Globant went from a 30%-plus digital-engineering compounder to ~2% growth in a single year, and the stock fell ~61% from $97 to ~$37 as the market priced two fears at once: that client tech spending froze, and that generative AI commoditizes the headcount-billed services model entirely. It now trades at 0.77x sales, ~5.5x EBITDA, and a ~16% free-cash-flow yield — distressed multiples for a profitable, cash-generative, founder-led firm. The question is whether the growth stall is cyclical or the first sign of obsolescence.
Globant (GLOB) is a Latin America-born, founder-led digital-engineering services firm — 31,000 engineers building software for clients like Disney and Santander. After years of 20-30% growth it hit a wall in 2025: revenue grew just ~2%, EPS fell ~39%, and management cut guidance repeatedly as client discretionary spend froze. The stock collapsed ~61% to ~$37, near its 52-week low, with the added overhang that AI coding tools threaten the labor-arbitrage model the whole industry runs on. Yet it still generated $261M of free cash flow (~16% yield) at a 0.77x-sales valuation, and every analyst target sits above the price. This is the deepest-value, highest-uncertainty name in the AI-services-disruption cohort.
Full thesis
Globant is a genuine value-versus-value-trap debate with an AI twist. The financials still work — $261M of free cash flow, ~35% gross margins, modest leverage, a founder-CEO who built the company from Argentina — but the growth engine seized up, decelerating from double digits to ~2% as client spending froze and AI fears spread. At 0.77x sales and a ~16% FCF yield, the market is pricing a permanent impairment of the headcount-billed services model. The bull case is that the slowdown is cyclical and AI is an accelerant — more software demand, AI-implementation projects, and Globant's 'AI-native' positioning. The bear case is that AI structurally deflates billable hours faster than new demand fills them. Probability-weighted fair value lands around $45 — roughly 20% above the current price and below the Street's ~$55 consensus, reflecting that the growth stall is real, not just sentiment. The level that turns it compelling is near the $32-35 lows, where the cash-flow yield alone underwrites the wait.
Globant: Cheap Enough to Matter, If AI Doesn't Eat the Hours
A 30%-plus digital-engineering compounder that hit a wall — growth to ~2%, EPS down 39%, the stock off 61% from $97 to $37 — as the market priced both a frozen client budget and the fear that AI commoditizes the whole headcount-billed model. At 0.77x sales and a 16% cash yield, the only question that matters is whether the stall is cyclical or terminal.
The Bottom Line
The simplest version of the Globant thesis: this was a premium digital-engineering compounder — 31,000 engineers building software for the likes of Disney and Santander, growing 20-30% a year — that hit a wall in 2025 when client budgets froze and the market simultaneously decided that AI would gut the headcount-billed services model. The stock fell almost two-thirds. What's left is a still-profitable, still-cash-generative, founder-led business trading at distressed multiples, where the entire investment case reduces to one unresolved question: is the growth stall a cyclical air pocket, or the first crack in a model AI is about to make obsolete?
The two variables that actually matter over the next 12-18 months are: (1) whether revenue growth re-accelerates off the ~2% floor as discretionary spending thaws and AI-implementation work ramps — or stays stalled — and (2) whether Globant's "AI-native" positioning (AI Pods, Globant Enterprise AI, augmented coding) translates into the company winning AI work rather than being disintermediated by it. Everything else — the LatAm talent base, the studio model, client concentration — is secondary to whether AI turns out to be the thing that fills Globant's pipeline or the thing that empties it.
Where We've Been
GLOB spent 2025 in a staircase down — each guidance cut taking another leg out of the stock as a former 30%-grower repriced as a no-growth, AI-threatened services firm. From a high near $97 it slid all the way to a $32.50 low, shedding ~61%, and sits near those lows now, well below its $42 50-day and $55 200-day averages. The fundamentals didn't merely decelerate; the growth engine actually stalled, which is why this drawdown was worse than the rest of the cohort's.
