Stock Analysis · Bulls vs Bears

WhiteFiber, Inc. WYFI

🐂 Bulls vs Bears 🐻Report #212 Oct 2026
★ Willem owns this stock · 1.8% of the portfolio
70
high risk▲ +5 vs last
🐂 9vs11 🐻
Bears lead
50%
▲ Chance of the biggest tailwind the forecasts miss
70%
▼ Chance of the biggest headwind the forecasts miss
The verdict

The tenant side of the story got stronger. Nscale now has $1.48B of cash, a $3.36B pre-IPO round and an investment-grade-rated loan for its North Carolina servers. The funding side got weaker: the 10-Q now warns that existing resources "may not be sufficient" to fund both the pipeline and near-term debt, the NC-1 project loan is still unsigned three quarters late, and the parent is openly weighing share sales. The…

🎙️ Bulls vs Bears podcast — WhiteFiber, Inc.
Alex and Sam debate every bull and bear argument · 14:36
0:0014:36
⤓ Download

Written by Claude (Anthropic’s AI) from public sources · Willem owns WYFI · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 9

  1. 1The ten-year, $865M NC-1 lease has started billingStrong
  2. 2The tenant's balance sheet is now much strongerStrong
  3. 3Contracted revenue is worth more than the whole companyMedium
  4. 4The cloud business now runs on multi-year, prepaid contractsMedium
  5. 5GPU rental prices rebounded in 2026, lifting new contract pricing (new)Medium
  6. 6The powered NC-1 site can grow to 99 MW and beyondMedium
  7. 7The August convertible bought time at a 25% premium (new)Medium
  8. 8The stock trades near its IPO price, with untouched analyst targetsWeak
  9. 9Retrofitting old factories is proven at MTL-3 and repeatable at NC-2/NC-3Medium

🐻 Bear arguments 11

  1. 1The company's own 10-Q flags a funding gapHigh
  2. 2The NC-1 project loan is three quarters late and still unsignedHigh
  3. 3Cash burn is deep, and prepayments flatter operating cash flowHigh
  4. 4One tenant holds about 93% of the backlogMedium
  5. 5Underlying cloud revenue is shrinking behind a one-off feeMedium
  6. 6Shares are up 18% in eleven months, with more overhangMedium
  7. 7Bit Digital is now openly weighing sales of its stake (new)Medium
  8. 8Related-party bridge loans cost about 50% a yearMedium
  9. 9Dated promises slip, quarter after quarterMedium
  10. 10A small player whose tenant is also a competitorMedium
  11. 11Power tariffs are tightening in North Carolina and QuébecLow

Bull arguments show how well they survive the devil’s advocate (strong / medium / weak); bear arguments show the damage they could do after the counter-argument (high / medium / low).

Bull case

🐂 Bull #1The ten-year, $865M NC-1 lease has started billing

  • The point: The single contract that turns WhiteFiber from a GPU reseller into a landlord has moved from paper to invoices.
  • Evidence:
    • Enovum NC-1 Bidco signed the Services Agreement with Nscale in November 2025. The initial service order is about $865M over 10 years for 40 MW of IT load, including escalators and one-off charges, with electricity passed through [1].
    • The September deck says initial billing has started [4].
    • On the Q2 call (2026-08-12) about 20 MW was available. Management said full 40 MW run-rate billing would come "by end of August" and that the switchgear problems were solved [12,13].
    • Bit Digital's CEO repeated on 2026-08-13 that WhiteFiber "expects to reach full contracted run-rate billing later this month" [16].
    • Colocation RPO by year at 2026-06-30 [1]:
      • H2 2026: $36.1M.
      • 2027: $93.5M. 2028: $94.8M. 2029: $95.3M. 2030: $94.0M.
      • Thereafter: $519.2M.
  • Why it matters: About $93–95M a year of contracted colocation revenue from 2027 is roughly the size of WhiteFiber's entire LTM revenue ($94.5M) [4]. Steady rent of that kind is what lenders will finance against.
  • 😈 Devil's advocate:
    • Full 40 MW billing is still not confirmed in any filing. The last hard data point is "initial billing" [4].
    • "Full ramp" is "subject to completion… and customer acceptance" [4]. The ramp has already slipped four times (Bear #9).
  • Verdict: Strong The contract is the best-evidenced asset. The Q3 10-Q should show whether billing reached run-rate (about $7M a month implied).

🐂 Bull #2The tenant's balance sheet is now much stronger

  • The point: The main fear in report #1 was that Nscale, a loss-making pre-IPO company, could fail to pay. Nscale's own filing now shows a far better-funded tenant.
  • Evidence (Nscale S-1, 2026-09-18 [19]):
    • Cash $1,478.4M and debt only $137.4M at 2026-06-30.
    • A $1.2B delayed-draw North Carolina GPU facility (SOFR+2.375%, matures December 2031), rated investment grade.
    • $103.4B of active and contracted contract value, with a weighted life of 5.7 years. This includes Microsoft (up to $43.8B) and Anthropic (up to $44.6B).
    • It then raised $3.36B in a pre-IPO convertible round led by Third Point, with Nvidia committing $1B (2026-09-25) [20].
    • The IPO filing (NYSE: NSCL; Goldman, JPMorgan, Morgan Stanley) is public, but there is no price range yet (2026-10-08) [21].
  • Why it matters: NC-1 rent is about $86M a year. That is small against Nscale's cash and financing, so the risk of a missed payment in the next 2–3 years looks low.
  • 😈 Devil's advocate:
    • Nscale lost $1,020.1M in H1 2026 on $140.6M of revenue, and its biggest customer was 52% of revenue [19].
    • The S-1 does not name WhiteFiber, Madison or the NC-1 end user [19].
    • Nscale's contracts average 5.7 years, while WhiteFiber's lease runs 10. The second half of the lease relies on Nscale finding a new end user.
  • Verdict: Strong for the next 3 years, Medium beyond that. The tenant can pay today. Whether it will want this building in 2031–2035 is unknown.

🐂 Bull #3Contracted revenue is worth more than the whole company

  • The point: The market value is below the contracted revenue already signed.
  • Evidence:
    • Total RPO is $1,008.0M at 2026-06-30: colocation $932.9M and cloud $75.1M [1].
    • Not yet in RPO: more than $540M of new cloud contracts signed since May [1,4]:
      • BaseTen: $165.2M over 3 years.
      • Prime Intellect: $108.2M (Vera Rubin).
      • Iceland: $87.5M over 5 years.
      • Paris: more than $160M over 5 years.
      • A $16.0M order.
    • Market value is about $674M (45.12M shares × $14.94 on 2026-10-12) [5,27].
    • B. Riley (a banking client of WhiteFiber) values NC-1 alone at about $1.56B (20x about $78M of net operating income) [report #1].
  • Why it matters: If the contracts are delivered and financed at sensible rates, equity value should follow the contracted cash flows.
  • 😈 Devil's advocate:
    • RPO is revenue, not equity value. Enterprise value is about $0.86–1.0B (estimate) once you add about $466M of debt and subtract cash after Q3 burn.
    • The cloud contracts need hundreds of millions of dollars of GPUs.
    • Two of the last three big cloud customers ended early (Bear #5).
    • Interpretation: the "backlog > market value" headline overstates the gap.
  • Verdict: Medium The contracted base is real, but much of it has to be funded before it can be earned.

