Target Hospitality
A landlord and caterer for construction crews: it puts up whole villages of rooms and kitchens next to remote data centres, power plants and mines, and runs them.
The scores
Growth3/4 ↑
Revenue fell from $564M in 2023 to $321M in 2025, then rose 39% in Q2 2026 on one driver, data-centre and power villages.
Business quality2/4 →
Owned beds win fast contracts, but rivals and builders' own camps can do the job, and demand ends with each build.
Financial position3/4 ↓
Net debt was small in June, but about $320M of building is borrowed on a $660M line to 2031.
Execution record3/4 ↑
Replaced the lost government revenue within eighteen months and raised 2026 guidance twice; the large villages have yet to open.
Per share capture4/4 →
Diluted shares fell from 105.3M to 99.5M in two years, a $30M buyback followed, and the build uses debt, not shares.
Potential3/4 ↑
The upside is committed: signed minimums, dated openings and funding, but bounded by contracts ending from 2027 to 2031.
Risk profile2/4 →
Two customers at a third of revenue, a termination right on the largest contract, and Dilley cancellable on 60 days' notice.
Valuation3/4 ↑
At 4.7 times the exit-2027 pace the price pays for the plan and about one round of renewals, not more.
Evidence quality3/4 →
Contract terms and results are filed, but counterparties, termination fees and renewal odds are not disclosed.
If it works 1.3× · if it fails you keep about 25%.
The Verdict
We rate Target Hospitality Neutral. The business is solid: it owns thousands of rooms it can move to a new site fast, and in 2026 data-centre and power builders signed over $1.7B of contracts. The price is fair. It already pays for those contracts and for about one round of renewals, while the contracts house builders who leave when the building is done. As a bet it is priced for success: if the plan works the whole company is worth only a little more than today, and if it stalls shareholders keep about a quarter. The view hinges on whether any data-centre village outlives its build.
The long caseBuilding AI data centres and power plants in empty places needs thousands of beds fast, and Target already owns them.
The evidence under this - 5 claims
- Net debt rises with the $313M to $333M of second-half capital spending (C20), so company value on the year-end balance sheet will be higher.
- Trailing EBITDA is H2 2025 plus H1 2026, from released quarterly figures, not a filed annual total.
- Minimum revenue understates contracts with variable revenue and options, such as the Hub's $20M to $40M a year and extensions to 2035 (C07).
- Assets outlive contracts and can be moved: the beds from the lost 2025 government contract now serve WHS (S05).
The company
A company building a data centre or a power plant in empty West Texas needs thousands of workers on site, and there is nowhere for them to sleep. Target brings in modular rooms it owns, adds kitchens, laundry and security, and runs the whole village day and night for one bill.
Its biggest new village, in North Texas, is for about 4,000 people; another, for AI infrastructure, is for about 3,370. Without it, the builder runs its own camp or pays crews to drive from motels hours away.
The evidence under this - 3 claims
The filing does not say what the termination fee covers in every case, only that in certain circumstances it compensates invested capital and contractual economics.
How it makes money
Every contract is a lease on the rooms plus a fee for meals and services, with a minimum the customer pays whether the beds are full or not. In the June 2026 quarter services were 61% of revenue and rent 36%. Customers often pay in advance, which funded much of the building: first-half cash from operations was $111.0M. The money stops when the contract ends, unless it is renewed.
The evidence under this - 4 claims
Where the money comes from, stream by stream
| Stream | Share of sales (Q2 2026) | How it earns it |
|---|---|---|
| Services income | 61% | Meals, housekeeping, security and community management at the villages, billed per person or per bed. |
| Specialty rental income | 36% | Rent on the modular rooms under lease agreements with minimum commitments. |
| Construction fee income | 3% | Building the Lithium Nevada hub, now substantially complete. |
Who runs it
Brad Archer has run the company since 2019 and led the switch from government and oilfield work to data-centre villages after the 2025 contract loss. Directors and officers own 3.8%, and he and the commercial chief bought shares in September. His long-term pay rides on profit before interest, tax and depreciation, share returns against other small companies and share-price targets. TDR Capital, the controlling owner since 2019, sold most of its stake in 2026.
The evidence under this - 3 claims
Why it wins, and who could beat it
Why customers come back
Target's edge is size in a market of small operators. It owns thousands of rooms across Texas, New Mexico and Nevada, so it can move idle beds to a new site: the Pecos data-centre village needed under $15M of new capital. A rival's annual report names it among the few firms that own and lease units at scale.