GLOB share price · Jun 2025 → Jun 2026 (approximate monthly)
The catalysts that moved it
| Date | Catalyst | Significance |
|---|---|---|
| 2021–2023 | Revenue compounds from $1.3B to $2.1B (~20-37% growth) | The premium-growth era that earned a 30x+ multiple |
| Through 2025 | Repeated guidance cuts as client discretionary tech spend freezes | The credibility breaker — a growth name that kept lowering the bar |
| 2025 | AI-disruption fear hits all of IT services (Accenture, EPAM, Endava, Globant) | The existential overhang: does AI gut the labor-arbitrage model? |
| FY2025 | Revenue +1.6% to $2.45B; EPS falls ~39% to $2.33 | The wall, in the numbers — growth and earnings both rolled over |
| FY2025 | Free cash flow rises to $261M (~16% yield) despite lower earnings | The cash engine kept running — the foundation of the value case |
| Mid-2026 | Stock near $32-37, 52-week lows; all analyst targets above the price | Deep-value territory with a uniformly more constructive sell-side |
Four Argentines and a Bet on Digitally-Native Talent
Globant was founded in Buenos Aires in 2003 by four entrepreneurs who bet that Latin America's engineering talent — culturally close to the US, in the right time zones, cheaper than Silicon Valley — could build world-class software for global brands. Two decades later it's a 31,000-person firm, still run by co-founder Martin Migoya. The talent arbitrage that built it is exactly what AI now threatens.
Martin Migoya and his co-founders started Globant in the wreckage of Argentina's 2001-02 economic crisis, with a simple insight: there was a deep pool of skilled, lower-cost engineering talent in Latin America, and a growing appetite among US and European companies for "digital" software work — not legacy IT maintenance, but the building of modern consumer-facing products. Globant positioned itself as the "digitally native" alternative to the old offshore outsourcers: nearshore, design-led, and aimed at the innovative end of the work rather than the commodity end.
The organizing idea was the "Studio" model — specialized pods (AI, cloud, gaming, data, cybersecurity, design, and dozens more) that could be assembled into cross-functional teams for a client. It let Globant pitch deep expertise in a vertical or technology while staffing flexibly from a large, fungible talent base. Marquee clients — Disney famously among the largest — anchored the credibility, and the company rode the 2010s wave of enterprise digital transformation from a small Argentine shop to a NYSE-listed firm worth, at its 2021-2025 peak, well over $10B.
Sell talent, billed by the hour, at a margin. Strip away the studios and the design language, and Globant's economics are a services firm's: hire engineers at a cost, bill them to clients at a markup, and grow by adding more billable heads. It's a good model — it scaled to $2.5B in revenue at ~35% gross margins — but it ties revenue to headcount. That linkage was the engine of the growth story for twenty years. It's also precisely the linkage generative AI threatens: if AI lets a smaller team do the same work, the headcount-to-revenue relationship that Globant compounded on could run in reverse. The origin story and the existential risk are, once again, the same story.
Migoya remaining as CEO matters. Founder-led services firms tend to defend culture and talent better through a downturn, and Globant has leaned hard into an "AI-native" identity — reorganizing around AI Pods, building proprietary tools (Globant Enterprise AI, augmented coding and testing platforms), and pitching itself as the firm that delivers AI transformation rather than gets erased by it. Whether that's genuine repositioning or narrative depends entirely on whether it shows up in the revenue line.
How the Money Actually Works
A services model: bill engineers to clients at a markup, run it across industries through the studio structure, and convert the margin into surprisingly strong free cash flow. The growth line is where the story broke. The cash-flow line is why the stock isn't a zero.
Globant earns revenue by staffing teams of engineers, designers, and specialists onto client projects — digital products, platform modernization, data, cloud, and increasingly AI implementation — billed largely on a time-and-materials or managed-team basis. Gross margin runs around 35%, typical for a high-end digital-engineering firm, with the spread between what it pays talent and what it bills clients driving the profit. The business is geographically diversified in delivery (Latin America, India, and elsewhere) and in clients (North America the largest market, with meaningful Europe and LatAm exposure).