🐂 Bull #4The cloud business now runs on multi-year, prepaid contracts

  • The point: Since May, cloud deals run 3–5 years. The large ones are prepaid or financed at project level.
  • Evidence:
    • BaseTen: 1,392 B300 GPUs, 36 months, start November 2026 [1].
    • Iceland: 576 B300 GPUs, 5 years, start December 2026, with a revenue share on top [1].
    • Prime Intellect: 576 Vera Rubin GPUs, 36 months, start Q2 2027 [1].
    • Paris: 12 months of fees prepaid. The term sheet is binding, and the project-level financing "would not be guaranteed by WhiteFiber, Inc." [1].
    • Deferred revenue rose from $79.6M to $143.1M in six months [1].
    • Management says the new cloud contracts give more than $200M of annualised revenue once fully deployed [12].
  • Why it matters: Prepayments plus non-recourse debt shrink the equity each GPU cluster needs, and lock in pricing for 3–5 years.
  • 😈 Devil's advocate:
    • After the initial term, each order "automatically renews for successive one-month periods unless terminated by either party" [1].
    • The France financing still had no definitive documents at 2026-08-12. Material terms were "subject to ongoing negotiation" [1].
    • No Paris start has been confirmed since the September target.
    • The customers are young AI companies.
  • Verdict: Medium The structure is better than in 2024–25, but none of the four big deployments is confirmed live.

🐂 Bull #5GPU rental prices rebounded in 2026, lifting new contract pricing (new)

  • The point: GPU rental prices bottomed in late 2025 and then rose, which supports pricing on WhiteFiber's new contracts and on its legacy fleet.
  • Evidence:
    • SemiAnalysis: the 1-year H100 contract price went from a low of $1.70 in October 2025 to $2.35 in March 2026, about +40% [36].
    • Silicon Data's B200 index was about $5.62 entering September, +27.6% year to date (secondary) [38].
    • Implied pricing on WhiteFiber's new contracts (TCV ÷ GPUs ÷ hours) [1]:
      • About $4.5 per GPU-hour on BaseTen's B300s.
      • About $3.5 on Iceland's B300s.
      • About $7.2 on Prime Intellect's Vera Rubin GPUs.
  • Why it matters: Higher prices raise the return on each GPU dollar and make prepayments easier to negotiate.
  • 😈 Devil's advocate:
    • In 2024–25 hyperscaler H100 prices fell about 33% in a year ($9.34 → $6.26) and marketplace prices went below $2 [37].
    • A forward market (Kalshi, secondary) prices B200 about 9% lower by March 2027 [38].
    • The cycle can turn fast.
  • Verdict: Medium It is a real tailwind today, but it is cyclical and partly reflected in the contracts already signed.

🐂 Bull #6The powered NC-1 site can grow to 99 MW and beyond

  • The point: Grid power is the scarce input, and NC-1 has a contractual path to more of it.
  • Evidence:
    • Duke has delivered 54 MW gross [4].
    • The Capacity Agreement targets 99 MW by May 2029 [1].
    • Management says about 200 MW more is under utility evaluation, and the site could reach about 300 gross MW over time [12,14].
    • Nscale holds priority rights on future capacity [4].
    • CBRE: primary-market vacancy is 1.4%, 80.4% of capacity under construction is pre-leased, and asking rents for 10+ MW deals rose 6.7% in H1 2026 [35].
  • Why it matters: Each MW at an existing powered site is cheaper and faster than greenfield.
  • 😈 Devil's advocate:
    • On the Q2 call, management said Phase 2 "is not imminent" [12].
    • Duke has given no delivery schedule.
    • Duke's large-load settlement (2026-10-06) would apply to new agreements signed after 2026-06-01 for loads of 50 MW or more at 80%+ load factor. It requires a 10-year minimum term, minimum bills at 75% of contract demand, and upfront funding of dedicated facilities [32].
    • The existing Duke agreement can be cancelled by either party on 60 days' notice [1].
  • Verdict: Medium The power path exists, but the next tranche is a 2027–28 story and may cost more.

🐂 Bull #7The August convertible bought time at a 25% premium (new)

  • The point: WhiteFiber raised long-dated money without selling shares at the low.
  • Evidence:
    • $310M of 5.00% notes due 2032-09-01, with a conversion price of about $33.84, a 25% premium [2].
    • The first holder put is 2030-09-06 [2].
    • About $180M of net cash remained after exchanging $198.15M of the 2031 notes [4].
    • No convertible matures, and no holder put falls, before 2029-02-06 [1].
    • There is no going-concern language, and disclosure controls were judged effective at 2026-06-30 [1].
  • Why it matters: The conversion price is more than twice today's price, so it only dilutes if the stock recovers. It takes away a 2026–27 refinancing wall.
  • 😈 Devil's advocate:
    • The exchange cost about $118.5M of cash plus about 6.3M new shares (+16% at once) [2,5].
    • Total convertible principal is now about $342M against a $674M market value.
    • At recent burn rates (Bear #3), about $180M covers roughly two quarters of capex.
  • Verdict: Medium It bought 2–3 quarters, not a funded plan.

🐂 Bull #8The stock trades near its IPO price, with untouched analyst targets

  • The point: The share price has reset. Expectations on the stock are low, while estimates have not been cut.
  • Evidence:
    • $14.94 on 2026-10-12, against a 52-week high of $46.87 (2026-06-22) and an IPO price range of $15–17 [27,22].
    • 11 analysts, 9 Buy, average target $38–40, range $32–50. No downgrade other than Weiss in six months [22,23].
    • FY2027 consensus EBITDA is about $131M (StocksGuide; Needham $175.2M) [25,26].
    • Estimate: EV/FY2027 EBITDA is about 6.6–7.6x (EV $0.86–1.0B).
  • Why it matters: If NC-1 and the cloud contracts deliver, the multiple is low for contracted data-centre cash flow. Peers with hyperscaler leases trade far higher.
  • 😈 Devil's advocate:
    • Targets that ignore a 68% fall usually come down after the fact.
    • The FY2027 estimates (revenue $261–296M) assume every deployment arrives on time.
    • Short interest is about 13% of float and beta is about 6.3 [28,22].
    • The September–October slide (about $20 → $14.94) has no identified cause, so a reason may still surface.
  • Verdict: Weak to-Medium. The stock is cheap if delivery happens, but it is cheap for reasons Bears #1–#3 explain.

