TASK · TaskUs, Inc. · Stock research

Current Thesis

The leg on offer is a mix shift, not a growth story at the consolidated line. On 2026-08-05 TaskUs reported Q2 CY2026 service revenue of $308.855M against a $297.51M consensus and adjusted EPS of $0.33 versus $0.29 — revenue landed $10.9M above the high end of the company's own guide. Inside that, AI Services did $66.1M, +25.8% YoY, the seventh consecutive quarter above 25% growth, and management guided that line to more than 30% YoY in Q4, driven by autonomous-vehicle, autonomous-delivery and robotics customers. AI Services is 20% of revenue. Trust & Safety is 25% and is the segment the market assumes large platforms are automating away. Digital Customer Experience is the remaining 55% and grew 6.4%.

So the buyer is underwriting one thing: that a ~20%-of-revenue line compounding at 25–30% out-grows the decay in the other 80% fast enough to re-rate a business the sell side currently prices between $6 and $13. The equity is a busted buyout — stockholders declined to approve the $16.50/share Blackstone-and-founders take-private on 2025-10-08 and the agreement was mutually terminated the next day with no break fee — and it has since paid out a $3.65/share special dividend and levered up to do it.

Bullish and bearish views on TaskUs, Inc.

The model's bull view on TaskUs, Inc. (TASK), in brief: AI Services $66.1M, +25.8% YoY in Q2 CY2026 (reported 2026-08-05) — seventh straight quarter above 25%. The bear view: Q3 is guided below consensus. $300M–$302M against a $306.413M estimate — roughly +0.8% YoY at the midpoint. The raise was to the full year; the next quarter was marked down. Margin is going the wrong way. Adjusted EBITDA was $57.7M at an 18.7% margin, down from 21.0% a year… Both cases follow in full.

Bull Case

• AI Services $66.1M, +25.8% YoY in Q2 CY2026 (reported 2026-08-05) — seventh straight quarter above 25%. Growth is described as led by mobility, logistics and travel: autonomous vehicles, autonomous delivery, robotics. That is physical-AI data operations, a different demand pool than text-annotation contracts.
• Q4 AI Services guided above 30% YoY — an explicit acceleration commitment made on the 2026-08-05 call, with Q3 expected near the Q2 rate. It is a dated, falsifiable claim management put in writing.
• Beat and raise. FY2026 revenue guidance moved to $1.220B–$1.240B from $1.210B–$1.240B against a $1.234B consensus; adjusted free-cash-flow guidance midpoint also rose $5M, to $110M–$120M. Adjusted FCF was $36.4M in the quarter, $140M on a trailing-twelve-month basis at 57.9% conversion.
• Client base broadening. Accounts contributing more than $10M annually rose from 17 to 21. Excluding the largest customer (~20% of revenue), YoY growth was roughly 15% versus 5.0% reported.
• Cash and price. Cash stood at $180.3M at quarter-end. The 16-analyst median target sits at $9.50 with a $6–$13 range (stockanalysis.com forecast page, retrieved 2026-08-08); Wedbush kept an Outperform on 2026-08-06 while cutting its target to $9.

Bear Case

• Q3 is guided below consensus. $300M–$302M against a $306.413M estimate — roughly +0.8% YoY at the midpoint. The raise was to the full year; the next quarter was marked down.
• Margin is going the wrong way. Adjusted EBITDA was $57.7M at an 18.7% margin, down from 21.0% a year earlier. Cost of services rose 11.7%, more than double the 5.0% revenue growth. FY adjusted EBITDA margin is guided to about 19.0%.
• The balance-sheet cushion was spent. The $3.65/share special dividend (~$333M aggregate), declared 2026-02-25 and paid 2026-03-25, was funded with a $500M term loan and a $100M revolver. Net debt/adjusted EBITDA went from 0.1x to 1.3x.
• Concentration. One customer is approximately 20% of revenue. Trust & Safety, 25% of the business, faces automation at the same large platforms that buy it.
• The buyout floor is gone. $16.50/share was refused by stockholders on 2025-10-08 partly on the argument that Street targets then averaged $18.50. The median target today is $9.50.

Setup & Price Structure

On the pipeline's split- and dividend-adjusted series, the last completed daily close was $7.30 on 2026-08-07, 38.5% below the adjusted 52-week high of $11.87, with a 3-month return of +13.3% and RSI(14) at 64.5. Because the $3.65 special dividend already went ex, the adjusted drawdown is a total-return figure — unadjusted quote services show a 52-week high of $18.39 (Investing.com, 2026-08-05), and the gap between the two series is the dividend. Levels here use the adjusted basis.

