This is investment research, not personal financial advice.

Globant (NYSE:GLOB) fell from a $40.99 regular close to a $35.50 after-hours indication, a 13.39% drop, after its 13 August results showed Q2 revenue almost unchanged at $614.4 million and cut the full-year outlook. The share-price move was company-specific. Accenture, Cognizant and EPAM moved by roughly one percentage point or less in the same extended session, while Globant's customer count, headcount and adjusted margin all contracted (Yahoo Finance 2026; Globant 2026a).

The market did not merely mark down a quarterly miss. It repriced the transition from selling engineering time to selling AI-assisted outcomes. Globant says its new AI Pods can generate about ten percentage points more gross margin than traditional time-and-material work. Yet the reported business still lost 77 customers above $100,000 of annual revenue, removed 2,673 employees year on year and produced a lower gross margin. The 13.39% reaction looks roughly proportionate: the post-release price no longer assumes a quick return to the old growth rate, but it still requires AI to restore moderate growth and a cleaner owner-cash margin.

Seventy-seven missing names changed the meaning of flat revenue

Q2 revenue rose by only $0.2 million from the prior-year quarter. Read alone, that looks like stagnation. The operating ledger is more revealing. Customers generating more than $100,000 over the latest twelve months fell from 981 to 904. Accounts above $1 million slipped from 339 to 331. Total employees fell from 30,084 to 27,411, and technology, design and innovation professionals fell from 28,097 to 25,632 (Globant 2026a).

A smaller workforce can be evidence of productive automation. It can also be the residue of weak demand. Globant's margins point toward the second interpretation, at least for now. Gross margin declined from 35.4% to 33.9%, while adjusted operating margin declined from 15.0% to 13.2%. Revenue per quarter-end technology professional rose to about $23,971, but that rough ratio uses a point-in-time workforce against a quarterly revenue flow and ignores utilization, subcontractors, currency and project mix. More revenue per employee did not become more gross profit per dollar of revenue.

The guidance reset made the same point. Management now expects 2026 revenue of $2.428 billion to $2.462 billion, compared with its previous $2.462 billion to $2.508 billion range. Adjusted operating margin guidance moved down by half a percentage point to 13.5% to 14.5%, and adjusted diluted EPS guidance fell from $6.10-$6.50 to $5.75-$6.15 (Globant 2026a; Globant 2026c).

The reported customer loss needs one qualification. Globant changed its broad active-customer threshold from more than $10,000 of trailing revenue in its older annual reports to more than $100,000 in current materials, so the 1,731 customers disclosed for 2024 cannot be placed beside the 944 reported for 2025 as though every departure were real. The Q2 comparison is clean because both periods use the same $100,000 threshold. Those 77 missing names are not a definition change.

Globant still sells hours while promising outcomes

Globant is a single-segment technology-services company. It combines software engineering, design, consulting, cloud work, data modernization and digital marketing for large enterprises. The economic unit has historically been a person-hour. Time-and-material contracts represented 78.9% of revenue in 2023, 71.0% in 2024 and 66.7% in 2025. Fixed-price work rose from 18.3% to 28.0% over the same interval; licenses, resales and other revenue rose from 2.8% to 5.3% (Globant 2025).

That mix shift matters because an hourly contract returns much of an AI productivity gain to the customer. If ten engineers complete work that once required fifteen, fewer hours get invoiced unless prices rise. Globant's annual filing says rates cannot always be increased automatically to match wage inflation, never mind a sharp reduction in labor content. A fixed-price or outcome-priced contract changes the split: the customer pays for a result, and Globant can retain some savings if delivery works.

AI Pods are the proposed bridge. They are monthly, token-metered packages marketed around output or consumption instead of staffing. Glob.AI annual recurring revenue reached $52.8 million in June, up about 60% from March, with an exit-2026 objective of at least $110 million. Management reported a $436.8 million AI Pod opportunity pipeline, adoption by 45% of its top twenty accounts and gross margins about ten percentage points above traditional time-and-material delivery (Globant 2026b).