🐂 Bull #9Retrofitting old factories is proven at MTL-3 and repeatable at NC-2/NC-3

  • The point: Converting old industrial sites that already have grid connections is faster and cheaper than building new.
  • Evidence:
    • MTL-3 was delivered in about 6 months at under C$9M per MW [4].
    • The company targets $8–10M per gross MW, "up to 40% cheaper than greenfield" [4].
    • NC-2/NC-3: two ex-Unifi plants in Yadkin County for $60M cash. They start at 60 MW gross and could reach about 200 MW. Ready-for-service is targeted for Q3 2027, and there are non-binding LOIs with investment-grade credit support [7].
  • Why it matters: If NC-1's model repeats, value compounds per site.
  • 😈 Devil's advocate:
    • NC-1's cost per MW is still undisclosed, and management declined to give a capex number [14].
    • NC-2/NC-3 still need an energy study and separation approvals before closing. Closing is about end-October at the earliest (estimate) [7].
    • 60 MW at $8–10M per MW is $480–600M of capex, against a $674M market value.
  • Verdict: Medium for the method, Weak for the funding.

Bear case

🐻 Bear #1The company's own 10-Q flags a funding gap

  • The point: WhiteFiber states it may not be able to fund both its pipeline and its near-term debt without raising more capital.
  • Evidence:
    • Q2 10-Q: "our existing capital resources may not be sufficient to fund both this pipeline and our near-term debt maturities without accessing additional capital, whether through indebtedness, equity or equity-linked securities, project-level financing, or otherwise" [1].
    • Working capital swung from +$85.2M to −$28.6M in six months, "primarily driven by the classification of certain indebtedness as current" [1].
    • The 10-K said that without more debt WhiteFiber "will not have sufficient funds to retrofit NC-1" or reach 99 MW [9].
    • The NC-2/NC-3 purchase ($60M) and four GPU deployments are still to be funded [7,1].
  • Why it matters: Each extra dollar raised at about $15 a share dilutes heavily. Each dollar borrowed adds to an already large debt load.
  • 🛡️ Counter-argument:
    • The August convertible added about $180M net [4].
    • Cloud deployments are meant to be funded by prepayments and project debt [1].
    • Management cites positive adjusted EBITDA and operating cash flow [1].
    • There is no going-concern language.
  • Severity: High The company says it in its own words, and the counter-argument depends on two financings (NC-1, France) that have not closed.

🐻 Bear #2The NC-1 project loan is three quarters late and still unsigned

  • The point: The financing that was supposed to return equity and repay the bridge has slipped every quarter.
  • Evidence:
    • Targets so far [47,15,14,12,4]:
      • "Q1 2026" (November 2025).
      • "Q2 2026, taken longer than expected" (2026-03-26).
      • "Moving toward a near-term financing solution" (2026-05-14).
      • Lender exclusivity, with "no assurance… on favorable terms or at all" (2026-08-12).
      • Still "expected" (deck, 2026-09-14).
    • B. Riley's "within weeks" (2026-09-10) has passed without an 8-K. EDGAR shows no filing after 2026-09-14 [17].
    • Management wants a structure "financeable from day one" and is "solving for the lowest cost of capital" [12].
  • Why it matters: Until it closes, NC-1 equity (about $150M or more) is trapped, the 50%-effective bridge stays outstanding, and new raises are more likely.
  • 🛡️ Counter-argument:
    • The tenant's own NC facility got investment-grade ratings (Bull #2), which should help lenders get comfortable.
    • An earlier summary cited a target of about 75% loan-to-value on the campus [47].
    • A delay to improve terms is not a failure.
  • Severity: High It is the single most important missing piece. Each further slip raises the odds of an equity raise.

🐻 Bear #3Cash burn is deep, and prepayments flatter operating cash flow

  • The point: Free cash flow is deeply negative. The positive operating cash flow comes from customer prepayments, not from profits.
  • Evidence (from filings [1,10,11,9]):

    • Quarterly free cash flow (operating cash flow minus capex and deposits):

      Quarter Free cash flow
      Q3-25 −$32.5M
      Q4-25 −$52.5M
      Q1-26 −$165.9M
      Q2-26 −$89.7M
      LTM ≈ −$341M on $94.5M revenue

    • H1-26 operating cash flow of +$89.1M included +$63.6M of deferred revenue. H1 net loss was −$27.0M.

    • H1 G&A was $32.6M against $50.7M of revenue. That included $11.0M of stock compensation, about $2.2M of bad debt and a $5.0M software write-off.

    • H1 cash interest paid was only $32K against $8.0M of interest expense. Most interest is accrued or paid in kind.

  • Why it matters: Prepayments are revenue spent in advance. They turn into service obligations with no matching cash later.
  • 🛡️ Counter-argument:
    • Most of the capex is NC-1. Once NC-1 is complete and financed, colocation earns about $93–95M a year at high margins (Bull #1).
    • Interpretation: burn should fall sharply in 2027.
  • Severity: High in 2026–27, Medium after that.

🐻 Bear #4One tenant holds about 93% of the backlog

  • The point: Almost all contracted value depends on one counterparty.
  • Evidence:
    • Colocation RPO is $932.9M of $1,008.0M total (93%) [1]. Nscale is the "sole contracted tenant" at NC-1 [1].
    • One customer was about 63% of H1 2026 revenue. There are seven customers, and three orders were discontinued in H1 (about $5.1M impact) [1].
    • Nscale's own largest customer was 52% of H1 revenue [19].
    • Nscale's contracts average 5.7 years, against WhiteFiber's 10-year lease [19].
  • Why it matters: A payment problem, renegotiation or non-renewal at NC-1 would hit most of the enterprise value at once.
  • 🛡️ Counter-argument:
    • Nscale is now well-funded (Bull #2).
    • Electricity is passed through.
    • The new cloud contracts spread exposure across four or five more names in 2027.
  • Severity: Medium (it was High in report #1). Tenant credit improved, but concentration did not.

🐻 Bear #5Underlying cloud revenue is shrinking behind a one-off fee

  • The point: Q2's +54% revenue headline hides a fall in recurring cloud revenue.
  • Evidence:
    • Q2 revenue of $28.84M included a $12.3M termination fee from the Initial Customer [1].
    • Revenue excluding the fee was about $16.5M, against $18.7M a year earlier (−11%).
    • Cloud revenue excluding the fee was about $11.5M, against $16.6M (−31%) (calc from [1]; Needham reached the same $11.5M [26]).
    • The fee was amended to $15.7M after quarter-end and was "outstanding as of the date of this Form 10-Q" [1].
    • DNA Fund still owed most of its $7.3M (only $2.2M collected) [1].
  • Why it matters: The legacy fleet is losing revenue faster than new contracts arrive. Collection risk on exit fees adds to the gap.
  • 🛡️ Counter-argument:
    • The freed GPUs were "redeployed… to other customers" [1].
    • The new contracts start from September 2026 to Q2 2027 and should more than replace the lost revenue if delivered.
    • If the extra $3.4M of the fee is collected, it is upside (Tailwind 4).
  • Severity: Medium

🐻 Bear #6Shares are up 18% in eleven months, with more overhang

  • The point: The share count is rising fast, and the convertibles add more.
  • Evidence:
    • Shares outstanding went from 38.26M (2025-11-12) to 45.12M (2026-08-25), +18% [11,5].
    • Fully diluted is about 49.6M if both convertibles settle in shares, or about 51.9M at the maximum make-whole rate (+10% to +15%) [4].
    • The zero-strike call (5.9M shares) only offsets dilution in 2031 [1].
    • The RSU plan has 4.0M shares authorised and 1.23M granted [1].
  • Why it matters: At $15 a share, every $100M of new equity is about 6.7M shares (+15%).
  • 🛡️ Counter-argument:
    • Both convertible prices ($25.91 and $33.84) are far above the share price.
    • The company can settle the 2032 notes in cash [2].
  • Severity: Medium High.