The weakness is time. That rival finds operating crews are a fifth to a quarter of building crews, so the need peaks during construction.
The evidence under this - 3 claims
The threat
The rivals are regional and private operators, motels and RV parks, and the builders themselves, who can run their own camp. Civeo, the one listed rival, left the US. For the Dilley detention centre the risk is the government moving people elsewhere.
The evidence under this - 4 claims
The 10-K describes the rivals as small independents and RV parks rather than funded firms (C35), and no rival is named in either.
The rivals: CVEO, GEO, CXW, and 3 more that are private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| Private | Regional and private-equity-owned workforce lodging operatorsPriced against it: Private | Workforce Hospitality Solutions (WHS) | On sale | Private |
| CVEO | CiveoPriced against it: Same product | Workforce Hospitality Solutions (WHS) | On sale | Same product |
| In-house | Developers and contractors housing their own crewsPriced against it: In-house | Workforce Hospitality Solutions (WHS) | On sale | In-house |
| Not listed | Small independent lodges and RV parks in the PermianPriced against it: Not listed | HFS-South | On sale | Not listed |
| GEO · CXW | Other detention operators for ICEPriced against it: Not checked | Government | On sale | Not checked |
How big it could get
How big this could get
US private spending on data-centre construction ran at a $84.95B yearly rate in August 2026, up 73% in a year. What nobody publishes is how much of that goes to housing crews, so the market cannot be sized from outside. Management puts it at about $18B across data centres, power and mining, a figure no outside source checks.
The evidence under this - 2 claims
It is built from historical hospitality-service shares applied to announced projects, and no outside body publishes the figure.
Who has signed
Seven contracts carry minimums: two hyperscaler villages in Texas, an AI infrastructure village, two power-plant villages, the first data-centre village and a lithium mine hub. The largest customer can walk away, in some cases paying a fee, and the government can cancel Dilley on 60 days' notice.
The evidence under this - 7 claims
The filing does not say what the termination fee covers in every case, only that in certain circumstances it compensates invested capital and contractual economics.
The results
The numbers
Sales grew 12% over the twelve months to Jun 2026, faster than the −17% in the year to Dec 2025.
Operating margin is −11%, down from 43% in the year to Dec 2023.
Running costs take 25¢ of each sales dollar, up from 22¢ in the year to Dec 2021.
Shareholders' money lost 9.8% a year, down from −4.7% in the year to Dec 2021.
The business brought in $170M of cash from operations, 49% of sales, before capital spending, which the filer does not tag.
It owes $38.6M more than the cash it holds, about 0.2 years of the cash its operations bring in.
The share count rose 3.7% in 4.5 years, and pay in shares costs 2.2% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-11-01. Fiscal years end in December.
| Dec 2021 | Dec 2022 | Dec 2023 | Dec 2024 | Dec 2025 | 12 mo to Jun 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $291M | $502M | $564M | $386M | $321M | $347M |
| Sales growth | +29% | +72% | +12% | −31% | −17% | +12% |
| Profit | ||||||
| Gross margin | 35% | 49% | 56% | 46% | 13% | 14% |
| Operating margin | 13% | 35% | 43% | 28% | −11% | −11% |
| Net margin | −1.6% | 15% | 31% | 18% | −12% | −11% |
| What growth costs, per sales dollar | ||||||
| Selling, general and admin | 16% | 12% | 10.0% | 14% | 18% | 19% |
| Pay in shares | 1.7% | 3.8% | 2.0% | 1.9% | 2.4% | 2.2% |
| Cash | ||||||
| Cash from operations | $105M | $306M | $157M | $152M | $74.1M | $170M |
| Balance sheet | ||||||
| Cash and short-term investments | $23.4M | $182M | $104M | $191M | $8.35M | $6.07M |
| Debt | $332M | $331M | $179M | $182M | $3.76M | $44.7M |
| Net cash | −$308M | −$149M | −$75.5M | $8.89M | $4.59M | −$38.6M |
| Returns and shares | ||||||
| Return on capital | - | 25% | 34% | 14% | - | - |
| Shares (diluted) | 96.6M | 100M | 105M | 101M | 99.5M | 100M |
What growth costs
Growth here costs capital rather than running costs. Each new village is bought and built up front, about $500M in 2026, and then earns a high margin with little added overhead, so profit grows much faster than sales. The risk is that the capital is borrowed against contracts that end in a few years.