The revenue trajectory tells the whole story: $1.30B → $1.78B → $2.10B → $2.42B → $2.45B across FY2021–FY2025. For four years that's a compounder. The fifth year — +1.6% — is the wall. Growth decelerated from +37% in 2022 to +18%, then +15%, then to essentially flat, as enterprise clients froze discretionary project spend and the AI uncertainty made buyers hesitant to commit to large engagements. EPS followed, falling from $3.82 to $2.33 as operating margin compressed on the deleverage. That deceleration chart — growth collapsing toward zero — is the single most important visual in the Globant story.
Free cash flow actually rose while earnings fell. Despite the ~39% drop in net income, Globant generated $261M of free cash flow in FY2025 — up from $221M in FY2024 — as working capital and capex turned favorable. Against a ~$1.6B market cap, that's a ~16% free-cash-flow yield (and a ~19% yield on a trailing-twelve-month basis). Net debt is modest at under 1x EBITDA, and the company began buying back stock. A services firm that converts profit to cash this efficiently, with a clean balance sheet, doesn't go to zero on a growth stall — it has the financial room to wait out a cyclical trough or restructure toward an AI-leveraged model. The cash flow is what separates "deep value" from "falling knife" here.
The honest caveat: GAAP EPS of $2.33 puts the trailing P/E around 16x, but Globant — like its peers — reports a meaningfully higher adjusted EPS (excluding intangible amortization, stock comp, and acquisition costs), on which the multiple is mid-single-digits. The cleanest, least-disputable valuation anchors are the ones that don't depend on which earnings number you trust: 0.77x sales, ~5.5x EBITDA, and a ~16% FCF yield. Those are distressed multiples. The question the rest of this report works through is whether they're cheap, or cheap for a reason that proves permanent.
Net revenue · FY2021–FY2026E
Year-over-year revenue growth · FY2022–FY2026E
FY2025 at a glance
| Metric | FY2025 | Note |
|---|---|---|
| Revenue | $2,455M | +1.6% YoY |
| Gross margin | ~35% | Services economics |
| Operating income | $172M | 7% margin, down from 9%+ |
| Net income | $103M | EPS $2.33 (GAAP) |
| Free cash flow | $261M | ~16% yield, up YoY |
| Net debt / EBITDA | ~0.8x | Modest leverage |
What you're buying at 0.77x sales
At under one times revenue and a 16% cash yield, the market is pricing Globant as a structurally impaired business — a services firm whose model AI is about to deflate. If that's wrong, and the stall is a cyclical demand air-pocket that thaws, the re-rating from these multiples is large.
The discipline is in the sizing and the entry. This isn't NICE, where growth merely decelerated — Globant's growth went to nearly zero, so the cushion is the cash flow, not the trajectory. You buy this for the asymmetry off a washed-out base, not because the fundamentals are reassuring. They aren't yet.
How Wide Is a Services Moat, Really?
IT services moats are real but shallower than software moats — built on relationships, domain knowledge, talent, and switching friction rather than data gravity or network effects. Globant's are decent. The honest question is whether AI erodes them faster than the company can deepen them.
Embedded client relationships and switching friction
Once Globant's teams are embedded in a client's product organization — knowing the codebase, the systems, the institutional context — replacing them mid-flight is disruptive and risky. Large enterprises don't re-bid strategic engineering relationships casually; there's real friction and accumulated knowledge that makes the incumbent vendor sticky. It's not the multi-year lock-in of mission-critical software, but it's meaningful, and it's why services revenue tends to recur even without contracts that guarantee it.
Talent, scale, and the studio expertise
A 31,000-person talent base with deep benches in specific technologies and verticals is genuinely hard to assemble — recruiting, training, and retaining engineers at scale is the core operational competence of a services firm, and Globant has spent two decades building it across Latin America and beyond. The studio model packages that talent as differentiated expertise (gaming, AI, data, design) rather than generic staff augmentation, which supports higher-value, higher-margin work.
