Bitdeer Technologies

Phase A — Understand the business

Lens 1 · Company Overview

Bitdeer is a vertically integrated Bitcoin-mining group that has become its own ASIC designer and is now attempting a third act as an AI/HPC datacenter operator. Three things are happening inside one P&L, and the dossier only makes sense if you hold all three:

1. Self-mining — the core. Bitdeer runs its own fleet of Bitcoin miners in its own datacenters and sells the hash to mining pools (FPPS mechanism) for BTC. $396.0M revenue in 2025, 63.9% of total, +142.8% YoY ``. Revenue here is mechanically BTC price × hash rate × uptime − network difficulty.
2. SEALMINER hardware — Bitdeer designs its own mining chips (SEAL01→SEAL04) and rigs (SEALMINER A1→A4), uses them internally, and sells them to third parties. Rig sales went from $0.6M (2024) to $108.3M (2025), 17.5% of revenue — a brand-new business line that exists because the SEAL02/A2 reached mass production ``.
3. Datacenter services + the AI pivot — General Hosting, Membership Hosting, Cloud Hash Rate (all legacy, all shrinking), plus a nascent AI Cloud business ($6.8M in 2025, ~$69M ARR by mid-2026 ``) that is converting Bitcoin sites to NVIDIA GPU colocation/cloud.

Customers: Self-mining "customers" are mining-pool operators (Bitdeer mainly uses FPPS pools). Hardware customers are third-party miners. AI-cloud customers are enterprises renting GPU capacity. The legacy hosting/cloud-hash-rate book is rolling off as old rigs are decommissioned post-halving ``.

Contract structure — important: Self-mining contracts with pools have a duration of less than a day and continuously renew; there is no take-or-pay, no recurring lock-in, no backlog. The revenue is spot-BTC exposure, full stop. Hosting/membership carry monthly consumption-based fees + variable yield kickers ``. The AI-cloud book is where the only "contracted, recurring" revenue could come from — and it is the smallest segment.

Scale (the asset base is the real story): 10 operating datacenters across the US, Norway, Bhutan, Ethiopia, Malaysia, 1,744 MW of electrical capacity in use as of 2026-03-31, with a pipeline to ~3,003.5 MW, and 78.1 EH/s of hash rate under management ``. The power footprint — not the current AI revenue — is the asset the bull case is really underwriting.

Lens 2 · Supply Chain

Map upstream → Bitdeer → end customer, named:

• Foundry / chip fabrication (the chokepoint): SEAL01–04 chips are designed in-house but fabbed at advanced process nodes by third-party foundries (the 20-F repeatedly cites reliance on suppliers "to produce the advanced process nodes used in our chips" and warns of allocation risk ``). The filing does not name TSMC/Samsung explicitly, but advanced-node Bitcoin ASICs at sub-15 J/TH are a TSMC-class process. This is the single-source dependency — if foundry allocation slips, the whole vertical-integration thesis slips. (The SEAL04 delay that triggered the securities suit is exactly this risk materializing — see Lens 10.)
• GPU supply (for the AI pivot): NVIDIA H100, H200, B200, GB200 NVL72, and GB300 NVL72 systems, procured via individual POs from NVIDIA and OEMs ``. No long-term GPU supply agreement disclosed — they buy on POs into a constrained market. Contrast IREN, which has NVIDIA as a strategic investor (see Lens 7).
• Mining-rig peripherals / containers: sourced via Bitdeer Inc. (fka Bitmain Inc.) supply agreements, e.g. Kemet Foil Manufacturing (copper foil) ``.
• Power (the other chokepoint): electricity is 33.9% of total cost of revenue ($273.0M in 2025) ``. Sites are sited for cheap/stranded power: Bhutan hydro (600 MW), Norway hydro (309 MW), Ethiopia (50 MW), Texas/Ohio (US grid + planned on-site gas in Alberta). Power contracts (e.g. Clarington's 570 MW under contract with a local utility) are the gating item for the AI build-out — and one (Clarington) is under litigation (American Heavy Plate Solutions, see Lens 10).
• Bitdeer → mining pools (BTC) / hardware buyers / AI-cloud tenants.
• Construction partners: Data Center Installations AS (Norway, Tydal AI conversion EPC, open-book cost-plus) ``.

