Xometry, Inc. (XMTR) Q2 2026 Earnings Call Transcript

Good day, and thank you for standing by. Welcome to Xometry's Q2 2026 Earnings Call. Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Shawn Milne, Vice President of Investor Relations. Please go ahead.

Good morning, and thank you for joining us on Xometry's Q2 2026 earnings call. Joining me are Sanjeev Sahni, our Chief Executive Officer; and James Miln, our Chief Financial Officer. During today's call, we will review our financial results for the second quarter of 2026 and discuss our guidance for the third quarter and full year 2026. During today's call, we will make forward-looking statements, including statements related to the expected performance of our business, future financial results, strategy, long-term growth and overall future prospects. Such statements may be identified by terms such as believe, expect, intend and may. These statements are subject to risks and uncertainties, which could cause them to differ materially from actual results. Information concerning those risks is available in our earnings press release distributed before the market opened today and in our filings with the U.S. Securities and Exchange Commission, including our Form 10-Q for the quarter ended June 30, 2026. We caution you to not place undue reliance on forward-looking statements and undertake no duty or obligation to update any forward-looking statements as a result of new information, future events or changes in our expectations. We'd also like to point out that on today's call, we will report GAAP and non-GAAP results. We use these non-GAAP financial measures internally for financial and operating decision-making purposes and as a means to evaluate period-to-period comparisons. Non-GAAP financial measures are presented in addition to and not as a substitute or superior to measures of financial performance prepared in accordance with U.S. GAAP. To see the reconciliation of these non-GAAP measures, please refer to our earnings press release distributed today and our investor presentation, both of which are available on the Investors section of our website at investors.xometry.com. A replay of today's call will also be posted on our website. With that, I'd like to turn the call over to Sanjeev.