The moat's vulnerability — stated plainly
Here's where intellectual honesty matters: a services moat is only as durable as the demand for the service. If AI tools let clients accomplish the same outcomes with dramatically fewer billed hours — whether Globant's hours or anyone's — then relationships and talent scale matter less, because there's simply less work to be embedded in. Globant's bet is to get ahead of this by becoming the firm that implements AI for clients (selling the picks and shovels of AI transformation), converting the threat into the demand driver. That's a credible strategy, but it's a strategy, not a moat. The moat protects the relationship; it doesn't protect the hours.
The Industry, and the Squeeze From Both Ends
Digital-engineering services is a large, fragmented, cyclical market caught in the middle of an AI transition that could expand it or hollow it out. Globant competes against bigger consultancies above, cheaper offshore giants below, and — newest and least understood — AI tools that compress the work itself.
The competitive set
Above Globant sit the scaled players: Accenture (vastly larger, with the consulting relationships and AI-services push), and the digital-engineering specialists EPAM, Endava, and Thoughtworks — direct peers that have suffered the same de-rating and growth stall, which tells you this is a sector problem, not just a Globant problem. Below sit the offshore heavyweights — TCS, Infosys, Cognizant, Wipro — competing on price and scale for the more commoditized work. Globant has historically differentiated on the design-led, innovation end, but that's exactly the discretionary work that froze first when budgets tightened.
The cyclical layer: discretionary spend froze
Much of what Globant sells is discretionary — new product builds, digital initiatives, transformation projects — the first line items cut when CFOs get cautious. The 2025 stall was, in significant part, a demand cycle: clients delaying projects amid macro uncertainty and, ironically, amid their own indecision about how AI changes their software roadmaps. If that's the dominant factor, it's cyclical and reverses. The hard part is that it's tangled up with the structural AI question, making it genuinely difficult to tell how much of the stall is "later" versus "never."
The structural layer: AI as competitor and tool
The newest competitive pressure isn't a company — it's the AI tooling itself. Code-generation and agentic AI tools let a client's existing developers (or a much smaller vendor team) produce more, faster, threatening to compress the billable hours at the heart of the model. The optimistic read, which Globant and its peers all argue, is that AI simultaneously creates demand — every enterprise now needs to build AI into its products and operations, and most lack the in-house capability, which is precisely what a services firm exists to provide. Whether the demand-creation outruns the hours-compression is the unresolved question the whole sector is trading on, and Globant — small, levered to discretionary work, and already stalled — is the most exposed expression of it.
Does AI Eat the Hours, or Fill the Pipeline?
This is the entire investment case. Globant's collapse and its potential recovery both run through one question: whether generative AI deflates the billable-hours model faster than it creates new AI-transformation demand. Management's answer is to bet the company on the second half of that sentence.
If AI writes the code, you need fewer engineers — and Globant bills engineers. The services model monetizes human hours. As code-generation and agentic tools make developers dramatically more productive, the number of billable hours required to deliver a given project falls. Clients capture that productivity as lower bills or in-house capability; vendors see revenue per project shrink. In the structural version of this thesis, the entire IT-services industry faces years of deflation, and a small, discretionary-exposed, already-stalled firm like Globant is squarely in the path. The ~2% growth in 2025 isn't an air pocket in this reading — it's the first year of a new, lower-growth regime.
The bull case is that AI is the biggest demand catalyst services has ever seen. The argument: every enterprise on earth now has to rebuild its products, operations, and data infrastructure around AI, and the overwhelming majority lack the in-house talent to do it. That is an enormous, multi-year wave of new work — and a services firm with deep engineering benches and an "AI-native" posture is exactly who gets hired to deliver it. Globant has leaned all the way in: reorganizing around AI Pods, building Globant Enterprise AI and proprietary augmented-coding and testing platforms, and pitching itself as the partner that helps clients adopt AI rather than the vendor AI replaces. In this reading, AI compresses the hours per legacy project but explodes the number of projects, and net demand rises.
