Forty thousand, seven hundred twenty-three shares. That is what Intesa Sanpaolo reported in BlackRock's iShares Bitcoin Trust on June 30. On March 31 the number stood at 646,809. A 93.7% reduction. The headline writes itself: Italy's largest banking group is running from Bitcoin. Then the options table enters the evidence. The held-call row collapsed from a notional 2,496,500 shares to 18,000 โ a 99% decline. And a new put row materialized: 500,000 shares of IBIT notional. The market reaction was predictable โ sell first, ask questions later. But forensics does not run on emotion. The logic held until the ledger lied. This is not the anatomy of an exit. Leverage removed. Insurance added. Same asset. Same bank. Opposite risk profile. The numbers moved as reported. The strategy moved in the other direction.
Intesa Sanpaolo is not a regional experiment. It is Italy's dominant banking group, managing close to a trillion euros in assets. Its digital-asset timeline reads as measured capitulation, not speculation. In July 2024, it underwrote Italy's first on-chain digital bond on Polygon โ USD 25.6 million. In January 2025, it executed its first direct Bitcoin purchase: 11 BTC for approximately USD 1.03 million. A proof-of-life. Later that year, it stood up a dedicated desk for digital-asset options, futures, and spot ETFs.
That desk is the lens for this filing. A 13F is a static snapshot โ one day, June 30, rendered as a table. Institutional investment managers with at least USD 100 million in qualifying assets must disclose their equity holdings quarterly; the options table accompanies the equity rows under SEC rule. When a bank operates an options desk, the boundary between proprietary inventory, client facilitation, and hedging blurs. The snapshot is a shadow of the structure.
The broader market frames the move. Spot Bitcoin ETFs posted a record monthly net outflow of roughly USD 4.5 billion in June. July reversed with USD 172.4 million in inflows. August has added another USD 170 million. IBIT remains dominant, with nearly USD 61 billion in cumulative inflows. Ethereum spot ETFs, meanwhile, have drawn renewed institutional attention after staking was enabled โ a regulatory shift that turned a passive product into a yield product. Intesa's filing captured the quarter-end moment of maximum panic. Yet the bank did not simply sell Bitcoin. It re-engineered its exposure.
Start with the numbers.
| Instrument | March 31 | June 30 | Change | |---|---|---|---| | IBIT shares | 646,809 | 40,723 | -93.7% | | IBIT held-call notional | 2,496,500 | 18,000 | -99.3% | | IBIT put notional | 0 | 500,000 | new | | iShares Staked ETH ETF | 116,200 | 349,600 | +200.9% | | Bitwise SOL Staking ETF | 2,817 | 7 | -99.8% |
The bank reported 40,723 IBIT shares. Its held-call position implied 18,000 notional shares. Its new put position implied 500,000. That put is more than twelve times the reported equity holding. That imbalance is the first forensic red flag. You do not purchase put premium equal to half a million shares to hedge a forty-thousand-share position. Either the put protects exposure held outside the 13F โ total return swaps, structured notes, prime brokerage inventory โ or it functions as a synthetic overlay. Both readings converge: the bank's true Bitcoin footprint is larger than the equity row suggests.
Based on my audit experience reconciling institutional filings against on-chain custody addresses, 13F tables are an incomplete map. Large banks relocate exposure through derivatives that never surface in the equity section. The put row is the tell. Someone authorized significant premium for downside protection. Sellers do not buy puts. Holders with defined-risk mandates do. A held-call row does not reveal strike price or expiration. It reveals only the notional share count of the underlying contract. That missing granularity is where the real risk lives.
A prudent analyst asks what this structure says about the bank's market view. A long call is a leveraged bet on appreciation. A long put is a bet on decline โ or protection against it. Holding both in the same instrument is a volatility trade, not a directional trade. But the magnitude matters. The call position collapsed by ninety-nine percent; the put position exploded from zero. That directional asymmetry is the signature of a bank that no longer wants upside leverage but is unwilling to abandon the asset class. It wants to be long Bitcoin without being long Bitcoin's volatility.
The ETH expansion reinforces the diagnosis. The iShares Staked Ethereum Trust ETF position rose from 116,200 to 349,600 shares โ a 200% increase. Staking alters the economic character: a staked ETH wrapper produces yield, roughly three percent in current conditions. Bitcoin produces nothing. In a bear market, yield is oxygen. The bank converted a zero-yield, high-volatility profile into a yield-bearing instrument with a softer beta envelope. Trace the hash, ignore the hype. The rotation is not away from digital assets. It is a rotation toward carry.
The Solana line completes the picture. The Bitwise Solana Staking ETF fell from 2,817 shares to seven. Seven is not an allocation. It is residue โ the dust left behind to avoid the administrative friction of a zero balance. SOL unwound. ETH tripled. BTC restructured. Deliberate portfolio-level rebalancing, not a panic.
Why would a conservative bank choose this structure over a sale? Tax treatment explains part of it. Selling realizes gains or losses and triggers the accounting event; a derivative overlay does not. The bank can hedge without surrendering the option to reclassify later. Client facilitation explains another layer. An options desk holding calls and buying puts may be warehousing volatility exposure for clients who want Bitcoin-linked instruments without equity ownership; the held rows do not distinguish proprietary positioning from market-making inventory. Capital efficiency seals the argument. Shifting from leveraged calls to protective puts lowers downside sensitivity. A lower volatility footprint means less regulatory capital reserved against the position. For a systemically important bank, that is the strategy, not a detail.
There is also what the filing does not show. The options table reflects one day. June 30 tells you nothing about July's USD 172.4 million reversal or August's USD 170 million. The market read the headline. I read the structure. Every exploit is a history lesson in slow motion. In 2022, Terra taught me to track exit liquidity before trusting narratives. In 2020, Compound taught me to test the twelve-second windows people assume are safe. This 13F has its own window: the gap between the reported share count and the derivative overlay that redefines it.
Let me steelman the bulls. The conventional read โ bank cuts IBIT by 94%, bearish signal โ misses what the options table actually says. If Intesa were bearish on Bitcoin, the rational trade would be simple: sell the shares, skip the options. Instead, the bank left a residual equity position, collapsed its long-call leverage, and added a put floor. That is the behavior of a holder managing downside, not a seller fleeing the asset. Silence in the logs is the loudest scream. The macro timing supports this. June's outflows belonged to retail capitulation. July brought institutions back. BlackRock's own clients reportedly sold roughly USD 60 million of IBIT in a single week while buying over USD 20 million of the spot Ethereum fund โ the same rotation Intesa executed. The bank was early in the quarter, not wrong in direction. The market has a habit of mistaking institutional hedges for exits; it did the same in 2022, when funds that protected downside before the Terra depeg were read as sellers. On-chain forensics later showed the hedgers were the largest accumulators after the reset.
The next 13F is the audit. If the put persists at scale, Intesa has converted its Bitcoin exposure into a defined-risk structure anchored near current prices โ an instrument that behaves more like fixed income than digital asset. If the put unwinds and the call row re-expands, this filing was a tax-harvesting exercise disguised as risk management. Either way, the 93.7% headline number was never the story. The options table was. The chain remembers what you forget. So do the filings โ if you read the rows instead of the spin. Read the rows. The risk was never the asset. It is the structure around it.