Executive Summary

Q3 revenue grew 65% YoY to $20.2M but came in below S&P Global consensus of $22.0M; EPS of -$0.47 missed the -$0.10 consensus as public-company costs and elevated non-cash SBC weighed on profitability . Values retrieved from S&P Global.*
Adjusted EBITDA was $2.3M vs $5.6M a year ago, reflecting higher costs post-IPO; gross profit rose 90% YoY to $12.7M on strong cloud and initial colocation contribution .
Management disclosed a wind-down of a customer arrangement representing ~+$21M annualized cloud run-rate post-quarter end; the capacity is being re-marketed to higher-quality counterparties, implying a short Q4 underutilization before new contracts begin contributing early Q1 2026 .
Key catalyst: imminent NC-1 anchor signing (two finalists, upsized scope, stronger pricing/duration) and financing plan (~75% LTV) for the 99MW campus; MTL-3 began recognizing revenue in October with Q4 contribution of “slightly more than $2M” .

What Went Well and What Went Wrong

What Went Well
- NC-1 commercialization advanced with “very firm proposals” from two highly creditworthy counterparties; terms improved on both price and duration. “I will be countersigning the one who signs first.”
- MTL-3 completed on schedule and is “fully operational and generating revenue” for Cerebras; Q4 contribution expected to be slightly >$2M, showcasing rapid retrofit execution .
- Strategic discipline on cloud contracting despite price-led competition: “We will not compromise on counterparty strength or deal structure simply to announce a transaction.”
What Went Wrong
- Missed Street on revenue and EPS; EBITDA materially below consensus, driven by ~$11.3M non-cash SBC, public-company readiness costs, and a $2M service credit recognized in Q3 cloud results . Values retrieved from S&P Global.*
- Announced wind-down of a customer arrangement (~$21M annualized run-rate) post-quarter end; management expects short-term Q4 underutilization as capacity is reallocated .
- Operating expenses rose sharply to $34.7M vs $13.1M in the prior year (public-company and growth investments), compressing adjusted EBITDA to $2.3M vs $5.6M in Q3’24 .

Financial Results

Headline results vs prior periods and S&P Global consensus

Metric	Q3 2024 (YoY base)	Q2 2025	Q3 2025 Est.*	Q3 2025 Actual
Revenue ($M)	12.28	18.70	22.02*	20.18
Diluted EPS ($)	-0.01	-0.2335*	-0.1025*	-0.47
Adjusted EBITDA ($M)	5.64	3.30	6.82*	2.28

Values retrieved from S&P Global for estimate columns and Q2 EPS.*

Quarterly progression (oldest → newest)

Metric	Q1 2025*	Q2 2025	Q3 2025
Revenue ($M)	16.77*	18.70	20.18
Net Income ($M)	1.43*	-8.83	-15.75
Diluted EPS ($)	N/A*	-0.2335*	-0.47
Adjusted EBITDA ($M)	5.87*	3.30	2.28
Total Operating Expenses ($M)	14.72*	27.82	34.68
Cash & Equivalents ($M)	9.08*	16.40	166.5

Values retrieved from S&P Global where asterisks appear.*

Q3 2025 segment and margin detail

Metric	Cloud	Colocation	Total
Revenue ($)	18,032,898	1,692,280	20,179,766
Other Revenue ($)			454,588
Electricity Costs ($)	631,099	288,797	919,896
DC Lease Expense ($)	1,380,553	166,114	1,546,667
GPU Lease Expense ($)	3,454,308	—	3,454,308
Wage Expense ($)	—	92,494	92,494
Other Segment Items ($)	848,588	127,542	976,130
Segment Gross Profit ($)	11,718,350	1,017,333	12,735,683
Approx. Gross Margin (%)	~65%	~60%	~63%

Additional P&L items (Q3 2025)

G&A expense: $21.3M; includes ~$11.3M non-cash SBC
Total operating expenses: $34.7M
Cash & equivalents: $166.5M at 9/30/25

Guidance Changes

Metric	Period	Previous Guidance	Current Guidance	Change
G&A (incl. SBC)	Q4 2025	Not provided	~$10M	New
CapEx	Q4 2025	Not provided	“Increase materially” on NC-1 development; potential discretionary GPU spend	New
MTL-3 Revenue	Q4 2025	Not provided	“Slightly more than $2M”	New
NC-1 First Revenue	2026	Prior indicated Q1’26 initial deployment	Initial deployment Q1’26; revenue from May 2026	Refined timeline
NC-1 Anchor Contract	Near term	“Weeks” to formalize	“Very near term”; 2 finalists; improved pricing & duration; potential ~40MW+ initial scope	Upgraded terms; timeline extended
Cloud ARR mix	Q4 2025 → Early Q1 2026	Not provided	Wind-down of ~+$21M annualized run-rate; short Q4 underutilization; new contracts to contribute early Q1	Transition disclosed

