In the autumn of 2017, when "digital gold" was still a pitch-deck slogan rather than a settlement layer, I sat through a week of token launches that all promised the same thing: the monetary properties of gold, minus the vault. Eight years on, the actual gold market is quietly delivering the thesis those decks only sold โ and it has arrived with a $4,900 price target that has almost nothing to do with inflation.
On September 12, Goldman Sachs reiterated that figure for the end of 2026 and flagged its risks as skewed upward. The number itself is unremarkable; sell-side targets usually are. What matters is the mechanism packaged alongside it: a feedback loop of official-sector accumulation on one side and a crowded options book on the other. Crypto has been staring at that same loop for three years without naming it. Tracing the sentiment pivot from 2017 to today, the trade has changed hands. Gold is no longer priced as a hedge against rising consumer prices. It is priced as the neutral reserve asset that Bitcoin spent a decade insisting it already was.
That is the part worth unpacking carefully.
To be precise about the source: the Goldman note, as relayed through media, carries exactly four load-bearing claims. A $4,900 end-2026 target with net upside risk. Sustained and strong central bank demand. Recovering ETF inflows sitting on top of elevated call-option positioning. And a tail scenario in which re-emerging Fed rate-hike expectations force traders to unwind hedges, engineering a drawdown "more severe than usual."
That is the entire scaffold. No spot price, no real-yield anchor, no dollar index, no quantified purchase data. Four sentences of thesis and one mechanism โ and the mechanism is where the actual value sits.
The macro frame matters because of what it says about gold's job description. For four decades, gold's primary function in a portfolio was to offset inflation and, more recently, deeply negative real rates. That function is being quietly replaced. When the official sector โ central banks, sovereign funds, reserve managers โ becomes the marginal buyer, the asset stops trading on CPI prints and starts trading on reserve policy. The difference is structural. A CPI-driven market does not care who owns it. A reserve-driven one does, and it is deliberately slow, deliberately political, and entirely indifferent to your quarterly target.
This reclassification is not a Goldman invention. It is the observable residue of reserve diversification that accelerated after 2022, when the freezing of sovereign assets moved the question of dollar dependence from an academic footnote to a balance-sheet problem. Gold, with no counterparty and no jurisdiction, is the one instrument that mathematically cannot be sanctioned. That property โ not yield, not scarcity โ is what the official bid is paying for.
The precedent is not theoretical. When a meaningful slice of one sovereign's reserves became unusable overnight in 2022, every reserve manager on the planet ran the same calculation: how much of my balance sheet depends on a jurisdiction I do not control? For most, the answer was uncomfortable. Gold was the only line item a sanctions order could not delete, and its share of global reserves began to climb. Bitcoin occupies the same jurisdictionless category, but it fails the reserve test on two counts that have nothing to do with ideology โ its drawdown history would be a political liability, and its liquidity, deep by crypto standards, is thin by sovereign standards. That gap is the whole argument, and it is structural rather than cyclical.
The most under-discussed line in the note is the one about option positioning, because it describes a machine rather than an opinion.
When call open interest in gold runs hot โ and Goldman explicitly calls it elevated โ the dealers who sold those calls are structurally short gamma. As spot rises, they must buy the underlying to stay delta-neutral. As spot falls, they must sell it. The result is a market that amplifies whatever direction it is already moving in. The same book that fuels a rally becomes the fuel for the crash, because the dealer's hedging response is mechanically symmetric and its consequences are not. There is no sentiment in this loop, only arithmetic. The algorithmic truth behind the token narrative is that what looks like conviction is frequently just compulsory flow.
I have watched this exact structure operate in crypto. During the DeFi Summer of 2020, I spent three weeks reverse-engineering lending mechanics that everyone described as "liquidity" and almost nobody described as "leverage" โ the distinction mattered then for precisely the reason the options loop matters now. On Deribit, where the bulk of genuine BTC option liquidity still sits, dealer gamma behaves identically, except the perpetual funding rate adds a second amplifier that gold simply does not have. When crypto traders talk about "momentum," they are usually describing dealer hedging they cannot see and cannot price.