Thanks, Shawn. Good morning, and thank you for joining our Q2 earnings call. Our accelerating growth and record Q2 results demonstrate the growing strength of our AI-native marketplace. Q2 was a record quarter for Xometry across many fronts, including revenue, gross profit and adjusted EBITDA. Q2 revenue increased 41% year-over-year to a record $229 million. Marketplace revenue growth accelerated to 45% year-over-year, driven by broad-based strength across many verticals, improving conversion rates, increasing adoption by new buyers and growing wallet share of existing buyers. We continued to see strong enterprise growth. Q2 marks our fourth consecutive quarter of accelerating revenue growth. Our AI models are effectively optimizing conversion, buyer growth and market share gains. On top of the revenue growth, Q2 adjusted EBITDA improved $10.2 million year-over-year to $14.1 million. We delivered record results in the first half of 2026, and we are again raising our outlook for the second half of this year. The strength we saw in Q2 has continued in Q3, and we are off to a strong start in the quarter. We expect strong compounding growth and operating leverage to continue through the second half of 2026 and for years to come. I reiterate this point every quarter because it remains fundamental to our long-term growth thesis. We have penetrated less than 1% of our massive $275 billion custom manufacturing TAM. This is a market that remains largely offline today, still run on email quotes and multi-day turnarounds. The majority of it is still sourced within a 30- to 50-mile radius of a factory floor. To me, that gap is the opportunity. We are in the opening innings of a digital transformation and our ability to capture this massive untapped opportunity gives us strong confidence in our long-term growth trajectory. Beyond the shift from offline to online, we see a long-term tailwind in physical AI. The rapid growth of robotics, autonomous systems and defense platforms creates a new category of customers needing fast on-demand parts. Our marketplace is well positioned to capture this demand. The four consecutive quarters of accelerating growth are direct evidence that the product-led strategy put in place last year is working. We are defining the e-commerce playbook in custom manufacturing and raising the experience bar for buyers and suppliers everywhere. We will structure our calls going forward to provide updates on our key focus areas. First, I will talk about establishing Xometry as the infrastructure for custom manufacturing. Our confidence in driving the next S-curve of our growth stems from our ability to apply our core AI models to a decade-plus of proprietary data. The intelligence we derive from our data gives our customers the confidence in manufacturability of their part, the instant pricing options we present and Xometry's ability to source the part optimally. The depth of our data and the intelligence is unrivaled in custom manufacturing. That same intelligence is now extending beyond our marketplace, embedding natively into partnership environments like Siemens Design Center, proof that intelligence itself — not just the marketplace around it — is the infrastructure other platforms want to build on. More broadly, we see this as Xometry becoming the connected layer between design intent and actual physical production. Siemens is the first but not the only place we expect to integrate our intelligence. Deep technical integration is a critical pillar of our product-led growth strategy. Establishing seamless AI-native digital threads that remove friction from the custom manufacturing workflows will be an ongoing focus. We are making excellent progress on the Siemens collaboration with both teams moving at the pace we expected. We are building real integration, bringing Xometry's manufacturability and pricing insight directly into Siemens Design Center. In addition to the work within Design Center, there's also active collaboration between Siemens Supplyframe and Thomas, extending the value both teams can bring to their customers. We will share more as we hit milestones, and we expect the Siemens partnership to positively impact our 2027 operating results. Our second focus area is improving our AI-native marketplace experiences. Our customer and supplier online journeys are rapidly defining the e-commerce playbook in custom manufacturing. As I've shared before, one of our core beliefs is that the B2B buying experience should be every bit as good as what people experience in their personal lives. The days of clunky B2B software, multistep checkout processes and waiting days for an email quote are simply over. What we are seeing is a generational shift in who is making manufacturing purchasing decisions. The engineers, procurement buyers and supply chain lead roles are now full of dynamic digitally native individuals. They expect the same frictionless experience at work that they have in their personal lives. When they find that Xometry can deliver, they become Xometry champions inside their organizations. In Q2, we made significant upgrades to our proprietary AI models that power our marketplace. Specifically, we enhanced model capabilities in costing, sourcing and process recommendations. These comprehensive upgrades deploy a new generation of high-capacity interconnected models that span the entire manufacturing journey. The upgrades enable additional data insights across five elements of our data architecture: geometry, manufacturability, certifications, supplier capability and production outcomes. The smarter the platform gets, the faster we can turn complex engineering inputs into manufacturing decisions. As part of our AI model upgrades, we launched a new generation cost prediction model that considers a greater breadth and depth of inputs, including geometry, material, finish and whether the part is a stand-alone part or one of several in a job. This granular input leads to more accurate pricing. The model understands the specific parameters each part requires, delivering an approximately 15% improvement in CNC cost prediction accuracy. In addition, in Q2, we launched a new adaptive sourcing model that uses our proprietary data to price jobs dynamically for our