WhiteFiber: A Deep Value Investing Study

Cost structure and business nature. In 2025 the main direct costs of the cloud business were power, data center leases, GPU server leases, third-party customer support, and other operating costs; the main direct costs of colocation were power, leases, data center staff wages, and other costs. This is not an asset-light software business; it is a capital-heavy, power-heavy, equipment-heavy, financing-heavy infrastructure business. In 2025 segment gross profit was $47.63 million, but depreciation and amortization that year ran to $23.44 million, G&A to $52.51 million, and operating loss came to $26.82 million. WhiteFiber's "surface gross margin" does not equal freely distributable cash flow.

Industry stage and competitive landscape. Demand on the industry side is real. JLL projects global data center capacity will expand substantially between 2025 and 2030, driven mainly by AI and cloud; Brookings likewise notes that the growth of generative AI is raising reliance on high-performance, low-latency data centers. WhiteFiber positions itself as a "vertically integrated" AI/HPC infrastructure provider, trying to capture profit at both the cloud-compute and data center layers. The other side of the sector: leading rivals are expanding at scales far larger than WhiteFiber's. CoreWeave disclosed Q1 2026 revenue of $2.078 billion, a revenue backlog of $99.4 billion, and active power above 1GW; Nebius reported Q1 2026 revenue of $399 million, cash of $9.298 billion, and continues to expand aggressively; and in May 2026 Google and Blackstone announced a new TPU cloud company with an initial equity commitment of $5 billion and plans to bring 500MW of capacity online in 2027. So while WhiteFiber's industry is buoyant, the competitive barrier is not "rivals cannot get in" but "who has more capital, customers, power, chips, and execution capability."

Moat assessment. Through a Buffett-style lens, I do not currently see a wide moat. WhiteFiber has some factors that could become advantages: renewable-energy environments in Iceland and Quebec, management's retrofit-first data center development approach, the NC-1 capacity arrangement with Duke Energy, and the signed 10-year 40MW Nscale contract with total contract revenue of roughly $865 million in theory. But these look more like project resources and execution opportunities than a structural moat that durably supports pricing power and keeps competitors out. The brand is still weaker than the leaders, network effects are very limited, switching costs exist but are not irreversible, and distribution is not particularly exclusive; what truly matters remains capital, power, customer relationships, and execution. On balance, the moat looks like it is "trying to form" rather than "already formed."

Would I hold if the stock market closed for 5 years. At today's price and the company's current state, my answer is: no. Not because the industry lacks prospects, but because WhiteFiber has not yet shown long and stable enough real cash flow and governance quality for me to be comfortable acting as a "business owner" without market liquidity.

Scores. Business understandability: 3/5. Industry attractiveness: 3/5. Moat strength: 2/5.

Management and Capital Allocation

Is management trustworthy. WhiteFiber CEO Sam Tabar and CFO Erke Huang continued to serve concurrently at parent Bit Digital for a period after the IPO, and under a transition services agreement both were expected to spend up to about 30% of their working time on Bit Digital matters. WhiteFiber also clearly discloses that potential conflicts may arise between Bit Digital and WhiteFiber over personnel, transactions, service pricing, M&A, and capital arrangements. For long-term minority shareholders, this means the governance structure is controlled-shareholder in character from the outset, not a fully independent, owner-oriented setup.

Equity alignment. As of March 1, 2026, the CEO held roughly 147,000 shares, the CFO about 88,000 shares, and total management and directors (8 people) only 268,000 shares; meanwhile Bit Digital held 27.04 million shares, about 70.5% of the voting power. The party that truly controls the company's fate is the parent, not a structure of high co-ownership between management and public shareholders. From the standpoint of aligning incentives with minority shareholders, this is not an ideal configuration.

Capital allocation record. The company's three core capital moves over the past two years are: acquiring Enovum in Q4 2024 to enter colocation; raising capital in the 2025 IPO; and issuing $230 million of 4.5%-coupon convertible notes due 2031 in January 2026, netting roughly $102.5 million, while simultaneously spending $120 million to buy zero-strike call options. These moves show management clearly knows it needs to "finance fast, expand fast, secure sites fast, and build fast," with a capital-allocation logic of typically grabbing scale and capacity first, then fighting for future profit. This is not uncommon in growth infrastructure, but it does not fully align with the per-share intrinsic-value growth that Buffett-style rational capital allocation pursues.