The evidence under this - 4 claims
- Two weeks earlier the exit-2027 targets were $700M and $260M (S07), so the figure moved with one new award.
- An exit pace is one quarter annualised, not a year's result, and it excludes the cost of the capital spent to reach it.
The last report (10 Aug)
Quarter ended June 2026, filed figures.
In the June 2026 quarter revenue rose 39% to $85.5M and profit before interest, tax and depreciation rose more than fivefold to $18.2M. The new villages did it: the WHS segment kept $19.4M of every $36.3M it billed. The oilfield villages slipped, with 70% of beds filled against 76% a year earlier.
Two weeks later a new $250M hyperscaler village raised the 2026 plan again.
What it did to the thesisIt proved the new villages earn far more per bed than the business they replaced, and that customers keep signing. It left open the question the view turns on: what happens to each village once its building is finished.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $-0.09 against $-0.11 expected; $-0.09 as filed, with no adjustments.
The evidence under this - 3 claims
What management did with the money
It repaid its costly 2025 bonds, bought $30M of stock when its largest shareholder sold in September, and is spending about $500M in 2026 on new villages, paid for by customer advances and a new $660M credit line rather than new shares.
Insiders in 2026Q3: 8 purchases and 2 sales, from their Form 4 filings.
The evidence under this - 4 claims
The price, and where we disagree
In short. Fair. The price pays for the signed contracts and about one round of renewals, but nothing proves yet that a data-centre village outlives its build.
The shares fell 36% from late September to $13.67 on 30 October, with no news from the company. The fall came as TDR Capital sold and handed out shares.
The next test is close: the September-quarter results on 5 November should show whether the two largest villages took their first workers on time. Behind it all, data-centre building in the US is still rising fast, though Texas is auditing every project waiting for grid power.
The evidence under this - 7 claims
The filing does not say what the termination fee covers in every case, only that in certain circumstances it compensates invested capital and contractual economics.
Meta, OpenAI and Amazon said they would comply, and the order targets grid connection, not construction under way (S19).
How we read the multiple
With its debt, the company costs about $1.41B. What has to be true:
1. Management plans a $300M yearly pace of profit before interest, tax and depreciation by late 2027; the price is about 4.7 times it. 2. The signed contracts have roughly $2.0B of revenue left, worth roughly $1.0B of profit before overhead. 3. That is less than today's price plus about $320M still to spend, so renewals are needed.
It has lost money in each of the last four quarters under accounting rules.
Against its own history. Buyers pay far more than they usually have for this company's profit, because they are paying for villages that have not opened yet rather than for today's earnings.
The evidence under this - 5 claims
- Net debt rises with the $313M to $333M of second-half capital spending (C20), so company value on the year-end balance sheet will be higher.
- Trailing EBITDA is H2 2025 plus H1 2026, from released quarterly figures, not a filed annual total.
- Two weeks earlier the exit-2027 targets were $700M and $260M (S07), so the figure moved with one new award.
- An exit pace is one quarter annualised, not a year's result, and it excludes the cost of the capital spent to reach it.
- Minimum revenue understates contracts with variable revenue and options, such as the Hub's $20M to $40M a year and extensions to 2035 (C07).
- Assets outlive contracts and can be moved: the beds from the lost 2025 government contract now serve WHS (S05).
Priced against its rivals
Only Civeo sells the same thing to the same buyers as a large part of what it does. A price comparison needs at least two, so the price is judged against its own history instead.
The bet: what it pays if it works
The bet is on the whole company: the large data-centre villages open on time and reach management's 2027 pace from contracts already signed.
Even then, valued as a business whose contracts end in a few years, the company is worth only a little more than today, so the bet is expensive. If the villages slip and nothing renews, shareholders keep about a quarter of what they hold, because the building is paid for with debt.
| EBITDA in two to three years | Contracted village beds needed | At the stated multiple |
|---|---|---|
| $200.00M | about 7.4K at 100% kept | 0.9× |
| $300.00M (the big win) | about 11.1K at 100% kept | 1.3× |
| $400.00M | about 14.8K at 100% kept | 1.7× |
To pay 2× (a Fair bet) it would need about $469.10M of EBITDA (about 17.4K contracted village beds); to pay 5× (a Cheap bet), about $1.17B of EBITDA (about 43.4K contracted village beds), at the same multiple.
