it just has once again to do with the mix of deployments. So just before I even get into the deployments, right, the $22.5 billion coming off at $22.7 billion had a healthy amount of revenue in the quarter, you decrease it.

沒錯。225 億美元以及未來 12 個月內 15% 的分段（banding），就是我們對外揭露的數字。我認為這再次與部署組合（mix of deployments）有關。所以在我談部署之前，225 億美元是從 227 億美元下來的，本季確認了相當健康的營收，因此會把它扣減下來。

As you know, we have pricing adjustments when we redo the backlog plus the fact that we added Southern Glazer's. But it really then comes down when you're tracking it, it just really comes down to the 11 deployments we are putting in, in the quarter, just what those pricing adjustments were.

如你所知，我們在重新計算積壓訂單（backlog）時會有價格調整，再加上我們新增了 Southern Glazer's。但真正追蹤時，主要還是取決於本季我們正在導入的 11 個部署，以及這些價格調整的幅度。

So it could be lumpy at any given time. I think the more promising thing is that despite the revenue that we're generating every single quarter, our backlog still remains very, very stable. And as we said before, that backlog still doesn't include the contract for the 400 back of store systems.

所以在任何特定時間點都可能會有波動。我認為更令人振奮的是，儘管我們每一季都在產生營收，我們的積壓訂單仍然非常、非常穩定。而且如我們先前所說，這個積壓訂單仍未包含 400 套店後（back of store）系統的合約。

And it looks like revenue deployments have been falling now for at least four quarters. Is that kind of a mix effect? What is the kind of -- it seems a little odd that system sizes in the Walmart business is actually going up, right?

而且看起來每次部署所帶來的營收已經至少連續四季在下降。這是組合效應嗎？因為看起來有點奇怪，Walmart 業務的系統規模其實是在變大，對吧？

Agreed. But it also just has to deal with at what point in the cycle we are in the installation phase, right? Because the revenue is going to come in as we get closer to month 13 forward. So there is a little bit of lumpiness, but I think it's better instead of just focusing on one given quarter, if you look at the multiple of the quarters and where we are and what the expectation is going forward, given that the fit banding is at 15% for the next 12 months.

同意。但這也與我們在安裝階段所處的週期位置有關，對吧？因為營收會在我們更接近往後第 13 個月時才會進來。所以確實會有一些波動，但我認為與其只看單一季度，不如看多個季度、我們目前的位置，以及在未來 12 個月 15% 分段的前提下，往前的預期是什麼。

So does that mean that you'd expect revenue per deployment to start going up again? Or will it continue to sort of trend down?

所以這是否意味著，你會預期每次部署的營收會再次開始上升？還是會繼續呈現下滑趨勢？

I think in the coming -- we don't guide to backlog. I think the expectation for the next quarter, given the guide we gave, that's really where our expectation. But of course, we're always looking to not only have stable backlog, but to increase our backlog.

我想在接下來——我們不對積壓訂單提供指引。我認為對下一季的預期，就以我們給出的指引為準，那就是我們的預期所在。但當然，我們一直在努力不僅維持積壓訂單穩定，也要增加積壓訂單。

Well, revenue is going up. We expect revenue to go up.

嗯，營收正在上升。我們預期營收會上升。

Yes. And we do expect revenue to go up, of course. That's where the 15% comes in.

是的。而且我們當然也預期營收會上升。這就是 15% 的由來。

Colin Rusch, Oppenheimer & Company.

Colin Rusch，Oppenheimer & Company。

Now with the ARMS platform purchased, can you talk a little bit about the opportunity to start introducing new offerings with semi-automation or robots that are more interactive with humans and existing assets that might be a little bit lower barrier to entry for some of the customers that you might want to grow with?

現在你們已收購 ARMS 平台，能否談談是否有機會開始推出一些新產品/方案，例如半自動化，或是與人類及既有資產互動性更高的機器人；這些方案對某些你們希望拓展合作的客戶而言，可能價格更低、導入門檻也更低，讓他們先起步，之後再逐步走向全自動化系統？

I'm not sure I understand your question.

我不確定我理解你的問題。

I'm just looking for a sense of opportunities that you guys could bring to market that would be a little bit lower price for customers, a little bit lower barrier to entry to get them started as they move towards fully automated systems.