Chokepoints: (1) advanced-node foundry allocation for SEAL chips; (2) NVIDIA GPU allocation for AI cloud; (3) grid power + interconnect timing for the 1,259.5 MW pipeline. All three are external and all three are tight in 2026.

Lens 3 · Competitive Advantages (moats)

What's genuinely defensible:

• Self-fabbed ASICs at industry-leading efficiency. SEALMINER A4 (SEAL04) hit 9.45 J/TH in April 2026 — among the best published figures in the industry, improving on the A3's 12.5 J/TH and the A2 Pro's 14.9 J/TH . Bitdeer is now "the third credible ASIC player" alongside Bitmain and MicroBT (Whatsminer), with Canaan (Avalon) trailing ``. Owning the chip means owning your own fleet cost curve and a hardware-sales revenue line — a structural edge over miners who buy from Bitmain at retail.
• Power portfolio. 1.7 GW in use, ~3 GW pipeline, much of it cheap hydro (Bhutan/Norway). Power + interconnect is the binding constraint for both Bitcoin and AI — this is the optionality the equity trades on.
• Vertical integration. Chip → rig → datacenter → mining/hosting/AI under one roof; few peers span the whole stack.

Why the moat is thinner than it looks:

• Self-mining has no moat. It is a price-taker on a commodity (BTC) with a continuously-renewing sub-daily "contract." The only edge is cost-per-TH, which the entire industry is racing to compress — Bitdeer's own A4 forces faster replacement cycles that erode everyone's edge, including its prior fleet ``.
• The ASIC edge is a treadmill, not a castle. Bitmain and MicroBT will respond; ASIC leadership is measured in months, and the SEAL04 delay lawsuit shows the roadmap is not a guarantee.
• The AI moat is mostly absent today. Bitdeer is renting NVIDIA GPUs on POs with ~$69M ARR; it has no anchor hyperscaler contract. IREN (NVIDIA + Microsoft deals) shows what an AI moat looks like, and Bitdeer doesn't have it yet (Lens 7).

Bargaining power: weak with foundries and NVIDIA (it needs them more than they need it); moderate with mining pools (commoditized); improving with hardware buyers as SEALMINER gains share.

Lens 4 · Segments

Revenue by segment, all `` (US$ thousands):

The story the table tells: Bitdeer has deliberately collapsed its capital-light, recurring service businesses (cloud hash rate, hosting) and reallocated everything into self-mining + selling rigs — i.e. it has increased its raw BTC-price beta and its capex intensity at the same time. Cloud Hash Rate (a hedge that smooths BTC volatility) went from 18.4% of revenue to 0.3%. The 2025 revenue growth is almost entirely (i) more hash pointed at self-mining (25.1 EH/s avg vs 7.5 EH/s) plus higher avg BTC, and (ii) a brand-new hardware line. Geography is not segment-disclosed in the income statement, but the asset base skews Bhutan/Norway (hydro) + US (Texas/Ohio) ``.

---

Phase B — Measure performance

Lens 5 · Earnings Result

FY2025 income statement (US$ thousands, ``):

Read it carefully — the "profit" is an accounting mirage. The $65.6M FY2025 net profit is entirely manufactured by a $365.0M "other net gains" line, of which $444.9M is a non-cash gain on the fair-value change of derivative liabilities (the convertible-note conversion features + Tether warrants) — a number that swings violently with Bitdeer's own share price and flips to a loss when the stock rises. In 2024 the same line was a −$507.5M loss. Strip it out and the operating reality is: gross profit of just $61.0M on $620M revenue, against $245M of opex (R&D + G&A + selling). The core business lost money before the derivative magic.

The honest scorecard is management's own non-IFRS ``:

• Adjusted EBITDA: $35.2M (2025) — DOWN from $37.4M (2024) and $93.5M (2023), despite revenue +77%. That is the tell: revenue nearly doubled and adjusted EBITDA shrank, because gross margin collapsed from 21% to 10%.
• Adjusted loss: −$229.9M (2025) vs −$51.3M (2024) and +$15.2M (2023). On a clean basis the company is loss-making and the loss is widening.

Margin moves & why: Gross margin fell to 9.8% because (i) electricity cost +51.9% to $273.0M as the fleet scaled, (ii) mining-rig depreciation +326% to $83.0M (more rigs + a useful-life cut from 2–5y to 2–3y effective July 2025, accelerating the charge), and (iii) low-margin hardware sales (COGS $89.3M on $108.3M revenue ≈ 18% gross) ``. R&D doubled to $153.9M — the ASIC bet is enormous relative to a $61M gross-profit base.