The print itself was messy intraday: shares fell 3.36% in the 2026-08-05 regular session and traded up 5.06% to $6.64 after hours on the release (Investing.com). The adjusted close two sessions later at $7.30 sits above that after-hours reference, so the post-earnings advance is the structure currently at risk. RSI at 64.5 says the move is extended but not yet at the 70 threshold.

Life-cycle: MATURING. The AI-data-services narrative is not new attention — Q2 CY2026 was the seventh consecutive quarter of >25% AI Services growth (2026-08-05), meaning the segment story has been running since 2024. It is still working at the segment line while flow into the equity moderates: Wedbush cut its target to $9 on 2026-08-06 while keeping Outperform, and the 16-analyst median of $9.50 spans $6 to $13. "Industrials Stocks Moving" and a "Whale Alerts" options-activity list — rather than fresh thematic coverage. That is well-known-and-still-working, with the incremental bid thinning, not a narrative in first discovery.

Crowding / positioning observables (stated, not judged): RSI(14) 64.5 with a +13.3% three-month return into a print; TASK named in a Benzinga unusual-options-activity screen dated 2026-08-06, the session after earnings; sell-side targets moving down (Wedbush $9, 2026-08-06) while ratings stay constructive; a $6-to-$13 target spread across 16 analysts, which is unusually wide dispersion for a $1.2B-revenue services business; and no Form 4 or other insider filing in the filings feed covering the trailing window here. The next scheduled company event is roughly three months out, so there is no imminent earnings date compressing positioning.

Catalyst Calendar (next 30 days)

• 2026-08-14 — Q2 2026 13F deadline. The only dated observable inside 30 days. Reveals whether the event-driven holders who voted down the $16.50 take-private on 2025-10-08 are still on the register or have exited into the post-termination drawdown.
• ~2026-11-04 (est.) — Q3 CY2026 results. The first test of the Q3 guide ($300M–$302M) and of whether AI Services held near the Q2 growth rate as management indicated. Date is an estimate from the company's recent early-November reporting pattern; not confirmed.

There is no confirmed company-scheduled catalyst inside the next 30 days.

Elapsed catalysts

• ~2026-02 (est.) — Q4 CY2026 results. Where the "AI Services above 30% YoY" commitment made on 2026-08-05 is graded. (passed 4d ago)

What Would Change Our Mind

The structure that matters is the post-earnings advance built between the 2026-08-05 release and the 2026-08-07 close. Give that back and the market has decided the beat-and-raise did not change the pricing of a business guided to +0.8% YoY revenue in Q3: a weekly close below $6.60 on the adjusted series surrenders the move off the $6.64 after-hours print entirely.

Separately, the mix-shift leg breaks on its own terms if the Q3 report (~2026-11-04, est.) shows AI Services growing below roughly 25% or management walking back the >30% Q4 commitment — that is the single number the whole re-rating case rests on. A third condition: FY adjusted EBITDA margin guided under the current ~19.0% while cost of services keeps compounding faster than revenue would mean the AI work is being won on price rather than on scarcity.

On the other side, the case strengthens if the largest customer's ~20% revenue share falls while consolidated growth holds, or if the count of $10M+ accounts extends beyond the 21 reported for Q2.

Correlation Notes

• BPO / customer-experience peers. TASK trades in the same de-rating cohort as the listed outsourcers — Concentrix, TTEC, Genpact, Teleperformance — where the shared bear thesis is that conversational AI compresses seat-based revenue. Read-across is high on any peer guidance cut, and TASK's Trust & Safety and DCX segments (80% of revenue combined) carry that beta regardless of what AI Services does.
• Private AI-data comparables. The valuation anchor for the AI Services line sits in private markets (Scale AI, Surge, Mercor). There is no listed pure-play proxy, so the segment gets no independent multiple until it is large enough to disclose separately in guidance.
• Physical-AI capex. Q2 AI Services growth was attributed to autonomous vehicle, autonomous delivery and robotics customers. That ties the accelerating line to AV/robotics program funding rather than to LLM training budgets — a different and less-correlated cycle than the hyperscaler capex complex.
• Rates. The $500M term loan funding the March 2026 special dividend puts 1.3x net leverage on a floating-rate structure, adding a rate-sensitivity channel that did not exist when net leverage was 0.1x.
• Single-client platform spend. With one account at ~20% of revenue, moves in that platform's content-moderation and outsourcing budget transmit directly; the Q2 disclosure that ex-largest-client growth was ~15% versus 5.0% reported quantifies the drag.