我只是想了解，你們有哪些機會可以把一些產品帶到市場上，讓客戶以較低的價格、較低的導入門檻先開始使用，並在他們邁向全自動化系統的過程中逐步升級。

Yes. So I think the ARMS software is something that we could sell to a customer. The company actually doesn't make anything except software. So we could sell that to customers and introduce our software. The other thing is that the Fox robots, these are $100,000 machines.

是的。所以我認為 ARMS 軟體是我們可以賣給客戶的東西。這家公司實際上除了軟體之外不做任何產品。因此我們可以把它賣給客戶，並導入我們的軟體。另外一點是，Fox 的機器人，這些是 10 萬美元的機器。

And so I think that is -- the way I look at it is, we will become -- that our hardware will continue to grow. Our sales are going to continue to grow, but we will become much more of a software-centric company that's selling machines that basically perform for what we want our software to do.

因此我認為——我看待這件事的方式是，我們將會——我們的硬體會持續成長。我們的銷售將會持續成長，但我們會變得更像一家以軟體為核心的公司，銷售的機器基本上是為了實現我們希望軟體完成的功能。

So for instance, some of the Fox customers as a company, it's no secret, it's DHL, one of the largest 3PLs in the world. They really like the Fox robots. They want us to help them manage the dock. They may never buy a Symbotic system. But if you sell -- I don't know, you sell 20,000 of these $100,000 machines, that's a pretty good sale.

舉例來說，Fox 的一些客戶作為公司，這不是祕密，就是 DHL，全球最大的三家 3PL 之一。他們非常喜歡 Fox 機器人。他們希望我們協助他們管理碼頭作業。他們可能永遠都不會購買一套 Symbotic 系統。但如果你賣——我不知道，比方說你賣出 20,000 台每台 10 萬美元的機器，那就是一筆相當不錯的銷售。

So I'm not saying we sell that to DHL, but it's a huge market out there, and it's a much easier point of entry. Your question is appropriate because the last two weeks, we've had two major potential customers, retailers who are interested in automation, great companies, well-known names. And they're really looking at how they can enter into the automation space without a lot of experience.

所以我不是說我們會把這賣給 DHL，但外面有一個巨大的市場，而且這是一個更容易切入的切入點。你的問題很恰當，因為在過去兩週，我們有兩個主要的潛在客戶——對自動化有興趣的零售商——很棒的公司、知名的名字。而他們真正關注的是，在缺乏大量經驗的情況下，他們如何能進入自動化領域。

And so we can sell them a very small system. We can sell them a small system and a dock system. So that's one of our focuses is to get some of these very large customers in with an entry-level product. So it could be a single one-in and a one-out sell. That could be in the tens of millions, low tens of millions number. And so yes, that's what we're looking at.

因此我們可以賣給他們一個非常小的系統。我們可以賣給他們一個小型系統以及一個碼頭系統。所以我們的重點之一，是用入門級產品把這些非常大型的客戶先導入。因此它可能是一個單一的一進一出（one-in、one-out）的銷售。那可能是數千萬美元等級、低數千萬美元的金額。所以是的，這就是我們正在看的方向。

And not to mention that the back of store system will be...

更不用說，店後端系統將會是……

The back of the store system is another opportunity.

店後端系統是另一個機會。

Perfect. And then there's certainly been a lot of investment around perception technology. And notably, one of the LiDAR vendors is now selling LiDAR with color capability and functional safety. I'm just curious about how much leverage you might get from those sorts of perception solutions into simplifying bot design and optimizing performance and how we should think about the adoption cycle in some of those newer perception technologies going forward?

很好。另外，在感知技術方面確實有大量投資。特別是，其中一家 LiDAR 供應商現在正在銷售具備彩色能力與功能安全（functional safety）的 LiDAR。我只是好奇，這類感知解決方案能在多大程度上讓你們在簡化機器人設計、優化效能方面取得槓桿效益，以及我們應該如何看待未來這些較新感知技術的採用週期？

Yes. So a lot of -- so there's a number of people, some of us, that are doing LiDAR, slow-moving bots to interact with people, I won't mention names of companies, but you know who they are. What we're doing is bots with LiDAR that are fast moving and weigh a lot.

是的。所以有很多——有一些人，包括我們在內，正在做 LiDAR、低速移動、與人互動的機器人；我不會提公司名字，但你知道他們是誰。我們在做的是配備 LiDAR、移動速度快而且重量很大的機器人。

The change in technology and the reason we will expect to have LiDAR on all our bots within the next, I don't know, two years on the outside is that these LiDAR used to cost four years ago, they were $5,000, now they're under $500. And so they become very affordable for our bots and then it really enables our software.

