Sam Bankman-Fried loses appeal as crypto's largest fraud conviction stands

A federal appeals panel in Manhattan has upheld Sam Bankman-Fried's fraud conviction and 25-year sentence, closing the highest-profile criminal chapter in crypto's first decade and opening a quieter, harder one. The ruling, handed down on June 12, rejected the defence argument that the FTX founder had been handed an unfair trial because customers and lenders were eventually made whole by the bankruptcy estate. The appellate judges found that whether the money came back was beside the point: Bankman-Fried had taken it under false pretences, and a jury had convicted him of that fact beyond a reasonable doubt.

The legal architecture of the case now looks settled. What remains unsettled is what the verdict, plus the parallel collapse of the regulatory perimeter elsewhere, signals about the post-SBF era. The appeal loss lands in the same week that Warsaw moved to tighten its own crypto perimeter through a separate regulatory front, and as the BRICS bloc continues to construct alternative rails for digital settlement outside the dollar system. Read together, the message is not that crypto is being legitimised or delegitimised. It is being re-territorialised: sorted into national lanes, scored by national courts, and priced by the regulatory weather in each capital.

The Manhattan ruling, in plain terms

The Second Circuit's brief order was unsentimental. Bankman-Fried's lawyers had argued that because FTX's bankruptcy estate recovered enough customer funds to pay most claimants back at bankruptcy prices, the criminal conviction could not stand: the alleged harm, in their framing, had been undone. The court disagreed. Fraud under US federal law is a crime at the moment of taking, not at the moment of loss. The recovery of customer money by professional liquidators several years after the fact does not retroactively erase the act of taking it. The 25-year sentence, imposed by Judge Lewis Kaplan in March 2024, survives.

Bankman-Fried still has procedural options. A petition for rehearing en banc, and ultimately a cert petition to the Supreme Court, remain on the table. But the substantive ground has narrowed dramatically. The factual record from the November 2023 trial, where jurors heard Caroline Ellison describe a deliberate scheme to use customer deposits to plug holes at Alameda Research, is now appellate precedent within the circuit. Future defendants in similar posture know what the playbook looks like.

The political weather around the case has also shifted. The prosecutors who brought the case, led by Damian Williams at the Southern District of New York, made clear during the trial that FTX's collapse was not a market accident but a confidence game. The appeals court has now ratified that framing.

What the appeal was actually about

It is worth separating two arguments that travelled together through the public conversation and travelled separately through the courtroom. The first was substantive: that the trial judge had admitted expert testimony and customer-loss figures that overstated the harm, and that the government's theory stretched the wire fraud statute past its intended use. The second was procedural: that the late addition of a campaign-finance charge, and the government's cross-examination of Bankman-Fried on issues the defence had not opened, deprived him of a fair shake.

The court handled both with economy. On the campaign-finance charge, the panel pointed out that Bankman-Fried had testified on direct examination about his political donations and his intent in making them, opening the door to the government's cross. On the loss calculations, the court treated the bankruptcy recovery as irrelevant to the question of whether the taking had occurred. None of these were close calls in the published order. The defence's strongest card was always going to be the customer recovery, and the court cut that card in half with a single paragraph.

The global map is hardening along different vectors

Here is where the story stops being about one man and starts being about the architecture forming around his absence. In the same week that the Second Circuit ruled, the Polish government signalled it would move on a domestic crypto licensing regime that treats retail platforms more like broker-dealers than like software vendors. The framing in Warsaw is consumer protection: Polish retail savers, the argument goes, deserve the same disclosures and capital buffers that Warsaw requires of any other financial intermediary. The political logic is uncontroversial inside Poland's regulatory establishment, and it cuts across both the governing coalition and the opposition.

A thousand miles east, the BRICS grouping has been pressing the other direction. Member-state central banks have spent the last eighteen months publishing working papers and pilot programmes for cross-border settlement in non-dollar digital instruments. The geopolitical premise is openly stated in Russian and Chinese commentary: dependence on dollar rails is a chokepoint that can be tightened by Washington, and any serious multipolar settlement system needs an alternative rail. Whether those instruments are denominated in CBDCs, in tokenised commodity baskets, or in private stablecoins issued by non-US entities, is a technical question. The political direction of travel is not.

The American ruling and the Polish regulation point the same way. Both treat crypto as finance, full stop, and apply the existing toolkits of finance to it. The BRICS experiment treats crypto as infrastructure for an alternative financial order. Both projects are now operating simultaneously, with very little overlap, and each one is now harder to dislodge than it was a year ago.

The SpaceX signal hiding in plain sight

One under-noticed piece of context landed this week. SpaceX's IPO disclosures revealed a $1.3 billion bitcoin treasury reserve, treated by the company as a corporate balance-sheet asset rather than as an operating business. The framing from analysts has been that the largest private balance sheet now entering public markets is, in effect, a treasury test case: a publicly traded company holding bitcoin through a full cycle will show public-market investors, quarter by quarter, what mark-to-market on a corporate crypto position actually looks like.

That matters because it changes what a regulated exchange is. The tradable instrument a regulated venue will list next year may not be a retail stablecoin or a memecoin. It may be a corporate bond, a money-market fund share, or a treasury bill issued by a company that holds bitcoin on its balance sheet. The supervision problem then becomes the supervision problem of capital markets in 1985: counterparty risk, custody, disclosure, and the integrity of the price-discovery mechanism. Those are problems the existing regulatory apparatus was built to handle. Bankman-Fried's crime, at its core, was the failure of those mechanisms in a venue that was not built to enforce them.

What survives the appeal

Bankman-Fried will almost certainly serve his sentence. The narrow legal question that survives is whether the Supreme Court takes the case, and on what ground. The broader question that survives is what the ruling actually says about the line between a failed business and a criminal enterprise. The court's answer is unambiguous on the facts of this case: when a firm solicits customer funds on the representation that those funds are segregated and available on demand, and then routes them into a related trading vehicle, the misrepresentation is the crime, regardless of what happens later.

The industry is now digesting that line. Compliance officers at offshore venues are rewriting risk memos. In-house counsel at US-domiciled platforms are checking the language of their terms of service against the trial record. The next wave of crypto enforcement will not look like FTX: it will look like ordinary securities fraud wrapped in a token wrapper, prosecuted with the FTX trial as the cited authority. That is the world the Manhattan appeals court just ratified.

For the global map, the lesson is the same one that the Polish veto and the BRICS pilot projects have been teaching in their own languages: the regulatory perimeter around digital assets is no longer a single perimeter. It is a patchwork of national perimeters, each calibrated to a different theory of what crypto is and what it is for. The Manhattan ruling is the American theory, in writing. It will outlast the man it was written about.