EXLS · ExlService Holdings, Inc. · Stock research

Current Thesis

The leg being bought remains the unwinding of an AI-eats-BPO de-rating: EXL repriced from a seat-based services book that agentic AI deletes into a vendor that grows because enterprises are deploying AI. Two things have changed since 2026-08-12. First, the financing overhang cleared — on 2026-08-18 the company closed a five-year, up-to-$1.0B senior secured facility with PNC Bank as administrative agent ($400M term loan plus revolver capacity to $600M, maturing 2031-08-18, accordion equal to the greater of $470M or trailing-four-quarter EBITDA), simultaneously terminating the Citibank 2024 Credit Agreement under which $381.2M of borrowings sat largely in the current column at 2026-06-30. Second, price left the shelf it had held for two weeks: $34.89 on 2026-08-07 and $34.90 on 2026-08-11 gave way to a $37.53 close on 2026-08-21, with RSI(14) at 71.1, the shares up 27.2% over three months and 16.6% below the $45.00 52-week high. The fundamental frame is unchanged from the 2026-07-28 print — data-and-AI-led revenue $362.6M against $278.8M a year earlier, ~61% of the $594.8M total; digital operations down to $232.2M from $235.7M; GAAP net income $64.5M versus $66.1M. What is different is that both named near-term catalysts are now resolved or distant: the refinancing happened, and nothing else is scheduled before a Q3 report expected late October.

Bullish and bearish views on ExlService Holdings, Inc.

The model's bull view on ExlService Holdings, Inc. (EXLS), in brief: The refinancing landed larger and longer, 2026-08-18. The bear view: GAAP earnings went the wrong way. Net income $64.5M vs $66.1M in Q2 2025 on 15.6% revenue growth; reported diluted EPS $0.42 vs $0.40 came from retiring 5.8M shares in H1, so the per-share line improved while the profit pool did not. The legacy book is contracting. Digital… Both cases follow in full.

Bull Case

• The refinancing landed larger and longer, 2026-08-18. Capacity rises from a $600M limit to $1.0B, the maturity moves to 2031-08-18, and the company described greater covenant flexibility; Bank of America, JPMorgan Chase and TD Bank were joint lead arrangers. The $381.2M of borrowings at 2026-06-30 — a $290.0M revolver and a $91.2M term loan — no longer face a near-dated agreement.
• The AI line is the company. Q2 2026 data-and-AI-led revenue $362.6M vs $278.8M (Q2 2026 Form 10-Q, filed 2026-07-28); H1 2026 $704.2M. Roughly 30% growth on ~61% of the mix is the number the disruption argument has to beat.
• Beat and raise on 2026-07-28. Revenue $594.8M vs $573.9M consensus; adjusted EPS $0.59 vs $0.55. FY2026 revenue guided to $2.390–2.415B from $2.300–2.330B, adjusted EPS to $2.25–2.29 from $2.18–2.23 against a $2.23 consensus.
• The raise is mostly organic. Organic constant-currency growth guidance moved to 13–14% from 12–13% on the 2026-07-29 call, with the iMerit contribution disclosed separately at $28–32M of 2026 revenue.
• Margin held through the mix shift. Q2 adjusted operating margin 19.7%; adjusted EBITDA $128.4M at 21.6%.
• executed while the de-rating was still in force.
• Published targets sit above the last close. Nine tracked targets, consensus $44.38, range $39–$48: J.P. Morgan $48 (2026-07-30), Baird upgrade to Outperform at $45 from $35 (2026-07-30), Needham $45 (2026-07-29), Barrington $43 (2026-07-30), Jefferies $39 reiterated (2026-07-31).

Bear Case

• GAAP earnings went the wrong way. Net income $64.5M vs $66.1M in Q2 2025 on 15.6% revenue growth; reported diluted EPS $0.42 vs $0.40 came from retiring 5.8M shares in H1, so the per-share line improved while the profit pool did not.
• The legacy book is contracting. Digital operations revenue $232.2M vs $235.7M. If pricing deflation is real, that is the line where it appears first.
• A bigger facility is an M&A mandate, not cash. The 2026-08-18 agreement raises borrowing headroom by $400M plus an accordion of at least $470M. EXL agreed to buy iMerit Technology on 2026-06-22 for a reported $310M ($170M cash, $140M contingent) — that deal already supplies the wedge between the 14–16% reported and 13–14% organic constant-currency FY2026 guides. Further debt-funded deals widen it.
• The delivery-labour check is unavailable. The Q2 2026 10-Q discloses no headcount, so revenue per employee — the direct read on whether AI is deflating the cost base faster than the price base — cannot be computed until the annual report.
• Coverage has nothing left to convert. Five target actions landed between 2026-07-29 and 2026-07-31 and none since; as of 2026-08-21 all nine tracked ratings are Buy or Strong Buy and the consensus $44.38 has not moved while price rose to $37.53.
• Concentration. Insurance was the largest vertical at $197.8M of Q2 revenue, tying group growth to P&C and life carrier technology budgets.