The steps it has to take
- Done Four AI and power villages signed and earning revenue
- In progress Data Center Hub and AI Infrastructure Community take first workers due 2026-12-31
- Not started Data Center Hub complete at about 4,000 people due 2027-06-30
- Not started A data-centre contract renewed past its first term due 2027-09-30
The detail behind the label
The evidence under this - 5 claims
- Two weeks earlier the exit-2027 targets were $700M and $260M (S07), so the figure moved with one new award.
- An exit pace is one quarter annualised, not a year's result, and it excludes the cost of the capital spent to reach it.
- Net debt rises with the $313M to $333M of second-half capital spending (C20), so company value on the year-end balance sheet will be higher.
- Trailing EBITDA is H2 2025 plus H1 2026, from released quarterly figures, not a filed annual total.
Where we disagree with the market
An AI building boom that doubles Target's sales by 2027 and turns it to a profit, with every analyst positive.
The contracts are real, but they house building crews and end with each build, so the price is fair rather than cheap.
The market may be right that builders will keep coming: data-centre construction is still rising and Target keeps winning contracts, some built from idle beds at little cost. If villages move from one finished campus to the next, renewals look like new contracts.
Against us, the only evidence on what follows a build comes from a rival, and Target has never disclosed a data-centre renewal either way.
What analysts say, and the evidence
The evidence under this - 4 claims
What could go wrong, and how we would know
The risks
- A few big customers. Two customers were a third of first-half revenue, and the largest new contract lets the customer walk away, sometimes paying a fee.
- Government cancels Dilley. The detention-centre contract, the most profitable per bed, can be cancelled on 60 days' notice; a similar contract ended that way in 2025.
- Texas slows data centres. The governor ordered an audit of every data centre waiting for grid power, and most of the new villages are in Texas.
- Borrowing to build. About $320M of building is still to pay for in 2026, mostly borrowed against contracts that end within a few years.
- A large seller. The former controlling owner still holds about 14.6% after selling most of its stake this year.
The evidence under this - 6 claims
The filing does not say what the termination fee covers in every case, only that in certain circumstances it compensates invested capital and contractual economics.
Meta, OpenAI and Amazon said they would comply, and the order targets grid connection, not construction under way (S19).
What we still do not know
- Does any data-centre customer renew its village once the campus is built, and at what size?
- What would the AI infrastructure customer owe if it ended its contract early?
- Is Target trying to sell the Dilley detention centre?
The evidence under this - 4 claims
The filing does not say what the termination fee covers in every case, only that in certain circumstances it compensates invested capital and contractual economics.
What would change our view
- Neutral→AttractiveA price that pays only for signed contracts
It needs all of:
- ○The company is valued at 9 times its 2026 profit plan, before interest, tax and depreciation.checked daily
- ○Quarterly profit before interest, tax and depreciation reaches $40M as the large villages fill.checked at the next results, due by 5 Nov
- Neutral→AttractiveProof the villages outlive their builds
It needs all of:
- ○Target announces that a data-centre or power customer renewed or extended a village beyond its first term.watched as it happens
- ○Quarterly profit before interest, tax and depreciation reaches $60M.checked at the next results, due by 5 Nov
We would also be wrong if
- The North Texas hub or the AI infrastructure village opens later than the end of 2026.
- The government cancels the Dilley contract.
- Profit before interest, tax and depreciation for 2026 falls below $105M.
What is coming
- 5 November 2026 Q3 2026 results: Whether the Data Center Hub and the AI Infrastructure Community took first occupants and the 2026 guide holds.
Every dated event ahead
- 5 November 2026 Q3 2026 results: Whether the Data Center Hub and the AI Infrastructure Community took first occupants and the 2026 guide holds.
- 30 June 2027 Data Center Hub full completion: Whether the largest village reaches about 4,000 people on time.
- 30 September 2027 First term of the initial Data Center Community beds ends: The first test of whether a data-centre customer renews after its build.
Since we wrote this
- 1 NovThesis writtenNeutral, medium convictionWritten at the owner's request four days before the Q3 2026 print, after the shares fell 36% from September.