suppliers. The new model incorporates an upgraded partner suitability score for each job measured against partners' specific machine capabilities, quality and on-time shipping history. By matching jobs to the right machines, we ensure our suppliers receive a curated flow of better-fitting work, which strengthens our network overall. Lastly, we launched an upgraded context-aware AI process recommender. It reads a part's industry application to anticipate needs like tighter tolerances for aerospace components. By factoring in material and geometry, it closes a real gap for first-time customers. Buyers are now accepting the recommendation more than 85% of the time with the biggest gains coming from exactly those first-time customers. I will now share advances we made in our third focus area, expanding our buyer and supplier networks. Our focus is on becoming the most expansive custom manufacturing e-commerce platform for buyers and suppliers. We delivered strong active buyer growth during the quarter. Active buyers increased 20% year-over-year to over 89,000. Our marketing teams have strengthened our martech and personalization capabilities, helping drive almost 4,000 net adds. Their efforts, combined with our AI model optimizations are driving conversion rate improvements and continued robust buyer growth. At the same time, these efforts are also steadily decreasing our marketing costs as a percentage of revenue, which is improving marketplace unit economics. There are still millions of potential buyers to convert. We expect continual improvements in marketing technology and AI models will enable us to steadily increase our conversion rates and net adds. Our partnership with Siemens and our other initiatives to become the infrastructure layer in custom manufacturing will further accelerate these trends. In Q2, we further strengthened our U.S. injection molding offering. We added new auto-quotable materials to expand choices for our buyers. We also made free on-demand design-for-manufacturability consultations schedulable directly on the platform, connecting customers with our injection molding experts from the very first quote. Additionally, we launched self-service one-click reordering for injection molded parts. Customers can generate new quotes that automatically carry forward configurations, specifications and files from the original order, routing them directly back to the original supplier holding the tool. Small thing on the surface, but it is exactly the kind of friction we are trying to strip out everywhere because friction is the only thing standing between a first order and a habit. We are ever more focused on expanding our global supplier network and improving our supplier experience. Our global network of over 5,000 active suppliers across 50 countries remains a significant strategic advantage, giving buyers unmatched speed, capacity and resilience. In Q2, we added capacity in newer international markets, providing our buyers more choice and flexibility, including India and Vietnam. We continue to expand domestically as we recently referenced in our Texas market press release. Additionally, we are investing in new categories, adding suppliers with specific advanced capabilities and certifications. Furthermore, we are supporting our high-performing partners in obtaining additional certifications to meet growing production demands. Alongside expanding supplier breadth, we continue to expand the depth of our relationship with each supplier, driving up average volume per supplier and becoming more integral to their long-term success. Our fourth focus area is deepening enterprise engagement. We continue to deliver robust enterprise growth. Our Q2 revenue from our larger customers increased by more than 40%. We delivered a record net addition of 175 accounts with greater than $50,000 spend. As Xometry becomes more embedded in our customers' workflows, we are seeing continued wallet share gains and more predictable spend. We ended 2025 with four accounts spending at least $10 million annually. We expect more accounts to cross that threshold in 2026. This is driven by multiyear production programs across key end markets. For example, this deeper integration is exactly what we saw with a major enterprise robotics leader who faced a critical build deadline. They turned to us with a massive challenge, nearly 200 complex parts, including metal structures and tight tolerance CNC components, all needed within a short delivery window. We brought the strength of our marketplace to bear and leveraged 40 network suppliers to deliver the full build on time and on spec. This success proved that our platform is the go-to solution for high-stakes, large-scale hardware programs directly resulting in significant follow-on production work. The final focus area I will share is leveraging services opportunities. Our services offerings under Thomas helped us engage with suppliers who aren't yet in the marketplace ecosystem. Thomas has built the largest digital sourcing network in North America for industrial manufacturing, providing buyers access to over 500,000 suppliers. We are focused on improving our monetization on the Commerce platform and leveraging the VOS network to supplement Xometry's supplier capacity. We made strong progress on our Thomas platform in Q2. We have completed the transition to our new ad platform and redesigned search experience. These are already yielding improvement in monetization. We are building on this momentum by launching new AI-powered tools and processes for our Thomas Marketing Services. With the traction we are seeing across advertising and marketing services, we are ever more confident about inflecting the revenue curve and bringing our services offerings back to revenue growth year-over-year in the second half of 2026. In conclusion, I'm very excited about the road ahead. And before I hand it over, I want to thank our entire team. The stellar results we are reporting today are the outcomes of a lot of hard work across product, technology, sales, marketing and operations. I'm proud of the pace at which our teams continue to execute our product-led growth strategy. I will now turn the call over to James for a more detailed review of Q2 and our business outlook.

