2026-11-12 · 2026-11 / week-2

Rumble's Quake revenue is real; the post-print price still leaves no bounded short

Rumble's Quake revenue is real; the post-print price still leaves no bounded short

Summary: RUM's first full quarter with Quake AI brought $68.2 million of revenue and positive adjusted EBITDA, but the company still used $113.8 million of operating cash over nine months and reported an $89.9 million Q3 GAAP loss. Its Nov. 11 close rose 16.6% after results. A reference-only 12-month map has a flat base and a 77.9% upside stress, so the short is Reject / no trade.

Run field Value
Published November 12, 2026, 09:08 Singapore time
Research cutoff November 12, 2026, 09:08 Singapore time
Scope U.S.-listed short opportunities
Decision Reject / no trade
Reference RUM Class A, $5.96 regular-session close on Nov. 11, 2026, 4:00 p.m. ET; reference only, not an entry.

Opportunity Ranking

Rank Candidate Direction Mispricing Evidence Freshness Catalyst Window Positioning 10/5 Status Tradeability Main Rejection Risk
1 RUM Short Q3 revenue scale and the first adjusted-EBITDA profit sit beside weak cash conversion, a lower Q4 revenue range, and a capacity schedule partly pushed beyond 2027. Q3 release and Nov. 11 post-result close. Q4/FY26 results, then 2027 capacity updates; exact next-report date not verified. Oct. 30 short data are stale for the result and float estimates conflict. Reject Close reference only; no locate, borrow, spread, depth, or exit review. Quake demand and the 16.6% post-result close could support further rerating or a squeeze.
2 INMD Short Revenue and GAAP profit eased, but management reaffirmed guidance and held $512.7 million of cash, securities, and short-term deposits. Q3 release and Nov. 11 close. Q4/FY26 results; exact date not stated in the release. Oct. 30 short data show 2.27 million shares and 4.72% of float; settlement predates Q3 results. Reject No short case with a bounded downside map. Cash, stable gross margin, and affirmed guidance limit the evidence for a sharp downside.
3 ELMT Short High short interest contrasts with 55.6% Q3 revenue growth, record backlog, and a U.S. government investment announcement. Q3 release and Nov. 11 close. Backlog conversion and government-funded capacity. Oct. 30 short interest was 1.585 million shares, 15.65% of reported float; settlement predates Q3. Reject No short case with bounded adverse risk. Strong results and a large backlog can sustain the rerating; shares gained 8.4% on Nov. 11.
Control SPY Short Broad-market reference only; no issuer-specific catalyst supports a standalone short. Nov. 11 close. None identified for this screen. ETF positioning is not a company-specific thesis. No setup Liquid benchmark, but no differentiated short mechanism. SPY rose 1.32% on Nov. 11.

Selected research screen: RUM ranks first for information value, not as a trade candidate.

Why this one now: This is the first full reported quarter after the Northern Data acquisition. It replaces the old deal-supply question with an operating one: can GPU revenue produce cash after the costs and capital needed to keep the fleet useful?

What should surprise the reader: The acquired compute business supplied more than 70% of Q3 revenue. The issuer separately said Rumble revenues rose only $3.6 million year over year. The growth is real, but it does not show that the pre-acquisition business suddenly accelerated.

Why RUM Leads the No-Trade Screen

The strongest bearish case is that the market is valuing a large GPU fleet before reported cash generation supports the valuation. Quake AI contributed $68.2 million of RUM's $96.6 million Q3 revenue, while the company reported $113.8 million of operating cash use in the first nine months. Management also said a portion of the previously disclosed 250 megawatts of unmonetized capacity targeted for 2027 may move beyond 2027 because of market conditions. These are company-reported facts, not proof that customer demand will fail. RUM's Q3 release

The counterargument is strong. RUM said its 22,000-NVIDIA-GPU fleet ran at approximately 85% utilization during the quarter. Quake generated $68.2 million in its first full quarter, and Q3 adjusted EBITDA was positive $4.4 million. Depreciation is largely non-cash in the current period, and a high depreciation charge can reflect valuable acquired infrastructure. The post-result close rose 16.6%. A short here would lean against real demand and a strong tape. The counterargument is why this screen ends without a position.

