TH: Trough Earnings and the Data Center Workforce Pivot

Target Hospitality (TH) filed its FY2025 10-K on March 11 showing exactly what you'd expect from a company that lost its best contract: revenue down 17%, EBITDA down 73%, net loss of $37M. The PCC government contract — $168M/year at 83% margins for housing immigrants in West Texas — was terminated in February 2025. That was the company's golden goose. It's dead.

What the 10-K also shows is that management didn't sit around mourning. They pivoted into workforce housing for AI data center and critical minerals construction, created a new segment (WHS), and signed $678M+ in minimum contracted revenue across six contracts. Two of those contracts — worth $152M combined — were signed in March 2026, after the balance sheet date, and appear in the subsequent events footnote. Three analysts cover this stock. Their mean target is $10.50. They haven't modeled the March contracts.

The question is whether TH at $9.70 is a trough-earnings turnaround or a value trap where an irreplaceable miracle contract gets replaced by structurally lower-margin business.

The Setup

The numbers tell the story of the trough:

Two years ago this company generated $344M in EBITDA at 61% margins. The PCC contract was doing the heavy lifting — immigration processing housing at government-guaranteed minimums with near-zero marginal cost. When that contract terminated, margins collapsed from 51% to 17%.

But the balance sheet survived. Management redeemed $181M in 10.75% senior secured notes in March 2025. TH is now debt-free with $183M in liquidity ($175M undrawn ABL + $8.3M cash). Interest expense dropped from $17M to $6M. No going concern. Clean EY audit. No material weakness.

The balance sheet is the floor. This company cannot die from financial stress. That's not a thesis — it's a constraint that removes the worst-case scenario from consideration.

The Pivot: WHS Segment

Three contract clusters make up the new WHS segment:

1. Lithium Nevada / Thacker Pass ($175M total, $111M minimum, through 2027)

TH is housing 700+ construction workers building the largest lithium mine in the US. Engineering is 93% complete, 950 workers on site ramping to 1,800 at peak in 2026. The DOE released the first $435M of a $2.26B loan even under the Trump administration (taking 5% equity in both LAC and the project itself — unusual terms, but the deal proceeded). GM owns 38% with a 20-year offtake. This project is past the point of cancellation.

The counterparty risk here is not default — it's pace. If lithium prices stay depressed, Phase 2 never happens. TH's contract covers Phase 1 construction only. After 2027 completion, those assets need redeployment. But through the contract period, this revenue is solid.

2. Data Center Community ($134M minimum, through May 2028, SW United States)

TH owns 1,050 beds housing workers building a data center. Customer identity is undisclosed — NDA suggests a major hyperscaler. Initial 250-bed facility was built in under 60 days (demonstrating rapid deployment capability), then expanded by 800 beds. Deferred revenue of $9.3M at year-end confirms advance payments received.

3. Power Generation — West Texas and Nevada ($187M minimum combined)

This is where it gets interesting. The West Texas Power Community contract ($129M, 47-month term, 1,400 workers) supports a "multi-gigawatt power-generation project for hyperscale AI-driven data-center development." The filing says TH needs only $2-5M of incremental capex because they're using existing infrastructure — the same Pecos-area modular units that used to house immigration workers.

We cross-referenced the contract description against announced projects. The customer is almost certainly the Chevron / GE Vernova / Engine No. 1 consortium building a $6.5B, 2.5-5 GW power plant in Reeves County, TX — directly adjacent to TH's existing Pecos infrastructure. Chevron filed tax abatement applications in Reeves County, FID expected early 2026, operations targeting 2027. The "multi-gigawatt" language matches perfectly.

Two additional contracts: Pecos NG Power Plant ($23M, 26 months, 400 rooms) and a Northern Nevada Power Community ($35M, 25 months, 250 beds). Both leverage existing infrastructure with minimal capex.

The Demand Wave Is Real

This isn't a one-customer story. At least 15 GW of data center power capacity has been announced in the Permian Basin alone:

• Chevron: 2.5-5 GW, Reeves County ($6.5B)
• Pacifico GW Ranch: 7.65 GW, Pecos County
• Poolside Project Horizon: 2 GW, CoreWeave anchor tenant
• Bolt/TPL (Eric Schmidt): 1+ GW across Permian
• IREN Sweetwater: 2 GW

Each GW-scale project needs 1,000-5,000 construction workers at peak. Remote Permian Basin locations have zero local housing stock. Contractors cite labor availability as the primary growth constraint. Workers are being imported across state lines.

Civeo (CVEO), TH's closest public comp, independently confirms the trend. CEO Bradley Dodson cited "accelerating data center construction activity" as a 2026 growth driver on the Q4 2025 earnings call (March 3, 2026). CVEO added new 10-K language about "growing demand for accommodations in connection with data center construction projects." But CVEO has zero disclosed data center contracts. TH has $286M+ in signed deals. First-mover advantage is real — TH has the beds, the infrastructure, and the contracts.

The Margin Question — Where the Thesis Lives or Dies

Here's the problem. WHS generated $96.8M in revenue at 21% adjusted gross margin in 2025. The PCC contract ran at 83%. The gap is enormous, and it's not going to close entirely.

