Platinum's fundamentals are real. The price already reflects them — and then some.

Thesis

Platinum is in a structural supply deficit. South African shafts collapsed from 81 to 53 since 2008. Above-ground stocks depleted 46% over three consecutive deficit years. Lease rates peaked above 25% (vs 1-3% normal). Chinese jewelry demand is inflecting as gold fatigue at $5,000+/oz drives substitution to platinum at a 2.41:1 discount ratio. GFEX launched platinum futures in November 2025, creating institutional infrastructure that didn't exist before. Pandora, the world's largest jeweler by volume, is shifting from silver to platinum plating globally in 2026-2029.

The supply side can't respond. All three SA miners (SBSW, ANGPY, IMPUY) are prioritizing shareholder returns over expansion. Implats is actively withholding 400,000 oz of strategic inventory. No one is building new shafts.

PPLT — the physically-backed platinum ETF — is the vehicle. It holds 1,413,099 audited ounces as of December 31, 2025, up 26.7% YoY. AUM tripled from $1.0B to $2.9B. Gross creations hit $913.6M in 2025 (4.1x 2024). Zero redemptions in December despite a 17-year price high. Institutions accumulated and held.

The thesis is right. The trade, at current prices, is not.

What the 10-K Shows

PPLT filed its 10-K/A on March 3, 2026 (amendment added KPMG's ICFR opinion). The filing is confirmatory — no surprises, no structural changes, expense ratio unchanged at 0.60%.

The net-new signal lives in three places:

1. January 2026 price range. PPLT hit $261.62 peak and $182.62 low within the same month. That's a $79 range — 43% of the low — in 31 days. The January low nearly touched our $175-185 entry zone. The current price of $193 sits 26% below the peak.

2. Flow asymmetry. In 2025, creations ran 2:1 over redemptions (618,811 oz in vs 314,238 oz out). December — the month PPLT hit its annual high — had zero redemptions. Institutions didn't sell the top. And post-year-end through February 26, another 200,000 shares were net created despite the correction. The money that came in during 2025 is still there.

3. JM 2024 supply/demand. Johnson Matthey's audited data shows a 774 koz deficit in 2024 — the fourth consecutive deficit year. Jewelry demand sits at 1,375 koz, barely half the 2015 peak of 2,746 koz. Investment demand surged +193% to 525 koz in 2024 — before GFEX, before the Chinese jewelry catalyst, before the price doubled. The institutional accumulation began before the narrative.

What the 10-K does NOT show: any risk factor language around GFEX concentration, Chinese demand dependency, or hydrogen demand collapse ($543M+ in electrolyzer write-offs across CF, CMI, and BE in 2025). This is ETF filing boilerplate lag, not sponsor endorsement.

Factor Decomposition

Here's where the thesis gets uncomfortable.

Regressing PPLT against GLD, SPY, and PALL (1Y daily returns):

R-squared = 0.697. PPLT is 42% palladium. The reported idio volatility of 47% (from simple SPY-only regression) is misleading — it drops to 30% once you model PGM sector properly. Well below the 75% threshold.

The platinum-specific factors where we have genuine informational edge — Chinese jewelry substitution, SA shaft closures, stock depletion, Pandora — explain only 30% of the variance. The other 70% is PGM sector beta, gold monetary premium, and market exposure we don't have unusual insight into.

Total edge contribution across all factors: approximately 51%. This is a PGM sector bet with a platinum-specific overlay, not an idiosyncratic play.

Rolling correlation with GLD has surged from 0.36 six months ago to 0.76 today. PPLT is becoming MORE correlated with gold, not less — meaning it's increasingly redundant with existing GLD exposure.

The Bear Evidence

Three things weaken the bull case from current levels:

SBSW calls the rally speculation. Their own market analyst, Kleantha Pillay, said on the February 20 earnings call: "Platinum along palladium 2025 largely driven investments speculation rather than fundamental industrial requirements." She warned explicitly of post-GFEX-settlement price correction: "Post, get price correction. Nature investment demand, unfortunately." Producers have every incentive to talk prices up. When a producer talks them down, listen.