Now the ETF piece. Goldman notes that ETF inflows are recovering. In gold that means GLD and its peers; in crypto it means the spot Bitcoin vehicles that have become the same instrument class wearing a different ticker. The marginal buyer of both assets is increasingly an allocator running a model โ an RIA rebalancing on a schedule, a multi-asset fund expressing a macro view, a treasury desk parking duration. That buyer does not hold conviction in the cultural sense. That buyer holds a mandate and a rebalancing band. Which means the flow is more predictable, more mechanical, and far more capable of reversing in unison than the retail-flow narrative suggests.
The ETF channel amplifies this. A spot gold ETF and a spot Bitcoin ETF are functionally the same primitive: a wrapper that lets a mandate-constrained allocator express a view without custody. That wrapper removes friction, and friction removal cuts both ways. It makes the inflow easier in a bull tape and the outflow easier in a drawdown, because the allocator who can buy in an afternoon can sell in an afternoon. When both products sit inside the same multi-asset model, they get rebalanced by the same risk budget. A rally that pulls both up on the same day is not confirmation of a shared thesis; it is evidence of a shared buyer.
Here is where the two assets genuinely diverge, and where the laziness of the "Bitcoin is digital gold" framing becomes expensive.
If gold is being repriced as the neutral reserve asset, the question for crypto is whether Bitcoin is a complement or a substitute. My read, after a decade of watching allocators make this decision badly, is that they are two different trades wearing the same costume. At the reserve-manager level, gold wins the mandate almost by default โ it has no volatility excuse to make, no drawdown narrative to defend in front of a parliament, no halving to explain. At the portfolio level, Bitcoin captures the convex tail: the scenario in which the dollar system's problems are faster and messier than the official sector can absorb with metal alone. Gold is the position you hold when you want to be right slowly. Bitcoin is the position you hold when you want to be right violently. Conflating them produces neither.
Then there is the Fed tail, which is the sharpest part of the note. If rate-hike expectations re-emerge, traders unwind hedges and the drawdown is outsized. Note the transmission channel: gold and Bitcoin are now priced off the same real-rate discount. A finite, non-yielding, jurisdictionless asset is worth, in the end, the present value of its scarcity โ and that present value is set by the discount rate. Gold and Bitcoin have quietly become two expressions of a single trade: the market's willingness to pay for an asset that produces nothing, priced against a rate that produces everything.
That shared discount rate is also why the crypto market's reflexive faith in "halving cycles" keeps disappointing. A halving changes supply. It does not change the discount rate. When the discount rate is moving against you, supply discipline is a rounding error, and the same options structure that magnifies the upside magnifies the unwind.
Tokenized gold makes the crossover literal. PAXG, XAUT, and their cousins let a crypto-native desk hold the official-sector trade inside a wallet โ a small market today, but a structurally telling one. What it proves is not that gold is coming on-chain, but that the boundary between "crypto portfolio" and "macro portfolio" has dissolved. The composability that made DeFi legible in 2020 is now welding TradFi's oldest asset to crypto's newest plumbing. That welding cuts both ways: it imports gold's official bid into crypto's risk engine, and it exports crypto's volatility into a market that reserve managers thought they controlled.
Now the honest part, because a Goldman target deserves scrutiny before it deserves reverence.
The entire forecast rests on central bank demand being "sustained and strong" โ and nowhere in the note is that treated as a variable to be tested. It is treated as a fact to be assumed. That is a circular argument dressed as analysis: gold reaches $4,900 because central banks keep buying, and central banks keep buying because gold is going to $4,900. The single largest assumption in the thesis is also the one with zero quantitative support in the note itself. Until monthly purchase data confirms it, the target is a hypothesis wearing a price tag.
The second blind spot is the one crypto investors are worst positioned to see. Elevated call positioning is described as an upside catalyst, but the same open interest is the machinery of the drawdown. When positioning is crowded, the exit is not a slope; it is a cliff. Following the code trail from hack to recovery taught me that systems fail at the seam, not at the center โ and the seam here is the options book, where a single shift in the Fed's expected path can convert fuel into firewood.
So the interesting question is not whether gold reaches $4,900, and it is not whether Bitcoin eventually out-reserves it. Both are path-dependent claims about a discount rate nobody has ever forecast well.
The interesting question is this: when the official sector is the marginal buyer of the world's oldest neutral asset, and the world's newest one is priced off the same rate, which of them is actually hedging whom โ and what happens to both the first time a central bank decides the trade is too crowded?