技術的變化，以及我們預期在未來——我不知道，最多兩年內——讓所有機器人都配備 LiDAR 的原因是：四年前這些 LiDAR 的成本大約是 5,000 美元，現在已經低於 500 美元。因此對我們的機器人而言變得非常可負擔，並且真正賦能我們的軟體。

So where other people are using LiDAR for basic like a Kiva bot or something that moves slow, follows the line, meant to be used with humans. What we're really doing is putting LiDAR on bots. It's like a self-driving vehicle that want to go fast.

所以當其他人把 LiDAR 用在基本款、像 Kiva 機器人之類的用途——移動很慢、沿著線走、設計上是要與人一起使用。我們真正做的是把 LiDAR 裝在機器人上。這就像一台想要高速行駛的自動駕駛車輛。

And so we're really trying to have bots that are now combined with ARMS and AI, really getting much closer to within our structure, a lifestyle facility that really, we don't -- we may go long periods of time before humans actually have to go in and interact with a bot. That kind of technology does not exist out there for warehouse automation. And that's our goal.

因此我們真的在努力打造現在結合 ARMS 與 AI 的機器人，在我們的架構內更接近一種「生活型」設施（lifestyle facility）：我們——人類可能可以很長一段時間都不必真的進去與機器人互動。這種技術在倉儲自動化領域目前並不存在。而那就是我們的目標。

Derek Soderberg, Cantor Fitzgerald.

Derek Soderberg，Cantor Fitzgerald。

Wondering, Rick, if you can expand on the ARMS acquisition a bit. You talked about a little bit in the prepared remarks and during the Q&A. I was wondering how you'll monetize that. Is that going to be a subscription or bundled through kind of the systems price? And is this more for Exol? Or is the plan to deploy this at your large existing customers as well?

Rick，我想知道你是否能多談一點 ARMS 的收購案。你在事先準備的發言以及問答環節中都有稍微提到。我想知道你們會如何將其變現。這會是訂閱制，還是會以某種方式綁在系統價格裡？而這主要是給 Exol 用的嗎？還是也計畫在你們現有的大型客戶那邊部署？

No. We will deploy this with -- as an option for all of the Symbotic customers, including Exol. So it will be a software add-on.

不是。我們會把它作為選配，部署給所有 Symbotic 客戶，包括 Exol。所以它會是一個軟體加購模組（add-on）。

Got it. And Rick, could you just talk about where this acquisition kind of started? Was this something customers were asking about? And then just high level, I was curious if you think eventually a large retailer might, in a sense, cede control of the distribution facilities to Symbotic or Exol as you sort of really fully automate the supply chain here, if maybe it makes more sense for you guys to take on the facilities and they would just pay you per case or any of those types of conversations happening?

了解。Rick，你能談談這次收購最初是怎麼開始的嗎？這是客戶在詢問的事情嗎？另外從高層次來看，我也好奇你是否認為，最終某個大型零售商可能在某種意義上，會把配送設施的控制權交給 Symbotic 或 Exol；當你們把供應鏈真正完全自動化之後，也許由你們接手設施、他們按箱（per case）付費會更合理——是否有在進行這類對話？

Yes. So Exol is definitely getting those inquiries, and we've been funneling them through Exol. We also have a number of sites, a number of customers where we sold them a system and then Symbotic runs the system at a cost per case. So -- and ARMS just means that if we were to do that with this kind of maintenance that we would charge the customer and our operating costs would be lower.

是的。所以 Exol 的確正在收到這類詢問，而我們一直把這些需求導流到 Exol。我們也有一些據點、一些客戶：我們賣給他們一套系統，然後由 Symbotic 以每箱成本（cost per case）的方式來營運該系統。所以——而 ARMS 的意義在於，如果我們以這種維護模式來做，我們會向客戶收費，而我們的營運成本會更低。

And so we would be the beneficiary, both of the software and the more efficiencies. What ARMS does is it creates a database combined with the operating system, which is inherent in every Symbotic system, but it says to somebody that -- so these -- everybody, all the maintenance people in the front of the structure or working there have a handheld device.

因此我們將同時受益於軟體以及更高的效率。ARMS 的作用是建立一個資料庫並與作業系統結合；作業系統本來就內建在每一套 Symbotic 系統中，但它會告訴某個人——也就是說，這些——每個人，所有在結構前端或在現場工作的維護人員，都有一台手持裝置。

And it would say, lift 606 in here and geolocated in this particular part of the building. And remember, some of these buildings are 1 million square feet. This lift has a failed valve, I need you to go there. Here's a picture of what it should take to fix it. I've already checked before you go. This is what AI does. I already checked these two parts are in inventory.