Thanks, Sanjeev, and good morning, everyone. Our record results for the second quarter underscore the continued scaling and increasing efficiency of our marketplace, driving both accelerated growth and expanding profitability. Revenue growth accelerated for the fourth quarter in a row. This accelerating top line was paired with yet another quarter of improved adjusted EBITDA profit margins. These achievements demonstrate that our marketplace is becoming the essential infrastructure for a predominantly offline and fragmented industry. Driven by our record Q2 results and strong start to Q3, we are raising our revenue and adjusted EBITDA guidance for the full year. Q2 revenue grew 41% year-over-year to $229 million, a 500 basis point sequential acceleration from Q1. Q2 marketplace revenue was $215 million and services revenue was $13.9 million. Q2 marketplace revenue increased 45% year-over-year, a 500 basis points acceleration from Q1, driven by broad-based strength and adoption across the marketplace as we continue to capture significant market share. Q2 active buyers increased 20% year-over-year to 89,557 with a net addition of 3,976 active buyers, the highest number of net adds in 10 quarters. Strong Q2 net additions were driven by our product-led growth strategy, optimization of AI models for market share and efficient marketing, including personalized pricing initiatives. Q2 marketplace revenue per active buyer increased a robust 21% year-over-year, primarily due to increasing wallet share. We view accounts with at least $50,000 spend as the top of the enterprise funnel. In Q2, the number of accounts with last 12-month spend of at least $50,000 on our platform increased 23% year-over-year to 2,039 with a record 175 quarterly net adds. Enterprise investments continue to show strong returns. Our enterprise strategy focuses on our largest accounts, each of which we believe have $10 million plus in potential annual account revenue. Services revenue was up slightly quarter-over-quarter as we stabilize the core advertising business. We expect services revenue will return to year-over-year growth beginning in Q3. We are focused on improving engagement and monetization on the platform, which remains a leader in industrial sourcing, supplier selection and digital marketing solutions. Q2 gross profit was $87.5 million, an increase of 34% year-over-year. Q2 marketplace gross profit dollars increased a robust 42% year-over-year. Q2 gross margin for marketplace was 34.7%, flat quarter-over-quarter. We are focused on driving marketplace gross profit dollar growth through the combination of top line growth and gross margin expansion. Our AI models are effectively optimizing conversion, buyer growth and market share gains. In Q2, this led to accelerating active buyer net adds, accelerating revenue growth and increased leverage on marketplace advertising, with spend down 220 basis points year-over-year. In fact, over the last year, our marketplace revenue growth has accelerated nearly 20 points to approximately 45%, driven by our optimization efforts. We expect marketplace gross margins to be higher in the second half of the year than the first half of the year, and we expect this trend to continue, expanding further into our target range of 35% to 40%. Our commitment to strong discipline and rigor in capital and resource allocation across all teams while continuing to invest in growth initiatives is reflected in our Q2 operating costs. Total non-GAAP operating expenses for Q2 were $73.2 million, a 19% increase year-over-year, which is less than half the growth rate of our revenue. Q2 operating costs as a percentage of revenue improved 590 basis points year-over-year. In Q2, sales and marketing decreased 320 basis points year-over-year to 13.2% of revenue. This reflects improving enterprise sales execution, efficiency of AI models as well as the strength in martech and personalization capabilities that Sanjeev mentioned earlier. Marketplace advertising spend was a record low 3.4% of marketplace revenue, reflecting our optimization efforts, driving improved marketplace unit economics. In Q2, operations and support decreased 90 basis points year-over-year to 8% of revenue. We are focused on driving increasing automation with AI across operations and support. Q2 adjusted EBITDA was $14.1 million compared with $3.9 million in Q2 2025. Q2 adjusted EBITDA improved $10.2 million year-over-year, driven by strong revenue growth, gross profit and operating efficiencies. Alongside accelerating revenue growth, we delivered expanded adjusted EBITDA margin of 6.2%, a 380 basis point increase compared with 2.4% in Q2 2025. Q2 U.S. segment adjusted EBITDA was $17.4 million, a $10.6 million improvement year-over-year. Q2 U.S. segment adjusted EBITDA margin was 9%, a significant increase compared to 5.1% a year ago, driven by strong gross profit dollar growth and operating expense leverage. Our International segment adjusted EBITDA loss was $3.3 million in Q2 2026 or 9.6% of revenue, a 140 basis point improvement from a loss of 11% in Q2 2025. At the end of the second quarter, cash and cash equivalents and marketable securities were $517 million, including $248 million raised in our follow-on offering completed on June 2 and $50 million from the Siemens investment announced on May 7. Our successful follow-on offering will support key organic growth initiatives and our selective tuck-in M&A strategy. We generated $17 million in operating cash flow in the first half of 2026, driven by strong operating leverage and working capital efficiency. In the second quarter, we invested approximately $13 million in cash CapEx, almost entirely software related, reflecting our technology investments in the platform and accelerating product rollouts. We are focused on improving cash flow conversion given our asset-light model and limited capital spending. Our disciplined execution has led to strong revenue and gross profit growth in our AI-native marketplace, coupled with significant operating leverage and increased operating cash flow generation. We are focused on strategically balancing future investment with a relentless pursuit of operating leverage given the vast market opportunity and our low penetration rates. As we rapidly approach a $1 billion run rate this year, we have a clear trajectory for improved adjusted EBITDA margins while sustaining our investment in growth. Now moving on to guidance. We are raising our outlook for 2026, which includes higher revenue and profitability for the full year. For the second half of 2026, we are now expecting revenue growth of approximately 30% year-over-year and incremental adjusted EBITDA of 20%. For the third quarter, we expect revenue in the range of $234 million to $236 million or 30% to 31% growth year-over-year. We expect Q3 marketplace growth to be approximately 33% year-over-year, driven by ongoing momentum from our growth initiatives. We expect Q3 services revenue to be up modestly year-over-year. In Q3, we expect adjusted EBITDA of $16 million to $17 million compared to $6.1 million in Q3 2025. For the full year 2026, we are raising our revenue growth outlook to 33% to 34% from 27% to 28%, driven by approximately 37% marketplace growth. This equates to an approximately 8-point acceleration over full year 2025 growth. For the full year 2026, we expect adjusted EBITDA in the range of $60 million to $62 million. Before we open up to questions, I want to recognize our team. The results we've discussed today reflect their execution, and I'm equally excited for what those results make possible going forward. We have real momentum, a large market in front of us and a team that has demonstrated it can deliver. That combination gives us genuine confidence in what's ahead. With that, operator, can you please open up the call for questions?

