Globant S.A. (GLOB) Earnings Call Transcript
August 13, 2026
Earnings Call Speaker Segments
Thanks for joining Globant's Q2 2026 earnings call. Arturo Langa here — I handle Investor Relations for the company. Everyone dialing in today is in listen-only mode. Once the presentation wraps up, we'll open things up so you can ask questions. Just so you're aware, this session is being recorded and is streaming live on YouTube. You ought to have gotten the earnings release by now. If it hasn't reached you, you can grab one from our site, investors.globant.com. To kick things off, we'll hear from our Chief Executive Officer, Martín Migoya, our Chief Technology Officer, Diego Tartara, and our Chief Financial Officer, Juan Urthiague, and after that a Q&A in which our Chief Revenue Officer, Fernando Matzkin, will join them. Before getting started, I want to flag that certain remarks made during today's call might count as forward-looking statements. That covers our business and financial outlook as well as responses to a number of your questions. Statements of that kind carry risks and uncertainties, which are laid out in the company's earnings release and other SEC filings. Keep in mind that our financial statements follow IFRS accounting rules. On today's call, we'll also report non-IFRS, or adjusted, measures — that's how we track performance internally, and it's the simplest way to stack Globant up against others in the industry. A reconciliation between IFRS and non-IFRS measures appears at the end of the press release we put out on our investor relations website announcing this quarter's results. I'll hand the call over to Martín Migoya now.
Hello, everyone, and thanks for being here this afternoon. I'd like to talk about a transformation we're leading — a new way of creating value for clients, of executing our projects, and of setting prices — and we're already seeing that approach build momentum. For more than 20 years, we've driven digital reinvention for top global companies, creating the software, products, and platforms their operations depend on, carried out by committed high-achieving teams and billed either as fixed-scope contracts or on a time-and-materials basis; that business is still Globant's foundation. Twelve months back, I walked you through AI Pods — an additional revenue channel native to AI, which rests on that same base yet gets charged according to what we actually produce and the worth we provide, or by usage, instead of the hours we invoice. Since AI Pods generate greater amounts of work at better margins for roughly the same price, the figure at the top of our income statement may fail to reflect what's really happening beneath it. As that expands, it will make sense to look at yearly recurring revenue for each employee, the margins on AI Pods, and how deeply clients are penetrated, together with the overall revenue number. Prior to continuing, allow me to be exact regarding two labels you'll hear throughout today's call. Glob.AI is the platform we made available to the market just last week. On the platform sit AI Pods — service units operated through AI agent workflows and overseen by our people. Glob.AI ARR is what I'll call the revenue they generate. Keep all three in mind: the platform, the pods, and that figure. Revenue for this quarter rose about 60%, reaching $52.8 million. By the end of the year, we project Glob.AI's ARR will exceed $110 million. Here's how I'll go through it, in that sequence: the model, the metric that captures it, the sources of growth, how to interpret our reported revenue while both models operate simultaneously, and where results are already visible. Right now, the technology services sector overall is expanding at rates that are basically flat. But flat is just an average, and averages conceal what's actually happening. Within that flat sector, we've identified a growth runway: AI native services. That area is expanding because it captures what enterprises are looking for from AI more effectively than the conventional hours-based model. Clients notice the distinction, and an increasing number are shifting their budgets to match, while more of them are deciding to partner with us using this new approach. As enterprises strip away one layer of complexity after another—infrastructure, platform, software—some are starting to strip away business services too. We refer to this as service as software. In the same way the cloud reshaped software infrastructure and brought in revenue that was both predictable and recurring, Glob.AI accomplishes that same thing for professional services. You simply switch on the outcome, and then you pay according to your consumption, with Globant's experts included right in the offering; this unlocks budgets that were previously out of our reach. Each year, an estimated more than $6 trillion gets spent across the global professional services industry — roughly four times the size of the IT services market that Globant has traditionally been measured against. That explains our deliberate decision: we respect the market we currently serve, while simultaneously expanding our offerings into a larger total addressable market. Our clients are being steered by us toward this new model. 45 of our clients have now embraced our AI native delivery system of AI Pods across numerous projects of theirs. At this point, nearly all of what we deliver includes AI. That is not what we tally as Glob.AI ARR — that figure only reflects revenue delivered and billed differently based on the output, value, or consumption clients receive, rather than the hours behind it. This is a rigorous metric, and intentionally so. Growth in this number does not signal AI tacked onto pre-existing work. What it signals is the business model itself transforming. As of June, Glob.AI ARR hit $52.8 million, climbing from $32.8 million in March — approximately 60% growth within one quarter. When you put a standard engineer alongside AI, the productivity gain we've measured is 30% greater. Our pipeline now stands at $436.8 million, up from $352 million during Q1. Within our 20 largest accounts, 45% have adopted it. Under this approach, gross margins land almost 10 percentage points above what traditional delivery gives us. As things stand, we expect to close out 2026 with at least $110 million in Glob.AI ARR. That figure is the yardstick we hold ourselves to. How the market should value Globant is increasingly tied to AI-native revenue, which we monitor every quarter. These changes have left their mark on Globant as a whole as well. On a run rate basis, revenue per head at Globant came to $95,800, a 9.7% increase year-over-year; the same talent is producing more value, and we are capturing it. Three of the largest demand waves in our industry are captured by Glob.AI ARR. Those are core modernization, experience debt, and agentic process transformation. On earlier earnings calls, we walked you through them. What has shifted is that all three increasingly convert via AI Pods. Allow me to walk through each. First, core modernization. Across the world's 2,000 largest public companies, there is a technical debt backlog ranging from $1.5 trillion to $2 trillion. It once required a big team billing hours across months. Now we can deliver it as an outcome faster; that explains why clients move to the new model here first. Second, experience debt. Every customer-facing surface needing to be rebuilt for an AI-first world. Multi-month rebuilds are being turned into same-week releases by our Vercel and Claude-powered AI Pods. Third, agentic process transformation. The biggest opportunity — rethinking how a business operates around agents. Since the value lies in the transformed process rather than the hours, outcome pricing fits best here. Now, a few words about where we stand currently and how to read the top line while this transition plays out. Turning to Q2, our top line came in at $614.4 million, which landed inside the guidance we provided, representing a 1.2% increase quarter over quarter and a return to modest growth versus the prior-year period. Revenue from AI Pods accounts for about 2% of our overall sales right now, and by year-end we anticipate that figure climbing to 4%. That growth pocket I referenced previously — demand for AI Pods — remains visible; we are deliberately speeding up these migrations, accepting a near-term hit to revenue. The reason is that, in the long run, it delivers greater worth to the customer through outcome-based consumption that can be predicted, and greater worth to Globant via improved margins plus exposure to more advanced engagements. During this quarter, repeat clients generated 96% of our revenue, and our top 20 and top 50 accounts expanded 6.6% and 6.9% year-over-year, respectively; the concentration of that expansion sits where Glob.AI and AI Pod adoption runs deepest. Data and AI Studio now ranks as our second-biggest studio measured by revenue, representing nearly 11% of sales with growth of almost 35% year-over-year. More and more, our AI studios pair AI-native offerings with conventional staff augmentation, bringing their expertise to bear. That top 50 expansion signals a solid grasp of what these customers' sectors require. Acting as the distribution engine is our core business, delivering Glob.AI and the AI Pods built atop it to major enterprises through ties cultivated across 20 years. Both pipeline and bookings remain at a solid level. Their mix is increasingly weighted toward AI, data, cloud, and integration engagements. With all of that said, the quarter found us navigating a more challenging climate.