Dividends, buybacks, M&A, and dilution. WhiteFiber has no explicit dividend commitment and no substantive buyback plan; at this stage cash goes mainly to reinvestment and construction. On the other side, the company recognized $16.87 million of stock-based compensation expense in 2025 and another $7.317 million in Q1 2026; at year-end 2025 there were still 2.791 million shares available for future grants under the 2025 Plan. Share count also rose from 27.04 million at year-end 2024 to 38.34 million at year-end 2025, an increase of about 41.8% in one year. This does not mean the company did something wrong, but it does mean the bar for "per-share value" growth has been materially raised for shareholders.

Management candor. One point to credit: the company does not dodge the risks. In its 10-K and 10-Q it explicitly states the initial-customer service suspension, the prepayment and receivable negotiations, customer concentration, dependence on Bit Digital, and future financing needs. The disclosure itself is candid. But candid disclosure is not completed verification; investors still need to see these issues "resolved," not merely "explained."

Score. Management and capital allocation: 2/5. Governance conflicts, control-shareholder dominance, elevated dilution, and capital allocation still in a "prove-it" phase are the main reasons I do not give a higher score.

Financial Quality and Owner Earnings

First, the core financial table.

Note: 2024 and 2025 data are from WhiteFiber's 2025 10-K; Q1 2026 data are from the Q1 2026 10-Q; some margins and free cash flow figures are my calculations from the raw data.

Revenue growth is real, but the quality still needs verification. In 2025 revenue rose from $47.64 million to $79.16 million, with cloud services up from $45.73 million to $68.75 million and colocation up from $1.36 million to $8.91 million; Q1 2026 revenue rose further to $21.92 million, with colocation growing faster, mainly attributable to MTL-3 coming fully online. The growth itself is not a problem, but this is still the "fast ramp off a small base" stage, and growth depends heavily on customer deployments, GPU installs, site energization, and contract delivery.

Margins are not pretty. In 2025 segment gross profit reached $47.63 million, but with D&A of $23.44 million and G&A of $52.51 million, operating profit turned to -$26.82 million; in Q1 2026 G&A rose again to $17.77 million, which the company discloses consists mainly of employee stock-based compensation, salaries, and $6.20 million of professional and consulting fees, widening the quarterly operating loss to $11.02 million. This shows WhiteFiber's problem is not "no revenue" but rather that "the loss between revenue and what actually reaches shareholders is still large."

Cash flow looks fine on the surface but must be unpacked. Operating cash flow was $45.66 million in 2025, clearly above the net loss, and Q1 2026 operating cash flow was still positive. At first glance that is good, but once unpacked: 2025 operating cash flow was driven by $23.44 million of D&A, $16.87 million of stock-based compensation, and a $48.75 million increase in deferred revenue; Q1 2026 was supported by a $64.30 million increase in deferred revenue, while receivables rose $67.81 million. The company also discloses that the large Q1 receivable increase relates to new customers, non-recurring charges, and the timing of billing and collection; the deferred revenue increase is mainly because customers made $65.70 million of prepayments for future cloud and data center services. My read: WhiteFiber's cash flow is not fictitious, but it is still far from "the kind of steady, freely distributable inflows a mature business produces."

Free cash flow is very poor. In 2025, capex plus equipment/project deposits consumed $268.4 million, producing free cash flow of roughly -$222.7 million; in Q1 2026 the company invested another $169.2 million in fixed assets, for quarterly FCF of about -$139.9 million. This shows the company at this stage is not "growth producing cash" but "growth continuously consuming cash, dependent on capital markets and customer prepayments." From the standpoint of permanent capital loss, this is a very important red line.