Your first question comes from the line of Andrew Boone with Citizens.

I wanted to ask about demand. It sounds like AI has been a key unlock on the platform. Can you just expand upon that? How do we think about AI in terms of unlocking incremental demand? What specifically is it helping with? And then just as a bigger picture, how do we think about AI and model improvements and that rolling through the model on a go-forward basis? And then secondly, on Siemens, I'd love to hear more broadly just how partner conversations are going outside of Siemens. It feels like it's a step function change for functionality for CAD platforms, broadly speaking. What has that done as you guys now look more broadly across the industry? How are those conversations going?

Thanks for the question, Andrew. Let me take the questions one after the other. First, as you can see, we are clearly super excited about the phenomenal outcomes our teams delivered: marketplace revenue accelerating to 45% growth and 43% in the first half year-over-year. All of that is broad-based. We see growth happening across categories, across new and existing buyers, conversion rate improvements and just a broad-based increase in adoption of the platform as the go-to place for custom manufacturing. To me, this is all rooted in the product-led growth strategy we've been working on and we've been talking about over the last year. So that product-led strategy is truly bringing together the AI models that actually are no longer independently doing pricing, sourcing and manufacturability. They're all interconnected models that are truly focused on driving the real big outcomes, which is the marketplace revenue, the profitability, the gross profit and then the EBITDA. And as they do that, we want to make sure that we are continuing to get new buyers excited about joining the platform. You saw active buyers were up 20%, net adds in the last quarter were the highest net adds in the last 10 quarters. All of that to say that the product-led strategy that we put in place is the real core and the AI models are not just independent models now. They are an ecosystem that's powering the growth that we have. To your second question on Siemens, we are making excellent progress on that collaboration. Like I said in my prepared remarks, the two teams are actually working together to a timeline on integration; we are really fundamentally changing the way price for custom parts shows up inside a CAD tool. That's really net new. The most exciting part is that we are getting continuous approaches by customers who want to be part of the initial group experiencing this new way of working, and in some cases even wanting to provide feedback so they can actually help shape what gets rolled out. And I think this is really exciting because we see the same asks come to us not just in the U.S., but also in EMEA where we have a strong platform. So lots more exciting things to come. We'll share more milestones as we go along, but we expect clearly the Siemens partnership to positively impact our FY '27 results.

Your next question comes from the line of Brian Drab with William Blair.

I'm just going to ask one question for now. Sanjeev, can you just elaborate on what changes you're making in the pricing algorithm? And you've been leveraging the millions and millions of parts and the data associated with them to improve that algorithm for years. Is this a step function change in the algorithm? What can you just elaborate on what exactly is changing and what this could do for gross margin in the marketplace longer term?

Thanks, Brian. I think I'll start by saying first that we expect the marketplace gross margin to expand further into the target range of 35% to 40%, driven by all of these changes that we are making. What you're seeing is actually the proprietary data that we've had for a decade-plus of parts that we produce. We are more extensively mining that data to find behavior patterns, find conversion rate opportunities and optimize those models such that every part that gets produced helps make the next quote stronger, which means that the feedback mechanisms on the actual production floor are making their way back to when the next part is priced on the site. So we already know if we've got a part that actually was more complicated than anticipated, what should that do to the pricing of the next part, what kind of questions we should be asking the customer. So it's really becoming an intelligence engine. And I think the exciting part there then is how we are taking that intelligence engine and embedding it in places like Siemens in the future to truly drive the future of manufacturing.

Your next question comes from the line of Ron Josey with Citi.