Pressure from geopolitics in the markets we've newly entered, oil prices swinging and dragging on travel, plus decisions taking longer to close in North America. Juan is going to walk you through a full-year outlook that we've revised. We introduced Glob.AI last week, and through that launch we're extending this very model to any enterprise via one self-service platform. Services that are AI-native, with pricing tied to output and value or else to consumption, are now within reach for a larger slice of the market, not only our biggest accounts. Let me show you how this works in practice. Glob.AI serves as one place an enterprise goes to locate, deploy, and begin consuming an AI Pod, skipping a discovery process that drags on for months and a lengthy ramp-up. A client logs in, describes their technological opportunity in everyday words, and the platform pulls from Globant's whole network of technological solutions and partnerships, suggests AI Pods, and lets clients begin building that very day. It closes the distance between mental throughput and arriving at sound business choices. Clients retain sovereignty over the models they choose and where those run; these Pods get built alongside the companies that define this technology. Certain AI Pods are engineered with name partners — secure code review alongside Anthropic, digital twin engineering on NVIDIA Omniverse, prototype to product with Vercel, and enterprise integration with Salesforce and MuleSoft. The platform also operates across the wider model ecosystem: Anthropic, OpenAI, Google, Azure, AWS, NVIDIA, Meta, among others. I'm pleased to share that Sarab Narang is coming aboard as Glob.AI's CEO. Sarab is a seasoned AI and technology executive carrying over 23 years of experience. He arrives from ServiceNow, where he drove commercialization of its AI business and before that held senior AI leadership positions at AWS, where he contributed to building and scaling AI platforms, among them Amazon SageMaker and Amazon Bedrock. Earlier on, he established KPMG's AI and machine learning practice. Nothing I've covered thus far functions absent the right partners; back in June, we announced a multi-year alliance with Anthropic, making us a preferred services partner in the Claude Partner Network. We've moved fast since signing. Several Claude-powered AI Pods are in production already. They were on display at our Globant Tech Summit in July. We're training thousands of Globers on Anthropic tools, and an active joint pipeline exists with several large financial institutions, airlines, and e-commerce companies. A year on from our initial partnership, OpenAI has designated Globant a selected partner in its new partner network. With Vercel, clients can ship AI-built applications natively in one click, converting multi-month projects into deliveries within the same week. We launched Vercel-powered AI Pods together, agentic units that design, develop, and modernize enterprise digital products on Next.js. FIFA is using AI Pods powered by Glob.AI to scale its digital ecosystem into a continuous, personalized experience for football fans around the world. Its key platforms recognize fan preferences across competitions and, powered by AI Pods, use real-time data to generate new experiences year-round. Three years have now passed since we launched our partnership with British Airways, and together we've been delivering a platform designed for speed and ongoing innovation. British Airways hit a significant milestone on that transformation path in June, launching their new mobile app after thorough testing aimed at making each part of the customer journey easier and more intuitive, so it serves as a real-time travel companion. The feedback from customers has been positive, underscoring the better user experience, especially the live flight notifications feature. That's only the beginning; we're expanding the partnership with additional features driven by our AI Pods model, speeding up what comes next; over in the Gulf region, we're engaged with one of its biggest financial institutions, constructing its first agentic bank. With our AI Pods model powering it, intelligent agents will operate across acquisition, onboarding, servicing, and risk, transforming how the bank runs and how its customers interact with it. GUT posted a strong Q2 2026, capped by yet another remarkable showing at the Cannes Lions International Festival of Creativity in June. The network took home 22 Lions, among them a third straight Grand Prix for longtime client Mercado Libre — the first agency-client partnership ever to reach that milestone at the festival. GUT also rolled out fresh work for Google Chrome and Ray-Ban Meta, built the world's first clay bar for Stella Artois at Roland-Garros during the French Open, and unveiled RIMOWA's For a Lifetime of Lives platform, honoring craftsmanship via tales of longevity and evolution. We're intentionally constructing a services business that's meaningfully different, with an expanding revenue base beneath it, and with a figure I've asked you to hold us to each quarter. This coming chapter also calls for disciplined choices now, including calls on our cost base to fund the transition and safeguard our margins. I don't take those lightly, and I'm thankful to our teams for the determination they're demonstrating. Thanks to our clients, our partners, and our Globers everywhere, who build this alongside us daily. With that, Diego will reveal the machine underneath. Thank you.