The balance sheet is not so fragile as to be immediately dangerous, but it is far from comfortable. At end-Q1 2026 the company had about $80.11 million in cash and restricted cash, $352.6 million in shareholders' equity, $222.3 million in convertible debt carrying value, and $443.7 million in total liabilities. Short-term solvency looks fine, but it faces ongoing high capex and construction needs, so the real risk is not "no money right now" but "whether it must keep financing over the next few years, and at what cost." Because TTM EBIT is still negative, net debt/EBITDA and interest coverage offer no sense of safety.

Accounting quality and comparability. Three points here. First, WhiteFiber's historical financials come from the Bit Digital carve-out, and before the IPO many corporate-service costs were parent allocations, which naturally weakens the standalone comparability of historical periods. Second, the company has sales-type leases and net investment in leases, which makes a small portion of revenue closer to interest income than to traditional service revenue. Third, stock-based compensation is sizable, and if investors simply add all of it back to profit they will overstate true owner earnings. On balance, I see no clear sign of fraud, but I cannot treat this as already "clean, simple, and mature" financial statements.

Owner Earnings estimate. Following Buffett's approach, I care more about "cash attributable to shareholders and distributable without harming competitive position." Net income in 2025 was -$24.68 million; adding back $23.44 million of D&A brings it near breakeven; but if you do not treat stock-based compensation as a "free lunch" and further deduct what I conservatively estimate at least $15 million to $25 million of maintenance capex, then 2025 owner earnings should still be negative or near zero. Given the wider Q1 2026 operating loss and continued heavy project investment, I lean toward putting WhiteFiber's current conservative owner earnings at -$20 million to $0 per year, not a positive number. So the question "how many times owner earnings does the current valuation represent" has, on a conservative basis, the answer: meaningless, or extremely expensive.

Valuation and Margin of Safety

Start with a static read of the current valuation. At a $25.52 share price and roughly $980 million market cap, combined with end-Q1 2026 cash of $75.78 million, $222.3 million of convertible debt, and my estimated TTM revenue of about $84.32 million, WhiteFiber currently corresponds to roughly 11.6x P/S and 13.4x EV/S. At the same time, with TTM EBIT negative and TTM FCF deeply negative, metrics like P/E, EV/EBITDA, and P/FCF are either distorted or meaningless. For a company that has not yet proven owner earnings to be positive and stable, this valuation is not cheap.

Method one: discounted owner earnings. Because current owner earnings are insufficient to capitalize directly, I can only produce a "forward, explicitly assumed" valuation rather than pretend this is a precise answer. All the valuations below are assumption-driven, not facts.

The core premises behind these scenarios are: the Nscale contract delivers smoothly, deployed GPUs reach higher utilization, customer concentration drops markedly, G&A as a share of revenue falls, and future financing is completed without heavily dilutive discounts. If these conditions cannot all be met, the conservative or neutral scenarios are more representative. At today's $25.52, the share price sits roughly in the optimistic-scenario range and leaves no margin of safety for conservative investors.

Method two: relative valuation. Comparing WhiteFiber with several larger, more mature, or at least more market-validated AI infrastructure and related players makes this clearer. CoreWeave currently has a market cap of about $52.599 billion, 2025 revenue of $5.131 billion, Q1 2026 revenue of $2.078 billion, and a revenue backlog of $99.4 billion; Nebius had Q1 2026 revenue of $399 million, cash of $9.298 billion, and debt of $8.450 billion; Applied Digital had $2.1 billion of cash and restricted cash and $2.7 billion of debt at the end of fiscal Q3 2026, with quarterly revenue of $126.6 million. In other words, WYFI's sales multiple today is already close to, and not clearly lower than, those of several peers with larger scale, stronger financing capacity, and more customer resources; this does not say those peers are cheap, only that WYFI is not cheap either.

Method three: asset or liquidation value. At end-Q1 2026 WhiteFiber's shareholders' equity was $352.6 million; stripping out goodwill of $19.81 million and intangible assets of $12.43 million leaves a net tangible book value of about $320.4 million, or roughly $8.3 per share. But this is only book value, not liquidation value. Its fixed assets, construction-in-progress deposits, and specialized data center assets are worth a great deal in good times, and may discount heavily in bad times; with convertible debt and lease liabilities ahead of equity, book value's protection for the share price is limited. My view is that WhiteFiber's "hard assets" provide some floor, but that floor is far softer than it looks on the books.