🐻 Bear #7Bit Digital is now openly weighing sales of its stake (new)

  • The point: The 59.9% parent has moved from "long-term holder" to "considering" sales and covered calls.
  • Evidence (Bit Digital Q2 call, 2026-08-13 [16]):
    • CEO Sam Tabar: "we do not intend to sell White Fiber shares this year." On selling part of the stake to fund buybacks: "definitely something we're considering", but not in 2026.
    • It is evaluating out-of-the-money covered calls, which need the shares to be registered. Its CFO said it was "coordinating with WhiteFiber on possible registration statements later this quarter".
    • Bit Digital trades at about a 40% discount to its NAV.
    • Bit Digital holds 27,043,750 shares (59.9%) [5]. Its lock-up with Barclays for the convertible ends about 2026-10-17 (60 days from the offering memorandum, estimate) [5].
  • Why it matters: About 27M shares could come to market against a free float of about 18M. Covered calls cap the upside and add hedging pressure.
  • 🛡️ Counter-argument:
    • "Registration creates flexibility. It is not a step towards exiting" [16].
    • There has been no sale in 2026, and the pledge runs to year-end.
    • Bit Digital needs WhiteFiber to stay valuable, because the stake "very much helps the share price" [16].
  • Severity: Medium The risk is mostly a 2027 overhang, but the registration statement could arrive within weeks.

🐻 Bear #8Related-party bridge loans cost about 50% a year

  • The point: The parent lends to the subsidiary at punitive terms, and the same CEO runs both companies.
  • Evidence:
    • The Bit Digital Capital facility (up to $100M, increasable to $150M) has these terms [6,1]:
      • Interest 9.5%.
      • 3% fee taken off each draw up front.
      • The lender must get back at least 1.1x each advance.
      • Maturities of 90–180 days.
    • The effective rate was 50.6% in Q2 [1].
    • $80M has been drawn. The first $50M had a 90-day maturity (about 2026-08-24, extendable 30 days). Whether it was repaid or extended: not found.
    • $20M of the loan was assigned to B. Riley Securities, which also publishes research on WhiteFiber [6].
    • Other links [1]:
      • Sam Tabar is CEO of both companies.
      • Bit Digital charged $2.4M under the Transition Services Agreement in H1.
      • On 2026-08-10 WhiteFiber bought $1.0M of SAIHEAT shares. Four Bit Digital managers, including CFO Erke Huang, invested personally in the same deal.
  • Why it matters: Value can leak to the parent and its circle, and capital decisions may favour Bit Digital.
  • 🛡️ Counter-argument:
    • Independent committees and fairness opinions (Needham, Seaport) covered the loan [16].
    • The bridge is meant to be repaid from the NC-1 loan.
    • At $80M, the cost is about $8M of fee and return over its life, which is small against the NC-1 lease.
  • Severity: Medium

🐻 Bear #9Dated promises slip, quarter after quarter

  • The point: Almost every dated milestone has slipped.
  • Evidence:
    • NC-1 billing dates [15,9,14,12,4]:
      • "Full 40 MW May 31" (2026-03-26).
      • 27 MW in April plus 27 MW in May (10-K).
      • "Slightly later than June 1" (2026-05-14).
      • "End of August" (2026-08-12).
      • "Initial billing" (deck, 2026-09-14).
    • Paris: July → end of September [44,1].
    • The NC-1 loan: see Bear #2.
    • Capacity: 76 MW promised for YE2026 in the 10-K, now about 70 MW [9,4].
    • Q1 G&A was guided "slightly above $11.4M" and came in at $17.8M [15,10].
    • In its favour: the Q1 cloud revenue guide was met, there are no restatements or material weaknesses, and controls were effective [10,1].
  • Why it matters: 2027 estimates assume four GPU deployments plus NC-2/NC-3 work happen on schedule.
  • 🛡️ Counter-argument:
    • Most slips were 1–3 months and blamed on suppliers (switchgear) or customers (change orders).
    • No contract has been lost because of a delay.
  • Severity: Medium

🐻 Bear #10A small player whose tenant is also a competitor

  • The point: WhiteFiber is tiny in a market of GW-scale players, and its GPU cloud sells a commodity.
  • Evidence:
    • About 54–65 MW gross online in 2026 [4]. Peers are far larger:
      • Applied Digital: about 1,410 MW contracted, $36.2B of 15-year leases (secondary) [40].
      • Hut 8: 245 MW, 15 years, $7B, backed by Google [41].
      • Galaxy Helios: 133 MW delivered to CoreWeave [40].
    • Those leases are backed by hyperscalers. WhiteFiber's is backed by a neocloud.
    • The 10-K concedes WhiteFiber does "not have the resources to compete with larger providers" [9].
    • Nscale, its tenant, also sells GPU cloud [19].
  • Why it matters: WhiteFiber has less negotiating power on rents, financing costs and GPU supply.
  • 🛡️ Counter-argument:
    • CBRE expects demand to shift toward 25–75 MW inference sites, which is WhiteFiber's size [35].
    • Vacancy is 1.4% [35].
    • Speed of delivery at retrofits is a real edge.
  • Severity: Medium

🐻 Bear #11Power tariffs are tightening in North Carolina and Québec

  • The point: Both of WhiteFiber's power markets are introducing data-centre-specific terms.
  • Evidence:
    • Duke settlement, NCUC E-7 Sub 1329, filed 2026-10-06 and still pending approval [32]:
      • Minimum term of 10 years (or 5 years plus the ramp).
      • Minimum bills at 75% of contract demand.
      • Liquidated damages on exit.
      • Upfront funding of dedicated facilities.
      • It applies to new agreements from 2026-06-01. The existing NC-1 agreement appears to be grandfathered (interpretation).
    • Hydro-Québec "CD" tariff, Régie docket R-4333-2026 [33,34]:
      • About C$0.135/kWh all-in (derived), for sites of 5 MW or more.
      • Hearings continue on 2026-12-07 to 12-11 and 12-15 to 12-18, so the requested 2026-11-01 start will slip.
      • Existing customers phase in at 62% → 100% by April 2030.
  • Why it matters: These terms raise the cash cost and lock-in of NC expansions, and could squeeze Montreal margins if power is not fully passed through.
  • 🛡️ Counter-argument:
    • The NC-1 lease passes electricity through [1].
    • MTL-1 (4 MW gross) falls below the threshold.
    • Montreal is a small share of revenue.
  • Severity: Low

Tailwinds NOT in the forecasts

These are not in current guidance (management gives none beyond MW targets) or, as far as public sources show, in analyst consensus. Consensus assumptions are not published, so "not in consensus" is an interpretation for each.