Earnings Call Themes & Trends

Topic	Previous Mentions (Q2 2025)	Current Period (Q3 2025)	Trend
NC-1 commercialization	Multiple parties; aiming to formalize “in weeks”; some asked for 40–99MW; Q1’26 delivery targeted	Two finalists; “very firm proposals”; better pricing and term; no pause in site work; aiming to sign imminently	Strengthening demand; upgraded terms
Financing strategy	Engaged debt advisors; targeting ~75–80% LTC; exploring banks/insurance/private credit; 200–300 bps over SOFR indicated	Target ~75% LTV; align with anchor contract; financing for full 99MW campus	Consistent; moving toward execution
Cloud contracting discipline	Focus on multi-year, take-or-pay; avoid short-term public cloud-like deals	Not competing on price; reject “irresponsible” pricing to win logos; focus on performance/reliability/software	Discipline reiterated
Technology differentiation	First to market with scheduled-fabric Ethernet; +30% perf vs benchmarks	DriveNets cluster performance “unmatched”; cross-DC workload white paper targeted Q1; V1 site identified	Advancing roadmap
Montreal facilities	MTL-3 to recognize revenue in Oct; MTL-2 flexible for colo or cloud	MTL-3 operational; Q4 rev slightly >$2M; MTL-2 remains flexible/priority after NC-1	Execution, monetization progressing
Demand backdrop	Robust pipeline; retrofit speed/cost advantage (2x faster, 40% cheaper)	Demand “far greater” than a month ago amid market volatility; runners-up seeking 2026/27 capacity	Strength broadening

Management Commentary

“We expect to finalize and anchor agreements in the very, very, very near term.”
“We have very firm proposals in hand... demand continued to build... put upward pressure on both pricing and term length.”
“On the Cloud side, we continue to scale deliberately... compete on performance, reliability, and software.”
“We will not compromise on counterparty strength or deal structure simply to announce a transaction.”
“Completing [MTL-3] in under six months... is a testament to the speed and precision of our development and operations team.”

Q&A Highlights

NC-1 status and scope: Two finalists, both “competing extremely hard”; potential to take full site; timeline extended by diligence and added bidders; pricing/duration improved; development work has not paused .
Financing approach: Plan to finance ~75% of NC-1 with long-dated asset-backed credit matched to contract duration; counterparty quality paramount to terms .
Cloud pivot: Winding down >$20M ARR customer; capacity being redirected to better-credit counterparties; expect short Q4 underutilization; identified new customers with comparable or better terms starting early Q1 .
Competitive posture: Management rejects uneconomic GPUaaS pricing; sees peers taking “irresponsible” deals to announce logos; WF emphasizes performance, reliability, and software differentiation (DriveNets, cross-DC orchestration) .
Montreal ramp: MTL-3 now revenue-generating; Q4 contribution slightly >$2M; MTL-2 reserved for flexible use (colo or internal cloud/R&D) .

Estimates Context

Q3 2025 vs S&P Global consensus: Revenue $20.18M vs $22.02M*; Diluted EPS -$0.47 vs -$0.10*; EBITDA -$8.13M vs +$6.82M* — a broad miss, with management citing a $2M service credit in cloud, elevated non-cash SBC (~$11.3M), and public-company/scale-up costs . Values retrieved from S&P Global.*
Forward estimate implications: Temporary Q4 underutilization from the ARR wind-down may pressure near-term revenue, partially offset by on-demand pools; MTL-3 adds >$2M Q4 revenue tailwind; signing of NC-1 anchor (with improved pricing/duration) would likely drive estimate revisions for 2026–2027 .

Key Takeaways for Investors

Near-term catalyst: NC-1 anchor contract signing (two finalists, upgraded terms) should clarify financing and 2026–2027 revenue trajectory .
Short-term headwind: Q4 underutilization due to ARR wind-down, but management expects new higher-quality contracts to begin contributing early Q1; monitor cloud run-rate updates .
Execution proof points: MTL-3 delivered on time and now producing revenue; validates retrofit playbook and underpins a growing colocation platform .
Cost normalization: G&A expected around $10M in Q4 including SBC after an SBC-heavy Q3; watch operating leverage as scale builds .
Balance sheet: $166.5M cash provides runway for NC-1 development and selective GPU procurement; financing expected to target ~75% LTV on NC-1 .
Strategy/tone: Management remains disciplined on cloud pricing and counterparty quality; technology differentiation (DriveNets, cross-DC) aims to compete on performance, not price .
Risk/reward: Estimate miss is backward-looking; stock likely keyed to NC-1 signing/financing milestones and visibility into 2026 revenue ramp; monitor customer mix shift and Q4 utilization .

S&P Global disclaimer: Items marked with * are values retrieved from S&P Global (Capital IQ) consensus/financials.

WhiteFiber, Inc. (WYFI)·Q3 2025 Earnings Summary