Setup & Price Structure

• The gap. $30.53 close on 2026-07-28 pre-print; shares jumped 17.9% on 2026-07-29 after the beat and raise. That gap has not been tested since.
• The shelf. Roughly two weeks of range between the 2026-08-07 close of $34.89 and the 2026-08-11 close of $34.90 — the reference support the current leg lifted off.
• The breakout. The 2026-08-18 facility headline preceded a $37.53 close on 2026-08-21, a post-print high, with RSI(14) 71.1 and price 16.6% under the $45.00 52-week high.
• Crowding observables, stated plainly: RSI(14) at 71.1 with the shares up 27.2% over three months; no published target change in the three weeks from 2026-07-31 to 2026-08-21 while price advanced, so the move is running ahead of a static $44.38 consensus rather than being pulled by fresh estimate revisions; unanimous Buy/Strong Buy ratings leave no bearish coverage to flip; Jim Cramer flagged EXLS on 2026-07-17 and told viewers to wait for better prices, which is a retail-attention marker that predates the print. No earnings date falls inside the next 30 days.
• The narrative is maturing. The accelerating phase was 2026-07-28 to 2026-07-31 — the print, the 17.9% gap and five target actions in three sessions. Since then the narrative has kept working (the 2026-08-18 refinancing extended it) on visibly thinner sell-side flow: zero new targets in three weeks, coverage already saturated on the bull side. Well known, still working, moderating flow.

Catalyst Calendar (next 30 days)

• 2026-08-23 → 2026-09-22: no scheduled company event. The refinancing that anchored the September window closed early, on 2026-08-18. There is no earnings date, no guidance update and no scheduled investor day inside the 30-day window; anything that moves the name in that period is unscheduled — a deal announcement drawn on the new facility, a rating action, or sector read-through from IT-services peers reporting in the interim.
• ~2026-10-27 (est.): Q3 2026 results. Third-party earnings calendars point to 2026-10-27; a company scheduling release is the confirmation and had not been verified as of 2026-08-21. First test of the 13–14% organic constant-currency guide, of whether data-and-AI holds above ~61% of revenue, and of whether digital operations declines faster than the -1.5% YoY recorded in Q2.
• ~2026-10-27 (est.): Q3 2026 Form 10-Q. EXL filed its Q2 10-Q the same day as the print. The Q3 balance sheet is where the $381.2M stops being current-classified and the new term loan's rate and covenant terms become checkable line items.

What Would Change Our Mind

The structure at risk is the August base. Losing them would say the credit-agreement leg was a headline pop rather than a re-rating, and would put the untested 2026-07-29 gap toward $30.53 back in play. A daily close below $34 is the gradeable break.

On fundamentals, three observables would flip the read independent of price. A Q3 print (est. 2026-10-27) that guides organic constant-currency growth below 13% would say the July raise was one quarter, not a trend. Digital operations declining materially faster than the -1.5% YoY of Q2 while data-and-AI growth slows from ~30% would say the mix shift is substitution rather than expansion. And a debt-funded acquisition announcement large enough to widen the reported-versus-organic spread beyond the disclosed $28–32M iMerit contribution would say the new $1.0B facility is being used to buy the growth rate.

On positioning, if the 2026-10-27 print arrives with the $44.38 consensus still unchanged since 2026-07-31 and price back in the mid-$34s, the upgrade cycle is spent and the narrative has moved from maturing toward saturated.

Correlation Notes

• AI-disruption-of-services complex. EXLS trades with the question of whether agentic AI compresses or expands enterprise services spend; its own datapoints cut both ways (data-and-AI $362.6M growing ~30%, digital operations $232.2M shrinking), so it tends to move with peer prints and with hyperscaler commentary on enterprise AI deployment rather than with software multiples.
• USD/INR and offshore-labour policy. Delivery is India-weighted, so currency and visa/wage policy feed directly into reported growth and margin; the company reports constant-currency growth separately for that reason (Q2 revenue +15.6% reported, +15.9% constant currency).
• Insurance carrier technology budgets. Insurance at $197.8M was the largest Q2 vertical, so October results from large P&C and life carriers are a read-through into the Q3 print.
• Rates. The 2026-08-18 facility is a floating-rate bank instrument; with capacity at $1.0B plus an accordion, the cost of incremental drawings tracks short-rate policy rather than the equity narrative.