Thanks, Martín, and welcome, everybody. Martín has just described our strategic vision for Glob.AI and the fundamental way it reshapes how customers obtain our offerings. What I want to cover today is the engine underneath — the technology stack, how operations actually run, and the head start that lets this delivery approach succeed at scale. Under the old professional services approach, you trade people's hours or bespoke builds, which means engineers keep resolving identical core issues over and over. Glob.AI ends that pattern by functioning as an asset-based engine. Inside the platform, more than 20 years of enterprise engineering and sector-specific domain expertise has been encoded into curated, battle-tested playbooks. These workflows use agents, they're deterministic, they're documented, and they're designed to hit production-grade reliability. Because these agentic assets come in modules and have been validated, we get significant compounding value across different industries. Consider a workflow for IT root cause analysis that we built for an airline client — within a few weeks rather than many months, you can shift it to pharmaceutical supply chains or media distribution pipelines. Our direct way of monetizing this compounding IP is AI Pods, which take customers from a challenge stated in natural language through to deployment ready for production, without the typical friction. When we introduced Glob.AI, we noted that prompting a raw LLM leads to a lot of wasted tokens, logic that's hallucinated, expensive retry cycles, and additional hours for people to supervise. Glob.AI tackles this through optimization built on structured, deterministic processes. Before any AI agent lifts a finger, context gets gathered, architecture gets charted, and data gets prepped automatically on our platform, with automated quality gates enforced at every step. Through improvements to the orchestration layer, we guarantee that every token used yields output that is verifiable and ready for production. At the same time, control is what enterprise adoption hinges on. Glob.AI's architecture intelligently routes across over 140 LLMs, delivering full model independence so customers are never tied to one provider. Critically, each transaction stays locked inside the client's dedicated token vault. That ensures absolute token sovereignty. Client data is never exposed or leveraged to train third-party models, which lets organizations safely compound their own institutional intelligence over time; we are seeing this platform model's impact directly in our operational results. By embedding specialized AI agents into continuous delivery workflows under expert supervision, we are reshaping the software engineering life cycle. This leverage lets us separate gains in output, velocity, and delivery from growth in headcount. On top of that, it allows us to add nonlinear and recurring revenue into the mix, with structurally stronger unit economics that will eventually reshape the business. That is the structural fingerprint of a company climbing the value chain. Now, I prefer showing to telling. Let me walk you through a brief Glob.AI tour. It begins in a simple way. Once you've picked out your plan along with the AI Pods — whether through the catalog or by chatting with the Glob.AI agent — the projects module becomes available to you. At that point, you're ready to begin. There are two levels on which Glob.AI functions: the organization and the projects. The organization acts as the foundational layer, and it's where you configure the tools and preferences that extend into all of your projects. Actual work happens within projects, and every project carries its own goals in addition to one or more AI Pods running simultaneously at different price points, which consistently account for both the tokens consumed and the human supervision the project requires. So let's kick things off with the organization level. The first thing your company needs is background information, and you can deliver that directly from the systems it already depends on. Connect GitHub, and it gains the ability to view, modify, and submit pull requests within your repos. With Google Workspace, you can gather signals out of Gmail, Calendar, and Drive. Atlassian maintains synchronization with your Jira tickets and Confluence pages. Background information may also arrive by way of a document or a description. After you're inside your project, everything resides in one place. At the top, there's a snapshot of your project that pulls together the overall totals, how many AI Pods are currently active, your goals, how many tokens you've consumed, what you're spending right now versus your cap, and the start button that kicks everything off. Right beneath that, four key metrics also show up on your personalized dashboard. Pick a time frame and you can see spend, tokens used, total tasks, and artifacts created. Below all that is a brief summary along with your active goals, so the work items currently underway stay easy to see. Finally, there's your AI Pod fleet — the specialized teams handling your delivery. Over on the right side you'll find your communication channel, the Glob.AI agent. Just tell us what you're looking for in everyday words, or start by picking one of the options we've suggested below. You're also welcome to bring up your forward deployed engineer — that's the Globant specialist who's been assigned to your project — along with the direct line you have to the person supervising each and every step. Whenever you need to reach out, a single message is all that separates you from your expert. Goals represent the units of work that your AI Pods will carry out, and you make them by just informing the agent of what you're trying to accomplish. Repositories can be connected so the agent can look straight at your code base and analyze it. You might explain what you need using ordinary language, or you could attach a document and have the agent pull the requirements out of it. In any case, it's the agent that puts forward the objectives. Your job is to look them over and give approval, and whenever adjustments are needed, you can communicate that straight to it or loop in your expert — who is always on hand within the chat and fully informed about your project. Once everything looks right, you simply say start. With that one move, your project becomes locked in and is passed along to the forward deployed engineer, who oversees each stage of everything that follows. Artifacts serves as the shared space where your project files live. In this spot, you'll discover all the documents, specs, ABRs, or implementation summaries your AI Pods have generated, each one ready to be previewed in just one click. Both you and your FDE are able to upload files here, so everything stays together in a single location. A portion of the artifacts will show up straight in your GitHub repository too, either as commits or as pull requests. Once you select your name, you'll be able to access the notification settings. How Glob.AI reaches out is entirely your choice — email, WhatsApp, or Telegram — and you decide which events trigger each of those channels. That covers approvals waiting on you, tasks you need to complete, freshly delivered items, or run status updates. Selecting your organization brings you to a complete control panel containing four tabs: overview, which holds all the general information; members, the place where your team is managed; billing, where you get the full financial picture; and last, usage, which details everything consumed during the current cycle. The foundation of Glob.AI is not an abstract idea — it reflects what we have consistently shown in actual deployments; during prior calls, we explained how the initial AI Pods launches delivered significant efficiency improvements, such as accelerating drug discovery work at PharmaMar by a factor of 15, reducing supply chain contract timelines at YPF by 40%, or bringing legacy migration schedules down from 14 months to 2. Glob.AI carries such weight at this moment because those customized, high-value outcomes are no longer isolated projects. We have converted that experience into offerings available in our regular product lineup. With strong co-engineering partnerships alongside hyperscalers and model providers backing us, Glob.AI converts those established enterprise results into a repeatable, on-demand capability that any client can reach from day one. We've constructed the platform underneath, put the governance framework in place, and demonstrated the economics at scale. All of what I've just laid out is what accumulates behind a single figure, Glob.AI ARR; we're eager to pursue this coming chapter alongside all of you. Thank you so much.
