Sigma Lithium

Phase A — Understand the business

Company Overview

Sigma Lithium Corporation is incorporated in Canada (Toronto registered office), operates entirely in Brazil, reports in USD under IFRS, and trades as SGML on Nasdaq . Auditor is **Grant Thornton Auditores Independentes, São Paulo** . Shares outstanding: 111,402,979 as of 2025-12-31 ; corroborated at 111.40M .

How it makes money: one revenue line — sale of lithium-oxide (spodumene) concentrate, grade nominally SC5.5, invoiced on Chinese SC6 CIF benchmarks with grade/price adjustments. Nameplate at Grota do Cirilo (Greentech Plant 1) is 270,000 t/yr of concentrate (~38–40kt LCE) . FY2025 revenue **US$110.01M** (down from US$151.35M in 2024) .

Customers/counterparties: historically Glencore (multiple 22kt shipments 2023–24, 50% prepaid at a premium), Yahua, and — critically — Mitsui, which terminated its 2019 pre-agreement in Jan 2023 . In 2025 Sigma signed **US$146M of new offtake prepayments**: US$96M for 70,500t (2026 delivery) and US$50M for 40,000t/yr over three years . The commercial model has shifted from spot-to-Glencore toward prepayment-financed offtake — a tell that traditional working-capital debt is scarce.

Contract structure: no take-or-pay of note; pricing floats with the SC6 benchmark, so revenue is fully exposed to the lithium cycle. This is not a recurring-revenue or contracted-margin business — it is a commodity price-taker.

Commercial-layer files for critical-materials are all marked missing in the Step-0 context; Lens 1–4 are therefore web/filing-grounded, not KB-grounded.

Supply Chain

Upstream inputs → Sigma → end customer, named:

• Deposit/inputs: self-mined pegmatite ore (Grota do Cirilo, Minas Gerais). Reserve 77.0 Mt @ 1.40% Li₂O (P&P), resource ~109 Mt @ 1.39%, ~25-year mine life at 520ktpa ``. Grade is genuinely high for hard rock — a real physical advantage. Reagents (dense-media/flotation) and diesel are the main consumables; DMS ("dry-stacking", low water) is the ESG pitch.
• Processing: Greentech Plant 1 (270ktpa, operating); Greentech Plant 2 under construction to reach 520ktpa ``.
• Logistics chokepoint: trucked ~500 km to the Port of Vitória (Espírito Santo) → shipped to China. Inland trucking + a single export port is the physical single point of failure.
• Mining contractor: work is contracted out — and Sigma abruptly switched mining contractors in late Oct 2025, demobilising the mine for "restructuring," which triggered the November crash ``. Contractor dependence is a live operational risk, not a theoretical one.
• Buyers (converters): Glencore (trading intermediary), Yahua (Chinese converter), plus the unnamed 2025 offtake counterparties. End demand = Chinese lithium-carbonate/hydroxide plants → cathode → EV/ESS batteries.

Chokepoints/single-source: (1) one mine, (2) one export port, (3) contractor-operated mining, (4) demand concentrated in Chinese converters. Four serial single points of failure — the supply chain is thin by design (it's one asset).

Competitive Advantages (moats)

What's real:

• Grade + cost. Sigma reports CIF-China cash cost ~US$442–458/t in 2025 and an all-in sustaining cost of ~US$592/t on 2026 guidance ``. In hard-rock terms that is genuinely first/second-quartile — driven by the 1.4% grade. This is the one durable edge.
• Reserve scale + life. 25-year, 77Mt reserve is a large, long-life orebody — optionality most juniors lack.
• Jurisdiction/logistics. Brazil (vs. Australia/Africa) is closer to Atlantic markets and outside the China–Australia axis; a modest geopolitical-diversification premium.

What's marketing, not moat:

• "Triple/Quintuple Zero Green" branding. ESG positioning is a soft differentiator; spodumene is a fungible commodity priced on Li₂O content and impurities, not on carbon story. Converters pay for chemistry, not virtue. The green premium is real but thin and cyclical.
• Bargaining power is weak. A single-asset ~110ktLCE-scale producer selling into concentrated Chinese buying power has little pricing power; the shift to prepayment offtake is the buyers extracting terms.

Verdict on moat: the moat is a cost-curve position on one orebody, not a franchise. It protects margins at mid-cycle prices and evaporates at trough prices — which is exactly what 2024–25 demonstrated (a net loss despite "low cost"). Cost leadership among hard-rock peers, but no moat against the price.