Balance-sheet flags (these matter more than the P&L):

• Operating cash flow: −$1,738.7M in 2025 (vs reported net profit +$65.6M). The gap is a $786.9M inventory build + $408.2M prepayments for SEALMINER mass production, plus $560.5M of revenue that is non-cash (BTC received and not sold) ``. The company is consuming cash at an extraordinary rate to fund the chip ramp.
• Cash collapsed from $476.3M (2024) to $149.4M (2025) ``. Plus crypto $83.1M + crypto-receivables $135.6M.
• Borrowings ~$1.0B, including three convertible tranches and $521.8M from BIT Group, a related party ($275.0M current) — see Lens 10/13 ``.

Guidance: Bitdeer does not give formal revenue/EPS guidance (the guidance.csv is empty; the 20-F gives capacity timelines, not financial targets). Forward visibility is therefore poor and BTC-dependent.

Q1-2026 print (post-filing, the most important update) ``:

• Revenue $188.9M, +170% YoY (vs $70.1M Q1-25). Self-mining $146.9M; avg self-mining hash rate 63.2 EH/s (+551% YoY); hash rate reached ~65 EH/s by March.
• Gross LOSS of −$39.0M (cost of revenue $228M). Net loss −$159.5M (vs +$105.3M Q1-25, which had a derivative tailwind).
• This is the whole thesis in one print: Bitdeer scaled hash rate ~8x into a BTC price that fell ~50% from its Oct-2025 high — and self-mining went gross-margin-negative. Mining more BTC at a falling price with rising difficulty = a loss. The chip edge did not save the unit economics.

Market reaction: BTDR at ~$17.88, market cap ~$4.16B (2026-06-29), 52-wk range $6.92–$27.80 . The stock has been a high-beta BTC proxy with violent swings (Lens 8). Bitcoin itself is **~$59,860 (2026-06-29), down ~$48.5k YoY, lowest since 2024** — the macro backdrop is actively hostile.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts in the research layer (transcripts/ empty). From web ``:

• Management focus has shifted from "we are a Bitcoin miner" to "we are an AI/HPC infrastructure company that mines Bitcoin." The recurring phrases on recent calls: power capacity (1.7 GW → 3 GW), SEALMINER efficiency leadership, datacenter conversion to AI, GB300 NVL72 deployments, and AI Cloud ARR (~$69M, "90% GPU utilization").
• Tone trajectory: through 2025, increasingly promotional on the AI optionality and the SEAL04/A4 efficiency record. The Q1-2026 call had to pivot to defense — explaining a gross loss, the BTC drawdown, and managing the SEAL04-delay litigation narrative. The company "missed Q1-2026 EPS projections" ``.
• What they've stopped saying: the cloud-hash-rate / hedging language is gone (that business was killed). The framing is now all-in on self-mining scale + AI conversion.
• Executive churn: "Two Bitdeer executives resigned as Jihan Wu pushes company transformation" `` — a yellow flag on the AI-pivot execution and on Wu consolidating control.

Lens 7 · Comps

Bitcoin-miner / mining-AI cohort.

5-yr avg ROE: n/a (the cohort is too young / too loss-making for a meaningful 5-yr ROE; Bitdeer itself only listed via SPAC in 2023).

The comp lesson is the whole bear/bull pivot: the market is not paying for power capacity or hash rate anymore — it is paying for signed AI compute contracts. IREN, a former pure BTC miner of similar pedigree, is worth ~4–5x Bitdeer because it converted its power into NVIDIA and Microsoft contracts. Bitdeer has the power (1.7 GW, ~3 GW pipeline) and the conversion plan but only ~$69M of AI ARR and zero hyperscaler anchor. The entire re-rating case for BTDR is "become IREN"; the entire bear case is "you're still MARA/RIOT with a chip division and a related-party balance sheet."

