Alight: $101 million of free cash flow becomes minus $55 million after the tax payments

Alight reported $101 million of free cash flow for the first half of 2026. That is operating cash flow of $152 million minus $51 million of capital spending. In the same six months it paid $136 million under its tax receivable agreement and $9 million of finance lease principal, which leaves minus $44 million. Charge stock compensation at its cash cost, $11 million, and it is minus $55 million. Divide the first number by the $246 million market value and it looks like a 41% yield for six months. That is what you get if you stop reading there.

At $8.60, the October 20, 2026 regular session close, the shares trade 20.0% below my conditional Base value of $10.75, an upside of 25.0%. The conditional Bear value is $0.00, 100% below the price, and the Bull value of $44.61 is 418.7% above it. Confidence is grade D (10/25), the required margin of safety (MOS) is 50%, and the indicative cap is $5.37. I do not call it a buy zone, because the financing and acquisition facts that would make it one are missing. The verdict is Hold.

Valuation route: a leveraged benefits administration company, valued on the cash it leaves for common shareholders after interest, investment, compensation and tax receivable payments, not on adjusted EBITDA or reported free cash flow. Primary method (65% weight): five years of owner cash flow plus a value at the end of year five. Secondary method (35%): 7.5 times economic EBITDA with the full debt bridge. Classification: standard company, 2 of 12 on the scorecard. All values describe the business as it stood before the Abett acquisition, because that deal's price and funding are not disclosed. Treat the numbers as conditional research, not an appraisal of the company today.

What Alight is

Alight runs benefits administration for large employers: health, retirement and leave programs, plus the technology and call centers behind them. In the second quarter, revenue was $511 million and 92.2% of it was recurring. That recurrence is the franchise. Employers do not move such a program lightly, and a failed migration is painful for the client.

It is not a software subscription business, and I do not value it as one. Only part of revenue is subscription-based, the reported gross margin was 27.8% in the second quarter, and I could not verify the annual recurring revenue, retention or customer acquisition figures that a software multiple would need. Recurring does not mean growing: recurring revenue fell from $492 million to $471 million in the second quarter, and project revenue rose from $36 million to $40 million, which cushioned the total. Adjusted EBITDA fell from $127 million to $92 million over the same comparison. Management reaffirmed the 2026 outlook on October 14: revenue of $2,078 million to $2,098 million and adjusted EBITDA of $400 million to $415 million.

The company also changed shape. In 2024 it sold its payroll and HCM business, so older revenue and cash flow figures are not comparable, and 2025 included large goodwill impairments. Those are not cash, but they tell you that earlier expectations were too high.

The price, and what happened this week

The shares closed at $8.98 on October 16, fell 11.5% to $7.95 on October 19, and rose 8.2% to $8.60 on October 20. I found no company announcement for those days. The October 14 filing, which I did read, says the company had not reached a satisfactory definitive refinancing and was not actively pursuing one for now, and it reaffirms the 2026 outlook. A stock that moves 10% in a day on no news is a thin, nervous equity layer, and that is the main fact about it. The $8.60 is the October 20 regular close, not an intraday quote. The history chart in the screenshot, down about 96% from the start, is split-adjusted and gives no value anchor.

The company reports third quarter results after the US close on November 3, with a call at 4:30 pm New York time (10:30 pm in Prague). None of it is in this analysis.

Which cash to value

Reported free cash flow is operating cash flow minus capital spending. It leaves out two real claims: finance lease principal, and the payments under the tax receivable agreement (TRA). The TRA is a contract that passes part of certain tax savings to other parties. Alight classifies those payments as financing, so they never touch free cash flow, but the cash leaves all the same. Stock compensation is the third item: it costs nothing in cash but dilutes owners, so I charge it at a cash equivalent of $22 million a year going forward.

The operating cash flow figure also includes a receivable reduction of about $47 million, so the first half should not be doubled into a yearly forecast. The same bridge for earlier years: owner cash of minus $64 million in 2023, minus $93 million in 2024, plus $109 million in 2025 and minus $55 million in the first half of 2026. A growth rate across those years is not meaningful, because the endpoints change sign, and I do not print one.