Segments

our figures is empty , so no research-layer segment grounding exists. Functionally there is **one product segment** (lithium concentrate) and **one geography** (Brazil production → China-weighted sales). A secondary line emerged in 4Q25/1Q26: **~650,000 tonnes of "high-purity lithium fines"** — a lower-value byproduct/tailings stream sold to clear inventory during the restructuring, alongside ~5,000t of premium concentrate . This is not a durable second segment; it's inventory monetisation.

Trend: revenue decelerated hard — US$151.35M (2024) → US$110.01M (2025), −27% `` — as realised prices collapsed faster than volumes grew. Then a sharp 1Q26 re-acceleration (revenue +150% QoQ to US$42.3M) on the price rebound. Segment mix tells you nothing here; the price is the whole story.

---

Phase B — Measure performance

Earnings Result (latest print — Q1 2026)

The most recent print (Q1 2026, reported 2026-05-15) is the pivot the bull case rests on:

• Revenue: US$42.3M, +150% QoQ (vs. 4Q25), highest since 1Q25 (US$48M) ``.
• Net income: US$11.1M — a swing to profit ``.
• Margins (record): gross 61%, EBITDA 39%, net 26% ``.
• Volume: ~23,000t of concentrate sold ``.
• Realised price: US$1,790/t SC5 (≈US$2,150/t SC6) vs. US$630/t SC5 (US$756 SC6) in 3Q25 — a ~2.8x jump in realised price in two quarters ``. This is the entire beat: price, not operations.
• Debt: total debt US$134M, −21% YoY; trade-finance debt cut to US$19M . Cash rose to **~US$28M by 2026-05-15** from a razor-thin US$6.2M at 2025-12-31 ``.
• Market reaction: shares surged on the print ; the stock is up **+138% over the trailing 52 weeks** off the November low .

Flag vs. own history: the record margin is real but entirely price-driven and freshly recovered from a base of near-insolvency (US$6.2M cash in Dec-25, BofA flagging late vendor payments in Aug-25). One good quarter after a 29% two-day crash is a fragile foundation, not a trend.

Earnings Calls (sentiment trend)

transcripts/ is empty ``; sentiment is web-derived and labelled.

Tonal arc across the last ~4 calls ``:

• 3Q24–4Q24: confident — "record production," "significant cost reductions," Investor-Day triple-capacity plan.
• 1Q25–2Q25: defensive but disciplined — "cost outperformance," "deleveraging," "on-target production." Management pivoted the narrative from growth to survival/cost as prices fell.
• 4Q25 (post-restructuring): damage-control — "successful restructuring," mine remobilised, "resilience achieved throughout the down cycle." The abrupt contractor change is framed as a fleet upgrade "to match Greentech 3.0."
• 1Q26: victory-lap — "record profitability," "highest margins in history."

Recurring phrases: "Green/Quintuple Zero," "lowest-cost quartile," "self-funded/low-capex expansion," "climate pioneer." Things they stopped saying: the 2023-era "strategic review / final round" M&A language, and the US$15.3B NPV feasibility headline (see Lens 10). Sentiment is high-beta to the lithium price — the tone is a coincident indicator, not a leading one.

Comps

Multiples are `` with source/date, or n/a. Peers pulled from the lithium majors (index has no critical-materials peer set populated).

Read: SGML's EV/Sales 14.4x is extreme for a single-asset hard-rock producer — Pilbara (the cleanest comp) is a far larger, multi-asset spodumene pure-play at a ~$6B cap. On any normalised basis SGML trades at a scarcity/takeover premium, not a fundamentals multiple; the EBITDA multiple (585x) is meaningless on a trough-EBITDA denominator but signals the market is pricing forward recovery, not trailing reality. The honest comp is Pilbara — and SGML looks richly valued against it on an EV/resource-tonne and EV/Sales basis. (Precise per-tonne EVEnterprise valueWhat it would cost to buy the whole company: its market value plus its debt, minus the cash you would get with it. Often very different from market cap. comps: n/a.)

Stock-Price Catalysts (>5% moves, ~5-year pattern)

The tape reveals what the market actually reacts to for this name — and it is not quarterly operations:

• Feb 2023: Tesla takeover rumor (Bloomberg) — shares spiked; "Tesla weighing a bid" ``.
• Mar 2023: Grizzly Research short report — sharp drop; "bogus feasibility studies" ``.
• Nov 2023: "Strategic review advanced to a final round" — pop on deal hope ``; no deal ever closed — the single most important negative catalyst-that-wasn't.
• 2024: −64% for the year; 2025: −50%+ — the lithium-price collapse, the dominant multi-year force ``.
• Aug 2025: BofA flags vendor-payment delays / cash-flow stress ``.
• Late Oct–Nov 2025: abrupt mining-contractor switch → −29% in two sessions to ~$5.34 ``.
• Nov 2025: Paiva sells 56% of his holding at $9.84 ``.
• May 2026: Brazil waste-handling court ruling → −15% ``.
• Jun 9 2026: appeals court overturns the guarantee → +7% ``.
• May 2026: record Q1 print → surge; +138% trailing 52 weeks off the bottom ``.

