Margin of Safety

Margin of Safety

TaskUs at $6.07 trades at the midpoint of Evercore's conservative Case-B DCF ($5.10-$7.50) and about 37% of the $16.50 its own controllers offered to go private — a bid they revealed as a ceiling by walking with no fee rather than raising after a 92%-against minority veto.[1][2][3]

That single price does two jobs at once, and both are quantifiable. Sitting at the Case-B midpoint of about $6.30, the margin of safety is thin against Case B and wide against anything better: a move back toward Evercore's base "Case A" DCF of $12.40 to $20.20 a share (midpoint ~$16.30) [4] is roughly a double, contingent on the automation drag proving cyclical and adjusted free cash flow rebuilding to the guided $105–$115M [5]. The counter-fact belongs in the same breath: the rejected $16.50 is a ceiling for buyers, not a floor under the stock — it was struck on the pre-automation Case-A trajectory, and the same controllers walked rather than raise it. The margin of safety is real against the pre-automation forecasts, and roughly absent against the one that now describes the guide.

What the price buys

Source: share price and market data as of 31 July 2026, as reported; enterprise value derived from Q1 FY2026 total indebtedness of $491.6M [6] less cash of $152.3M [7], plus market capitalization; FY2025 adjusted EBITDA of $249.1M [8].

The equity is priced at roughly $565M against a business that earned $102.3M of net income and $249.1M of adjusted EBITDA in FY2025 [9] — about 5.5 times trailing earnings and a 13.1% trailing free-cash-flow yield. That is the surface a value screen returns: a former growth name down 68% from its three-year high, against a peer-group median drawdown of 46.7%. The screen is the reason to look; it is not the analysis.

The analysis begins one line lower on the balance sheet. In March 2026 the company borrowed to pay a $332.8M special dividend (Financials and Estimates), so the equity no longer sits on net cash. Enterprise value is now about $900M — market capitalization plus roughly $339M of net debt [10] — and the whole business changes hands at about 3.6 times FY2025 adjusted EBITDA. A cheap-looking equity multiple and a less-cheap enterprise multiple describe the same company; which one a buyer should weigh depends on whether the buyer inherits the debt, and here the buyer does.

Three anchors above the price, one below

The most useful valuation work on TaskUs was not done by the sell side. It was done by Evercore, the financial adviser to the special committee that evaluated the 2025 attempt to take the company private at $16.50 a share [11]. Its fairness analyses, disclosed in the merger proxy, are the closest thing the record holds to a full intrinsic-value study, and they bracket the stock from both sides.

Sources: Evercore financial analyses and comparable ranges, TaskUs merger proxy — public comparables [12], implied per-share ranges [13], DCF Case A [14], DCF Case B, premia and analyst targets [15]; merger price [16]; current mean target, consensus estimates as reported.

Four of the five backward-looking anchors sit well above $6.07. Evercore's discounted cash flow on management's base "Case A" forecasts produced $12.40 to $20.20 a share, using a 10.5%–12.5% discount rate and a 3.0%–5.0% perpetuity growth rate [17]. Its public-comparables work — a customer-experience peer set trading at about 5 times forward adjusted EBITDA — implied $12.20 to $15.60 across its two forecast-year bases [18]. Precedent BPO deals cleared at 6.8 to 7.6 times last-twelve-months adjusted EBITDA [19]. On any of these frames the current price is less than half of intrinsic value.

The fifth anchor is the one that matters. Evercore also ran the DCF on a conservative "Case B" forecast, and that produced an implied value of $5.10 to $7.50 a share [20]. Today's $6.07 falls almost exactly at its midpoint. The market is not pricing TaskUs at a discount to fair value; it is pricing TaskUs at the bank's own downside case, computed before the downside began to show in the numbers.

Why Case B is the live case

The Evercore ranges carry a date stamp that decides how to read them. They were struck in May 2025, against a $14.38 close [21], and Case A embedded the growth trajectory the company was still guiding to at the time. What has happened since is the automation cross-current (Automation Offset): management now guides FY2026 revenue to $1,210–$1,240M, up roughly 3.5% at the midpoint from 19.0% in FY2025, with adjusted EBITDA margin easing to about 19% [22], and its Trust and Safety line guided to decline [23]. A low-single-digit grower with a compressing margin is the Case B animal, not the Case A one. The stock has re-rated to the conservative forecast because the conservative forecast is now the base forecast.

That reframes the multiple comparison rather than erasing it. At about 3.6 times FY2025 adjusted EBITDA — closer to 4 times on the stricter, stock-comp-excluded basis Evercore used [24] — TaskUs still trades below the 5.0x–6.0x range Evercore applied to its customer-experience peers and well below the 6.8x–7.6x precedent-deal band. But those benchmarks are also May 2025 marks, and the peer set has de-rated with it: the comparison shows TaskUs is cheap relative to where CX outsourcers traded a year ago, not that it is cheap on a multiple the market is willing to pay today.

Sources: TaskUs current enterprise value derived from Q1 FY2026 filings [25] and FY2025 adjusted EBITDA [26]; peer and precedent multiples per Evercore analyses, May 2025 [27], [28].

What insiders said value was, in cash

The take-private attempt is a second, independent read on value — this one revealed by action rather than by model. Blackstone and the two founders, who together control the vote, offered $16.50 a share to buy out the minority [29]. The minority declined it: at the October 2025 special meeting the requisite stockholders did not approve the merger, and the deal was terminated with no fee paid [30]. Both sides of the control table put fair value above $16.50 — the controllers by bidding it, the minority by refusing it as too low.

That $16.50 is now $10.43 of air above the traded price, and it is tempting to read the gap as embedded upside. It should be read with care, in three parts. First, the bid was struck at a $14.38 share price on a Case-A trajectory; the same automation deceleration that pulled the stock to $6 would pull a re-bid well below $16.50. Second, the controllers did not raise their offer when the minority balked — they walked, then four months later paid themselves a $332.8M debt-funded special dividend [31], a route that returns cash pro-rata without paying minority holders a control premium. Third, the same insiders retain the votes to try again on their own timetable. The revealed-preference signal is genuine and it favors the long case; it is not a floor.

What has to be true

For the value reader, TaskUs resolves into a single, testable fork, and both prongs are visible in the cash flows rather than in the story.

The evidence that the drag is cyclical is the forward cash guide: management expects adjusted free cash flow of $105–$115M in FY2026, up from $73.7M, as the FY2025 capital-spending peak rolls off [32]. On the current price that is a high-teens forward free-cash-flow yield, and even the reduced 2026 consensus price target of $9.50 sits well above the tape. The evidence on the other side is that the guided margin steps down and the most automation-exposed revenue is guided to shrink — so the free-cash-flow rebuild depends on lower-margin AI Services work replacing higher-margin arbitrage volume faster than the volume erodes (Automation Offset).

Two facts bound the downside. The balance sheet, though no longer net cash, carries no maturities until the 2031 term loan and services roughly $31M of annual interest out of well over $100M of operating cash flow [33] — the chance of forced distress is low, which matters for a reader who wants bankruptcy risk near zero. And the price already embeds a conservative forecast, so the asymmetry is not symmetric: the margin of safety is thin against Case B and wide against anything better. What would settle the read is not a valuation debate but two or three quarters of data — whether adjusted free cash flow lands in the guided $105–$115M band and whether Trust and Safety stabilizes or keeps sliding.