Sanjeev, I wanted to ask a little bit more just about the improving conversion rates you mentioned on active buyers while also delivering that leverage in sales and marketing. Just talk to us about the drivers of the conversion rate improvements, especially as we think about the millions of clients to convert. I think you mentioned that. So just with conversion rates improving, sales and marketing showing leverage and millions to go, I just wanted to hear more about the approach here for sales and marketing. And then on the self-service for injection molding, you just talked about how that also has increased conversion rates. I wanted to hear a little bit more about the potential for other products to go self-service as you continue to integrate those capabilities.

Thanks, Ron. I'll start off and then have James chime in as well. What really is happening is as we think about the application of technology, I spoke a little bit about what we are doing in the marketing technology front or martech. We're actually using the signals we are getting from the customers' shopping behavior to identify what kinds of sales and marketing interventions should follow. In the same way, we are connecting the models. We are connecting the signals that the customers give us by amount of time that they spend on the platform, which page they leave, which page they continue to browse, to then use that to differentially plan our sales follow-ons or sales conversations. And I think that is what is making the change in us being able to really leverage the data we've had. But now with the scale and the strength we have, that's truly becoming a difference maker in being able to know when a customer is likely to convert and actually following a much stronger follow-on rhythm versus a customer who's a casual browser and then may not convert, and how do we resource that follow-on in a much more efficient way.

Yes. And thank you, Ron, for noticing on the leverage. I think we're really very happy here with what this is indicating in terms of improving unit economics, which bodes very well for continuing to grow into the large TAM ahead of us. So 4,000 net adds in the quarter — that was a record in the last 10 quarters — doing that while our AI models are optimizing our gross margin, gross profit, conversion and buyer growth. We also saw accounts above $50,000 at a record as well, net adds in the quarter of 175. So just as Sanjeev said, the work that we're doing on these AI models gives us a lot of confidence that we'll continue to see gross margin expand further over time into our target range of 35% to 40%, while also being able to support this strong top-line growth into our expanding opportunity.

And on self-service?

Ron, what was the question on self-service?

I can go back and ask the operator to put Ron back in the queue. Please go ahead, Ron.

Yes, Ron, could you just repeat your second question — just clarify the second question you had?

Yes. Just if you can hear me, hopefully you can. Just more insights on injection molding. I think you launched self-service there more recently. I wanted to hear about how that has changed the conversion rate to project and maybe you can collapse the time frames or compress the time frames and also how you think about self-service going forward for other projects?

Thanks for the question, Ron. The injection molding work that we are doing is truly transformational in the sense that we are not only taking a completely offline process that is a two-part process between the mold and the actual production, we are making a ton of progress in making sure the connective tissue between the mold and the actual final production is established from the beginning. The increasing adoption in injection molding is making a ton of great traction for us. The benefits of linking all of that upfront with the expert advice that I mentioned, and the reordering made super easy as one click, I think that's driving the awareness, and we are seeing that in the adoption numbers already. So lots more to come in that space. We actually are very excited about the tech improvements we've made there and we see similar self-service opportunities in other categories over time as we standardize more flows and improve auto-quoting capabilities.

Your next question comes from the line of Troy Jensen with Cantor Fitzgerald.

Gentlemen, congrats on great results here.

Thank you.

Thank you, Troy.

Sanjeev, I'd love to start with you. Could you just give us your thoughts on the M&A focus comment that you had? Curious if it's just kind of software tuck-ins or if you ever think you would need to own some more equipment?

Thanks, Troy. I can start. I think there's almost two elements to it. One is the tuck-in M&A for us are focused specifically on continuing to drive our moat. We've become more and more focused on driving a product-led growth strategy, which means ensuring that we have access to some of the best technology capabilities in this space and continue to drive growth that actually comes from that. So to that, we don't necessarily have a next update to give you. But as you can imagine, both thanks to the Siemens conversation and the point on M&A that we made during the equity raise, phones have been continuously ringing. There are conversations ongoing that we'd share as soon as we can. The focus is on technology, capabilities, talent and product add-ons rather than large capital-intensive equipment acquisitions.

Yes. And Troy, I think you know us, we're very disciplined in terms of our capital allocation. We feel that we're in a very strong position now in terms of our performance and in terms of our balance sheet, but we will remain disciplined. Building on Sanjeev's point, the marketplace model is clearly winning here and ways that we can continue to blend that and improve our offerings, improve our geographies, augment technology and talent that can help us on our roadmap are the types of tuck-ins that make sense. That complements also the approach that we're taking in terms of partnerships as well.