The truth is almost certainly somewhere in between, and the timing is everything. In the near term, AI is probably a headwind — it's adding to client indecision (why commit to a big build when the tooling is changing monthly?) and beginning to compress some hours, while the offsetting AI-transformation demand is still ramping. Over the medium term, the bull case is plausible: the demand wave is real, and firms that successfully reposition capture it. The risk is that the transition is messy and prolonged, the hours-compression bites before the demand-creation pays off, and a small-cap with stalled growth has to grind through several lean years to reach the other side.
What makes the bet ownable despite all that uncertainty is, again, the price. At 0.77x sales and a 16% free-cash-flow yield, the market has priced the bear case as the base case. You don't need the bull thesis to fully play out to make money — you need the stall to prove cyclical-ish rather than terminal, and the cash flow to keep funding the wait. That's a lower bar than "AI is unambiguously good for services," and it's the actual reason to look at Globant here rather than a conviction that the disruption fear is wrong.
SWOT
Where Globant is durable, where the thesis leaks, where the upside lives, and what would actually break it.
Strengths
- Strong cash conversion: $261M FCF (~16% yield) generated even as earnings fell — the floor under the value case
- Clean balance sheet: Net debt under 1x EBITDA leaves room to wait out a trough or restructure
- Founder-led: Co-founder Martin Migoya still CEO — culture and talent retention through a downturn
- Deep talent base and studio expertise: 31,000 engineers with differentiated vertical and technology benches
- "AI-native" positioning: AI Pods, Globant Enterprise AI, and augmented-coding tools — ahead of many peers on the pivot
Weaknesses
- Growth stalled, not just slowed: +1.6% in FY2025 with EPS down ~39% — the trajectory genuinely broke
- Headcount-billed model: Revenue tied to billable hours — exactly what AI threatens to compress
- Discretionary-spend exposure: The design-led, innovation work it specializes in is the first cut when budgets tighten
- Lost credibility: Repeated guidance cuts in 2025 eroded the management-execution trust a re-rating needs
Opportunities
- AI-transformation demand: Every enterprise needs to build AI in, and most lack the in-house talent — a multi-year wave Globant is built to serve
- Cyclical thaw: If the 2025 freeze was a demand air-pocket, growth re-accelerates off the ~2% floor
- Multiple re-rating: From 0.77x sales toward even 1.2-1.5x on any growth signal is a large move
- Buybacks at distressed prices: A clean balance sheet and 16% FCF yield make repurchases highly accretive here
- Consolidation: Sector weakness could make Globant either an acquirer of talent or an acquisition target
Threats
- Structural AI deflation: If AI compresses billable hours faster than new demand fills them, the model shrinks for years
- Prolonged demand freeze: Client indecision about AI roadmaps keeps discretionary projects on hold
- Client concentration: Large-client dependence means a single major-account reduction hits hard
- Offshore and Big-Four pressure: Squeezed between cheaper scaled offshore players and Accenture's AI-services push
- LatAm macro and currency: Argentina/LatAm exposure adds currency and political variability to delivery costs
Bull · Base · Bear
Twelve-month forward scenarios off a ~$37.50 starting price. The bear is weighted a touch heavier than the rest of this cohort because Globant's growth didn't just decelerate — it stalled — so the downside case is more live. The cash flow cushions, but the trajectory doesn't.
$62
What has to go right: Discretionary spend thaws and AI-transformation work ramps, pushing growth back to low-double-digits. Globant demonstrably wins AI projects, validating the "AI-native" pivot. The market re-rates it from 0.77x sales toward ~1.3-1.5x as a credible growth-services name again. This is roughly the Street's $54-78 target zone — the "it was cyclical, not terminal" outcome.
$45
The most likely path: Growth stabilizes in the low-to-mid single digits — better than the 2025 trough but no return to the old pace. Free cash flow stays strong, buybacks shrink the share count, and the multiple drifts modestly higher off a washed-out base. A cheap cash generator grinding back toward fair value without a full growth recovery.