The old RUM article covered expected Northern Data exchange-offer supply before the acquisition. That thesis is retired. The new evidence boundary is the acquired business's first full quarter, its cash requirements, the 2027 capacity schedule, and the first regular-session close after the Q3 release.

Due forecast audit: The May 6 note used a $7.31 reference and $8.75 invalidation level. ChartExchange's historical table shows RUM closed at $8.91 on May 26, crossing that level 20 calendar days after the note; it closed at $9.75 on May 27, after reaching $9.79 intraday. The May 27 close was 33.4% above the old reference. No order or execution record exists. This is a paper-path forecast audit, not trading P&L, and the later price decline does not undo the invalidation. RUM historical prices

The Market Price

StockAnalysis/S&P Global reports a Nov. 11 regular close of $5.96, up $0.85 or 16.63%, with a $5.05-$6.13 session range. The post-result session did not give a short seller a quiet entry. One ChartExchange historical table reports a $5.93 close and 5.09 million shares; StockAnalysis and ChartExchange's separate quote header show $5.96 and about 9.49 million shares. I use $5.96 as the reference and do not use the conflicting volume figures to claim execution quality. StockAnalysis price history · ChartExchange RUM history

After-hours observations are separate. ChartExchange showed $5.95 at 4:59 p.m. ET; StockAnalysis showed $5.84 at 6:16 p.m. and $5.90 at 7:47 p.m. ET. These vendor snapshots vary and are reference-only, not executable. The model uses the completed regular close.

StockAnalysis reports about $2.38 billion of market capitalization and $2.74 billion of enterprise value, using a 400.02 million share denominator. RUM's Q3 release lists 276.5 million Class A, 123.7 million Class C/ExchangeCo, and 95.8 million Class D shares outstanding at Sept. 30, as well as $358.8 million of convertible notes and $134.7 million of lease liabilities. Those denominators are not interchangeable. The release also says RUM plans to convert its existing universal shelf and Tether resale registration into automatic WKSI shelves. That creates registration capacity, not evidence of an issuance or sale. I do not use the vendor multiple as a fair-value estimate. The quote is consistent with investors assigning substantial value to Quake's growth path, but the available share and claim data do not support a clean reverse valuation. StockAnalysis statistics · RUM Q3 balance sheet

The Mispricing

Fact: RUM reported $96.6 million of Q3 revenue versus $24.8 million a year earlier. Quake AI contributed $68.2 million. The company said its Rumble video business increased revenue by $3.6 million, driven by a $5.6 million advertising gain partly offset by a $2.0 million decline in other services. Subtracting Quake's contribution leaves about $28.4 million of Q3 revenue outside Quake, close to the prior-year total. This arithmetic is a mix bridge, not an organic-growth forecast.

Inference: The market's immediate response rewards acquired compute revenue, utilization, and a positive adjusted-EBITDA quarter. The unresolved part is not whether RUM can book GPU revenue. It is how much cash remains after power, hosting, financing, maintenance, depreciation-replacement investment, and the next capacity build.

Fact: Q3 adjusted EBITDA was $4.4 million, but GAAP net loss was $89.9 million. Nine-month adjusted EBITDA remained negative $33.2 million. The company reported $113.8 million of nine-month operating cash use and $75.5 million of property-and-equipment purchases. Its nine-month cash balance declined by $103.5 million to $134.4 million. RUM's total liquidity figure of $142.3 million includes $7.9 million of Bitcoin. These are consolidated figures; the release does not allocate operating cash or equipment investment to Quake. They flag cash intensity, not proof of negative Quake unit economics. The adjusted metric is not cash flow; depreciation and other excluded items must be considered alongside reported cash use. RUM Q3 statements and reconciliation

Q4 revenue guidance of $88 million-$94 million is 2.7%-8.9% below Q3 revenue by simple arithmetic. The comparison is not a forecast of sequential contraction in the legacy business because Q3 was the first full Quake quarter and the release does not provide a complete segment margin bridge. It is a testable guide, not a short catalyst by itself.