它會說：這裡的升降機 606，並且地理定位在建築物的這個特定區域。而且要記得，有些建築物有 100 萬平方英尺。這台升降機有一個閥門故障，我需要你去那裡。這裡有一張圖片，告訴你應該如何修復。在你出發前我已經檢查過。這就是 AI 做的事。我已經確認這兩個零件在庫存裡。

So don't go to the lift and then go to the inventory room, go to the inventory room, get these two parts, go to the lift. The whole thing should take you 40 minutes. That -- we've been struggling with how do we make these maintenance systems more efficient. And we could sell this kind of system along with some of the Symbotic software to a lot of people in the world. This is the ultimate warehouse management maintenance system.

所以不要先去升降機那邊、再去庫房；而是先去庫房拿這兩個零件，再去升降機。整件事應該花你 40 分鐘。那——我們一直在苦思要如何讓這些維護系統更有效率。而我們可以把這類系統，連同部分 Symbotic 軟體，一起賣給全球很多人。這是終極的倉儲管理維護系統。

Greg Palm, Craig-Hallum.

Greg Palm，Craig-Hallum。

I wanted to go back to the OpEx and maybe honing a little bit more on R&D. I mean, in light of a lot of these kind of newer opportunities you have perishables and micro fulfillment, it was maybe a little bit odd to see R&D come down quite as much.

我想回到營運費用（OpEx），並且可能更聚焦一點在研發（R&D）上。我的意思是，鑑於你們有不少這類較新的機會，包括生鮮與微型履約（micro fulfillment），看到研發降幅有那麼大，可能有點不太尋常。

It doesn't sound like that might go up or, I think as you said, maybe more stable. So I guess, is that just -- are we really pairing things back? Or is that more kind of a reallocation of expenses? Just wanted to get a little bit more color there.

聽起來研發不太像會上升，或者如你所說，可能會更穩定。所以我想問，這只是——我們真的在縮減投入嗎？還是更像是一種費用的重新配置？想請你多給我們一些細節。

Yes. Greg, so let me step back. First and foremost, R&D expense quarter-over-quarter was flat. All the things that Rick mentioned are the things that we're going to get started on. So hence, when I said earlier, I expect overall OpEx to go up, my expectation is that between R&D and SG&A, we do expect a little bit of an uptick.

是的。Greg，我先退一步說明。首先，研發費用（R&D expense）季對季是持平的。Rick 提到的那些事情，都是我們即將開始著手的項目。因此我先前說我預期整體營運費用（OpEx）會上升，我的預期是研發與銷售、一般及行政費用（SG&A）之間，我們確實會看到一些小幅上升。

But as always, we want to maintain the ultimate flexibility in being able to increase our R&D. And that's where I make that comment of -- that's when the EBITDA margin staying flat quarter-over-quarter is really to give us that flexibility there.

但一如既往，我們希望保有最終的彈性，能夠提高我們的研發投入。這也是我提到——EBITDA 利潤率季對季維持持平，實際上是為了在這裡給我們那樣的彈性。

So I wouldn't say R&D has come down. It has stayed flat. We've gotten, call it, to a rhythm on the things we're investigating, but I expect a little bit of an uptick, not only in the fourth quarter, but in the quarters to come.

所以我不會說研發下降了。它是持平的。我們在正在研究的事項上已經建立起一種節奏，但我預期不僅在第四季，未來幾季也會有一些小幅上升。

Okay. That makes sense. And I guess maybe just shifting topics entirely just in light of the other news, Steve is joining the Board of Directors. I'm just curious, maybe you can give us some thoughts on, given his background, kind of what he brings to the table and how he might sort of help you scale to the next level.

好的。這就說得通了。我想也許完全換個話題，鑑於其他消息，Steve 將加入董事會。我只是好奇，或許你可以談談，考量他的背景，他能帶來什麼，以及他可能如何幫助你們擴張到下一個層級。

Yes. So I met Steve through one of my other Board members, they were on a Board together, spent a bunch of time with Steve. We're -- Steve is -- with his background, I think will be very, very helpful in helping us look at strategically M&A. We plan to be acquisitive.