Cash is counted once. Of the $215 million of corporate cash I treat $100 million as the working minimum, which leaves $115 million as a credit in the valuation, before whatever Abett cost. The $233 million of fiduciary cash is client money with an equal liability, so it adds nothing.

Shares and what comes before them

The cover of the quarterly report shows 26.419 million Class A shares on July 30 and 24,217 Class V shares, slightly above the June 30 figure I use. Seller shares with $250 and $300 vesting prices, after the split, are outside every case: they only vest at prices far from here. At $8.60, the market value is $227 million on June 30 Class A shares and $246 million on my 28.570 million count. I use the second figure. The October share count is not available to me.

Everything ahead of the common shareholder is much larger than the shareholder's own layer.

Add these up and the business is priced at an enterprise value of about $2,594 million, which is 6.4 times adjusted EBITDA at the guidance midpoint and 7.4 times my economic EBITDA, defined below. A plain screen using only basic shares, carrying debt and all cash shows 4.9 times. That is why a low EV multiple here overstates the bargain.

The TRA is not one clean number. Of the $463 million, $251 million is measured at fair value and $212 million is shown undiscounted, per the filing summary in my supplied analysis, and the company discloses up to $40 million of further disputed 2026 payments. I read the 10-Q only in part, so I did not re-read that footnote. I use $463 million as a proxy and test $350 million to $503 million. The company repurchased no shares in the first half, so the $216 million authorization adds nothing to value.

What the business can earn, year by year

The first forward year anchors on the guidance midpoints: revenue of $2,088 million and adjusted EBITDA of $407.5 million. The rest is my estimate, and none of it is company guidance. I let revenue slip in year two, rebuild margins slowly, let restructuring costs fade, keep capital spending near $105 million to $112 million, and raise interest from $105 million to $135 million, because the interest rate hedges expire in December 2026 and the 2028 debt has to be refinanced at some price. I add a one-time $30 million refinancing cost in year two.

Year one is 407.5 minus 22, 35, 105, 105, 30, 15 and 18, which is 77.5. That is 19.0% of the adjusted EBITDA it started from, before any TRA payment. The TRA is deducted once, as a lump sum in the bridge below, not again every year. Terminal owner cash is year five's 109.5 minus 20 of extra cash tax, or 89.5, because the tax savings of the transition years should not run forever.

Capital spending plus lease principal is about 5.9% of revenue, and the company does not split it into upkeep and growth. I treat all of it as necessary. Annual depreciation is about $138 million, so spending $105 million is not obviously excessive, and I do not add back any imagined growth portion.

The primary valuation: owner cash, then a value at year five

I discount the five yearly owner cash figures at 14%, a rate I chose for a leveraged company in this position, not a market consensus. At the end of year five I value the business at 8 times terminal owner cash of $89.5 million, which is $716 million, a future figure and not today's value. Today it is worth $371.87 million, and the five years of cash are worth $252.28 million, for a total of $624.15 million before the claims.

The exit carries 59.6% of the value before claims, and after the $463 million TRA is taken out the equity is much more dependent on it than that. Bank debt is not subtracted here, because interest is already inside every year of cash and I assume the debt is refinanced at the same size. That is a survival assumption, not a commitment from any lender. It is exactly the thing the October 14 filing did not deliver. If the refinancing fails, the answer is the Bear, not a smaller Base.

The secondary valuation: enterprise value

The second method starts from the business and then pays all claims. Economic EBITDA is first-year adjusted EBITDA less compensation and restructuring cost: 407.5 minus 22 minus 35, or 350.5. At 7.5 times, my assumption for a mature service business, the enterprise is worth $2,628.75 million. Subtract $2,000 million of bank debt, $45 million of finance leases and the $463 million TRA, add $115 million of cash and the $45 million note, and common equity is $280.75 million, or $9.83 a share. The 7.5 times is my estimate, equal to 6.45 times the guidance midpoint of adjusted EBITDA, not a peer median. A complete peer panel is not something I could verify, so I use no peer-derived target.

The two methods are built differently but are not independent. Both rely on the business holding its revenue, both pay the same TRA and both assume the same share count. When they agree, it shows my own inputs are consistent, and it earns no confidence. A balance sheet check is worse: common equity is mostly intangibles and tangible common equity is negative, so there is no asset floor under the shares.