The Positioning

ChartExchange reports 36.58 million shares short as of the Oct. 30 settlement, equal to 12.21% of its 299.67 million float estimate and 11.61 days to cover. Finviz reports the same short count but a 132.45 million float and 27.62% short float. The raw share count is the least ambiguous field; the float percentages are provider-dependent. Both settlement and average-volume inputs predate the Nov. 11 earnings move. ChartExchange short-interest table · Finviz RUM statistics

This is enough to treat squeeze risk as material, not enough to claim current borrow scarcity. Current locate availability, borrow fees, recall terms, intraday short flow, spread, depth, and exit capacity are unknown. A high reported short count can amplify an upside move; it does not establish that shorts must cover now.

The Catalyst

  1. Q4 and FY2026 results: The company has guided to Q4 revenue of $88 million-$94 million. The release date was not verified. Compare Quake and Rumble revenue, GAAP results, adjusted EBITDA, operating cash flow, property-and-equipment purchases, and the share count.
  2. 2027 capacity: RUM previously targeted 250 megawatts of unmonetized capacity for 2027, but says some portion may slip beyond 2027. The useful test is funded, installed, customer-occupied capacity with billings and cash collection, not megawatts announced.
  3. Northern Data ownership: RUM said it had reached about 98% ownership and expected to complete squeeze-out proceedings in early 2027. This may simplify the structure; it does not establish cash returns on the GPU investment.

The cheapest falsification test is the next issuer report: compare Quake revenue and costs with operating cash flow and equipment purchases, then check whether the 2027 capacity schedule is funded and contracted. If those facts show sustained cash generation, the short thesis should be retired.

Payoff and Probability Map

The map is a 12-month price-path sensitivity through November 2027, not intrinsic value or a price ceiling. The anchors are the Nov. 5 low, Nov. 10 pre-result close, and reported 52-week high. Scenario weights are subjective and uncalibrated. I assign 50% to the current level because a successful Q3 print and a strong close can persist without immediate follow-through; 20% to a return to the recent low if cash burn and capacity delays dominate; and 30% to a revisit of the prior high if utilization, GPU demand, and short covering sustain a rerating. The historical high is not an upper bound.

Scenario Probability Target / Level Return / Payoff from $5.96 Horizon Conditions Evidence Quality
Top Case 20% $4.49 +24.7% for a short By Nov. 2027 The market discounts Q4 guidance, operating cash use, and delayed capacity. Medium
Base Case 50% $5.96 0.0% By Nov. 2027 Q3 growth is accepted, but the next report does not yet establish a cash-generation path. Low
Bottom Case 30% $10.60 -77.9% for a short By Nov. 2027 Quake utilization holds, capacity is monetized, and the crowded short position faces further covering. Medium
Invalidation n/a No reliable price-only stop A close above the historical high requires a new map; it is not a cap. Any time Sustained positive operating cash after equipment purchases and evidence of funded, customer-occupied expansion would invalidate the short case. Medium
Payoff check Result
Probability-weighted expected price $7.06
Probability-weighted gross short return -18.4% before borrow, fees, slippage, and carrying costs
Current market level and timestamp $5.96, Nov. 11, 2026, 4:00 p.m. ET regular close
Primary instrument RUM Class A common stock, screening only
10/5 favorable base move 0.0%, below the required 10% decline
10/5 credible adverse move +77.9% to the historical-high stress anchor; further upside is possible
10/5 measurement basis Reference-only; not a verified entry
10/5 status Reject
Top-to-bottom gross reward/adverse ratio 0.32:1
Confidence Low; no defensible fair-value range, live borrow data, or calibrated scenario sample is available

The Reject decision does not depend on the subjective weights. Even if the low-price scenario were more likely, the base case fails the 10% hurdle, the upside stress exceeds 5%, the gross ratio is far below 2:1, and the weighted gross return is negative before costs.

The Kill Shot

The strongest counterparty case is that this is the start of a real compute business. RUM has more than 22,000 GPUs, reported 85% average quarterly utilization, and turned that capacity into $68.2 million of first-full-quarter Quake revenue. The $4.4 million adjusted EBITDA result may be an early operating-leverage signal; depreciation is non-cash in the period. Investors may reasonably give the company time to expand capacity, particularly if customers fund deployments.

The fragile short assumption is that rapid revenue growth will fail to turn into durable cash before capital claims and dilution matter. Q3 does not prove that assumption. Nor does it disprove it. The proper response to that uncertainty is a no-trade screen until operating cash, equipment investment, financing, and customer collections can be tied together.

