Revenue up 19%. Stock down 50%.

Today, I'm digging into TaskUs (TASK).

TaskUs is a digital outsourcing company. They staff the chat windows and support queues for tech companies, run the trust and safety teams that moderate content on social and marketplace platforms, and handle the data labeling that trains AI models. Most of that work is delivered offshore, heavily out of the Philippines and India, for clients that are mostly in the US.

They are the definition of a labor arbitrage play. But the market sure seems to be voting that this arbitrage is going to get crushed by AI and AI agents.

A few things that stood out to me:

• The same AI wave everyone says will gut customer-support outsourcing is actually TaskUs's fastest-growing line of business. Their AI Services segment (data labeling, annotation, model evaluation) grew about 59% in 2025 and now drives roughly 18% of revenue.

• Revenue reaccelerated to $1.18 billion in 2025, up about 19% after three flat years.

• The founders and Blackstone tried to take this company private at $16.50 a share in May 2025, and public shareholders voted it down in October. The stock now trades around $5.50. Ouch!

• A new $500 million term loan funded a $3.65 per share special dividend (about $333 million) in March 2026, taking a nearly net-debt-free balance sheet to roughly 1.5x EBITDA of net leverage.

What really jumps out is the gap between the business and the stock. On the numbers, 2025 was the best year in a while: revenue grew about 19% and operating margin pushed to 12%. Over that same period, the stock fell about 50% and now sits roughly three-quarters below its IPO price. The market isn’t pricing TaskUs’s current performance, it is pricing what it thinks AI does to this category over the next few years.

This is one of the more interesting companies I’ve reviewed. Not because their business model or moats are compelling, it’s the fact that they could easily continue to grow into more of a AI picks-and-shovels play vs. simply a distruptee.

The key takeaway for me here is that when you have a core competency (for them it’s hiring in their core markets and the systems for managing the work), try like hell to apply them to defensible, growing markets.

The big question here is if AI itself will begin to do the AI services work of data labeling, annotation, and model evaluation? Time will tell.

This will be an interesting one to revisit in a year or two!

With that, I'll see you tomorrow!

Nick

TL;DR

• TaskUs is a digital outsourcing company that runs customer support, content moderation, and AI data-labeling for tech-forward clients, delivered mostly from the Philippines and India.

• It makes money on services contracts, billing for trained headcount and outcomes, with revenue split roughly 56% Digital Customer Experience, 26% Trust and Safety, and 18% AI Services.

• Revenue reaccelerated to $1.18 billion in 2025 (up about 19%), led by a 59% jump in AI Services, and operating margin expanded to about 12%.

• The balance sheet was nearly net-debt-free at year-end, then they raised a $500 million term loan to pay a $3.65 special dividend, pushing net leverage to roughly 1.5x EBITDA.

• Stock-wise it has been brutal: TASK is down about 50% over the past year, and the whole CX-outsourcing group got repriced on AI fears (Concentrix down about 53%, TTEC about 56%), with TASK falling the least of the three.

• Operator takeaway: when a technology threatens your core, find the segment of that same wave you can sell into.

The 30,000-Foot View

TaskUs sells outsourced digital operations. A client (a fintech, a food-delivery app, a social platform, an AI lab) hands off work that is high-volume and people-intensive, and TaskUs builds and runs the team that does it. That work falls into three buckets: Digital Customer Experience (support and care), Trust and Safety (moderating user-generated content), and AI Services (labeling and evaluating data used to train AI models).

The economic engine is labor arbitrage plus process. TaskUs recruits, trains, and manages tens of thousands of frontline agents in lower-cost geographies, then bills clients for that capacity and the outcomes it produces. The real skill is standing up a high-quality team fast, keeping attrition low, and hitting service levels across tens of thousands of people. In a people business, attrition is the enemy of margins, so culture and employee experience are operating levers, not soft topics. What keeps clients around is embedding: once TaskUs is wired into a client's tools and playbooks, ripping them out is disruptive. But that lock-in is behavioral, not contractual armor, and clients can multi-source or pull work in-house.

Revenue mix by service line (FY2025):

• Digital Customer Experience: ~56%

• Trust and Safety: ~26% (derived as the residual vs. disclosed DCX and AI Services segments) [VERIFY against 10-K segment note]

• AI Services: ~18%

Delivery footprint by geography (FY2025):

• Philippines: ~54%

• Rest of World: ~22%

• India: ~13%

• United States: ~11%

Key Stats

• Market cap: ~$0.49B

• FY2025 revenue: $1.18B

• FY2025 gross margin: ~33% (reported), operating margin ~12%

• Employees: ~65,000

• 1Y total return: -50%

• Industry: IT services / digital outsourcing (BPO)

Company History

• 2008: Bryce Maddock and Jaspar Weir found the company, building an outsourcing shop aimed at high-growth tech clients rather than traditional enterprises.

• 2018: Blackstone takes a majority stake.