Growth and yield checks

The check that makes Alight look cheapest is the forward earnings multiple. A third-party forecast page, last updated October 5 and so older than the refinancing news, shows adjusted earnings per share of $10.00 for 2025 (split adjusted), $4.77 for 2026 and $5.22 for 2027. That is a 52.3% fall this year, so a growth-adjusted multiple on 2026 is not meaningful. Using only the 2027 rebound, the P/E is 1.65 and the PEG would be 0.175. I do not use it: the adjusted figure ignores the TRA, the basis of the estimates could not be verified, and a PEG of 0.175 on a falling business is a rebound distortion, not a bargain. I show it only to say why I disregard it.

On my Base, owner cash before TRA is $77.5 million in year one, a yield of 31.5% on the $246 million market value, 35.8% in year four and 44.6% in year five. These are before the remaining TRA payments and are not a dividend yield. My growth rate over the three annual steps from year one to year four is 4.3% a year, from $77.5 million to $88.0 million, or $2.71 to $3.08 a share. Over four steps to year five it is 9.0%, but the terminal figure is $89.5 million after higher taxes, which is why I do not extend it. A growth rate after TRA payments, and a yield after them, are unavailable, because the payment schedule is not public in a form I can use. Reliability: low.

Triangulating the values

The Base methods differ by 13.4% (difference over midpoint), under the 20% warning level, and that earns no confidence bonus because both depend on the same recovery. I gave the cash method 65% because TRA, investment and financing cash decide what is left for common owners, and the enterprise method 35% because it checks what the surviving franchise is worth. The weights were not tuned to the price. An equal split gives $10.53. The Bear is zero in both methods: before the floor the cash method is minus $501 million and the enterprise method minus $616 million. Zero is the limit of what a shareholder can lose, not a statement that the operating business is worthless.

Abett changes the picture in a way I cannot price. The company announced the acquisition on October 1 and has not disclosed the price, the funding or the acquired earnings. The $10.75 therefore assumes the deal changes nothing per share, which is not established. For unchanged shares, each $50 million of net value gained or lost moves the figure by $1.75: a loss of $50 million gives $9.00 and a gain of $50 million gives $12.50. A fair priced cash purchase adds no value but still uses cash that the working buffer needs. If the deal were funded by selling shares, both the new shares and the new cash count. As an illustration, not a forecast, selling $100 million of stock at $5 and keeping the cash gives $8.38.

What the price already assumes

The market does not need the business to collapse. It needs adjusted EBITDA about 1.1% below the middle of this year's guidance, or a lower exit multiple. That sounds like a small ask, and it is, which is the problem: because the equity layer is thin, a 1% shift in enterprise value of about $2.6 billion is worth roughly $0.90 a share. The market price could also reflect forced selling after the reverse split and the refinancing news, but I have no verified short interest or positioning data, so that is a hypothesis and not an input.

Bear, Base and Bull

The Bear combines lost sales, weaker margins and costlier refinancing, because weaker profit is exactly what makes lenders harder to deal with. The Bear owner cash is negative in years two to five, so the cash method gives common equity nothing, and the 50 million refinancing cost and higher interest make it worse. The Bull needs real recovery: sales up to $2.39 billion and adjusted EBITDA of $573.6 million by year five, without counting any Abett contribution. The same 14% discount rate and share count apply in all three. I attach no probabilities, so there is no probability-weighted value here, and I do not offer one. The Bear is a plausible absence of common recovery, not a forecast of bankruptcy.

Sensitivity and the case that flips the verdict

Across this grid the value stays between $7.81 and $14.15, so on cost of equity and the exit multiple alone the stock is never worth much more than 65% above the price. The inputs that matter more are the TRA deduction, the level of terminal owner cash and the enterprise multiple. One input at a time:

The label flips with one reasonable change. A 20% shortfall in owner cash, or half a turn less on the enterprise multiple, removes the whole discount. The action does not flip the other way: no single change here gives a price that clears a 50% margin of safety. The cap ranges from $2.53 to $7.35 across these tests, all below $8.60. A cap based on a conditional model is not sturdier than the model.