Good afternoon, everyone. In the second quarter, we hit our revenue guidance, sped up adoption of AI Pods, introduced Glob.AI, achieved sequential top-line growth, and kept our balance sheet in a sound position. Quarter over quarter, five of our eight verticals expanded, and notably, growth in our top 50 and top 20 cohorts ran well ahead of the company average. We also responded to the market volatility we observed by taking steps on our cost structure. I'll go over our results, then take you through the revised outlook. Revenue came to $614.4 million, inside our guided range, up slightly from a year earlier and 1.2% higher sequentially. Versus the prior year, Q2 revenues carried 80 basis points of FX tailwind. Geographically, Europe and Latin America each grew relative to the same period last year, by 6.8% and 5.9%, respectively. On the other hand, North America declined 2.4%, and new markets fell 17.7%. That new market segment created a consolidated drag of about 115 basis points on the year-over-year revenue growth number. Because of the continuing conflict, this particular geography ran into unanticipated project delays throughout the second quarter. Our cohort performance stayed the standout. Top 50 clients rose 6.9% year over year, top 20 rose 6.6%, and top 10 rose 4.4% — every one comfortably above company average, consistent with our 100- squared strategy. 16 of our top 20 relationships are posting positive year-over-year growth, and we keep scaling large deals signed recently. Adjusted gross margin stood at 36.5%, down somewhat as dollar weakness picked up, mainly hitting our biggest delivery center, Colombia, and our utilization stayed under target. Adjusted SG&A was 18.6% of sales, and adjusted operating margin reached 13.2%, under our guided range and caused by the margin impact. Reacting to these conditions and to best position ourselves for later expansion, we began a business optimization initiative in Q2. With this initiative, we secure the acquisition and retention of the capabilities our AI-focused strategy demands, while at the same time resizing our cost baseline to match the current market landscape. The primary actions in this plan involved a thorough review of our workforce to match skills and headcount with our strategic priorities, a consolidation of our global office footprint following an analysis of our facilities and lease contracts, and a strategic prioritization of our delivery centers to underpin future expansion. Tied to these actions, we took a one-time charge of $32.3 million in the second quarter; we anticipate certain actions carrying into Q3, which will be essential to safeguarding our profitability near term given the FX headwinds currently before us, and will be reinvested to power our growth engines, in particular our AI platform development and our people. Even with FX headwinds, we intend to lift margins through the added efficiencies planned for Q3 and by raising AI Pods in the mix, since they run at margins above company average. Adjusted net income was $60.3 million, with a 9.8% adjusted net income margin. Adjusted diluted EPS came in at $1.40. Our balance sheet is still a pillar of strength; when the quarter closed, we held $168.8 million in cash and short-term investments alongside $253.1 million in net debt. Free cash flow during the quarter totaled $12.6 million, while for the first half of 2026 free cash flow hit $48.7 million, which stands as a company record. Turning to capital allocation, the repurchase program our board approved in May for as much as $125 million across six quarters is now running. Given where the stock trades today, repurchasing Globant continues to rank among the best-returning uses of our capital, since the market values Globant like a legacy services business stuck in a weak cycle, when in reality we are the industry's fastest-scaling AI-native delivery platform. At the present valuation, the company trades at a double-digit free cash flow yield when measured on a normalized free cash flow basis. Now let me move to our outlook. Three external factors mainly shaped what we expected back in May, and all three are baked into our updated guidance. Back in May, the bottom end of our guidance assumed a meaningful worsening in our new markets region that, at that point, our forecast did not yet capture. That scenario came to pass, and our expectations for the region for the second half of the year have now shifted. Our dedication to the region is for the long haul, and key relationships there keep expanding. The sensible assumption as of today is that this environment continues over the near term. Second, we have felt certain knock-on effects from oil prices, which are weighing on the travel ecosystem. A number of our travel clients have eased the pace of their transformation programs so as to shield their own P&Ls, even though other players in that same industry are speeding up with us. In our view this represents a deferral of ramps, and we anticipate that this revenue comes back once industry volatility fades. Finally, we keep seeing drawn-out cycles when it comes to discretionary decision-making. Because of all of the above, we are updating our expectations for the second half of the year. For the third quarter of 2026, we now anticipate revenue in the range of $607 million to $615 million. We anticipate a non-IFRS adjusted operating margin in the range of 13.5% to 14.5%, plus an IFRS effective income tax rate within the 21%-23% range. Adjusted diluted EPS should land between $1.43 and $1.53 per share, assuming an average of 43.2 million diluted shares outstanding. As for the full year, we are updating our revenue guidance to a range of $2,428,000,000- $2,462,000,000, down from $2,462,000,000- $2,508,000,000 previously. On profitability, we now anticipate our adjusted operating margin for the full year in the range of 13.5% to 14.5%, driven by the increasing dollar weakness. The IFRS effective income tax rate should fall within the 21%-23% range. We now anticipate adjusted diluted EPS of $5.75- $6.15, assuming 43.6 million average diluted shares. We anticipate robust free cash flow generation during the second half, in line with our seasonality, and our capital allocation priorities remain the same. That buyback plan, plus our ongoing expansion of AI Pods. The optimization push we undertook during this quarter should show up in our margins by the time the year wraps up, putting us in position to start 2027 with a slimmer cost structure, all-time-high revenue per Glober, and our most profitable delivery model, AI Pods, nearing roughly 4% of revenue on a run-rate basis as the year ends. In closing, the shift toward AI Pods sped up, we set a productivity record, and we delivered strong results among our biggest clients. The robust appetite we're seeing for AI Pods confirms the industry thesis we hold, one we believe will change for the better; we will stay intensely focused on this delivery model transition so we can speed up these trends. Thanks for your ongoing backing.
Thanks, Juan, and hello to all of you. As we move into the Q&A portion of today's call, I'll begin by saying your name first. When that happens, kindly unmute your line, and then go ahead with your question. Once you've finished asking it, please mute your line again. I'd also like to request that each of you stick to one question plus one follow-up. Many thanks; keeping that in mind, our first question comes from Bryan Bergin at TD Cowen. Bryan, the floor is yours. Your line is open.
Alright, thanks. Hey, everyone. I'd like to ask about the business transition. You're posting solid sequential gains in Pods, and the target now sits at $110 million. My understanding is that's up from the earlier $60 million to $100 million. Given what you're observing, how long do you expect this transition phase to last as Glob.AI and the Pod model scale up, to the point where they can drive a re-acceleration in the company's overall trajectory? I suppose in connection with your updated revenue outlook for 2026, the midpoint of the constant currency forecast, I believe, is down just under 2%. What portion of that reflects deliberate impact from shifting to this engagement model, versus macro headwinds affecting the business?
Okay, I'll take the first one. That's a really important question. I'd say the move to the new model is something we're carrying out in a deliberate manner, and it's something we'll continue executing quarter over quarter. Honestly, the demand we've observed, the acceptance of the model, and the positive signals coming from the market are quite encouraging. It's still a small percentage, but we believe we'll keep accelerating this. If you ask me, if revenue from the new markets weren't affected, we'd already be on the positive growth side without needing to review the whole forecast for the year. So overall, I'd say it's a very positive moment, and it will accelerate growth a lot. Probably by the end of next year, we'll see a pretty strong effect of that kicking in. I can't say it right now. I don't think that's a piece of information we can describe right now in a very exact way. So with that caveat, I'd leave the second part to Juan.
Thanks. Hey, Bryan. Our full-year guidance midpoint now sits at $2,445. That works out to 0.4%. Currency gives us roughly a 70 basis point tailwind. So on an organic constant currency basis, we'd be looking at 1.1%. Looking at what changed in the guidance, most of it comes from lowering our forecast for the new market business. Right after our May report, there was a wave of news out of Saudi about budget cuts, project delays, and similar things. As you're aware, that's a market where we've been growing pretty well for the past few years, and we intend to keep doing that. A lot of deals are simply being pushed out or slowed, though none have been canceled. Conversations remain very interesting, so we're confident that market will bounce back before long. That accounts for half of the guidance revision. Roughly $10 million also ties in some way to what's going on there, since higher oil prices hit some of our travel and hospitality businesses, which meant lowering the forecast for the back half of the year for certain of those customers. The remainder is a bit of a blend — some assumptions we're making around certain migrations, plus the broader business climate and where things stand today. Again, I'd say it's also important to watch how the new business, the piece we're really pushing hard, is developing. Sure, it's still small, but once you start compounding quarter-over-quarter at 40%, 50%, it ramps up fast and becomes more meaningful. Like you mentioned, we'd been discussing $60 million- $100 million for this year. We're already past $52 million, and we have strong visibility into the second half since we're moving through that first early phase of trying, testing, and learning what working with an AI Pod actually involves. Among our top customers — 45% of the top 20, in fact — it's already in use. What that tells us is those customers are beginning to scale. So we're confident we can grow this business beyond $110 million by year-end.
All right. Thanks for walking us through that. I just want to dig a little deeper into the optimization that you carried out. Could you share what kind of savings you expect those initiatives to generate?