Lens 8 · Stock-Price Catalysts (>5% moves, last ~5y)

BTDR has only traded since its April-2023 SPAC merger (Blue Safari), so the window is ~3 years. Pattern ``:

• Bitcoin price is the dominant driver. The stock's 52-wk range ($6.92–$27.80) maps to the BTC cycle; the ~50% BTC drawdown from Oct-2025 to Feb-2026 dragged BTDR from the high-$20s toward the low-teens.
• SEALMINER chip milestones move the stock both ways — A4/9.45 J/TH launches were positive catalysts; the SEAL04 delay disclosure caused a single-day ~14% drop ($17.65 → $15.02) and triggered the securities class action ``.
• Financing & dilution events move it: convertible-note pricings, the Tether private placement (positive — a marquee crypto strategic), and equity offerings (the Feb-2026 raise priced at $7.94, a ~43% discount to the Nov-2025 $13.94 raise — a brutal dilution signal as BTC fell) ``.
• AI-pivot news (GB300 clusters, ARR updates, Norway/Tennessee conversions) move it positively but have not yet produced an IREN-style step-change because no anchor contract exists.

What the market actually reacts to: (1) BTC spot, (2) the SEALMINER roadmap (now legally contested), (3) AI-contract hope, (4) dilution. It is a levered, headline-driven BTC call with two embedded lottery tickets (chips, AI).

---

Phase C — Judge people & books

Lens 9 · Management

Jihan Wu (40) — Founder, Chairman, CEO. The central figure and the central risk.

• Track record: Co-founded BitMain (2013), the company that dominated Bitcoin ASICs for a decade. Founded BIT Group (2019, the Matrixport/Matrix Finance complex). Founded Bitdeer; spun it out of BitMain in Jan 2021; became CEO in March 2024 ``. This is a genuinely formidable operator who has built the most important hardware franchise in the industry — twice if SEALMINER succeeds.
• Skin in the game / control: Wu holds 69.5% of voting power with only ~19.5% economic ownership, via 44.4M Class V super-voting shares (10 votes each) held through Victory Courage Ltd → Cosmic Gains → a VISTRA-trusteed irrevocable trust (Wu = settlor + beneficiary) ``. Bitdeer is a Nasdaq "controlled company" — Wu can outvote all other holders combined. All directors+officers as a group: 2.8% of Class A, 69.8% of voting power.
• Capital-allocation history: Aggressive, debt-and-dilution-funded growth. In 2025 he raised ~$751M convertibles + $668M related-party loans + $351M equity to fund the SEALMINER inventory ramp ``. He bought back $65M of treasury shares and diluted via ATM + offerings in the same year — a mixed signal. The clean-EBITDA decline while revenue doubled suggests capital is being deployed into a low-return core (self-mining) at the worst point in the BTC cycle.
• Red flags (material): (1) Pervasive related-party financing — Wu's own company BIT Group is Bitdeer's largest lender (~$522M + BTC loans + structured products). The CEO sits on both sides of the company's funding. (2) Controlled-company governance — minimal minority-shareholder protection; index-inclusion opposition from advisory firms noted in the filing. (3) A securities class action naming Wu personally for the SEAL04 "on track" statements (Lens 10). (4) Executive resignations during the AI transformation ``.
• Archetype: Founder-operator/empire-builder with absolute control and a self-dealing-adjacent financing structure. Upside: visionary hardware founder. Downside: a one-man-controlled, related-party-funded vehicle at a cyclical low.

Other execs: Linghui (Matt) Kong (CBO, ex-BitMain), Haris Basit (CSO) — both named in the securities suit. Chao Suo (ex-BitMain HR). The bench is BitMain alumni — deep hardware DNA, thin on independent governance.

Lens 10 · Forensic Red Flags

Forensic lens — every figure labeled.

1. Earnings quality is the headline risk. The $65.6M FY2025 "net profit" is non-cash and reverses with the share price: $444.9M of it is a fair-value gain on derivative liabilities (convert features + Tether warrants) ``. This is negative earnings quality — the number flatters in a down-stock year and will hurt reported profit in an up-stock year. The economic truth is the adjusted loss of −$229.9M and operating cash flow of −$1,738.7M.

2. Cash flow vs earnings divergence is extreme. Net profit +$65.6M against operating cash flow of −$1.74B — a ~$1.8B gap, driven by an $786.9M inventory build and $408.2M of prepayments for SEALMINER production, plus $560.5M of revenue received in BTC and not converted ``. Inventory and prepayments are outrunning revenue by a wide margin — classic build-ahead risk if hardware demand or BTC disappoints (and Q1-26 showed BTC disappointing).