What Could Go Wrong

A short can lose sharply on GPU-demand news, an expanded Tether relationship, financing, capacity delivery, or another positive report. Reported short interest may be stale; an apparent decline in cash can reflect growth investment; Bitcoin changes total liquidity; convertible notes and share classes complicate per-share value. Gaps, halts, hard-to-borrow fees, recalls, borrow buy-ins, after-hours spreads, and weak exit liquidity can overwhelm a price thesis. There is no price floor or bounded loss for an unhedged short.

What Would Prove This Wrong

Retire the bearish research case if RUM reports repeated positive operating cash after property-and-equipment purchases, Quake unit economics after financing and operating costs, and a funded, occupied capacity schedule with customer collections. A price move alone does not prove the business case; a stronger business case does not automatically create a short entry.

Risk Audit

Risk Current evidence Effect on decision
Cash conversion Nine-month operating cash use of $113.8 million; equipment purchases of $75.5 million. Does not establish negative unit economics, but weakens the positive-EBITDA headline as a cash signal.
Capital structure Convertible notes, lease liabilities, warrants, multiple share classes, and conflicting vendor share counts. Blocks a precise per-share valuation bridge.
Positioning 36.58 million reported short shares; provider float percentages conflict and settlement predates results. Squeeze risk is material; live borrow is unknown.
Execution No current locate, borrow cost, recall terms, spread, depth, volume-quality, or exit-capacity evidence. Entry remains null and execution disabled.
Gap and event risk Earnings, capacity, Tether, or financing updates can move the stock outside regular hours. A stop cannot enforce a 5% adverse bound.
Market data RUM closing-price and volume snapshots differ across history pages; after-hours observations are stale and separate. Use the labeled $5.96 regular close only as reference.

Best Trade Strategy

No trade. Do not short RUM from this screen, and do not use options without a current chain, verified spreads, and maximum-loss analysis. Revisit after the Q4/FY2026 report and the next capacity update. A new screen would need a verified regular-session price, a supportable base decline of at least 10%, a credible adverse move no greater than 5%, at least 2:1 gross reward/adverse risk, positive expected value after borrow and trading costs, and verified locate, borrow, spread, depth, and exit capacity. None is established here.

Sources

Research Quality Scorecard

Criterion Score Evidence-based reason
Market disagreement 3/5 The stock rose sharply on an acquired compute revenue ramp while cash conversion remains weak, but the rally may be justified.
Evidence base 4/5 Fresh issuer release, post-result close, and published SI; no live borrow data or reconciled share denominator.
Positioning and flows 2/5 Short count is meaningful, but float denominators conflict and the settlement predates the report.
Catalyst path 3/5 Q4 results and 2027 capacity tests are observable, but the next report date and capacity timing are not confirmed.
Payoff architecture 2/5 Historical price anchors show severe adverse risk; no defensible intrinsic valuation or executable expression.
Invalidation discipline 4/5 Cash conversion and capacity tests are explicit; no false price-only stop is claimed.
Differentiated insight 4/5 The revenue bridge separates Quake's acquired scale from the small year-over-year gain in Rumble revenue.
Client value 5/5 The screen explains why a plausible bearish narrative still fails the short hurdle.
Total 27/40 Short note / no-trade screen only. The failed 10/5 gate controls the decision.

Bottom Line

The Q3 report proves that Quake AI can add revenue. It does not yet prove durable cash generation, and it does not make RUM a bounded short. The $5.96 close after a 16.6% rise, stale short-interest data, share-count uncertainty, and a broad upside path fail the Desk's 10/5 test. Reject / no trade.

AI Illustration Prompt

Create an editorial illustration of a GPU data-center aisle reflected in a Nasdaq closing screen, with a clear $5.96 RUM close and a small meter that separates revenue from cash generated. Place a commissioning calendar marked Q4 2026 and 2027 beside a server rack; show a portion of unmonetized capacity fading beyond the calendar edge. Use graphite, cool steel blue, restrained white, and one muted amber signal. The mood is analytical and skeptical, not apocalyptic or promotional. Keep the composition sparse, realistic, and readable, with a subtle but legible “The Mispricing Desk” watermark on the glass of the rack.