Price range judgment. Based on the three methods above, I give the following ranges: Conservative intrinsic value range: $4–8 per share. Fair intrinsic value range: $10–16 per share. Optimistic intrinsic value range: $20–30 per share. Ideal buy price range: $8–12 per share. Acceptable holding price range: $12–18 per share. Clearly overvalued price range: above $20, especially when near or above the current $25.52. These are my inferences and opinions, not financial-statement facts.

Margin-of-safety conclusion. WhiteFiber's biggest valuation vulnerability today is not "capital-market sentiment" but the question of whether future owner earnings turn positive and keep rising. If growth comes in below expectations, margins fail to improve, or the valuation multiple falls back from a high-growth premium, investors who bought "a good story" may still suffer real and lasting capital loss. I therefore consider the current margin of safety inadequate.

Risks, Comparisons, and Checklist

Most important risks and the bear case. The strongest bear logic is actually simple: WhiteFiber is not a "cash cow" that has proven its business model, but a growth infrastructure company that still needs continuous financing, continuous construction, and continuous new-customer wins to support its valuation. Once customer concentration recurs, NC-1 slips, power or install progress disappoints, or industry competition intensifies as incumbents and new entrants compress pricing, the current valuation could quickly lose support. Worse, the company's short independent operating history, governance under parent control, and elevated stock-based compensation and potential dilution will amplify the damage of execution missteps to per-share value.

The permanent-capital-loss scenario I worry about most. The prepayments, receivables, or termination fees the initial customer ultimately contributes come in below expectations; Nscale/NC-1 collections and power plans are delayed; the company keeps financing through high-cost debt or low-priced issuance; and large capex fails to convert into stable owner earnings. If these occur together, WhiteFiber could be quickly re-rated from a "high-growth infrastructure stock" into a "project stock with sunk capital, uncertain returns, and ongoing financing needs," and a 60%–80% permanent loss in the share price would not be alarmist. This figure is my risk inference, not company guidance.

What facts would overturn the investment judgment. If the following occur over the next 4 to 6 quarters, I would admit my cautious judgment was wrong: first, customer concentration falls clearly, with any single customer's revenue share dropping into a reasonable range; second, NC-1 energizes and bills on time, with deferred revenue steadily converting into high-quality revenue and cash; third, G&A as a share of revenue falls clearly and stock-based compensation is controlled; fourth, without heavily diluting existing shareholders, the company converts revenue growth into sustained positive owner earnings and free cash flow at least near breakeven. Conversely, if the initial-customer dispute widens, major long-term contracts slip, capex runs out of control, or Bit Digital related-party governance deteriorates, the entire investment thesis must be re-examined.

Comparison with other opportunities. Against its strongest peer competitor CoreWeave, WhiteFiber is clearly weaker on scale, customers, funding, backlog, active power, and market position; against a broad index like SPY, WhiteFiber offers highly concentrated single-company execution risk rather than diversified U.S. corporate earnings growth; and against the 4.67% U.S. 10-year Treasury yield on May 19, 2026, WhiteFiber's current price asks investors to take far more risk than a bond while not yet demonstrating sufficiently clear distributable cash flow. For an investor with a horizon of 10 years or more and a balanced risk appetite, I do not see it clearly superior to the index today, let alone clearly superior to a lower-risk asset.

Investment Checklist

The judgments above are based on the preceding financial, governance, competitive, and valuation analysis.

Final Investment Conclusion

【Final Rating】 Watch

【One-Sentence Investment Thesis】 WhiteFiber stands in a genuinely growing AI infrastructure sector, but today it looks more like a high-volatility growth ticket that must continuously deliver on four promises of construction, financing, customers, and governance, than a proven "good business" that durably creates owner earnings.

【Core Bull Case】 The sector's demand is strong, and the long-term expansion trend in AI and data centers is clear; revenue grew fast over the past two years, up about 66% YoY in 2025 and still up about 31% YoY in Q1 2026; the company has expanded from a single GPU cloud service into colocation and signed Nscale's 10-year 40MW long-term contract, showing it is not entirely without customer hooks and an expansion path; segment gross margin held roughly steady in the high-50% to low-60% range across 2024–Q1 2026, indicating the underlying business units are not without economics; and if future customer diversification and capacity delivery both succeed, the current business could move toward positive owner earnings.