Pattern: SGML trades on (1) M&A/takeover speculation, (2) the lithium price, (3) idiosyncratic Brazil/governance shocks — in that order of drama. Operations barely move it except when they signal distress. This is a headline/story stock, which is precisely why the valuation disconnects from trailing fundamentals.

---

Phase C — Judge people & books

Management

• Ana Cristina Cabral-Gardner — Founder-CEO & Co-Chair. Ex-Goldman Sachs MD, Head of LatAm Capital Markets; co-founded A10 Investimentos (2012/2013) with Marcelo Paiva; A10 acquired 100% of Sigma in 2015 and took it from greenfield to producer ``. Track record: genuinely built a producing mine from scratch and got Phase 1 to steady-state — non-trivial. But she is a banker/capital-markets operator, not a mine builder by background; the promotional, ESG-forward, capital-markets-savvy style is a double-edged sword (see red flags).
• Marcelo Paiva — Co-Chair, A10 portfolio manager, controls the largest shareholder (A10 FIA). Sold 56% of his personal holding at US$9.84 in Nov 2025 — the largest insider sale in 12 months ``.
• Skin in the game: insiders collectively own only ~1.6% (~US$17M) ``. Despite the "founders acquired 100%" origin story, direct insider alignment today is thin, and the most senior insider was a net seller into the trough. (The A10 fund stake is larger, but it is external LPs' money, not personal founder capital.)
• Capital-allocation history: self-funded Phase 1; secured a BNDES BRL 487M, 16-year loan at 7.45% covering ~99% of Phase-2 capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs. `` — a genuinely good, cheap, long-dated financing win. Offset by the balance-sheet brinkmanship (US$6.2M cash at Dec-25, late vendor payments) and the disruptive contractor switch. Mixed: shrewd financier, operationally jumpy.
• Founder vs. professional: founder-controlled, promotion-heavy. Implies vision and deal-making energy, but weak minority-shareholder protection and a habit of managing the story as hard as the mine.

Forensic Red Flags

our figures empty, so line-item forensics are web-grounded; labelled throughout.

• Cash flow vs. earnings: FCF −US$8.15M TTM against a −US$43.78M net loss and +US$2.58M EBITDA — EBITDA barely positive on a TTM basis means the "record margins" are a single-quarter, price-driven snapshot, not a run-rate. **Cash was near-critical (US$6.2M) at year-end 2025**.
• Vendor-payment stretching: BofA explicitly flagged increasing delays in vendor payments as a cash-flow-stress signal in Aug 2025 `` — a classic late-cycle liquidity tell. Watch payables/DPO in the exhibits (not on disk).
• Related-party structure: A10 (Paiva-controlled) provides administrative services to Sigma; CEO holds minority interests in related entities; transactions "reviewed by independent committees" ``. Founder-controlled + related-party service agreements + 1.6% insider float is a governance configuration that warrants scrutiny.
• Feasibility-study credibility: Grizzly's 2023 short alleged "bogus feasibility studies" and challenged the escalating NPV headlines (project NPV was marketed up to US$15.3B / 766ktpa in Jan 2023) ``. Sigma "vehemently denied" the report. No SEC enforcement resulted (see below), but the pattern of ever-larger resource/NPV headlines is a promotional flag to weigh.
• Reserve/resource inflation cadence: resource marketed 27% higher to 109Mt, reserve +40% to 77Mt, "world's 4th largest," "expected to reach 150Mt" . NI 43-101 (Canadian standard, explicitly **not comparable to SEC S-K 1300**, per the 40-F ) — a real, if standard-for-Canada, comparability caveat for US investors.
• Loss quality: net loss −US$50.19M FY2025 ; ROE **−49%** .