$26
What breaks it: The AI hours-compression thesis proves real — growth stays flat-to-negative, margins compress further, and the model is repriced as structurally impaired. The demand freeze persists, client concentration bites, and the multiple stays ~0.6x sales on a lower earnings base. The cash flow cushions the fall but can't reverse a shrinking business.
Price scenarios · Jun 2026 → Jun 2027
Time-Horizon Outlook
The near-term is about whether the stock has bottomed and whether guidance stops getting cut. The medium-term is about whether growth re-accelerates off the floor. The long-term is the binary: does Globant ride the AI-transformation wave, or get deflated by AI tooling?
Jun-Sep 2026
The most important near-term signal is simple: does management stop cutting guidance? After a year of downward revisions, a held or modestly-raised outlook would be the first step toward rebuilding credibility. Watch whether the $32-35 lows hold.
- Guidance stability — the credibility test after 2025's cuts
- Whether $32-35 holds as a floor
- Early evidence AI bookings are offsetting legacy-project softness
Sep-Dec 2026
The window where the growth rate either inflects off the ~2% floor or confirms a new, lower regime. This is where the "AI-native" pivot has to start showing up as revenue, not just slideware.
- Revenue growth re-accelerating vs. staying stalled
- AI-transformation work as a visible, growing revenue contributor
- Risk: another guidance cut confirms the structural-impairment read
The proof year
If the thesis works: growth back to high-single or low-double digits on AI-driven demand, margins recovering on operating leverage, the share count lower from buybacks, and the multiple re-rating off its distressed base toward a normal services valuation.
- Growth recovery and margin re-expansion
- AI work scaling into a meaningful share of revenue
- Risk: the recovery doesn't come and it settles into a low-growth value name
The binary
The long-run question is genuinely two-sided. The bull endgame: Globant becomes a primary delivery partner for enterprise AI transformation, with AI multiplying demand faster than it compresses hours, and the stock looks absurdly cheap in hindsight. The bear endgame: AI deflates the billable-hours model industry-wide and Globant shrinks into a smaller, lower-margin firm.
- Whether AI-transformation demand outruns hours-compression for services broadly
- Whether Globant is a winner or an also-ran in the repositioned industry
- Tail upside: a takeout, given the depressed valuation and strong cash flow
Risk Matrix
Ranked by what would actually move the stock 15%+ in either direction.
The Bottom Line, Revisited
Strip away the crash and what's left is a profitable, cash-generative, founder-led services firm trading at distressed multiples because its growth stalled at the exact moment the market decided AI might make its whole model obsolete. Cheap, yes — but cheap for reasons that are real, not imagined.
Globant is the hardest call of the cohort, and the most honest way to say it is this: it is unambiguously cheap, and it is cheap for reasons that might prove correct. The 16% free-cash-flow yield, the 0.77x sales multiple, the clean balance sheet, and the founder still at the helm are all real — and they're the reason this is a deep-value setup rather than a falling knife. But the growth didn't merely slow, it stalled to ~2%, EPS fell 39%, and management spent 2025 cutting guidance. The market isn't wrong that something broke; the debate is only whether what broke is cyclical or structural.
The entire case reduces to the AI question, and that question is genuinely unresolved — possibly a near-term headwind that becomes a medium-term tailwind, possibly a structural deflation of the billable-hours model. What tilts it toward "interesting" is that the price already embeds the pessimistic answer. The probability-weighted fair value lands around $45, ~20% above the current price, with a bull case worth 65% against a bear worth 31% — favorable asymmetry, cushioned by cash flow, but riding on a bet you should size honestly because you might be wrong.
This is a name for investors who can hold unresolved uncertainty and who understand the margin of safety here is the cash flow and the washed-out multiple, not the trajectory. The setup that turns it from "interesting" to "compelling" is the $32-35 lows, where the cash yield alone pays you to wait — and where, if the stall proves cyclical and AI fills the pipeline rather than emptying it, the re-rating is a multiple, not a bounce.