EXLS · ExlService Holdings, Inc. · Stock research

Current Thesis

EXL spent the first half of 2026 being priced as a casualty of agentic AI — a seat-based analytics and business-process operator whose billable headcount was assumed to be the thing AI deletes. The 2026-07-28 Q2 print argued the opposite: revenue $594.8M, +15.6% YoY (+15.9% constant currency), adjusted diluted EPS $0.59 against $0.49 a year earlier, and a raise to full-year revenue of $2.390–2.415B from $2.300–2.330B. Management lifted the organic constant-currency growth guide to 13–14% from 12–13% on the 2026-07-29 call and put data-and-AI-led work at roughly 61% of total revenue growing ~30% YoY. The narrative leg being bought is the conversion of a labour-arbitrage services book into an AI-deployment revenue stream that grows because enterprises are adopting AI, and the market is repricing that from a discount to something nearer a growth multiple. The stock closed 2026-07-28 at $30.53 and has traded in the mid-$34s since; the pipeline's basis is $34.89 on 2026-08-07, still 22.5% under the $45.00 52-week high.

Bullish and bearish views on ExlService Holdings, Inc.

The model's bull view on ExlService Holdings, Inc. (EXLS), in brief: Beat and raise on both lines, 2026-07-28. The bear view: One quarter is one quarter. TD Cowen cut its target to $39 on 2026-07-09, three weeks before the beat, and Jim Cramer on 2026-07-17 told viewers to wait for better prices. The bearish framing was live and recent; a single print reversed the tape, not the debate. The freshest… Both cases follow in full.

Bull Case

• Beat and raise on both lines, 2026-07-28. Q2 revenue $594.8M vs $573.9M consensus; adjusted EPS $0.59 vs $0.55 consensus. FY2026 adjusted EPS guided to $2.25–2.29 against a $2.23 consensus, up from $2.18–2.23.
• The raise is organic, not just acquired. Organic constant-currency growth guidance moved to 13–14% from 12–13% (Q2 call, 2026-07-29). The iMerit deal contributes a separately disclosed $28–32M of 2026 revenue, so the acquired piece is quantified rather than buried in the reported 14–16%.
• Mix has already turned. Data-and-AI-led revenue at ~61% of total, growing ~30% YoY (Q2 2026 investor materials, 2026-07-29). That is the number the disruption thesis has to argue against.
• Margins held through the mix shift. Adjusted operating margin 19.7% and adjusted EBITDA $128.4M (21.6% margin) in Q2 2026 — the AI pivot is not being bought with margin.
• Growth is broad across verticals. Q2 segment revenue: Insurance $197.8M, Healthcare & Life Sciences $158.0M, Banking/Capital Markets & Diversified $133.9M, International Growth Markets $105.1M. No single vertical is carrying the print.
• Sell-side capitulated in the same week. Needham raised its target to $45 on 2026-07-29; Baird upgraded to Outperform with a $45 target on 2026-07-30 — from a desk that had been on the sidelines.

Bear Case

• One quarter is one quarter. TD Cowen cut its target to $39 on 2026-07-09, three weeks before the beat, and Jim Cramer on 2026-07-17 told viewers to wait for better prices. The bearish framing was live and recent; a single print reversed the tape, not the debate.
• The freshest targets stop at the old high. Both Needham and Baird sit at $45 — precisely the 52-week high. The most bullish published work in the last two weeks describes a round trip, and the low end of the street range is $32.00, below the 2026-08-07 close of $34.89.
• Reported growth is flattered. FY2026 reported growth of 14–16% versus organic constant currency of 13–14%; the wedge is acquisition and FX. Q2 itself showed reported 15.6% against 15.9% constant currency, so currency was a modest drag.
• Debt sits entirely in the current column. At 2026-06-30 the balance sheet carried $381.2M of long-term debt in the current portion and $0 non-current, against $126.7M cash and $157.1M short-term investments. Refinancing terms land inside twelve months.
• Pricing risk is structural, not cyclical. If agentic tooling compresses per-outcome pricing faster than volume expands, a 15% grower with 19.7% margins re-rates in the other direction quickly. Nothing in the Q2 disclosure resolves that for 2027.