How the cap is set

Confidence D calls for a margin of safety of 50%, and I add nothing on top. Sales decline, refinancing, the TRA and the acquisition are already handled in the scenarios, the Bear and the confidence score, and charging them a second time would count one risk twice. The arithmetic cap is $10.75 x (1 - 50%) = $5.37. I do not call this a buy zone. The framework I use also allows a no-buy call when a financial risk cannot be measured, and that applies here: a lower quote alone does not repair unknown refinancing terms or acquisition funding. I also do not give a return to a buyer at $8.60, because that needs the timing of TRA payments and the funding plan, and I do not have either. A pre-TRA yield of 31.5% is not that return.

The dividend

There is none to count. The supplied analysis states that Alight replaced its regular dividend with other capital priorities in February 2026, and the company bought no shares in the first half. The $1 million dividend line in the first half cash flow does not establish a policy, and I do not annualize an old payment into a yield. This is not an income stock: cash first has to cover service delivery, interest, leases and the TRA.

Reasons to own

• A franchise that is hard to replace. 92.2% of second quarter revenue was recurring, and large employers change benefits administrators slowly.
• Real operating cash. The business produced $152 million of operating cash flow in six months, and the large impairments in 2025 did not use cash.
• A thin equity layer with big upside. If recovery, refinancing and disciplined deal funding arrive together, the Bull is $44.61, 418.7% above the price.
• The TRA may shrink. If the payments end up smaller or later than my Base, the stock gains: at a $350 million deduction the value is $14.70.
• An honest disclosure on October 14. The company separated its reaffirmed 2026 outlook from stale lender materials instead of blurring the two.

Main risks

• Refinancing. About $2,000 million is due in August 2028, the interest hedges expire in December 2026 and no satisfactory deal existed on October 14. A failed rollover can leave nothing for common shareholders.
• The TRA. It is not in reported free cash flow, its payment schedule is unclear, up to $40 million of 2026 payments are disputed, and a change of control can accelerate it. A takeover is not automatically a premium for shareholders.
• Falling revenue. Recurring revenue fell from $492 million to $471 million in the second quarter, and the second quarter gross margin was 27.8%.
• Capital spending and contract costs. Costs of implementing clients can be spread over seven or fifteen years, so adjusted profit can run ahead of cash.
• Abett. No price, funding or earnings are disclosed. Even a sensible deal can use liquidity the refinancing needs.
• Dilution. Awards, a rescue financing or a debt conversion can change ownership materially. A reverse split creates no value.
• Client cash. The $233 million of fiduciary funds cannot repay corporate debt, and the seller note is not cash on demand.
• Model risk. A small change in enterprise value can wipe out or multiply the common stake. No asset floor protects it.

Management execution

I score management credibility 8/20: execution 2/5, capital allocation 2/5, disclosure 3/5, alignment 1/5. Recurring sales and margins weakened while the commentary stressed outperformance, the dividend was replaced and no shares were bought, which conserves cash but also shows the room is small. The October 14 filing was candid about financing. I could not verify insider buying, ownership percentages or a multi-year record of guidance delivery, so the alignment score is a low estimate, not a finding. The outstanding awards and the transfer of tax benefits under the TRA also matter to minority shareholders.

Stock Analyza scorecard

On the economic scorecard Alight scores 2/12, a standard company in turnaround. The only points are 2/2 for recurring revenue. It gets 0/2 for returns above the cost of capital, 0/2 for organic growth, 0/2 for owner cash growth, 0/1 for the balance sheet, 0/1 for unit economics and 0/2 for the reinvestment runway. That is not an unfair sector metric: this is a nonfinancial service company, and owner cash is the right measure. The valuation score is 47/100: 28/40 for the discount to value, 0/20 for protection against the Bear, 10/15 for method agreement with no credit for independence, 2/10 for history, 6/10 for what the price implies and 1/5 for data and model quality.

Confidence score

The arithmetic and the sources can be good while confidence in the equity value stays low. Grade D reflects a range from zero to $44.61 and the dependence of common value on financing. It is not a prediction of insolvency. A narrower range would need new evidence, not a smaller discount rate.

What I would watch from here

The most important thesis breaker

Failure to refinance while keeping a business and a TRA cash bridge that still add up. The next dated test is the November 3 report: I would reconcile its cash, debt and TRA figures before changing any value here.