Sources 19
Sources - 19
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-03-11business description, segments, the Data Center Community contract, the DIPC terms and cancellation clause, competition, customers, employees, management's market estimatebacksC01C02C10C15C17C30C35C36C38
- S02primarycompany release2026-03-11FY2025 revenue, net loss, adjusted EBITDA and cash flow
Adjusted EBITDA (1) of $53.2 million.
backsC02 - S03primarycompany filing2026-05-11terms of the West Texas Power, Pecos Power, Data Center Hub and AI Infrastructure contractsbacksC07C08C09C30
- S04primarycompany release2026-05-11Q1 2026 revenue, net loss and adjusted EBITDA
Adjusted EBITDA (1) of $9.9 million.
read, but no claim rests on it - S05primarycompany filing2026-08-10Q2 2026 results by revenue type and segment, the four 2026 contracts, Lithium Nevada and DIPC terms, capital spending, debt, the new credit facility, customers, share countbacksC01C03C04C05C06C11C15C17C18C19C20C22C29C30C39
- S06primarycompany release2026-08-10Q2 2026 segment revenue, adjusted gross profit, beds, rates and utilisation, adjusted EBITDA, liquidity
With average WHS utilized beds surpassing 4,000 during the quarter, Target continues to demonstrate the depth of its operational platform
backsC04C05C13C19C30 - S07primarytranscript2026-08-10the leverage plan, the renewal rate, the Dilley question, competition and the pipeline
we expect leverage to decline as these communities open and anticipate exiting 2027 with net leverage well below 3x based on our current project schedule
backsC14C16C21C32C37 - S08primarycompany release2026-08-26the Pecos hyperscaler contract, the raised 2026 outlook and the exit-2027 target
Together, these factors position the Company to achieve annualized revenue exceeding $750 million and annualized Adjusted EBITDA (1) above $300 million exiting 2027.
backsC12C13C14C20C29C30 - S09primarycompany filing2026-09-09TDR's sale of 14.0M shares at $18.50 plus a 2.1M option, and the company's $30M repurchasebacksC22C23C40
- S10primarycompany release2026-09-10the number of shares repurchased
The Company purchased from the underwriters 1,693,599 shares of Common Stock (the "Repurchase Shares") at a price per share equal to the price per share paid by the underwriters to the Selling Stockholders in the Offering
backsC22C29 - S11primarycompany filing2026-10-26TDR's remaining 14,303,619 shares and the 22 October distribution in kindbacksC23
- S12primarycompany filing2026-10-01the October block sale noticebacksC23C40
- S13primarycompany filing2026-04-07beneficial ownership, the chief executive's pay, the pay measures and TDR's 2024 take-private proposalbacksC23C24C25C26
- S14secondaryvendor data2026-10-31Research pack (FMP market tier, FINRA short interest, Form 4 feed), frozen at content/TH/research-pack.jsoncontent/TH/research-pack.jsondaily closes, the analyst series, consensus estimates, insider trades, short interestbacksC03C24C27C28C29C40
- S15primarycompany filing2026-03-03Civeo Corporation Form 10-K for the year ended December 31, 2025 Civeo Corporationa rival's description of the market, the firms it competes with, and how headcounts fall after constructionbacksC31C35
- S16primarycompany filing2026-08-06CoreCivic, Inc. Form 10-Q for the quarter ended June 30, 2026 CoreCivic, Inc.the counterparty's account of the Dilley agreement and its termbacksC15
- S17primarygovernment data2026-10-01US Census Bureau, Value of Construction Put in Place, private construction seasonally adjusted annual rate (privsatime.xlsx) US Census Bureauprivate data-centre construction spending by month
Aug-26p | ... Data center 84950; Aug-25 | ... Data center 49052 (millions of dollars, seasonally adjusted annual rate)
backsC33 - S18primaryregulator2026-08-03Office of the Texas Governor, 'Governor Abbott Directs Comprehensive Data Center Audit' Office of the Governor of Texasthe Texas audit of data centres in the grid queue and the size of that queue
ERCOT is currently considering approximately over 474 gigawatts of requests to connect to the Texas grid
backsC34 - S19secondarynews2026-08-10Texas Public Radio, 'Gov. Greg Abbott touts tech company cooperation' Texas Public Radiohow the large data-centre builders answered the Texas audit
Gov. Greg Abbott is touting major technology companies' responses to his temporary moratorium on data centers
backsC34