All right, James, maybe just a follow-up for you. I'm assuming a lot of the profitability leverage is going to come from international as that shifts from being a drag to accretive. Can you talk us through investments you're doing there? Are you trying to accelerate that more than the U.S., timing to breakeven or any help would be great.

I'll kick off and Sanjeev can add. First, I'd point to the fantastic progress we've been demonstrating in the U.S. We've made a lot of progress on enterprise and on our product roadmap and playbook. In the last quarter, on a segment basis, our U.S. adjusted EBITDA margin was 9%, up from 5% a year ago, showing strong progress toward our longer-term target. On international, we're pleased with how the marketplace offering works globally. The unit economics are similar across the markets we're in, and the strength we get from being able to source globally continues to improve our offering. We are expanding in India, Turkey and Southeast Asia and the performance is encouraging. The plan is to take the playbook we've perfected in the U.S. and apply it to international markets, building on go-to-market and product approaches to drive similar trajectories. We remain disciplined in investment but will scale where we see returns and demand.

I totally echo James' point. You've seen us execute a product-led playbook in the U.S. and you're seeing strong revenue and adjusted EBITDA improvement. We're going to take the same playbook and execute in Europe and other international markets. You'll see that come through the numbers in coming quarters and years. We've perfected a playbook we want to now scale internationally.

Your next question comes from the line of Eric Sheridan with Goldman Sachs.

I want to come back on the first one to something you said in the beginning of the prepared remarks that I thought was really interesting. You talked about the rapid growth in new market segments like robotics and autonomous systems and defense platforms opening up pockets of market opportunity. Can you talk to us a little bit about how new verticals and innovation in the broader manufacturing landscape is opening up a wider array of potential customer growth for the medium to long term? Would love to go as far as you can in terms of digging deeper on that topic. And then secondarily, maybe for James, following on the equity offering you did recently, how should investors think about capital allocation and broadly balance sheet flexibility, strategic M&A and making sure you're making all the critical growth investments in the business over the medium to long term in terms of the balance you want to strike there?

Thanks for the question, Eric. Let me answer the first in two parts. One, in terms of momentum we are seeing on the platform, we've seen broad-based expansion in new customers and categories: CNC, additive, sheet and tube, injection molding — all of those are seeing growth. To your point, areas like robotics, autonomous vehicles and defense platforms are naturally finding their way to the online marketplace more easily. Part of the growth in active buyers reflects new buyer growth and traction from digitally native individuals in key roles who prefer the frictionless experience Xometry offers. In the physical AI space, robotics and autonomous vehicles are contributing significant traction from new buyers. The other side is becoming the infrastructure for custom manufacturing: the trend toward zero-touch manufacturing and digital threads from design to production. The intelligence we have can be embedded in other platforms, starting with Siemens, to help manufacturing move from multi-touch environments to zero touch. We see ourselves as a key provider of design and production intelligence across the manufacturing spectrum, which opens many longer-term opportunities.

And Eric, on capital allocation, over the last couple of years we've taken deliberate steps to ensure a capital structure and balance sheet that the opportunity ahead merits. We refinanced our convertible to give flexibility, completed a strong equity raise and have the Siemens investment. We ended the quarter with $517 million in cash. We moved from being cash burning a couple years ago to roughly neutral in the first half of this year, which puts us in a strong position in various operating environments. Our first focus is building shareholder value through sustainable profitable growth into the $275 billion TAM where we are less than 1% penetrated. Maintaining a strong balance sheet gives us flexibility for disciplined tuck-in M&A while we continue to invest in growth initiatives that deliver returns.

To add on M&A, we think about opportunities not just as adjacency but as ways to expand our e-commerce platform, deepen categories where we're strong, extend geographies where we already have presence, and add technical capabilities. It's focused on expanding the marketplace and accelerating traction.

Your next question comes from the line of Matt Swanson with RBC Capital Markets.

You've spent the last couple of years showing really strong growth despite having headwinds from PMI and the overall macro environment. As we're starting to see some of those indicators improve, are there any parts of the business that you're noticing a notable pickup that maybe you didn't know were seeing headwinds before, whether it be enterprise, discrete manufacturing or certain international geographies?