3. Crypto-asset accounting. Bitcoin (ex-USDC) is carried as an indefinite-lived intangible at cost, impaired down but not marked up under IAS 38 ``. So the balance-sheet crypto ($83.1M) understates value in up-markets and the income statement takes impairments in down-markets (a $3.8M crypto impairment + a $26.7M FV loss on BTC collateralized for the BIT Group loan in 2025). Asymmetric, and the BTC-collateral mark moves with the same price that drives revenue.

4. Related-party concentration (the big one). BIT Group (Jihan Wu's company) is the dominant counterparty: ~$521.8M of loans at FY-end, the BIT Assets Collateralized Loan (drawn $400M, ~9% + reference rate, secured by SEALMINERs + inventory + datacenter assets), additional $300M facilities, a BTC Collateralized Loan ($400M facility), a 6,000-BTC borrowing facility (post-FY, ~2,802 BTC outstanding at 3%), and a BIT Structured Product Agreement (puts/options/accumulators) ``. The company's solvency is materially entangled with its CEO's private vehicle, on terms set between related parties.

5. Depreciation policy change. Useful life of mining rigs cut from 2–5y to 2–3y effective July 2025, accelerating depreciation in H2-2025 ``. Conservative direction (good), but it depressed reported gross margin and should be remembered when modeling forward D&A.

6. SBC. Share-based payments $38.5M in 2025 (down from $45.5M in 2023) — material but not egregious; ~6% of revenue. Add-back inflates adjusted EBITDA.

7. Going concern / internal controls: Auditor (PCAOB ID 206) issued an unqualified opinion on both the financials and ICFR; no going-concern qualification ``. Management asserts liquidity is sufficient for 12 months — but that assertion leans on "opportunistic" future convertible/equity/related-party raises, i.e. continued access to capital markets and to BIT Group.

Regulatory findings (required sub-section):

• SEC enforcement (LR/AAER): None. Verified via SEC EDGAR EFTS — 0 Litigation Releases and 0 AAERs naming Bitdeer Technologies, 2021-06-29 → 2026-06-29 ``.
• Securities class action (material): Sakar v. Bitdeer Technologies Group et al., filed Dec-2025 in the U.S. District Court, SDNY, naming Bitdeer and execs Jihan Wu, Matt Linghui Kong, Haris Basit, alleging false/misleading statements that SEAL04 ASIC R&D was "on track" over a June 6, 2024 – Nov 10, 2025 class period; on the delay disclosure the stock fell ~14% ($17.65 → $15.02) . This strikes at the exact chip narrative the bull case relies on.
• Commercial litigation: American Heavy Plate Solutions, LLC v. Erie Creek / Whitetail Creek / Bitdeer Inc. (Jan-2026, Court of Common Pleas, Monroe County, Ohio) — alleges the planned Clarington, Ohio datacenter (570 MW, a key AI-pipeline site) violates a restrictive covenant; may delay power/construction ``.
• Non-SEC agencies (FTC/DOJ/etc.): web search surfaced no material consent decree, fine, or penalty beyond the litigation above ``.
• Net: No accounting-fraud enforcement. But a live SDNY securities-fraud class action on the chip roadmap + a site-level construction suit are both material and both hit the growth story directly.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection

No formal company guidance; the modeling task is unusually BTC-dependent. All outputs `` with arithmetic; no forecast.ts logged (watchlist mode — breadth only).

Bitdeer does not earn a stable EPS — earnings are dominated by (a) BTC-price-driven self-mining margin and (b) non-cash derivative swings. So I project adjusted EBITDA / clean operating outcome, not a GAAP EPS line (a GAAP EPS would be a meaningless function of the share price via the derivative remark).

Anchor actuals: FY2025 revenue $620.3M, gross profit $61.0M (9.8%), adjusted EBITDA $35.2M, adjusted loss −$229.9M . Q1-2026 annualizes to ~$756M revenue but at a **gross loss** .

Base (FY2026) — BTC range-bound ~$60–70k: Self-mining scales to 70–90 EH/s but at compressed/negative spreads; hardware sales soften as the whole industry's margins tighten post-halving; AI Cloud grows off a tiny base ($69M ARR → maybe $120–150M).

• Revenue ~$750–900M ``.
• Adjusted EBITDA roughly breakeven to mildly negative ``.
• Clean (adjusted) loss likely −$200M to −$300M as R&D stays ~$150M+ and depreciation steps up ``.