Those figures describe promise, not yet a reported business line. The company has not disclosed recognized AI Pod revenue, renewal rates, customer count, contribution margin, inference costs or the share retained by model and cloud partners. ARR is an annualized run rate. At $52.8 million, it equals about 2.2% of 2025 revenue. Even the $110 million exit objective would equal less than 5% of the existing revenue base. The new engine can grow quickly and still be too small to offset weakness in the old one.

The large accounts are resilient; the funnel is not

The strongest evidence for Globant's competitive position sits in established relationships. Existing 2024 clients generated 96% of 2025 revenue. The top-ten relationships had lasted an average of eleven years, and management said Q2 revenue from the top fifty clients grew 6.9%. North American and European enterprises face genuine work in data preparation, security, governance and system integration before generative models can run inside production workflows. That work favors an incumbent that knows the client's architecture and business rules (Globant 2025; Globant 2026a).

The counter-evidence is a narrowing customer base. The top client represented 8.9% of Q2 revenue, the top five 21.6%, and the top ten 30.6%. In 2025, those figures were 8.7%, 20.1% and 29.2%. Concentration is increasing while smaller accounts leave. That can improve near-term utilization if Globant focuses on bigger relationships, but it weakens the new-logo and mid-market funnel that used to replenish growth.

The sector comparison is harder on management's explanation than the macro backdrop. July employment in US computer-systems design and related services was 2.366 million, down 1.2% from a year earlier and 3.3% from January 2024. Labor demand is soft (FRED 2026). But Accenture grew Q3 local-currency revenue 3%, expanded operating margin to 17.0% and described more large AI transformations. EPAM grew Q2 organic constant-currency revenue 3.4% and expanded non-GAAP margin to 16.4%. Cognizant grew constant-currency revenue 4.1% and raised adjusted operating margin to 16.0% (Accenture 2026; EPAM 2026; Cognizant 2026).

AI is not compressing every services provider at the same speed. The peer evidence supports an anti-thesis too: enterprise AI can enlarge integration demand even as it reduces coding labor. Globant's nearshore engineering base, industry studios and long relationships could make the company a beneficiary. The present numbers say the transition is also company-specific. Peers are growing and protecting margins; Globant is losing customers and margin together.

Four years of capital growth produced less operating profit

The financial history shows why an earnings multiple alone is unreliable. Globant increased revenue by 38% from 2022 to 2025, but 2025 operating profit fell below the 2022 level. Net income attributable to owners fell to $102.9 million. Cash generation was better than accounting profit, yet acquisitions, stock compensation and repeated adjustments consumed a large share of it (Globant 2022; Globant 2023; Globant 2024; Globant 2025).

USD millions except ratios 2022 2023 2024 2025 1H 2026
Revenue 1,780.2 2,095.9 2,415.7 2,454.9 1,221.5
Net income to owners 148.9 158.5 165.7 102.9 38.7
Operating cash flow 197.5 318.5 248.7 301.2 84.7
Cash capex 95.4 126.8 113.6 89.5 36.0
Computed ROIC 13.6% 10.3% 9.0% 5.5% 4.5% annualized
Computed net debt / (cash) (337.2) (164.2) 136.5 116.4 253.1

Revenue, profit, cash flow and capex are source-reported. ROIC and net debt are author calculations. Capex combines cash purchases of property and equipment with intangible assets. The 1H cash-flow values are reconstructed from Q1 and Q2 releases because the Q2 release gives selected quarterly cash flows rather than a complete IAS 34 cash-flow statement (Globant 2026a; Globant 2026c).

The ROIC calculation keeps acquisition goodwill in the capital base. NOPAT is IFRS operating profit multiplied by one minus the effective tax rate. Invested capital is equity plus borrowings and lease liabilities, less cash and current investments, averaged across each year. NOPAT was $160.2 million in 2022 on $1.174 billion of average invested capital, a 13.6% return. In 2025 it was $128.3 million on $2.336 billion, or 5.5%. Annualizing first-half 2026 NOPAT of $55.1 million against $2.454 billion of average capital gives 4.5%.