Matt, it's Shawn. I'll take that. Certainly, we talked a lot about the last couple of years driving strong enterprise growth, which drove a big lift in our revenue growth rates, and Sanjeev's product-led growth strategy has added another layer to the growth. What we discussed on the call is that added momentum now around physical AI, which will drive long-term tailwinds. A lot of what you're seeing has been driven by the execution of the company. That said, we're now seeing manufacturing PMI turn positive for several months, and we can lean into that. You saw our active buyers this quarter were near a 10-quarter high. We're prepared to continue to gain outsized market share capture.

Yes, building on that, Xometry is built for this pivotal time in manufacturing. The choice for resilient and flexible supply chains, marketplace offerings and the value of an e-commerce experience continues to rise. These cross currents are areas Xometry was built to solve for, and we see that reflected in our results and opportunity.

That's super helpful. And then one follow-up on the Siemens partnership. We talked last quarter about how workflow integrations could supercharge enterprise adoption. Is there anything from these early customer conversations that is starting to help the way you think about your own R&D process and features or use cases to develop Xometry's enterprise capabilities inside and outside of Siemens in the future?

That's an excellent question. One of the most interesting outcomes of the Siemens collaboration has been an elevated understanding of what we do in enterprise conversations. Where we used to explain what a custom manufacturing marketplace can do, now we discuss specific opportunities: prototyping, production, special projects, and where customers' data is strongest and can drive insight. The nature of enterprise conversations is changing. Many customers already using design tools ask, 'When Xometry is in the design tool, could the flow do X, Y, Z?' Those inputs are extremely helpful as we work closely with Siemens' teams. Customer feedback helps define the outcomes and shape our product roadmap. So yes, early customer engagement is directly informing our R&D priorities and feature set.

Your next question comes from the line of Greg Palm with Craig-Hallum. The caller on for Greg Palm is Jackson Schroeder.

This is Jackson Schroeder on for Greg Palm. Just a follow-up to that PMI question: Is there any effect that's having on buyers and suppliers on the platform, both buyers expediting shipping more, seeing elevated traction, and suppliers being more or less price sensitive, how much extra capacity they have? And can you separate what you're seeing internationally versus the U.S.? Does international PMI ticking up possibly be a bigger benefit and help you grow international faster than domestic again to expand that as a percentage of sales?

Thanks for the question. We don't disclose specific customer-tier behaviors, but when you're growing at 45% year-over-year, you see broad-based strength. The marketplace play lets different customers choose what they need: some choose faster lead times, others prioritize price. We launched the ability to ship orders the same day in a couple of categories, which is exciting for customers needing speed. At the same time, other customers remain price sensitive, and our expanded sourcing in markets like India and Vietnam helps capture that demand. So growth isn't concentrated in one area; it's capturing both ends of the market. As growth picks up in EMEA and manufacturing focus strengthens in Europe, we see similar trends there and are ready with our playbook to capture that opportunity.

Your last question comes from the line of Ygal Arounian with Wedbush.

A question on the gross margins that were slightly lower year-over-year here in Q2 and then you talked about expectations to improve over the second half. I know you have the AI models driving better pricing over time. Is that what's driving the improvement over the rest of the year? Are there other factors? I just want to understand the components of gross margin through this year. And then on the guidance for the rest of the year and as we look into next year, is there any Siemens component in the guidance for this year? How are we thinking about the contributions into next year? Any updates on that at all?

Thanks for the question. The continued gross margin expansion into our target range of 35% to 40% is indeed driven by the AI models and the continuous improvement they deliver across multiple metrics. You saw marketplace gross margin over the first half is up about 100 basis points. The models drive better pricing, improved conversion and better supplier matching, which together boost gross profit dollars and margins. Q3 has started really strong and the indicators are trending well, so we feel very positive about the commitment to improve margins over the rest of the year. In terms of Siemens, there are some product-related costs embedded in our second-half guidance as the teams build for operating impact in 2027. We expect the Siemens partnership to have more visible operating impact in 2027.

Very pleased with the start to Q3. Marketplace gross margin over the first half is up about 100 basis points, and with more data, more suppliers and improved AI models working across the whole marketplace, we have good confidence in continuing to progress toward the 35% to 40% range over time. On Siemens, as Sanjeev noted, there will be some product-related investments and costs in the second half, which are embedded in our guidance. The expectation is that the integration and product work will contribute to operating results in 2027.

This concludes the question-and-answer session. Thank you for your participation in today's conference. This does conclude the program, and you may now disconnect.