Bull (FY2027) — BTC re-rates >$90k AND ≥1 anchor AI contract signed: self-mining spreads turn positive on the A4 fleet; AI conversions (Tennessee Q4-26, Washington GB300, Tydal) come online; the stock re-rates toward an IREN-style "AI infra" multiple on the 3 GW pipeline. Adjusted EBITDA could swing to +$300–500M and the equity could double-plus on the multiple alone ``.

Bear (FY2026–27) — BTC stays <$60k / drifts lower: self-mining stays gross-margin-negative; the $1B+ debt load (much of it related-party, ~9% cash cost) compounds; covenants/LTV on BIT Group collateral (SEALMINERs whose resale value falls with BTC) tighten; forced dilution at depressed prices (cf. the $7.94 Feb-26 raise). Adjusted loss widens past −$350M; equity retests the $6.92 low or below ``.

The single swing variable is BTC price, amplified by ~1.5–2x operating + financial leverage. This is structurally a levered Bitcoin call with two call options stapled on (SEALMINER share gains; an AI re-rate). Per SKILL provenance discipline, I will not invent a point EPS: n/a for a precise FY26/27/28 GAAP EPS; the base case is "clean operating losses persist absent a BTC recovery."

Lens 12 · Bull vs Bear

Bull case. Bitdeer is a vertically integrated power-and-silicon platform trading at ~$4.2B that the market is mispricing as a dying BTC miner. It owns (1) industry-leading ASICs (A4 at 9.45 J/TH — better than the Bitmain/MicroBT fleet most rivals run), giving it the lowest fleet cost curve and a hardware-sales annuity; (2) 1.7 GW of cheap power scaling to ~3 GW, the scarcest input in both Bitcoin and AI; (3) an AI-conversion pipeline (Tydal/Tennessee/Washington/Clarington) that, if it lands even one IREN-style hyperscaler contract, re-rates the whole equity by multiples — IREN proves a ~$4B miner can become a ~$20B AI-infra name in a year. Embedded surprise: an NVIDIA/Microsoft-class anchor deal, or a sharp BTC recovery that flips self-mining to fat spreads on the most efficient fleet in the world. Tether's ~19.7% strategic stake and Wu's BitMain pedigree are credibility anchors.

Bear case (permanent-impairment risks). (1) The core is structurally unprofitable at current BTC — Q1-26 proved that scaling hash 8x into a 50% BTC drawdown produces a gross loss; self-mining has no moat and no hedge after they killed cloud-hash-rate. (2) The balance sheet is a related-party house of cards — ~$1B debt, the largest slice from the CEO's own BIT Group, secured by SEALMINERs and BTC whose collateral value falls exactly when the company is stressed; a BTC leg-down triggers LTV/dilution spirals. (3) The bull narrative (SEAL04 chips) is in active securities litigation — an SDNY fraud suit alleging management lied about the chip being "on track." Pre-mortem (18 months out, thesis broken): BTC sat at $50–60k through 2026, self-mining bled cash, the AI conversions slipped (Clarington litigation, GPU/interconnect delays) and never landed an anchor tenant, covenants forced repeated raises at single-digit prices, and the stock made new lows while IREN/CoreWeave took the AI-infra mindshare. Multiples: on clean numbers BTDR is loss-making, so "the multiple is too high" is almost tautological — you're paying ~$4.2B for negative adjusted EBITDA and an option on two things going right at once.