Incremental ROIC is worse but noisy. From 2023 through 2025, year-end invested capital increased by about $1.031 billion while NOPAT fell by about $31 million. Acquisition timing, currency translation and restructuring distort any single-year ratio, so it would be false precision to call that a project return. The direction is still clear. Globant deployed more capital and earned less after-tax operating profit.

The adjustment bridge is now part of the business

At $35.50, Globant trades at about 6.0 times the midpoint of management's adjusted EPS guidance. That number is optically low because the adjustments are large and recurrent. In 2025 the company recorded $75.7 million of equity-settled stock compensation, $71.8 million of acquisition-related charges and $52.0 million of business-optimization costs. First-half 2026 added $36.9 million, $25.5 million and $32.3 million, respectively (Globant 2025; Globant 2026a).

A conventional cash bridge starts with operating cash flow and subtracts property and intangible capex. That produced $211.7 million in 2025. Subtracting stock compensation leaves $134.0 million. A deliberately strict bridge also subtracts all acquisition and optimization charges, leaving about $10.2 million. This strict figure over-penalizes cash because some acquisition charges are amortization, remeasurement or financing items already treated elsewhere. Ignoring them is equally misleading because charges tied to acquisitions have appeared every year, and optimization costs now span 2025 and 2026.

The latest-twelve-month bridge through June gives a usable range. Reconstructed operating cash flow was about $348.3 million and capex $79.3 million. After $73.7 million of stock compensation, the upper owner-cash measure is about $195.2 million. After acquisition and optimization leakage, the strict measure is about $99.6 million. The investable economic question is not which endpoint is correct to the dollar. It is whether the recurring leakage converges toward zero or remains a normal cost of running an acquisition-built company.

Balance-sheet risk is secondary. At June, cash and current investments totaled $168.8 million against $422.0 million of borrowings, producing computed net debt of $253.1 million before leases. Working capital was $417.4 million and the current ratio 1.84 times. Funded debt matures in May 2028; year-end 2025 filings reported compliance with covenants requiring at least three times interest coverage and no more than 3.5 times net leverage (Globant 2025; Globant 2026a).

Goodwill and acquired intangibles are the harder balance-sheet issue. Together they reached $1.908 billion at June, 59% of assets and 88% of equity. The 2025 auditor identified goodwill impairment as a critical audit matter. A revenue or margin reset can reduce valuation headroom long before it creates a cash crisis.

Repurchases cannot settle the acquisition record

Globant spent heavily to widen its geographic and capability base: about $272 million of total consideration on 2022 acquisitions, $451 million in 2023 and $529 million in 2024. Pentalog, GUT, Blankfactor, Exusia and several specialist consultancies brought customers, skills and new markets. Yet the acquisition-inclusive return on capital fell each year. Revenue growth slowed to 1.6% in 2025 while goodwill kept rising (Globant 2022; Globant 2023; Globant 2024; Globant 2025).

Capital allocation has since moved toward repurchases. The company spent $50 million in Q1 2026 and $25 million in Q2, completing its initial program, then authorized another $125 million. The share count has declined from the 2024 diluted average, but repurchasing shares while client count and cash conversion weaken can only create per-share value if the operating reset is temporary. It does not repair the returns on capital already committed to acquisitions.

Founder involvement offers continuity without dominant economic ownership. Chief executive Martín Migoya beneficially owned about 337,000 shares, and directors plus senior management owned 2.76% as of February 2026. Globant's foreign-private-issuer status also means executive compensation disclosure is less granular than a US domestic proxy. The 2025 filing reported aggregate executive cash compensation of $12.5 million, including $4.7 million of severance, and long-term awards largely in restricted shares. The compensation framework includes revenue and profitability, but public information does not let an outside reader reconstruct individual payout hurdles (Globant 2025; Globant 2026d).

A 6-times adjusted multiple still prices a recovery in owner cash

A services company with unstable adjustments is better valued on owner cash than adjusted EPS. The primary valuation here is a ten-year owner-cash DCF, cross-checked against guided adjusted earnings and enterprise value. It starts with the $2.445 billion midpoint of 2026 revenue guidance, subtracts $253.1 million of net debt and $60.4 million of non-controlling interest, and uses 43.228 million diluted shares for the per-share DCF. The market snapshot uses 43.180 million issued shares, hence the small difference in share bases.