Regulatory findings (required sub-section):

• SEC Litigation Releases: None. `` (EDGAR EFTS LR, 2021–2026).
• SEC AAERs: None. ``.
• Non-SEC / other: the material live matter is the Brazil (Minas Gerais) environmental dispute — a lower court (May 2026) raised a possible US$10M financial guarantee over waste-handling at Grota do Cirilo affecting the towns of Araçuaí and Itinga; a Minas Gerais appeals court overturned it on 2026-06-09, instead requiring Sigma to fund independent environmental monitoring . There was also a **2023 Brazil mineral-rights dispute** that "cast a shadow" on the expansion . Environmental/permitting/mineral-title risk in Brazil is the standing legal overhang.
• Item 3 (Legal Proceedings): not captured — lives in AIF Exhibit 99.1, not on disk ``.
• Net: No US securities-enforcement history; the real forensic risks are liquidity brinkmanship, related-party governance, promotional resource/NPV marketing, and Brazilian environmental litigation — not accounting fraud on the record.

---

Phase D — Project & stress-test

Forward Projection (base / bull / bear)

No our model create in --watchlist mode. EPS estimation is `` with arithmetic; the driver is the SC6 price × volume, and this is a swing-to-loss name, so the projection is scenario-dominated, not a smooth ramp.

Anchors: FY2025 revenue US$110M, net loss −US$50M ; Q1 2026 run-rate ~US$42M rev / US$11M NI at a ~US$2,150 SC6 realised price on ~23kt/quarter ; 2026 guided production 240,000t at AISC US$592/t ``; ~111.4M shares.

• Bear (SC6 ~$800, back toward the May-2026 spot): at ~$800 CIF against ~$592 AISC, cash margin is razor-thin; realised (SC5.5, lagged) can dip below all-in cost. Revenue ~US$140–170M, net loss/breakeven. FY2026E EPS ≈ −$0.20 to $0.00 ``.
• Base (SC6 ~$1,300–1,500 mid-cycle): 240kt at a blended ~$1,100–1,200 realised vs. $592 AISC → EBITDA ~US$120–150M, net income ~US$60–90M. FY2026E EPS ≈ $0.55–$0.80 . (Implies a forward P/E of ~15–22x at $12.37 — consistent with the "Forward P/E 8.11" figure if you use the more bullish street volume/price.)
• Bull (SC6 >$2,000 sustained + Phase 2 online): the Q1-2026 margin (61% GM) annualised on 240kt→ ramping 520kt is US$150M+ net income; FY2027E EPS > $1.50 with volume doubling ``. This is the case the current EV/Sales embeds.

The projection is a fan, not a line — the standard deviation of the lithium price swamps every operating variable. Base case ≈ $0.55–$0.80 FY2026 EPS, entirely contingent on price holding above ~$1,300 SC6.

Bull vs Bear

Bull case. Highest-grade, lowest-cost hard-rock orebody in the Americas; AISC ~$592/t keeps it cash-generative through most of the cycle. BNDES cheap 16-yr financing de-risks the Phase-2 capacity doubling to 520kt (→125kt LCE by 2027) — volume growth plus a price recovery is a double-lever. 25-year reserve life. Brazil jurisdiction + green branding = a structural takeover target (Tesla looked once; the sector is consolidating — Rio bought Arcadium). Q1 2026 proved the asset prints ~26% net margins the instant price cooperates. Street is Strong-Buy, PT ~$18–20 ``.

Bear case (permanent-impairment risks).

1. Single-asset, single-commodity, price-taker — no diversification, no pricing power; one orebody + one port + contractor-operated mining. A prolonged sub-$800 SC6 environment (the 2024–25 reality) produces sustained losses and forces dilutiveDilutionIssuing new shares, so each existing share owns a smaller slice of the same company. raises or distressed offtake terms.
2. Balance-sheet fragility — US$6.2M cash at Dec-25 and BofA-flagged vendor stretching show how close to the edge it ran. The recovery is one quarter old; a price relapse re-opens the liquidity hole.
3. Governance/promotional overhang — 1.6% insider float, Co-Chair dumping 56% into the trough, related-party admin agreements, a history of ever-escalating NPV headlines and a live Grizzly-style credibility question. Minority holders are along for the ride, not aligned with.

Pre-mortem (18 months out, thesis broke): SC6 stalled at ~$900; Phase-2 ramp slipped (contractor/permit/waste-monitoring friction in Minas Gerais); the offtake prepayments came due against under-priced volumes; cash tightened again and Sigma did an equity raise at a discount; the "takeover" never came because acquirers waited for a cheaper, cleaner entry. The stock round-trips to single digits.

Are multiples too high? Yes, on trailing reality — EV/Sales 14.4x and a net loss don't support $12.37 on fundamentals. The price is a forward-recovery + takeover option. If the option pays (price + M&A), it's cheap; if not, there is large downside to a Pilbara-style fundamentals multiple.