Stress test and independent check

I rebuilt the model from its stated inputs with separate code and reproduced: the five-year Base cash table and the $252.28 million of discounted cash; the $716 million exit and its $371.87 million present value; the $321.15 million equity and $11.24 a share; the $2,628.75 million enterprise value, $280.75 million equity and $9.83 a share; the 65/35 blends of $10.75 and $44.61; the 13.4% divergence; the reverse results of 6.38 times and $345.8 million; the sensitivity grid and the single-input table; the acquisition and share sale illustrations; the $5.37 cap; and the yields and growth rates. I found no construction error. In my first run a few tests differed because I had entered the cash credit and the enterprise bridge wrongly in my own code. After I corrected those inputs, they matched.

I confirmed against the sources: the closing prices of $8.98 on October 16, $7.95 on October 19 and $8.60 on October 20 (a screenshot of a broker quote agrees with $8.60); the October 14 filing wording on refinancing and the reaffirmed outlook; and from the quarterly report the $443 million long-term TRA, the $136 million of TRA payments, debt of $20 million current and $1,976 million long term, the $10 million of bank repayments, the cover share counts and the seller note terms.

I could not re-read everything. The quarterly report was too long for one fetch and the last part was cut off, so the TRA footnote split ($251 million and $212 million), the disputed $40 million, the debt maturity and hedge expiry details and the vesting thresholds of the seller shares come from the supplied analysis. The same applies to the second quarter and full year 2025 results, the Q2 receivable reduction, the Abett announcement, the dividend change, the third quarter date and the forecast page. They agree with what I did read, but I did not check each one line by line.

The forecast page also shows 2027 revenue of about $1.98 billion, below my Base year two figure of $2.05 billion, so my Base is the more generous of the two. The tests below were all run on the downside, and none of them changes the verdict.

The strongest case against Hold is the Bull of $44.61, a durable franchise with lower TRA cash and an affordable refinancing. I keep it in the table at full size and do not dismiss it because the stock is cheap on the chart. My answer is that it requires three things at once, a recovery, a refinancing and a sensibly funded deal, and none of the three is yet visible in the filings. The strongest case against owning at all is the Bear, in which lenders and the TRA take everything. The main limits: I do not know the October cash, debt, share count or TRA balance, the Abett terms, or how much of the first half cash flow was timing. Those are why confidence is D.

Final verdict: hold at $8.60

Alight may well be cheap if the business holds, if the lenders cooperate and if Abett is paid for sensibly. The data does not yet say that any of those will happen. A 96% fall on the chart is not a margin of safety, because in this capital structure a small change in operating value can erase common equity or multiply it. The $101 million of headline free cash flow is not cash for shareholders, the October 14 refinancing update was a negative, and the numbers above describe the company before Abett.

Verdict: HOLD at $8.60, watchlist for new money, no new buy. Conditional Base value $10.75, upside 25.0%, Confidence D (10/25), required margin of safety 50%, indicative cap $5.37, no actionable buy zone.

For holders, the model gives no reason to add and no instruction to sell; the right size depends on how much of the position you can afford to lose entirely, which I cannot know. For new money, it means waiting for a funded refinancing, the Abett terms and owner cash that stays positive after the TRA. The price allows for a weaker business. It leaves little room for a lender who says no.

Sources

Primary sources include Alight's second quarter 2026 quarterly report on Form 10-Q, the second quarter 2026 results and its SEC exhibit, the first quarter 2026 results, the 2025 fourth quarter and full year results and 2025 annual report, the October 14, 2026 current report on Form 8-K, the reverse split announcement, the Abett acquisition announcement and the third quarter reporting date. The earnings estimates used as a diagnostic come from a third-party forecast page last updated October 5, were not an input to any value, and their basis could not be verified.

The reference market price is $8.60, the October 20, 2026 regular session close, from the stockanalysis.com price history and the StockTitan quote page, which also give the October 16 and October 19 closes. Per share values use 28.570 million economic shares. Cash flows, the discount rate, exit multiples, the working cash minimum, refinancing costs, the TRA deduction and the weights are analytical estimates, not company guidance.