Case Five-year revenue path Owner-cash margin path WACC / terminal growth Value per share
Severe downside (3%) a year 2.0% to 3.5% 14%-16% / 1%-2% $1.40-$3.40
Bear 1% a year 3.5% to 5.0% 12%-14% / 1.5%-2.5% $12-$18
Base 4% a year 4.5% to 7.0% 10.5%-12.5% / 2%-3% $30-$45
Bull 8% a year 5.5% to 9.0% 9.5%-11.5% / 2.5%-3.5% $62-$97

The base midpoint uses an 11.5% discount rate and 2.5% terminal growth. Owner cash rises from $114 million in 2027 to $243 million in 2036 as the margin reaches 7%. Discounted explicit cash flow is about $942 million and discounted terminal value $930 million. After net debt and minority interest, equity value is about $1.559 billion, or $36.06 per diluted share.

That result is close to the $35.50 after-hours indication. Reverse valuation reaches the same conclusion. With 4% revenue growth for five years, a fade to 2.5% and an 11.5% discount rate, the price requires a mature owner-cash margin of about 7.1%. Alternatively, fixing mature margin at 7% requires roughly 4.2% five-year revenue growth. The market is not pricing permanent stagnation. It is pricing a moderate recovery from a poor first half.

A sensitivity table shows why the adjusted P/E can deceive. The table fixes the discount rate at 11.5%, terminal growth at 2.5% and first-year owner-cash margin at 4%.

Five-year revenue growth 4.0% mature margin 5.5% 7.0% 8.5% 10.0%
(2%) $13.01 $17.85 $22.69 $27.53 $32.37
1% $16.53 $22.43 $28.32 $34.21 $40.11
4% $20.66 $27.81 $34.95 $42.10 $49.25
7% $25.48 $34.11 $42.74 $51.37 $60.01
10% $31.10 $41.48 $51.86 $62.25 $72.63

Margin recovery matters at least as much as revenue. If acquisition and restructuring leakage keeps owner cash near 4% to 5%, low adjusted earnings multiples provide little analytical protection. If AI Pods move the mix toward fixed outcomes and 7% to 9% owner cash, the valuation changes sharply.

The reaction is proportionate because the proof is still ahead

The 13.39% after-hours drop erased the remaining presumption of a quick rebound, but it did not reduce the valuation to a no-recovery case. The independent base range of $30-$45 contains the post-release indication. The severe and bear ranges require sustained contraction or weak cash conversion; the bull range requires AI Pods to restore both growth and margins. That spread is wide because the company reports little about the new unit's economics.

Three dates can narrow it. Q3 revenue below $607 million would extend contraction; revenue above $615 million would begin to close the peer-growth gap. Adjusted operating margin outside the 13.5%-14.5% range would show whether workforce resizing is weakening or restoring delivery economics. By 31 December, Glob.AI exit ARR either reaches the stated $110 million objective or it does not. FY2026 cash flow in early 2027 will then show whether recurring adjustments have fallen enough for owner cash to approach the 7% margin embedded in the current price.

Confidence and source notes

Confidence is high in the filed financial history, balance sheet and customer-count comparison. Confidence is lower in the AI economics because ARR, renewals, model costs and contribution margin are management measures without segment disclosure. The Q2 release is unaudited and does not contain a complete IAS 34 interim report. The after-hours $35.50 figure is also an indication from thin extended trading, not an official closing auction. The next regular session will provide the first liquid price verdict.

The observable tension is precise. Globant's established client relationships still produce durable revenue, and AI Pods have early momentum. But the client ledger, gross margin and acquisition-inclusive return on capital are moving in the wrong direction. At $35.50, the market now assumes that transition produces roughly 4% medium-term growth and a 7% mature owner-cash margin. Q3 delivery and the December AI ARR milestone will show whether that assumption is becoming evidence.

References