Setup & Price Structure

Life-cycle: ACCELERATING, dated to 2026-07-28. The evidence is a step-change in attention rather than a maturing trend — the earnings gap from a $30.53 close, two target raises to $45 on 2026-07-29 and 2026-07-30, and a four-week gain of about 25.6% through 2026-08-06. Against that, the three-month return is only +12.3%, which places the entire quarter's advance inside the last four weeks and implies May–June was a drawdown. This is a recovery leg off a de-rated base, still 22.5% below the $45.00 high, so the incremental bid is not yet coming from a mainstream, fully-informed holder base.

Crowding and positioning observables, stated as observables:

• RSI(14) at 72.4 on the 2026-08-07 close of $34.89 — above the conventional overbought threshold.
• The 2026-07-28 earnings gap from $30.53 remains unfilled after roughly two weeks of trade; the shelf built in the mid-$34s is the whole post-print structure.
• Mean published target near $43.5 with a $32.00–$48.00 range (aggregated coverage, early August 2026); the low bound sits under spot.
• Coverage clustered into a three-day window (2026-07-29 to 2026-07-30) — flow-driven, and the kind of cluster that thins out once the note cycle is done.
• The trailing-30-day filing record shown carries no disclosed insider sales and no equity issuance into the move.
• No scheduled company event between now and the Q3 print, which historically lands in late October.

Structurally, $30.53 is the pre-print reference close and the floor of the gap; the mid-gap region around $32 is the first level at which the earnings reaction is meaningfully surrendered. $45.00 is both the 52-week high and the level two desks now target.

Catalyst Calendar (next 30 days)

• ~2026-08-10 (est.) — Q2 2026 Form 10-Q, on the 40-day large-accelerated-filer deadline for the quarter ended 2026-06-30. What it resolves: terms and maturity on the $381.2M current-portion debt, headcount versus revenue (the direct read on whether AI is deflating the seat model), and any Q2 repurchase activity not detailed in the 8-K.
• ~2026-10-27 (est.) — Q3 2026 results, outside the 30-day window and not yet scheduled as of 2026-08-08. First test of the raised 13–14% organic constant-currency guide.

Elapsed catalysts

• No scheduled earnings, product or regulatory event inside the window. The 2026-07-28 catalyst has come and gone; the next binary is the Q3 print. (passed 12d ago)

What Would Change Our Mind

The structure that matters is the 2026-07-28 earnings gap. Losing the lower half of it means the market has decided the beat-and-raise did not change the terminal question about seat-based pricing, and the two $45 targets published on 2026-07-29 and 2026-07-30 become stranded. Concretely: a daily close below $32 puts price inside the bottom of that gap and under the lowest published street target of $32.00. A close below the $30.53 pre-print reference would erase the reaction outright.

On fundamentals, the thesis breaks if the Q3 report (est. late October 2026) walks the organic constant-currency guide back below 13%, or if the reported-versus-organic wedge widens — that would say the raise was iMerit and currency rather than demand. A 10-Q showing headcount growing faster than revenue would cut against the central claim that AI is an accelerant here.

The narrative also fails without a price break if the window simply empties: eleven weeks with no scheduled event, an RSI of 72.4 and a coverage cluster already spent is a configuration where attention decays. If the theme rolls to a late-cycle, thinly-bid state before the Q3 print — coverage stops refreshing, the mid-$34s shelf erodes on declining range — the accelerating label is wrong regardless of where the fundamentals land.

Correlation Notes

• Inverse to the AI-infrastructure complex on disruption headlines. EXLS is one of the names that sells off when the story is "AI replaces services labour," even on sessions when compute and datacenter names are bid. The 2026 drawdown from $45.00 and the 2026-07-28 recovery both trace that framing rather than an index move.
• IT-services and analytics peer group. Moves with the enterprise-services cohort on budget-cycle datapoints; a peer guiding discretionary spend lower resets the multiple here before EXL reports anything.
• Insurance vertical beta. Insurance was the largest segment at $197.8M in Q2 2026, so P&C and life carrier technology budgets are a direct input.
• USD/INR and offshore delivery policy. Reported 15.6% versus 15.9% constant currency in Q2 2026 quantifies the FX channel; immigration and offshore-labour policy feed the cost side of the same delivery footprint.
• Rate sensitivity via the balance sheet. With $381.2M of debt in the current column at 2026-06-30, the refinancing cost tracks front-end rates more directly than it would for a term-loan-laddered peer.