  • NC-1 Phase 2: a long-term lease for the next ~45 MW gross — 50% chance within 36 months

    • Impact on revenue: +$63M a year, about +15–17% vs FY2028 consensus ($381–423M) [25]
    • Impact on EPS: +$1.24 EBITDA per share, about +24% vs FY2028 consensus EBITDA per share (about $5.25 = $237M ÷ 45.1M). EBITDA is used because EPS is near zero.
    • How we got the number (estimate): 45 MW gross × (40/54 IT ratio) ≈ 33 MW IT × $160/kW-month × 12 ≈ $63M; × 88% margin (B. Riley NOI/revenue) ≈ $56M ÷ 45.1M shares ≈ +$1.24
    • Expected value: 0.50 × $1.24 ≈ +$0.62 EBITDA per share
    • What it is: a second lease at NC-1 after Duke delivers power beyond 54 MW. Nscale has priority rights [4].
    • Why it's not priced in: RPO shows only Phase 1 [1]. Management calls it "not imminent" and gives no date [12].
    • Trigger to watch: a Duke delivery schedule, or a press release naming the Phase 2 tenant.
    • Why this probability: demand and Nscale's priority right make a lease likely. But "not imminent", the new Duke tariff terms and WhiteFiber's own funding needs push some outcomes beyond 36 months. That is why it is down from 55% in report #1.

  • NC-2/NC-3 close and sign a long-term tenant for the first 60 MW gross — 25% chance within 36 months

    • Impact on revenue: +$84M a year in the first full year (FY2029 at the earliest), about +20% vs the FY2028 consensus ($423M; no FY2029 consensus found)
    • Impact on EPS: +$1.59 EBITDA per share before any dilution, about +30% vs FY2028 consensus EBITDA per share
    • How we got the number (estimate): 60 MW gross × 0.74 ≈ 44 MW IT × $160 × 12 ≈ $84M; × 85% ≈ $72M ÷ 45.1M ≈ +$1.59. The $480–600M of capex would likely require new shares, which would cut this.
    • Expected value: 0.25 × $1.59 ≈ +$0.40 EBITDA per share (gross of dilution)
    • What it is: two ex-Unifi sites in Yadkin County, bought for $60M, ready-for-service targeted Q3 2027 [7].
    • Why it's not priced in: there is no closing, tenant or financing. The LOIs are non-binding [7].
    • Trigger to watch: the closing 8-K (closing is about end-October at the earliest), the energy study, a signed lease.
    • Why this probability: four steps must all go right (close, power, tenant, financing), and the company's record is to slip a quarter at each step.

  • GPU rental price strength lifts the legacy fleet on re-contracting — 40% chance during FY2027

    • Impact on revenue: +$9M, about +3% vs FY2027 consensus (about $266M)
    • Impact on EPS: +$0.16 EBITDA per share, about +6% vs FY2027 consensus EBITDA per share (about $2.90)
    • How we got the number (estimate): about $46M a year of underlying legacy cloud revenue ($11.5M × 4) × 20% price uplift ≈ $9M; × 80% incremental margin ≈ $7.4M ÷ 45.1M ≈ +$0.16
    • Expected value: 0.40 × $0.16 ≈ +$0.06
    • What it is: older H100/H200 capacity re-priced as contracts roll, after the 2026 rebound [36,38].
    • Why it's not priced in: consensus models new contracts, not re-pricing of the legacy fleet (interpretation).
    • Trigger to watch: cloud revenue per GPU in the Q3 and Q4 10-Qs, and the SemiAnalysis and Silicon Data indices.
    • Why this probability: prices are up about 40% from the low, but forward markets point flat to down for B200.

  • The amended $15.7M termination fee is collected in full — 60% chance by mid-2027

    • Impact on revenue: +$3.4M one-off (the amended $15.7M minus the $12.3M already booked), about +2.5% vs FY2026 consensus ($136.8M)
    • Impact on EPS: +$0.08 EBITDA per share, a one-off
    • How we got the number (estimate): $3.4M ÷ 45.1M ≈ +$0.075
    • Expected value: 0.60 × $0.075 ≈ +$0.045
    • What it is: cash from the Initial Customer under the amended termination agreement [1].
    • Why it's not priced in: the fee is unpaid, and DNA Fund paid only $2.2M of $7.3M, a precedent for slow collection.
    • Trigger to watch: the Q3 10-Q receivables note.
    • Why this probability: the amount was raised by agreement, which suggests willingness to pay, but this counterparty already walked away once.

  • Krambu 100 MW managed-services capacity starts generating fees — 30% chance of first revenue in 2027

    • Impact on revenue / EPS: not quantifiable. The fee structure, the funder and the site are undisclosed [39].
    • What it is: WhiteFiber is the exclusive operator of 100 MW of Krambu's liquid-cooled capacity from 2027. Customers fund the hardware [39,12].
    • Why it's not priced in: there is no contract value or customer.
    • Trigger to watch: a named offtaker or a disclosed fee.
    • Why this probability: the partner is an unknown private company with no disclosed financing.

Headwinds NOT in the forecasts

  • Another equity or equity-linked raise — 70% chance within 12 months

    • Impact on revenue: none
    • Impact on EPS: −$0.52 EBITDA per share, about −18% vs FY2027 consensus EBITDA per share ($2.90)
    • How we got the number (estimate): a $150M raise at $15 ≈ 10.0M new shares; $131M ÷ 55.1M ≈ $2.38 vs $2.90 → −$0.52
    • Expected value: 0.70 × −$0.52 ≈ −$0.36
    • What it is: an ATM, a follow-on or another convertible to fund NC-2/NC-3, GPU equity, or the NC-1 shortfall if the loan slips again.
    • Why it's not priced in: consensus per-share figures use about 45M shares (interpretation). Price targets of $32–50 imply funded growth.
    • Trigger to watch: an S-3 or ATM filing, a convertible launch, or the NC-1 loan slipping past Q4 2026.
    • Why this probability: the 10-Q itself flags the gap [1]. Four financings in eight months show the pattern. A closed NC-1 loan would lower the odds, which is why it is not 85%+.

  • One of the four new GPU deployments is late by 2 quarters or more, or terminated — 40% chance within 18 months

    • Impact on revenue: −$35M, about −13% vs FY2027 consensus ($266M)
    • Impact on EPS: −$0.47 EBITDA per share, about −16% vs FY2027 consensus
    • How we got the number (estimate): average contract ≈ $35M a year (Paris ~$32M, BaseTen ~$55M, Iceland ~$17.5M, Prime Intellect ~$36M); × 60% margin ≈ $21M ÷ 45.1M ≈ −$0.47
    • Expected value: 0.40 × −$0.47 ≈ −$0.19
    • What it is: a delay, renegotiation or termination at Paris, BaseTen, Iceland or Prime Intellect.
    • Why it's not priced in: the 2027 consensus step-up (+$125–160M) assumes all of them [25].
    • Trigger to watch: confirmation of the Paris start, the BaseTen go-live in November, and the Q4 cloud revenue step.
    • Why this probability: Paris has already slipped once, GPU supply is tight, and two of the last three big customers ended early.