Contrarian view (what the market refuses to see): the market is treating Sigma as a quality growth compounder in recovery; it is more honestly a binary, leveraged call on the lithium price wrapped in an ESG-marketing layer, run by promoters with 1.6% skin in the game. The green branding is doing valuation work that a fungible commodity shouldn't support.

Devil's Advocate (short-seller)

Dismantling the bull case:

• Revenue concentration = the lithium price. 100% of the thesis is one exogenous commodity Sigma cannot influence. If SC6 disappoints 20–30% (entirely normal for lithium), base-case EPS goes from ~$0.65 to a loss — a ~$1.30 EPS swing on a $12 stock. Growth "disappointing by 20–30%" here isn't a haircut, it's a sign flip.
• The moat is weaker than bulls think. "Low cost" did not prevent a US$50M loss in 2025. Cost-quartile leadership is meaningless when the price sits below everyone's all-in cost. The green premium is a few dollars a tonne, not a franchise.
• Most dangerous competitor bulls underestimate: the brine majors (SQM, ALB in the Atacama at $3–5/kg) and Chinese lepidolite/African spodumene swing supply — they set the marginal price and can crush the spodumene curve. Sigma is a price-taker to their price-setting.
• Worst capital-allocation / incentive signals: Co-Chair sold 56% of his stake at $9.84 (Nov-25) — insiders monetising near the low; related-party A10 admin services; 1.6% insider ownership; a disruptive contractor switch that torched 29% of the market capMarket capitalisationThe share price multiplied by the number of shares. What the market says the equity is worth.. That is not the behaviour of operators betting their own net worth on a coming boom.
• Assumptions that must hold for $12.37: SC6 sustained >$1,300; Phase-2 on time and on the BNDES budget; no further Minas Gerais waste/permit escalation; offtake prepayments not turning into under-priced delivery obligations; no dilutive raise. Several are outside management's control.
• The single scenario that permanently impairs: a multi-year sub-$900 SC6 trough (supply glut re-asserts) that forces distressed equity issuance or hands pricing to prepayment counterparties — permanent dilution/value transfer with no takeover bid to rescue it. Plausibility: moderate-to-high, given lithium already did exactly this in 2024–25.

Management Questions (ordered by information value)

1. At what sustained SC6 CIF-China price does Sigma generate positive free cash flow after interest and sustaining capex — not just positive gross margin? Show the breakeven.
2. Walk through the liquidity bridge: with US$6.2M cash at year-end 2025 and BofA flagging vendor-payment delays, what is the minimum cash you will hold, and what covenant/repayment schedule (BNDES + offtake prepayments) must you clear in 2026–27?
3. Why did the Co-Chairman sell 56% of his holding in Nov 2025, and why is aggregate insider ownership only ~1.6% for a "founder-led" company?
4. Detail every related-party agreement with A10 (scope, fees, term, who signs on Sigma's side) and how the independent committee prices them at arm's length.
5. What exactly drove the October 2025 mining-contractor change — cost, performance, or dispute — and what is the new contractor's track record and rate structure?
6. On the offtake prepayments (US$96M/70.5kt and US$50M/40kt-3yr): what price mechanism governs delivery, and what is the downside if spot falls below the implied prepay price?
7. Phase-2 to 520kt: give the month-by-month ramp, the remaining capex vs. the BNDES BRL 487M, and the completion date you'll commit to publicly.
8. How do you respond, today, to Grizzly's 2023 feasibility-study allegations — and how should US investors reconcile NI 43-101 headlines (US$15.3B NPV, 150Mt) with SEC S-K 1300 economics?
9. What is the current status and worst-case exposure of the Minas Gerais waste-handling / environmental proceedings and the independent-monitoring obligation for Araçuaí and Itinga?
10. Is the company for sale? What happened in the 2023 "final-round" strategic review, and under what price would the Board transact now?
11. What is your realised-price discount to SC6 benchmark (grade, impurities, freight), quarter by quarter — the true price you capture vs. the headline?
12. What share of 2026–27 volume is contracted vs. spot, and at what floor/ceiling?
13. What is the plan if SC6 sits at $800 for eight consecutive quarters — and is an equity raise on the table?
14. What is the all-in sustaining cost trajectory as you double volume — economies of scale vs. deeper-pit strip-ratio and Brazilian cost inflation?
15. What capital-return framework (if any) applies once Phase 2 is funded, or is all cash flow reinvested/deleveraged indefinitely?

---