• 2020: Renamed from TU TopCo to TaskUs, riding a surge in demand for content moderation and support from social and on-demand platforms.

• 2021 (June): IPO on Nasdaq under ticker TASK, with the stock around $20 in its first days on a dividend-adjusted basis.

• 2022 to 2023: Growth flattens as hypergrowth clients pull back and content-moderation work draws scrutiny over worker wellbeing.

• 2024: AI Services becomes a real growth driver as demand for training data and model evaluation takes off.

• 2025 (May): Maddock, Weir, and Blackstone propose taking the company private at $16.50 per share, valuing it around $1.62 billion.

• 2025 (October): Shareholders fail to approve the deal. The parties terminate the transaction with no termination fee, and TaskUs stays public.

• 2026 (February to March): TaskUs refinances with a $500 million term loan and pays a $3.65 special cash dividend (about $333 million), taking net leverage to roughly 1.5x EBITDA.

Show Me the Money

Standout financial features:

• Revenue reaccelerated to $1.18 billion in 2025 (up about 19%) after three years stuck in a narrow band, with AI Services driving it with 59% growth.

• Operating margin expanded to about 12% showing healthy operating leverage.

• It is a nice cash generator for a people business: free cash flow ran roughly $100 million in 2024 and about $74 million in 2025, with low capital intensity.

Financial Data

Stock Performance

Dividend-adjusted total returns through June 12, 2026 (incl $3.65 special dividend in March 2026):

The 5-year figure is essentially the return since IPO, and it is ugly: a day-one holder is down roughly three-quarters even after the big special dividend.

1-year peer comparison (dividend-adjusted total return):

This is a sector story, not just a TaskUs story. All three CX-outsourcing names got cut roughly in half over the past year as the market repriced the group on AI-automation fears, with TaskUs holding up the best of the three. I picked Concentrix as the scaled, diversified leader and TTEC as a closer-in-size pure-play, so the two bracket TaskUs on size and breadth.

The N.O.O.B. Nine: Competitive Powers

The Nerd Out on Business Nine is made up of Hamliton Helmer's famous "7 Powers" of competitive advantage (Scale Economies, Network Economies, Counter-Positioning, Switching Costs, Branding, Cornered Resource, and Process Power) combined with two of my own (Data Flywheel and Distribution Advantage).

Average Score: 2.6/5 - A process- and relationship-driven moat that is real but thin, the kind that rewards relentless execution over structural protection.

Memorable Marketing

TaskUs markets itself less like a back-office vendor and more like a culture-forward tech company. In outsourcing, the buyer's biggest fear is sloppy, high-attrition delivery that embarrasses their brand, so TaskUs sells the opposite: a "Ridiculously Good" service promise built on an employee experience that keeps the best frontline people in their seats.

Notable tactics:

• "Ridiculously Good" positioning: A simple, repeatable promise that reframes a commodity (outsourced support) as a quality and care story.

• Employer brand as a sales tool: Teammate wellbeing and culture is a client-facing pitch, because lower attrition means better, more consistent outcomes for the buyer.

• AI-forward repositioning (recent): Leaning into AI Services and internal tooling to reframe the story from "people who answer tickets" to "people who build and run AI."

Tactical takeaways:

1. Turn your biggest perceived weakness (commodity, high-churn labor) into your headline promise.

2. In a people business, your employer brand and your sales pitch are the same asset, so fund it like revenue.

3. When the narrative turns against your category, reposition toward the part of the story that is growing.

AI Uses & Opportunities

Current exposure:

• AI is the double-edged sword that defines this stock. On the threat side, generative AI agents are getting good at automating Tier-1 customer support, the bread and butter of the DCX segment.

• On the opportunity side, AI Services (labeling, annotation, model evaluation, red-teaming) is the fastest-growing part of the business, up about 59% in 2025. They also deploy AI internally to lift productivity per head and defend margins.

Future opportunities:

• Move up the AI value chain from raw labeling toward higher-value model evaluation and safety testing for frontier models.

• Sell "AI plus humans" managed services and push toward outcome-based pricing, getting paid for resolved issues instead of staffed hours.

• Expand trust and safety into AI-output moderation, since every company shipping generative features now needs someone watching what the models say.

Bumps in the Road

• AI automation of routine support is a structural demand risk for the largest segment, and the market is pricing it as such.

• Client concentration: a handful of large tech clients drive an outsized share of revenue, so a single loss or pullback bites hard.

• Offshore labor and FX exposure: wage inflation and currency swings in the Philippines and India flow straight through a labor-based cost structure.

• Governance overhang: insiders and Blackstone tried to buy the company cheap, minority holders blocked it, and the freshly added leverage thins the cushion.

Your Swipe File

• When a technology threatens your business (or ideally before it), try to bend your core competencies to tackle new markets. Like TaskUs has done with their AI Services division.

• Using debt to pay a special dividend can turn a short-term financial “gain” into long-term pain.

• Seek diversification in service offerings if it doesn’t dilute your core competencies.