  • Nscale renegotiates or partly defaults at NC-1 — 12% chance within 36 months

    • Impact on revenue: −$26M a year, about −6% vs FY2028 consensus (about $400M)
    • Impact on EPS: −$0.51 EBITDA per share, about −10% vs FY2028 consensus EBITDA per share ($5.25)
    • How we got the number (estimate): a 30% rent cut on about $86M a year ≈ −$26M; × 88% ≈ −$23M ÷ 45.1M ≈ −$0.51. A full default would be roughly three times this.
    • Expected value: 0.12 × −$0.51 ≈ −$0.06
    • What it is: Nscale seeks rent relief or exits, for example if its end user leaves after a 5–6-year term or the IPO fails.
    • Why it's not priced in: consensus treats the $865M as certain.
    • Trigger to watch: the Nscale IPO outcome, any amendment to the NC-1 lease in WhiteFiber's filings, and Nscale's customer concentration.
    • Why this probability: Nscale has $1.48B of cash plus $3.36B raised, so it is down from report #1. The risk sits mostly after 2029.

  • Hydro-Québec CD tariff costs absorbed at the Montreal sites — 30% chance within 24 months

    • Impact on revenue: none
    • Impact on EPS: −$0.05 EBITDA per share, about −2% vs FY2027 consensus
    • How we got the number (estimate): about 6 MW of load × 8,760 h × 90% × about C$0.065/kWh increase × 62% transition × 0.73 USD/CAD ≈ $1.6M; up to about $2.3M at the full rate ÷ 45.1M ≈ −$0.04 to −$0.05
    • Expected value: 0.30 × −$0.05 ≈ −$0.015
    • What it is: a new data-centre tariff in Québec [33,34].
    • Why it's not priced in: management has not discussed it.
    • Trigger to watch: the Régie decision after the December 2026 hearings.
    • Why this probability: utilities usually win most of a rate case, but contracts likely pass power through.

  • Bit Digital registers, hedges or sells WYFI shares — 45% chance within 18 months

    • Impact on revenue / EPS: not quantifiable. It does not change the company's earnings, only supply and demand for the stock.
    • What it is: a resale registration, covered calls on part of the 27.0M shares, or sales from 2027 to fund Bit Digital's buybacks [16].
    • Why it's not priced in: the "no sales in 2026" pledge anchors expectations.
    • Trigger to watch: an S-3 or S-1 resale filing, or a 13D amendment.
    • Why this probability: management called a sale "definitely something we're considering" and said registration filings were being coordinated "later this quarter" [16].

  • The Duke large-load tariff applies to the NC-1 expansion — 50% chance if Phase 2 proceeds

    • Impact on revenue / EPS: not quantifiable. Power is passed through, but upfront facility funding and minimum bills raise the capital need and the exit cost [32].
    • Trigger to watch: the NCUC order, and the date of any new Duke agreement for capacity beyond 54 MW.
    • Why this probability: a new agreement for 45 MW or more at a high load factor would likely meet the 50 MW / 80% test once summed with existing load. That is a legal question.

Investment Risk Score: 70/100 (high)

# Factor Weight Measured value Score Why
1 Liquidity & funding runway 15% Cash $56.1M (06-30) + about $180M net from converts + $26.2M RBC + $30M bridge ≈ $275M pro forma before Q3 burn (estimate); FCF −$90M to −$166M per quarter in H1 → runway about 2–3 quarters without new financing; $80M bridge at 90–180-day maturities; 10-Q: resources "may not be sufficient"; no going-concern language 78 Runway under 12 months on current burn and no closed long-term financing. Kept below 81+ only because there is no going-concern language and the NC-1 loan is in exclusivity.
2 Leverage 10% Debt ≈ $466M (converts $341.9M, bridge $80M, Iceland about $18M, RBC $26.2M), estimate; LTM adjusted EBITDA roughly $15–20M including a $12.3M one-off fee; Q2 adjusted EBITDA excluding the fee is negative 80 Large debt against negative recurring EBITDA. Long maturities on the convertibles (first put 2029) keep it off the top.
3 Dilution (12 months + overhang) 10% Shares 38.26M → 45.12M (+18% since 2025-11); fully diluted 49.6–51.9M (+10–15%); RSU pool 2.8M left 68 +10–25% band. The next raise (Headwind 1) would push it higher.
4 Profitability 10% LTM FCF ≈ −$341M on $94.5M revenue (≈ −360%); operating margin −22% to −72% per quarter; H1 net loss −$27.0M; gross margin before D&A about 59–65% 82 Deep losses on a cash basis. Better in 2027 if NC-1 bills in full, but not yet.
5 Revenue quality & concentration 10% Nscale ≈ 93% of RPO; top customer 63% of H1 revenue; Q2 included a $12.3M one-off; 10-year lease against a 5.7-year average tenant contract 80 More than 35% from one counterparty and one asset. The contracted, recurring nature stops it scoring higher.
6 Execution & track record 10% NC-1 billing dated 4 times, loan 4 times, Paris once, YE26 MW cut 76→70, Q1 G&A guide missed; no material weakness or restatement 72 Frequent misses and delays, but no control failure.
7 Competitive position & disruption 10% About 54–65 MW gross against peers of 245 MW–1.4 GW; 10-K admits the resource gap; tenant also a competitor; 1.4% market vacancy 62 Contested. Speed and small-site niche against scale disadvantages.
8 Cyclicality & macro sensitivity 10% ~100% AI-compute demand; H100 hyperscaler price −33% in 2025, then +40% in 2026; CoreWeave about $35B of debt shows neocloud leverage; colocation lease 10 years with pass-through 68 Highly cyclical end market, softened by the long NC-1 lease.
9 Regulatory, legal & geopolitical 5% Duke large-load settlement pending; Hydro-Québec CD docket; US Section 338 duties on Canadian goods; related-party dealings; no material litigation 45 Material but contained.
10 Valuation & expectations 10% Market value about $674M; EV ≈ $0.86–1.0B (estimate) ≈ 6.6–7.6x FY27 consensus EBITDA; targets $32–50 not cut; short interest about 13% of float; beta about 6.3; −68% from the peak with no clear cause for the latest leg 50 Undemanding on paper, but the estimates are fragile and the float is thin and volatile.

Overall: weighted average = 0.15×78 + 0.10×80 + 0.10×68 + 0.10×82 + 0.10×80 + 0.10×72 + 0.10×62 + 0.10×68 + 0.05×45 + 0.10×50 = 70.15 → 70.