WPIC forecasts a 2026 surplus. The World Platinum Investment Council's Q4 2025 Quarterly projects a 20 koz surplus in 2026 — the first in four years. Investment demand forecast -52%. Recycling +8%. Chinese jewelry demand -14% as H1 2025 stockbuilding normalizes. WPIC is THE authoritative source for platinum supply/demand. They themselves note the surplus doesn't fix structural issues, but the 2026 flow balance no longer supports the deficit narrative.

Hydrogen demand is dead. The WPIC 2030 forecast of 875-900 koz from PEM electrolyzers looks absurd after $543M+ in industry write-offs. CF chose CCS over electrolysis ($51M impairment). Cummins/Accelera took $458M in charges. Bloom Energy wrote down $34.6M. The Beautiful Bill Act limited 45V clean hydrogen credits. Blue/grey hydrogen (no platinum needed) is winning over green hydrogen (platinum-intensive). A demand pillar of the long-term thesis has collapsed.

Forward EV

Four scenarios, probability-weighted, 12-month horizon:

The EV math by entry price:

At $193, EV is negative. At the previously identified entry zone of $175-185, the Sharpe contribution is 0.01-0.07 — effectively noise against 48% annualized vol. The position doesn't clear an actionable Sharpe of 0.3 until $155, which requires a -20% drop from current and -40% from the January peak.

What the Options Market Thinks

The Jan 2027 options chain (310 DTE) tells a different story than ours:

P/C open interest ratio on Jan 2027: 0.25 (4:1 calls to puts). This is aggressive speculative call buying — consistent with SBSW's characterization of the rally as investment-driven. The options market is pricing a world where Chinese demand is structural, GFEX volumes persist, and platinum retests its January highs.

Either the market knows something we don't, or the speculative premium hasn't finished bleeding. Given a producer explicitly calling it speculation, and WPIC forecasting 2026 surplus, we lean toward the latter.

What Resolves the Doorway State

The central question — structural or cyclical — resolves in Q2-Q3 2026 through three independent information channels:

1. GFEX volumes. The January contract has settled. The March contract is settling now. If PT2605 (May) maintains elevated open interest, the infrastructure thesis holds. If volumes collapse after the first few settlement cycles, it was speculation.

2. JM 2025 annual report. Johnson Matthey will publish the definitive 2025 supply/demand balance in Q1/Q2 2026. If the deficit exceeds 700 koz and 2025 investment demand was 1,000+ koz, the structural case strengthens.

3. Pandora EVERSHINE rollout. Q1 2026 pilot across 30 Northern European stores. H2 2026 global rollout. Factory buildout in Thailand/Vietnam implies permanence. If early reception is strong, this becomes a genuine structural demand floor independent of Chinese speculation.

Until at least one of these resolves bullishly AND the price corrects to a level where the math works, PPLT is a watchlist position. The appropriate entry condition is dual-gated: PPLT at or below $180 AND at least one information trigger confirming the structural case.

Conclusion

The platinum supply thesis is real. Four consecutive deficit years, 46% stock depletion, >25% lease rates, and SA miner supply discipline are auditable facts. Chinese jewelry substitution is independently confirmed by three producers. Pandora's shift to platinum is a genuine structural catalyst for 2027+.

But the market knows this. PPLT ran from $83 to $262 in 12 months (+215%). The options market prices 59% probability of further upside. SBSW — a producer with every incentive to be bullish — calls it speculation. WPIC projects the first surplus in four years. Hydrogen demand, once a long-term demand pillar, has collapsed.

At $193, you're paying for someone else's call premium. The structural thesis needs a bigger margin of safety to compensate for 42% PGM sector beta you can't control, 30% idiosyncratic variance that barely clears noise levels, and a doorway state that's 55/45 on the central question.

Wait for information. Wait for price. If neither arrives, the market was right and we correctly avoided buying without edge.

LR: 0.7. The market is mildly overpricing PPLT relative to the evidence. Structural thesis confirmed but price embeds speculative premium that producer commentary, WPIC surplus forecast, and hydrogen demand collapse all suggest is too high. Divergence is moderate (our base case is ≈$188 vs market trading at $193 with heavy call skew), evidence quality is high (SEC filing, WPIC quarterly, miner earnings calls). Not a short — just not a buy at this price.

Evidence