Sure. Essentially, the situation is that we've been taking a look at our workforce and bringing it in line with where demand stands right now, as well as what the business currently requires given the new models sitting in front of us, and the skill sets needed for the fresh approach to service delivery that we've put in place. As a result, we needed to make certain adjustments within the organization. On top of that, when it comes to delivery centers, we optimized those as well. The hit during the second quarter came to about $32 million. For the third quarter, we anticipate somewhere in the range of $20 million to $25 million, and with that, the program wraps up for this year. Our view is that this will first of all help us cut costs, since otherwise a portion of that talent pool, or a portion of that talent within the pool, would have sat there with perhaps no way to assign them to new projects. Second, it helps us counter a huge FX headwind we're facing due to the weakness of the U.S. dollar. If you consider our biggest development center, Colombia, ever since the new president was elected there, it has gained close to 15%, and that hits our numbers enormously. We're going to counter that; we're going to put more money in. As we talked about on this call, we've just named a new CEO for our AI Pod — pardon me, the Glob.AI business — and we need to invest in that business since we believe it represents the company's future. That's where the money will go, and it will spare us costs we'd otherwise have incurred.
Okay. Thank you.
You are welcome.
Hello there.
Thanks a lot, Bryan. Our following question is being asked by Tien-tsin Huang of JP Morgan. Tien-tsin, the floor is yours.
Thanks, Arturo. On the optimization topic, I'd like to ask something. I simply want to confirm that I grasp it. As you put it, Martín, this choice is being handled by you with a great deal of seriousness. Which specific parts got affected? What portion of that came from the surprise you encountered during May, and what portion came from moving toward the new model? Or is the real driver more the delivery centers plus getting into closer alignment with certain FX and inflation patterns, as you mentioned regarding Colombia? I was hoping to get a clearer picture of that.
Hello there, Tien-tsin. How are you doing?
I'm doing fine.
There are various explanations behind that entire initiative, correct?
Yeah.
What we're seeing is this: on one hand, our shift toward that new model calls for specific kinds of forward deployed engineers, plus specific kinds of AI engineers, and those differ somewhat from the ones we previously had. That change in talent is one factor driving the optimization. Beyond that, we're also witnessing a shift in what the traditional business demands. Traditional business demand is leaving web UI testing behind and heading toward more data, more cloud, and implementation. That shift generated a need for new profiles we hadn't had much of previously, so we had to begin training and retraining. That's the consequence of it. On top of that, several programs and initiatives aimed at running the company more efficiently have had an effect. As you noticed, the rise in revenue per head carries a significant message within it, since we're getting more efficient at delivering our revenue; that push involves many different pieces. Juan, I don't know whether you'd like to add something there.
No, I'd say this is essentially a sort of restructuring — or perhaps not restructuring so much as giving our people new skills for the different kind of demand, and for the different kinds of services we're now delivering to customers via the AI Pods. A different skill set is definitely needed. On top of that, our margins have to be protected. We've got to ensure we counter all of the U.S. dollar softness hitting our Latin America business. And we also need to free up room for the investments Glob.AI will call for.
How much was the FX impact in the last year?
Suppose we examined each of the currencies across Latin America over the past.
5%.
Between eighteen months and 24 months, the total effect purely from currency movements comes to roughly 4 percentage points. Naturally, we've managed to push up revenue for each employee. That has allowed us to counterbalance a portion of it, and on top of that we achieved certain efficiencies during the prior year that gave us additional support; even so, how large this adverse force has been for us across Latin America is extremely, extremely substantial, and it is weighing on our margins.
Yes, I'm grateful for that. Since it's something you can't influence, it's smart to address it early. Quickly then, regarding the new model and Glob.AI, I liked what everyone and Diego had to say. When I look at Globant's past, the 50 squared account AI and the way it scales always comes to mind. Could you share some sense of what revenue per client might look like as you move deeper into your top 10, 20, 30, while you're learning? Is there a comparison or parallel we could make to Globant's growth under the earlier model and apply it to the new one? I'm just hoping to grasp better how this could scale past a few of the figures you provided for this year.
Yes. On the 100-squared program, traction looks quite strong. That customer group did grow — the top 50 was up around 7% or so, and that's really encouraging, correct? That's where we're rolling out these new offerings and the initial experiences across the past nine months of execution, or a year of execution, of our AI Pods. So that alone is quite encouraging. As for revenue size, suppose we hold gross margin steady, since gross margin on the AI Pods is much greater, and we have perhaps, say, $1.2- $1 on the new service line. Though not across every account. It varies greatly by each individual account. Keep in mind, you can't trace it back to the original model, because the original model was headcount.
Yeah.
It could be a set fee or based on hours and materials; under this fresh approach, you get tokens along with token oversight bundled into one rate, whether charged for each million tokens or for every output. That produces an entirely distinct kind of calculation, correct? Within this fresh calculation, there's room to fine-tune margins, to enhance oversight, to accomplish more using fewer resources, or perhaps, for certain accounts, additional oversight is required on particular projects, yet it's handled in a completely separate fashion compared to earlier. That's the reason I enjoy stating that this shift isn't merely some sandbox we launched. In my view, it represents what lies ahead for an enterprise. Shifting toward output, toward consumption, toward value under an entirely new framework compared to what came before. I'm not claiming the prior approach or the conventional approach is going to vanish. However, I am indeed saying this is a shift. Gradually, you can observe us building steam with this fresh delivery method, with this fresh method of billing our clients that stands entirely separate from the conventional approach. So drawing a comparison between those two matters can occasionally prove challenging. Truthfully, we don't have a lengthy track record. We've spent 12 months putting this into practice.
Yeah.
There are already certain indicators we have; those give us enough confidence to set a price for that million tokens, or for the outputs we're billing on, but plenty of additional developments still need to occur before we could reach the sort of determination you're looking for. What I can tell you is this: look, things are progressing quite well for us, and we're expanding quite well. It goes beyond what I myself anticipated. As I mentioned, $60 million- $100 million, and at this point it's no less than $110 million. My belief is that compounding will continue, since it makes tremendous sense for the customers we serve. A lot of the fresh initiatives I brought up during earlier earnings calls — regarding interface changes, process automation, extracting meaning from enormous volumes of information — every one of those demands an entirely different delivery approach. This new offering isn't only meant for those 100- squared customers; that's a wonderful aspect here, because it also exists to serve additional segments that could be smaller, and we're still figuring out how to pull that off. As we move forward, we'll continue broadening this. I don't know, Juan or Fer, whether either of you wants to chime in.
Actually, the final point you raised matters a great deal. Through Glob.AI, we're also exploring ways to expand our customer base, to serve clients across varying scales—something we previously couldn't manage in other segments—so we can speed up our revenue growth. As you mentioned, we've demonstrated that the 100-squared model works; it shows up clearly in how at least our 50 top customers have grown. The real challenge now is keeping up the growth of a client segment that behaves differently, correct? That segment requires a different approach to serving it. We think Glob.AI gives us a path to connect with a far broader set of customers in a simpler manner and one that's more sustainable.
Mm-hmm. Yeah. No, I appreciate the perspective. Good to see everyone.
Thanks.
Thanks, Tien-tsin.