Change vs last: +5 vs 65 (old 6-area method). Part of this is the method. The v3 table gives profitability and liquidity their own weight, and both are at the bad end here. Part is new evidence:

  • The 10-Q funding warning.
  • The fourth slip of the NC-1 loan.
  • Bit Digital's shift toward possible sales.

Tenant risk fell (Bull #2), which keeps the score from going higher.

Ranking check: WYFI sits between IPX (67) and GRRR (72).

  • It is riskier than DUOT (66) and IPX because it burns about $90–166M a quarter against a pro forma cash pile of about $275M, with no closed long-term financing.
  • It is safer than GRRR because it has a positive gross margin, a 10-year contracted lease with a well-funded tenant, and no near-term convertible reset.

Conclusion

  • What you are really betting on: that WhiteFiber can finance NC-1 and its four GPU deployments over the next 2–3 quarters without a large equity raise at about $15, so that roughly $95M a year of NC-1 rent plus about $200M a year of cloud revenue reach the per-share numbers.
  • The 2–3 things that will decide the outcome:
    1. The NC-1 project loan: whether it closes, how big it is, how much equity comes back, and whether the bridge is repaid.
    2. Whether the next dollar of growth capital comes from project debt and prepayments or from new shares.
    3. Delivery in Q4 2026 and Q1 2027: full 40 MW billing at NC-1, plus the Paris, BaseTen and Iceland go-lives.
  • Signposts:
    • Confirm the bull case:
      • An 8-K for the NC-1 loan showing a large cash return and the bridge repaid.
      • Q3 colocation revenue showing about $20M or more a quarter run-rate.
      • Paris and BaseTen confirmed live, with the France financing closed.
      • The Nscale IPO priced, and a Phase 2 lease or Duke schedule.
    • Break the bull case:
      • An ATM, S-3 or follow-on before the NC-1 loan closes.
      • The bridge extended again, or upsized to $150M.
      • Any NC-1 lease amendment or Nscale payment delay.
      • A Bit Digital resale registration followed by sales or covered calls.
  • New since the last report:
    • The tenant got safer and the company got tighter. Nscale has $1.48B of cash plus $3.36B raised [19,20], while WhiteFiber's 10-Q now says its resources "may not be sufficient" [1] and LTM free cash flow is about −$341M.
    • The parent's stance changed: selling WYFI to fund Bit Digital's buybacks is "definitely something we're considering" from 2027, and share registration is being prepared [16].
    • Underlying cloud revenue fell about 31% year on year once the $12.3M fee is stripped out [1]. Report #1's "committed parent" and "NC-1 loan frees equity" bulls were downgraded to bears or dropped.