【Core Bear Case】 Revenue is highly concentrated and the initial customer's service is in post-suspension negotiation, the biggest single-point risk; free cash flow remains deeply negative, with growth clearly dependent on capex, external financing, and customer prepayments; governance is not ideal, with Bit Digital holding about 70.5% of the voting power and executives serving dual roles during the transition period; stock-based compensation and potential dilution are elevated, with share count up sharply in 2025; and the current valuation is not cheap, with an insufficient margin of safety.

【Key Assumptions】 For the investment to hold, several hard conditions must be met: NC-1 and other expansion projects must be delivered on time and begin billing; customer concentration must fall markedly; deferred revenue must gradually convert into high-quality cash revenue rather than one-off prepayment froth; G&A and stock-based compensation as a share of revenue must fall; and follow-on financing must not be completed at heavy discounts and dilution. If any one of these fails to hold over the long term, the valuation should be marked down clearly.

【Fair Buy Price】 The cool-headed range I give is $8–12 per share; if you are willing to accept higher industry-growth assumptions, $12–15 per share can be viewed as a barely discussable range; above $20, I consider it largely reflecting optimistic expectations rather than a margin of safety. The basis is the three-scenario owner-earnings valuation, the limited protection of net tangible assets, and the fact that the current sales multiple is not low.

【Target Holding Period】 If the fundamentals are validated in the future, this kind of infrastructure company should be held on at least a 5–10 year framework; but as of now, I would rather place it on a watch list than directly lock it up for the long term.

【Expected Annualized Return】 Near the current price, my subjective estimate is: the conservative scenario could deliver an annualized return of -10% to -20%; the neutral scenario roughly -3% to +5%; and only the optimistic scenario, in which projects, customers, and financing deliver almost in sync, could push the annualized return to 12% to 18%. These are all inferences based on the preceding valuation ranges, not market consensus or company guidance.

【Maximum Loss Risk】 In the worst case, if the customer dispute, project delays, deteriorating financing, and a valuation de-rating all occur together, I would not be surprised to see the share price revert toward net tangible assets or even lower, and a 60%–80% permanent capital loss risk must be taken seriously.

【Tracking Metrics】 Going forward I will focus on: the single largest customer's revenue share; the final settlement amount and terms of the initial-customer dispute; NC-1's energization, install, and billing progress; the increase and release quality of deferred revenue; the speed of receivables collection; quarterly capex and financing arrangements; the G&A-to-revenue ratio; changes in stock-based compensation and total share count; whether segment gross margin holds; and whether sustained positive owner earnings or at least markedly improved free cash flow appears.

【Signals Triggering a Reassessment】 If the following occur, the thesis must be re-examined: customer concentration deteriorates again; the resolution of initial-customer receivables and prepayments comes in markedly worse than expected; long-term contracts slip or MW delivery is delayed; large further equity dilution occurs; high-cost financing appears again after the convertible notes; or project expansion's revenue fails to convert into profit and cash flow. Conversely, if the customer mix keeps improving, NC-1 delivers, FCF nears positive, and the G&A ratio falls, the valuation framework should also be marked up.

【Final Recommendation】 Look at WhiteFiber as a company, not as an AI-concept ticket. From the standpoint of a long-term business owner, my current verdict is: worth tracking continuously, but not worth forcing a "value investing" buy at the current price. If it can in the future convert its "growth story" into "real owner earnings, verifiable governance quality, and lower customer concentration," then even if the price is not the lowest, the risk-reward could be better; for today, the more disciplined choice remains: watch, and wait for facts to outweigh the story.

Open questions and limitations: The independent operating history is too short, lacking 5–10 years of comparable financials; maintenance capex is not separately disclosed, so owner earnings can only be estimated conservatively; some historical periods include parent-company allocations and carve-out effects; and many of today's bull arguments depend on future projects and contracts delivering rather than on long-term cash returns already accumulated.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.