是的。我透過我另一位董事會成員認識 Steve，他們曾一起在同一個董事會任職，我也花了不少時間與 Steve 交流。我們——Steve 以他的背景，我認為在協助我們從策略角度看待併購（M&A）方面會非常、非常有幫助。我們計畫採取併購成長。

We've built the balance sheet to be acquisitive. And so that's what we're working on. And so Steve is a perfect Board member for that. And his background when he was at Bain was in the tech sector.

我們已經把資產負債表打造到具備進行併購的能力。所以這就是我們正在推進的方向。因此 Steve 是非常合適的董事會成員。而他在 Bain 任職時的背景是在科技領域。

Michael Latimore, Northland Capital Markets.

Michael Latimore，Northland Capital Markets。

Two questions. I guess on the ARMS acquisition, how might you price that like the warehouse, how much might you charge for that or whatever metrics you use there? And then also in the third quarter, how much revenue came from just development revenue around micro fulfillment?

兩個問題。首先關於 ARMS 收購案，你們會如何為它定價？例如像倉庫那樣，你們可能會收多少費用，或是你們在那裡使用的任何衡量指標？另外，第三季有多少營收是來自微型履約相關的開發營收？

Okay.

好的。

I'll take the ARMS. I mean the ARMS will be a classic value pricing. If we can save somebody $1 million in warehouse maintenance, we're going to charge them a portion of that.

我來回答 ARMS。我的意思是，ARMS 會採用典型的價值定價（value pricing）。如果我們能替某個客戶在倉庫維護上省下 100 萬美元，我們就會向他們收取其中的一部分。

Yes. On the micro fulfillment side, the amount of revenue recorded in the quarter is in the high single-digit range, which is really kind of the average that I would expect going out.

是的。在微型履約方面，本季認列的營收落在高個位數的區間，這基本上也是我預期未來的平均水準。

Joe Giordano, TD Cowen.

Joe Giordano，TD Cowen。

Just quick, Rick, on Exol. I'm just curious what the final design looks like for this customer? What did you -- what did they decide to do in terms of trucks and who's responsible for that and how things are getting to and from the site. I think that was kind of up in the air potentially as a lot of different ways you can go. I'm just curious, we know how the inside of the building works, but how is the whole operation being kind of -- what's the flow sheet?

Rick，快速問一下關於 Exol。我只是好奇，這個客戶的最終設計看起來是什麼樣子？在卡車方面他們最後決定怎麼做、由誰負責，以及貨物如何進出場址。我記得那可能還未定，因為有很多不同的做法。我只是好奇，我們知道建築物內部如何運作，但整個營運是怎麼被——整體流程是什麼？

Yes. So you saw we made an announcement. We partnered with Manhattan on the software piece because so many people that we've talked to are already familiar with the Manhattan integration layer. We're also doing our own integration layer. But the inside of the building is pretty straightforward.

是的。你看到我們發布了一項公告。我們在軟體部分與 Manhattan 合作，因為我們接觸過的很多人都已經熟悉 Manhattan 的整合層（integration layer）。我們也在做我們自己的整合層。但建築物內部相當直接明瞭。

It will do -- we'll move pallets, we'll move cases, we'll do each picking. And we are both hired some of our own transportation people and also engaged with some potential transportation brokerage or transportation companies that actually can bring customers into us.

它將會——我們會搬運棧板（pallets）、搬運箱件（cases）、也會做單件拣選（each picking）。而且我們一方面聘用了自己的一些運輸人員，另一方面也與一些潛在的運輸經紀或運輸公司接洽，這些公司實際上可以把客戶帶到我們這裡。

So we will manage freight when the customers want us to manage the freight. We will have that capability, both in and out of the building, but probably more so out of the building.

因此，當客戶希望我們管理運費（freight）時，我們會管理。我們將具備這種能力，涵蓋進出建築物兩端，但可能更偏向於出庫端。

This concludes the question-and-answer session. I would now like to turn it back to Charlie Anderson for closing remarks.

問答環節到此結束。我現在想把時間交回給 Charlie Anderson 作結語。

Yes. Thanks, everybody, as always, for joining our call tonight. We really appreciate your interest in Symbotic and want everybody to have a good evening. Thanks so much.

是的。一如既往，感謝各位今晚參加我們的電話會議。我們非常感謝大家對 Symbotic 的關注，也祝大家今晚愉快。非常感謝。

Thank you for your participation in today's conference. This does conclude the program. You may now disconnect.

感謝您參與今天的會議。本節目到此結束。您現在可以掛線。