Open questions & data gaps

  • NC-1 loan and bridge. No 8-K after 2026-09-14 [17]. Whether the first $50M of bridge draws (due about 2026-08-24, extendable 30 days) were repaid or extended: not found. High importance; the Q3 10-Q (about mid-November) will show it.
  • Full 40 MW billing at NC-1. Not confirmed in any filing. High importance.
  • Paris start and France financing. Not confirmed. High importance for 2027 cloud estimates.
  • Cash at 2026-09-30. Not disclosed. The about $275M pro forma figure is an estimate before Q3 capex, so the runway in factor 1 rests on it.
  • Consensus assumptions. Sources conflict: FY2027 revenue is $261M, $266M or $296M, and FY2027 EBITDA is $131M (StocksGuide) against $175M (Needham). No source says whether Phase 2 or NC-2/NC-3 is included.
  • Why the stock fell from about $20 to $14.94 in September–October. No cause found. A general neocloud sell-off is the best guess (interpretation) [27].
  • Q2 call Q&A. Only summarised transcripts were available, so quotes are paraphrase-grade where marked.
  • Deck naming conflict. The September deck labels the $165M deal "Prime Intellect (Canada)". The 10-Q names BaseTen (1,392 B300, $165.2M) and Prime Intellect (576 Vera Rubin, $108.2M) [1,4]. The 10-Q is used.
  • Short interest. About 13% of float (MarketBeat/Equibles). Other float definitions give up to 45%. FINRA data was not checked.
Sources
  1. WhiteFiber Form 10-Q for Q2 2026 — SEC EDGAR — 2026-08-12 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026088026/ea0300877-10q_whitefiber.htm
  2. WhiteFiber 8-K: 5.00% convertible notes due 2032 and exchange — SEC EDGAR — 2026-08-21 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026092599/ea0302925-8k_white.htm
  3. WhiteFiber press release: closing of $310M notes (Ex. 99.3) — SEC EDGAR — 2026-08-21 — https://www.sec.gov/Archives/edgar/data/0002042022/000121390026092599/ea030292501ex99-3.htm
  4. WhiteFiber investor presentation, September 2026 (8-K Ex. 99.1) — SEC EDGAR — 2026-09-14 — https://www.sec.gov/Archives/edgar/data/0002042022/000121390026099424/ea030525701ex99-1.htm
  5. Bit Digital Schedule 13D/A on WhiteFiber — SEC EDGAR — 2026-08-25 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026093678/primary_doc.xml
  6. WhiteFiber 8-K: Bit Digital Capital delayed-draw term loan — SEC EDGAR — 2026-05-27 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026061575/ea0292280-8k_white.htm
  7. WhiteFiber 8-K: NC-2/NC-3 purchase agreement — SEC EDGAR — 2026-08-17 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026090492/ea0302334-8k_whitefiber.htm
  8. WhiteFiber 8-K: CFO and board changes — SEC EDGAR — 2026-08-03 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026084223/ea0299775-8k_whitefiber.htm
  9. WhiteFiber Form 10-K FY2025 — SEC EDGAR — 2026-03-26 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026034341/ea0278305-10k_white.htm
  10. WhiteFiber Form 10-Q for Q1 2026 — SEC EDGAR — 2026-05-14 — https://www.sec.gov/Archives/edgar/data/0002042022/000121390026056212/ea0288695-10q_whitefiber.htm
  11. WhiteFiber Form 10-Q for Q3 2025 — SEC EDGAR — 2025-11-13 — https://www.sec.gov/Archives/edgar/data/2042022/000121390025109964/ea0263547-10q_whitefiber.htm
  12. WhiteFiber Q2 2026 earnings call transcript — The Motley Fool — 2026-08-12 — https://www.fool.com/earnings/call-transcripts/2026/08/12/whitefiber-wyfi-q2-2026-earnings-call-transcript/ (read via summary tool)
  13. WhiteFiber Q2 2026 earnings call transcript — Investing.com — 2026-08-12 — https://www.investing.com/news/transcripts/earnings-call-transcript-white-fiber-q2-2026-revenue-jumps-54-as-stock-rises-166-93CH-4855417
  14. WhiteFiber Q1 2026 earnings call transcript — MarketBeat — 2026-05-14 — https://www.marketbeat.com/earnings/reports/2026-5-14-whitefiber-inc-stock/ (summary)
  15. WhiteFiber Q4 2025 earnings call highlights — Yahoo Finance — 2026-03-26 — https://finance.yahoo.com/markets/stocks/articles/whitefiber-q4-earnings-call-highlights-162109183.html (summary)
  16. Bit Digital Q2 2026 earnings call transcript — Webull/Benzinga — 2026-08-13 — https://www.webull.com/news/15396166761145344
  17. WhiteFiber EDGAR filing index (submissions) — SEC — retrieved 2026-10-12 — https://data.sec.gov/submissions/CIK0002042022.json
  18. Form 4, director Ichi Shih (sale 2026-05-29) — SEC EDGAR — 2026-09-01 — https://www.sec.gov/Archives/edgar/data/2042022/000121390026096352/xslF345X06/ownership.xml
  19. Nscale Ltd Form S-1 — SEC EDGAR — 2026-09-18 — https://www.sec.gov/Archives/edgar/data/2110365/000119312526395475/ck0002110365-20260918.htm
  20. Nscale closed a $3.36B pre-IPO round led by Third Point — The Next Web — 2026-09-25 — https://thenextweb.com/news/nscale-3-36bn-pre-ipo-round
  21. Nscale files registration statement for proposed IPO — StorageNewsletter — 2026-10-08 — https://www.storagenewsletter.com/?p=305824
  22. WYFI price target and analyst ratings — MarketBeat — 2026-10-12 — https://www.marketbeat.com/stocks/NASDAQ/WYFI/forecast/
  23. WYFI forecast — StockAnalysis — 2026-09-28 — https://stockanalysis.com/stocks/wyfi/forecast/
  24. WYFI forecast — TipRanks — 2026-10 — https://www.tipranks.com/stocks/wyfi/forecast
  25. WYFI consensus forecasts — StocksGuide / TickFlow — 2026-10 — https://stocksguide.com/en/forecast/WhiteFiber-KYG961151035 ; https://tickflow.io/stock/WYFI/forecast
  26. Needham Q2 note summary — Blockspace — 2026-08 (search summary only; page blocked) — https://blockspace.media/insight/whitefiber-cloud-contract-growth-needham-target-2026/
  27. Why WhiteFiber (WYFI) stock is down 5.20% — AAII — 2026-10-12 — https://www.aaii.com/investingideas/article/540271-why-whitefiber-inc8217s-wyfi-stock-is-down-520
  28. WYFI short interest snapshot — Equibles — 2026-10-12 — https://equibles.com/stocks/WYFI
  29. WhiteFiber slides 25.9% on financing and dilution concerns — Quiver Quantitative — 2026-08-19 — https://www.quiverquant.com/news/WhiteFiber+Slides+25.9%25+as+Financing+and+Dilution+Concerns+Pressure+Shares
  30. WhiteFiber slides 11.9% as AI-infrastructure rally reverses — Quiver Quantitative — 2026-07-14 — https://www.quiverquant.com/news/WhiteFiber+Slides+11.9%25+as+Recent+AI-Infrastructure+Rally+Appears+to+Reverse
  31. WhiteFiber stock price down 8.2% — MarketBeat — 2026-10-08 — https://www.marketbeat.com/instant-alerts/price-whitefiber-nasdaq-wyfi-stock-price-down-82-heres-what-happened-2026-10-08/
  32. Duke Energy large-load tariff settlement, NCUC E-7 Sub 1329 — via SELC — 2026-10-06 — https://www.selc.org/wp-content/uploads/2026/10/2026-10-06-Duke-Energy-LLT-Settlement.pdf
  33. Docket R-4333-2026 (Hydro-Québec CD tariff) — Régie de l'énergie — status to 2026-10-09 — https://www.regie-energie.qc.ca/fr/participants/dossiers/R-4333-2026/
  34. Google expert report (C. Cain), R-4333-2026 — Régie de l'énergie — 2026-08-11 — https://www.regie-energie.qc.ca/fr/participants/dossiers/R-4333-2026/doc/R-4333-2026-C-Google-0018-Preuve-Memoire-2026_08_11.pdf
  35. North America Data Center Trends H1 2026 — CBRE — 2026-08-27 — https://www.cbre.com/insights/books/north-america-data-center-trends-h1-2026
  36. The Great GPU Shortage: rental capacity — SemiAnalysis — 2026-04-02 — https://newsletter.semianalysis.com/p/the-great-gpu-shortage-rental-capacity
  37. H100 rental price over time — Silicon Data — 2025-12-21 — https://www.silicondata.com/blog/h100-rental-price-over-time
  38. GPU rental rates continue to price upwards — TechNews — 2026-08-04 — https://finance.technews.tw/2026/08/04/nvidia-a100-h100-and-h200-gpu-rental-rates-continue-to-price-upwards/
  39. WhiteFiber and Krambu partner on 100 MW — PR Newswire via Finviz — 2026-08-12 — https://finviz.com/news/381323/whitefiber-and-krambu-partner-to-bring-100mw-of-high-density-gpu-infrastructure-capacity-to-market-in-2027
  40. Applied Digital Form 10-K FY2026 and Galaxy Helios Phase I reports — SEC EDGAR / The Energy Mag — 2026 (summaries only) — https://www.sec.gov/Archives/edgar/data/0001144879/000114487926000048/apld-20260531.htm ; https://theenergymag.com/news/market-news/galaxy-digital-completes-initial-phase-of-helios-data-center-supplies-133-mw-to-coreweave
  41. Hut 8 signs 245 MW deal with Fluidstack — DatacenterDynamics — 2025-12-17 — https://www.datacenterdynamics.com/en/news/hut-8-signs-245mw-capacity-deal-with-fluidstack-as-part-of-multi-gigawatt-partnership-with-anthropic
  42. CoreWeave's debt hits $35bn — Capacity Global — 2026 — https://capacityglobal.com/news/coreweaves-debt-hits-35bn/
  43. Smoot-Hawley revived: Section 338 against Canadian trade — BLG — 2026-07-27 — https://www.blg.com/en/insights/2026/07/smoot-hawley-revived-a-never-before-used-depression-era-law-is-invoked-against-canadian-trade
  44. WhiteFiber Paris contract announcement — W.Media / 8-K Ex. 99.1 — 2026-05-21 — https://www.sec.gov/Archives/edgar/data/0002042022/000121390026060135/ea029187601ex99-1.htm
  45. Analyst notes (Needham, Roth, B. Riley, Clear Street) — TheFly via TipRanks — undated — https://www.tipranks.com/news/the-fly/whitefiber-price-target-raised-to-45-from-34-at-needham-thefly
  46. WhiteFiber consensus estimates — Investing.com — 2026-10 — https://www.investing.com/equities/whitefiber-consensus-estimates
  47. WhiteFiber Q3 2025 results — Barchart — 2025-11-13 — https://www.barchart.com/story/news/36109748/whitefiber-inc-reports-third-quarter-2025-results
Disclaimer. AI-generated research and podcast (Claude by Anthropic, voices by Google Gemini) for educational purposes — not financial advice and not a recommendation to buy or sell. Willem (MR.INVEST) owns this stock. Check the company’s own filings before making any decision.
← All Bulls vs Bears reports