Contrarian view (what the market refuses to see): Bulls price BTDR as a future AI-infrastructure company; the Q1-26 gross loss says it is still a sub-scale-margin Bitcoin miner that took on AI-sized leverage at a cyclical low — and its lender of last resort is its own CEO. The "AI optionality" is real but is being financed by a balance sheet that only works if BTC cooperates. The asymmetry is not obviously favorable here: the downside (BTC stays weak → dilution spiral) is concrete and near; the upside (anchor AI contract) is real but unsigned.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks the money machine: self-mining is a price-taker on BTC with sub-daily, no-lock-in "contracts." There is no recurring revenue, no backlog, no pricing power. When BTC fell ~50%, the most efficient fleet on earth still ran a gross loss (Q1-26) ``. Efficiency leadership doesn't create a moat; it just sets the marginal cost the whole network races toward.
• Revenue concentration / what shifts: 63.9% of revenue is one commodity's spot price ``. The hedge (cloud hash rate) was deliberately dismantled to chase BTC beta — so concentration increased into the drawdown.
• Why the moat is weaker than bulls think: Bitmain and MicroBT are not standing still; the ASIC lead is months, not years; and the SEAL04 delay (now litigated) shows the roadmap is not deterministic. The AI "moat" is ~$69M ARR on rented NVIDIA GPUs with no anchor — that's not a moat, it's a pilot.
• Most dangerous competitor bulls underestimate: IREN — same miner-to-AI pedigree, but it actually signed NVIDIA ($2.1B equity + $3.4B cloud) and Microsoft (~$9.7B) and is worth ~4–5x Bitdeer ``. Every quarter Bitdeer doesn't sign an anchor, IREN/CoreWeave take the scarce hyperscaler demand and the GPU allocation.
• Worst capital-allocation / governance: the related-party financing web — the CEO's BIT Group is the principal lender (~$522M + BTC + structured products), on related-party terms, secured by depreciating mining rigs. Combine with 69.5% super-voting control and a securities-fraud suit naming the CEO, and minority holders are passengers.
• Assumptions required for today's price: that BTC recovers materially AND the AI pivot lands an anchor AND the chip roadmap (under litigation) executes AND the balance sheet survives without crippling dilution. That's a four-way parlay.
• If growth disappoints 20–30%: with negative clean EBITDA and ~9% related-party debt, a revenue shortfall doesn't dent EPS — it accelerates the cash burn and the dilution spiral (the Feb-26 $7.94 raise, down 43% from Nov-25, is the template).
• Single scenario that permanently impairs: BTC range-bound $50–60k through 2027 → self-mining cash-flow-negative → BIT Group LTV breaches force collateral calls / dist­ressed equity raises → relentless dilution → equity makes new lows while the AI pipeline stays unfinanced. Plausibility: moderate-to-high given BTC is already at ~$59.9k and falling ``.

Lens 14 · Management Questions (ordered by information value)

1. The SDNY class action alleges you told investors SEAL04 was "on track" while it was delayed — walk through the actual chip-development timeline vs. your public statements, and what changed in your disclosure controls since.
2. Q1-2026 self-mining ran a gross loss. At what BTC price does self-mining turn gross-margin-positive on the A4 fleet, all-in (electricity + depreciation), and what % of your fleet is A3/A4 today?
3. BIT Group — your own company — is your largest lender. Quantify total related-party exposure, the LTV triggers, and what happens to Bitdeer if BIT Group calls or can't roll the BTC-collateralized facilities.
4. You've signed zero hyperscaler AI contracts while IREN signed NVIDIA and Microsoft. What is the realistic timeline and probability of an anchor tenant, and what's blocking it — GPU allocation, power timing, or counterparty trust?
5. Operating cash flow was −$1.74B in 2025 on a $786.9M inventory build. How much of that SEALMINER inventory is committed/sold vs. speculative, and what's your impairment exposure if hardware demand softens with BTC?
6. Cash fell to $149.4M. Lay out the FY2026 funding bridge without new related-party loans — how many quarters of runway at the Q1-26 burn, and at what share price are you willing to issue equity?
7. You killed the cloud-hash-rate hedge to maximize BTC beta right before a 50% drawdown. Defend that allocation decision and explain how you now manage BTC-price risk.
8. Which datacenter conversions (Tydal, Tennessee, Washington, Clarington) are firmly financed and on schedule, and how does the Clarington (AHP) litigation change the 570 MW timeline?
9. Depreciation life was cut to 2–3 years in July 2025. What does that imply about real fleet economic life and the embedded loss on rigs bought at the prior 2–5y assumption?
10. As a Nasdaq "controlled company" with 69.5% of votes in a family trust, what concrete minority-protection mechanisms exist on related-party transactions?
11. What is the steady-state gross margin and ROIC you target for the AI-cloud business, and how does it compete with CoreWeave/IREN on price and GPU access?
12. SEALMINER's edge is months-deep. What's the SEAL05+ roadmap and your foundry-allocation security to stay ahead of Bitmain/MicroBT?
13. Tether/Devasini hold ~19.7%. What does that relationship entail beyond capital — order flow, stablecoin treasury, governance — and are there conflicts?
14. Two executives resigned during the transformation. What happened, and how stable is the senior team executing the AI pivot?
15. Give a hard capital-allocation rule: at what BTC price or cash level do you stop growing self-mining and prioritize the balance sheet over hash-rate share?

---