The 21% was construction-heavy — TH was building the Lithium Nevada workforce hub and the initial data center community. Construction fee revenue is low-margin pass-through. As contracts transition from build to operate (services phase), margins should expand. Management explicitly guides to this: "margin improvement driven by WHS construction activity transitioning to higher-margin services operations."

But they also caveat it: "We cannot assure you that we will be able to deliver margin improvement through the effective servicing of the above mentioned contracts."

The honest answer is we don't know what WHS services margins will be. Historical PCC margins (83%) aren't applicable — different customer type, different contract structure, different competitive dynamics. Data center developers have more negotiating leverage than government immigration programs. Mining companies are cost-conscious. A realistic range for WHS services margins is 30-45%, but this is a guess, not a fact.

This is the doorway. Two patterns fit:

• 55% bull: WHS services margins normalize to 35%+, contracted pipeline delivers, 2027 EBITDA recovers to $120-150M
• 45% bear: WHS margins stay structurally low (22-28%), HFS-South ADR keeps declining, new contracts look good on revenue but don't replace PCC profitability

Q1 2026 earnings (May 11) is the first real data point. WHS margin above 28% = bull thesis on track. Below 25% = thesis in trouble.

The Factor Problem

Historical idiosyncratic variance is 43.3%. Way below the 75% target. The regression sees TH as an energy services small cap — correlated with oil activity, IWM, XLE.

But the forward thesis is structurally different from the backward regression. As WHS becomes 50%+ of revenue, the dominant driver shifts from Permian Basin energy activity to data center construction demand. The factor profile is migrating.

Seven factors drive returns from here:

Edge-weighted idiosyncratic variance: ≈29%. That's a sizing constraint. Where we have genuine informational advantage (WHS margin trajectory, balance sheet floor) represents only ≈30% of variance. The rest is systematic demand, energy cycles, and political risk.

Valuation: The Stock Already Moved

At $9.70, with ≈100.2M shares and no debt, enterprise value is ≈$964M. That's 18x trailing EBITDA ($53M) — expensive for a company losing money. At a 7x forward multiple (typical infrastructure services), the market is implying FY2026 EBITDA of ≈$138M. That's the bull case already priced in.

The entire C-suite bought stock in February 2026: CEO ($1.4M), CFO ($365K), General Counsel ($271K), and two officers. That's conviction. But the stock is up 22% in one week, RSI is 88, volume is 4.2x average. This is not a stealth accumulation — it's a crowded reaction to the 10-K.

Scenario analysis at $9.70 entry:

Probability-weighted EV: -17.8%. Negative. Asymmetry is wrong at this price — downside is 3x the upside.

Entry becomes mathematically interesting below $7.50 (EV turns positive at +7%). At $6.50, EV is +19% with bear case limited by the balance sheet floor. The balance sheet matters here — $0 debt, $183M liquidity, $530M total assets means the bear case is "disappointing growth company," not "existential risk." Liquidation value of modular housing inventory provides a floor around $3.50-4.00.

What We're Watching

May 11 — Q1 2026 earnings. The catalyst. First quarter with:

• WHS services revenue from Lithium Nevada (occupancy since Sept 2025)
• West TX Power and Pecos NG contracts active (started Mar/Apr)
• DIPC at full run-rate

Consensus: -$0.07 EPS. Low bar. WHS margin is the number. Above 28% = bull thesis lives. Below 25% = re-evaluate.

Unsolved gaps:

• Data center customer identity (SW US, $134M contract) — hyperscaler = strong, startup = risk
• Comparable transaction multiples for workforce accommodation peers
• HFS-South contract renewal terms and ADR trajectory

TDR Overhang

TDR Capital owns 65%. They tried to take TH private at $10.80 in 2024 — rejected. Stock is now $9.70, approaching that level. If data center contracts deliver, take-private becomes harder to justify at a lowball price. If the stock falls back to $5-7 during trough earnings, the risk resurfaces. CEO buying $1.4M of stock at current prices suggests management is not aligned with a lowball exit. We assign 15% probability to a renewed bid by year-end 2026.

Conclusion

The thesis is real. The setup is real. The price is wrong.

TH has a genuine trough-earnings inflection story with $678M+ in contracted revenue, a debt-free balance sheet, first-mover positioning in data center workforce housing, and a 15+ GW demand pipeline in its backyard. The entire C-suite is buying. CVEO independently confirms the demand trend. Chevron is almost certainly the West Texas Power customer.

But at $9.70, the market is already pricing $120-140M in forward EBITDA — the bull case. Entry here gives you 15% upside against 54% downside. That's not a trade.

This is a watchlist stock with alerts at $8.00 and $7.00, or a post-Q1-confirmation trade if WHS margins prove out and edge-weighted idio jumps from 29% to 50%+. The discipline is in waiting for the price to come to you, or for the data to justify the price.

The latent factor worth tracking: datacenter construction labor shortage as a recognized market factor. If workforce housing gets re-rated from "hospitality" to "datacenter infrastructure," TH's multiple expands. That's where the real edge might be — seeing a factor before it materializes. But that's a 2027 story, not a March 2026 entry.

Evidence