Appreciate it, Tien-tsin. Good to see you. Joining us now on the line is Maggie Nolan of William Blair. Maggie, the floor is yours.
Hi there. Thanks. Good to see you. I'd like to expand on something from one of those earlier questions. Let me think. Alright. Yeah. During your remarks, you mentioned that Glob.AI plus an emphasis on outcomes was unlocking budgets for you that previously weren't available; could you say more about where that growth originates? Who's doing the new buying? Is it expanding, or do you see the total addressable market expanding, and what's shifting your capacity to go further with clients?
So, that passage you brought up means this. Over many years, as I explained right at the start of this earnings call, we've built experiences and software products, and we've stayed very close to applying technology to craft experiences that connect emotionally with our customers. That's the heart of it, that's what makes us who we are. Then, with AI arriving, plenty of additional areas exist where it isn't only about crafting those experiences, but also running portions of those backends and processes that previously weren't something we could pursue. As I explained, with process automation and the shifting landscape of how AI gets used for essentially everything, including automating every process, building new org charts that mirror that automation, we think our AI Pods, which are now the AI Pod software that you'll find in Glob.AI if you visit, will develop into operations and will carry that same idea of an agent running something on your behalf while humans can examine the edge cases, with pricing structured per unit of that particular business case. Take travel, for instance: it would be per trip, or know your customer, per know your customer. That agentic work which this Glob.AI concept unlocks is far bigger than the software development life cycle we currently have within Glob.AI. What I mean is, with this fresh idea, this fresh concept, this fresh definition of how services and AI native services will be delivered, we can address far more than merely the original software development life cycle, and we can stretch our presence into AI Pods for operations, correct? That's what I mean regarding pursuing new budgets. There's also a market share element, since when you introduce this new approach and new method of charging, for each dollar you sell under this new model, you can capture perhaps another dollar, right, from that same budget, because you're performing tasks that other vendors used to perform. We're very enthusiastic about both of those things occurring simultaneously: expanding into additional areas and capturing more dollars for that work we historically did. Often, that was the reason we win and we won. There are plenty of instances I could point to—quite a few, actually. If that weren't the case, we'd have come out behind. This fresh way of defining both how work gets done and how we bill for it strikes me as a genuinely novel proposition for our clients; it went live last Thursday, and the response—the sheer volume of calls we're getting, the number of folks eager to learn how this Glob.AI model operates—represents, at its core, an enormous undertaking and an enormous intellectual undertaking to build something that hadn't previously existed, and we're the first to offer it. The entire team behind this development makes me incredibly proud. I'm incredibly proud that we've managed to bring it to market, and that we're able to win over our customers—and not merely a handful of cases anymore. It's going to exceed $100 million in ARR. That's an outright triumph, and I believe this is something you'll watch us push harder and harder on—deliberately moving toward requesting that our customers adopt a different model; that's going to take time; that's going to be a journey in its own right. It brings to mind other businesses that have reshaped the way they operate. On the whole, I think watching that shift unfold and steering it will be very thrilling.
Thanks, Martín. That's really useful. So then, clearly AI Pods is the growth engine here, along with Glob.AI, but EMEA had been anticipated to serve as a future growth engine for the company and has been a significant region over the last 2 years. You made it fairly clear that your outlook for that region was conservative or prudent given that conditions there have shifted. What I'd like to know is where, looking at it from an end market angle, whether that's a vertical or a geography, you're now focusing your attention as a possible growth engine across the coming 12 to 18 months, and what success metrics we ought to watch for there.
Thanks for that, Maggie. Here's how I'd put it: across our sector, growth has pretty much leveled out. If you compare us with other players, everyone's expanding at roughly the same rate. What we came to realize is that simply applying new technology to the way things have always been done doesn't get you there — you need a completely different way of delivering. So when we spot an area that's expanding rapidly and decide to focus our business on it, that's exactly the opportunity we're after. Of course, some sectors pick this up more quickly than others do. Financial services has given us a lot of wins, and so have media and entertainment. We've also had a lot of success with airlines. This approach boosts efficiency and speed — it's 30% faster. As the game shifts away from cutting costs—which is what most people have concentrated on with AI up to now—and moves instead toward bringing in revenue, which in my view is the intelligent application of AI, we're going to see a lot of sectors enter this area. What I'm trying to get across isn't simply a geography or a vertical, but a fresh approach to delivery that deserves your attention. That's the reason I'm asking you to keep us honest about that transition figure as we grow this business. This is where our efforts are focused—figuring out how to deliver far more efficiently. It isn't merely a matter of stacking headcount on top of AI. Having humans alongside AI generates a tremendous amount of junk, consumes enormous stretches of time just watching over that junk, and forces endless cycles of fixing things again and again. Once you introduce structure into that workflow — which is exactly what our Glob.AI delivery approach accomplishes — efficiency rises sharply, your results no longer hinge on whichever model you pick, and you gain the ability to scale. And Glob.AI itself? We're only at the starting line with it. Its scope stretches far beyond what I've described. Through Glob.AI we are working to remake Globant — the entire company — into an AI native organization. It serves as our AI native services arm. Let's take a look. Asking me to narrow this down to a single area, sector, or field isn't something I can do. The scope here is broad — extremely broad.
Understood. Thank you.
Very welcome.
Thanks, Maggie. We now take a question from Bryan Keane at Citi. Bryan, go right ahead — your line is live.
Hey, everyone. I just have a question, Martín. When you mentioned that you're deliberately choosing to shift additional work over to AI Pods, it sounds like that's putting a strain on your current business or cutting into that segment's core revenue. Could you speak to that downward price pressure and explain why it won't persist for a longer stretch as we move through this transition across the coming couple of years? Will we be stuck operating in negative revenue territory because of the deflationary effect that moving work toward the AI Pod model might create?
Look, essentially, many of our clients have been paying the same sum to gain greater productivity, correct? That's been true across the overwhelming majority of situations. My sense is that down the road this might shift to where first we have to win over procurement, we have to win over additional stakeholders, and that effort might take longer than simply playing the old-fashioned way, yet we're prepared to bear that cost. I believe the overall margin is going to be significantly stronger. Our ability to push margins even higher beyond where they stand today remains greater; so that deflationary situation is one we aren't observing at present. Should we have to bear the cost for it, we'll do so. I want to make that transition, and that's the reason I said, listen, we're shifting this over. Even knowing that with certain accounts there'll be somewhat reduced demand, I still think that on the whole the picture is going to look completely different.
Bryan, what I meant to say is that at times, based on what we've observed, convincing and persuading the procurement folks—and persuading our client—to make the switch while holding revenue at the same level with greater productivity can take more time, or considerably more. Occasionally we find ourselves in a situation where, yes, we know we want to migrate. Sometimes we're able to move more quickly when we're prepared to hand the customer certain efficiencies right away on price. Yet our view is that since this model delivers far greater efficiency and makes far more sense for them, we ought to be capable of either growing into other parts of the organization or taking market share away from competing vendors. That's because our conviction is that this is a model under which we need to deliver services. And if moving faster is possible, then we'll do that. That, I believe, is a central message we're putting forward here.
No, that's genuinely useful; so, following up on that, Juan, regarding revenue per head, it rose to nearly double digits. What portion of that comes from like-for-like pricing, or are you currently achieving somewhat improved pricing in the market? I'm just attempting to grasp that figure.
Several factors are at play here. The marketplace is competitive, correct? On certain occasions, we can complete work with fewer people since our delivery approach has become more efficient at present. When you examine overall headcount, it has declined by approximately almost 10% compared with a year earlier, while revenue per head has climbed by approximately almost 10% versus the prior year. Thus, our efficiency has improved. In certain instances, we have managed to secure some extra pricing. However, I would not treat that as typical, since the market remains extremely competitive at this moment. Even so, we have succeeded in raising revenue per employee, as our delivery has become more efficient. Sometimes, we bill under a new structure, using the AI Pod revenue or the Glob.AI revenue model. On other occasions, it might involve a fixed price, and because we can deliver more efficiently, revenue per head rises. It is not that we bill in the manner of Glob.AI, since it could be a fixed price, yet we still generate greater revenue per employee, because our delivery is more efficient. I believe that accounts for the steep rise. When you examine revenue per head three, four, five years back, it stood at roughly $60,000-$65,000 per employee. Today we are approaching above $90,000 and nearing $100,000.
Yes, got it. Thanks a lot.
Sure thing.
Hello, everyone.
Thanks a lot, Bryan. Now we'll take a question from Arvind Ramnani, who is with Truist Securities. Arvind, the floor is yours.
Hi there. Appreciate it, everyone, and good afternoon. I've got a few questions about Glob.AI. Martín, you mentioned that this is roughly where your efforts are concentrated. So a few things on that front. At present, how is the workload divided among OpenAI, Anthropic, and open weight models? And over the coming 18 months, what shift do you anticipate? Then, related to that, you also noted that many of the clients taking advantage of Glob.AI are ones you already had. But what number of clients, or what share, are essentially new clients who weren't previously using it? And the final thing on that: what would you say is Globant's distinctive value proposition? Does it center on enterprise context? Does it center on routing logic? What exactly is proprietary to it?
Yeah.
Your company? Yeah.
All right. Let me take the first one, then the last one, and I'll have Diego finish up. As you can see, we've just made public a partnership with Anthropic; we're very thrilled about what we can accomplish together. We're already witnessing some effect from that pipeline entering the scope of our work, which is quite exciting. I think models are going to keep evolving, and it's up to our customers to choose what they use. Looking at Glob.AI, we see something capable of using essentially any model at its core when it comes to developing or creating software, open weight models included, if that's what's needed. A large portion of our tokens gets processed through our own infrastructure and our own models, with open weight models used on many occasions. Some of our customers ask for that, while others say go all in with Anthropic or with OpenAI. So we have essentially complete independence on that particular matter. And naturally, tools will continue evolving and growing ever more sophisticated. As for the specific mode I'd like to explain regarding Glob.AI, Glob.AI is a services play. Essentially, we blend in a completely frictionless manner the building of the experience and the building of the software and the definitions you require with AI and with humans, and we bundle and pack it all into one price, and one price per consumption or per output. So essentially, what you saw, the demo, the video Diego showed us, is a video making it very clear that Glob.AI assists you with the definition, assists you with creating the specification, and then a set of agents gets triggered, a set of loops gets triggered, or a workflow gets triggered. Depending on what you need to do, those things differ, and this playlist has been curated and evolved over time. Depending on which playlist you're using, they'll require differing levels of supervision from humans watching what those agents are creating, but we bill you in an extremely transparent manner, either per million token or per output in the case of a user story or in the case of a. What I'm saying is that this coordination, that removal of friction when buying services, is the real thing we're providing, and that holds regardless of any model you might pick. That's why we say this is real AI native services, because it plays the same role as in the past when cloud didn't exist and we needed to create compute. You bought the servers, you bought the connectivity, you used to buy, or to hire the people to manage those servers, and then all at once someone coordinated everything beautifully, and ramping up infrastructure became so easy. Okay, so professional services still belongs to that earlier period, and through Glob.AI, what we're doing is building the AWS for services. You can very easily describe what your project is, set up the context, hook it up to your Jira, your GitHub, or whatever repository you prefer, and after that run the mission — and that mission gets carried out in a highly professional way, overseen by the finest individuals who can steer you toward the proper enterprise outcome that you're after. That idea, it looks simple, yet it's remarkably advanced; this is the next stage of services, and it isn't merely a platform. It can work with codecs, with our own coder. It can work with Claude Code. It can work with just about every AI tool available today. After that, it can be deployed across any kind of infrastructure; the key point, though, is that a person always coordinates it in a highly efficient way. No further team ramp-ups, no further lengthy procurement cycles for anything. No further confusion about what something is going to cost. This is something completely different, and that's the reason Diego played the video. But I don't want to step on your speech.
No, no, that's not the case at all. My apologies. I'd say it's a fair point, and completely legitimate to ask. Something we consistently run into across every enterprise engagement relates heavily to both accountability and repeatability. A frontier model won't give you accountability. You put in a request, you receive an output. Whether it's good or bad is on you. You deploy it, and whatever comes out is what you get. With Globant, accountability is what you receive. That's the reason humans are part of this. The next key point is that models aren't repeatable. If you turn to a frontier model, it performs. For generating output, it's extremely strong. It's impressive, and we're getting 100% of the value from that. Yet for planning — how do I actually carry something out — it comes down to how you phrase your prompt, the volume of information you supplied, and what the model's decision and reasoning process looks like. It burns a huge number of tokens on tasks that are sometimes simple and could probably have been handled for 10% of the cost, just to give one example; we've actually stepped away from that approach, and what you see there labeled as battle cards is essentially the formula. So what's the correct method — what's the right approach for delivering this kind of value? What's the right approach for a migration, a replatforming, an SAP S/4HANA migration, for instance? That blueprint contains a great many highly defined steps. Each step has a mandatory input — nothing less, nothing more — plus a mandatory output, along with oversight for it, known as a quality gate; the result is a repeatable system, one you can feed again and again and get identical results, and that's something AI won't give you; it makes enterprises lose their minds; you end up with no control over spending. Often, models are in fact operating at both ends — the definition and the testing — while what occurs in between resembles brute force, like trial and error, and that's frequently how graph engineering operates; so we've overhauled this entirely, and we rely on what we know, where humans genuinely contribute value, confirm everything is exactly right, and you deliver accountability for it. Going back to what you originally asked, I believe every company we speak with genuinely takes a lot of comfort and feels far more at ease with this approach. This is the kind of service we've been delivering for more than 20 years; it isn't about the output, it isn't the source code, it isn't the executable. It's about the process — capturing what the client truly requires, adding value on top of it, and making yourself accountable for that output, for that business impact.
Take this example from just a handful of days back. A customer over at Glob.AI had a fairly intricate project involving ERPs, APIs, and connections, and they required some form of architecture definition. That task landed with Glob.AI. Our team picked it up. Work begins being performed by the agents. Oversight takes place over the course of that work. We go back and forth with that customer 2 or 3 times, and in a record 48 hours we wrap it up. Accomplishing the identical thing the traditional way, even inside Globant, would have required a minimum of 2 or 3 weeks, on account of the meetings, the various items, and the scope gathering. Every one of those things, that entire process that used to be wildly inefficient, has now been squeezed into an extremely straightforward manner of getting it done; so the value generated by comprehending the customer more quickly and arriving at the accountable result, exactly as Diego was explaining, is the central point of what we are discussing today; that is the transformation I intend to bring about for the entire company, for all of Globant, for all of our customers. My view is that this produces significantly improved margins, revenue that is far more predictable, and revenue that is much more consistent and recurring. In my belief, it produces results for our customers that go far past merely applying AI alongside a group of engineers, and I believe it is the response that many others will adopt. We are not the only ones. In this vision we are starters, and we are innovators, yet things will not end here; so it is going to be fun.
Yeah. Just quick follow-up. Sorry.
Go ahead.
The next question comes from the line of Jonathan Lee from Guggenheim.
Go ahead.
Jonathan, please. Your line is open.
Excellent. It's a pleasure to meet with both of you, and I'm grateful that you're fielding my inquiries. Martín, I value the forward-looking perspective, and I realize that at present you don't detect any downward price pressure related to AI Pod work. Still, once that pressure does appear, what's your plan for countering it — especially in situations where customers return at the point of renewal asking for a bigger portion of the efficiency benefits? Which structural safeguards exist inside the Pod model that would allow you to maintain your rates once procurement, as it surely will, starts pushing back?
I'm not sure I caught what you were asking. Could you please say it once more?
So when you consider the deflationary pressure that isn't present right now, what occurs if it does show up and you have to fight it going forward? And does the Pod model have anything built into its structure that allows you to keep your pricing intact when procurement tries to negotiate you down?
Oh, well, look. My view is that if procurement pushes back, you're always going to have a negotiation. But that's true across every model out there, not only the AI Pod model. The key point, though, is that I think we're able to be far more efficient. Should the customer choose to do the same thing, it'll require less money. Should the customer choose to do more — which is what we're seeing in most situations — they'll pay the same amount. And here's the thing: customers are going to want to do a great deal more. In this new model, my sense is they'll wind up paying more money to generate far more, just as we've described previously. Once again, this comes down to how much software, how many solutions and processes have to be built or utilized in this new era. I don't think you can forecast this game so simply, by assigning a fixed value to the number of things that have to be done or that are required. Sorry. Since that number of things shifts all over the place and frequently rises as new needs arise, as I mentioned earlier, when those new needs come about, where you ultimately end up with the AI Pods, in my view, is going to be higher; and that's already proven out in the top accounts, which are expanding at 6% and 6.9%.
Plus, in that same situation, if a customer is looking to do the same thing and there’s a possible savings for them, the model operates with better margins compared to the traditional one.
That's right.
There may be situations where our revenue dips slightly on certain accounts while the dollar figure stays the same, but ultimately, here's our perspective: this is simply how services must be provided from now on. Should we execute with greater speed and quality than competitors, then even when we sacrifice some profit on projects that require trimming revenue a bit, more engagements will come our way, market share will be captured, and faster growth will return eventually. That's our mindset. And yes, we're prepared to accept a portion of those situations. Once more, this doesn't apply universally. Up to now, plenty of instances have shown it just continues to accumulate.
Yeah.
Additional effort. Yet in certain situations we've observed, should we accept a lower price, migration can be sped up; that's the point we're making right now. We're ready to move faster since it puts us in a stronger spot with the client, shields us from rivals, and we think it will bring more business to Globant moving ahead. Simultaneously, given that we're able to operate more efficiently, particularly as we grow, additional margin can be captured throughout. So several things are occurring concurrently. Modeling this remains difficult, even for us. Each day we're constructing models. But what we're noticing at minimum is obvious momentum, an obvious gain in margins, and clients are pleased and returning to expand the model. That's the way we view it.
Right. What we've discovered is a segment experiencing rapid growth, and that's where our opportunity lies. Our aim is to broaden that transformation and accelerate it. To put it plainly, that's the essence.
One more point, I'd say, is that everybody knows the current market has zeroed in on the cost-cutting machine and the operational side, and that's significantly harder. In my view it's a worse situation, since when you're included in the cost calculation, it becomes far more challenging, correct? That's the kind of pressure involved, and businesses might look to take back some of those efficiencies. As soon as markets begin to bounce back and shift toward a revenue-generating engine and quit ignoring products, for instance, the picture will certainly transform. We've witnessed this time and time again. When the product becomes the focus, you're looking to pack in fresh features, extra content, and so forth. So the goal isn't about the money you save, it's about improving what you offer and making it stand out ahead of the rest; so with luck we'll see that occur before long.
Appreciate the additional detail. As a quick follow-up, which macroeconomic environment is baked into the updated 2026 guidance? In other words, does that span presume things stay as they are right now, perhaps a leveling off in North American decision cycles, or further decline? And how much room for error sits inside the bottom end in case of additional weakening?
No, I'd say the scenario we view as most probable is the midpoint, which essentially already bakes in a meaningful hit from the Middle East along with our travel customers. Naturally, once more, should conditions improve somewhat on the macro front, particularly within the U.S., we could likely land somewhat above that span. Once more, should conditions worsen — we aren't observing that at present — the cushion exists purely to guarantee that this span is something we'll deliver regardless of what happens; yet the scenario we view as most probable, as it always has been, is the midpoint. Regrettably, certain factors have emerged that affected our prior guidance and lay outside our control; even so, the midpoint remains the scenario we view as most probable. We've got plans in place to land above it. We'll do our best to carry those out. Certain things simply aren't within our control.
Building on that, for the year as a whole, 90% of our revenues are already under contract; in certain markets that have taken a hit, especially the new ones, we've shifted the pipeline toward banking, toward public sector segments that haven't been hit as hard as entertainment and travel and hospitality, which is where our attention had been directed in new markets up to now; the pipeline is being rebuilt. Some of those fresh opportunities are being closed by us. We also anticipate that new markets took a pretty heavy blow during this quarter. A swift bounce back is also what we expect as the year draws to a close.
Much appreciated, Jonathan. Good to see you. That brings today's Q&A portion to a close. Thanks, everyone, for spending your time with us; at this point I'll turn things over to Martín for a few closing remarks.
Appreciate that, Arturo. And thanks to all of you for joining us today. I'm grateful for the support you keep giving, and I look forward to catching up with you on the call for our next quarter. Thanks a lot. Bye-bye.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Globant S.A. transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Globant S.A. earnings transcripts and 132,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.