Forty million shares. That's the shape of the number hiding inside the headline.
Larry Ellison — Oracle's co-founder, chairman, and still its single largest holder at roughly 41.5% of the company — has reportedly stood up a fresh 10b5-1 trading plan authorizing the sale of up to $8 billion of ORCL stock. At a share price near $200, that converts to something close to 40 million shares. Against Oracle's free float of roughly 1.6 billion, it is about 2.5%. Against Ellison's own ~1.19 billion share position, it is roughly 1.4%.
Crypto did not care. The wire ran it. The terminal printed it. It died in four hours.
Wrong call.
Alpha moves before the charts confirm the truth. And this filing — which touches nothing on-chain, mentions no protocol, and involves a database company founded in 1977 — is the cleanest live specimen of the exact mechanism that has drained DeFi retail for four straight years: the scheduled insider exit. TradFi built a compliance cage around that mechanism. DeFi copied the schedule and threw the cage in the bin.
Here is what the missing cage costs. In March 2024, Arbitrum's cliff unlock pushed 1.1 billion ARB into circulation — roughly 87% of the tokens already trading. One date. One afternoon. Roughly $2.3 billion of notional supply arriving at a fixed block height. No US-listed company on this planet could legally or practically do the equivalent. Ellison is moving 2.5% of float across a plan that must survive multiple federal filings. Arbitrum moved 87% of float across a single timestamp.
Both events got a blog post. Only one of them got a checkbox on a federal form.
The context you need before you price anything
Oracle is not a crypto company. It has an Oracle Blockchain Platform that the market has largely ignored since 2018. It runs the second-largest enterprise cloud franchise behind AWS and Azure, and since 2023 it has been the single biggest beneficiary of the AI infrastructure trade. Its remaining performance obligation — contracted future revenue — reportedly crossed $455 billion in the September 2025 quarter, a figure that grew triple digits year over year. The stock ran to roughly $345 a share in September 2025, briefly putting Oracle near a trillion-dollar market cap and Ellison near the top of the world's wealth rankings. It has since given back a large chunk of that move.
Into that, the plan. And that's the part most readers get backwards.
A 10b5-1 plan is not a sale. It is a pre-commitment structure. An insider adopts a written plan at a moment when they can certify they hold no material non-public information, and that plan then executes mechanically on a predefined schedule. The whole point is to remove discretion. The insider gives up the ability to time the market in exchange for a legal affirmative defense against insider trading claims.
The SEC tightened the rules on 14 December 2022, effective 27 February 2023, and the tightening is the interesting part. Directors and Section 16 officers now face a cooling-off period of 90 days after adoption, or two business days after the filing of the 10-Q or 10-K covering the fiscal quarter of adoption, whichever is later — capped at 120 days. Other insiders face 30 days. Overlapping plans are banned except in narrow cases. Single-trade plans are limited to one per twelve months. Directors and officers must personally certify in the plan that they are not aware of material non-public information. Form 4 filings now carry a checkbox disclosing reliance on a 10b5-1 plan, with a footnote identifying the adoption date. Form 144 filings go through EDGAR electronically. And under Item 408 of Regulation S-K, issuers must disclose the adoption and termination of trading arrangements in their quarterly reports.
Read that list again. Cooling-off. Certification. Overlap ban. Transaction-level reporting within two business days. Quarterly disclosure.
That is a disclosure architecture engineered specifically so that a retail holder can see the exit coming and price it.
Now hold it against the DeFi standard
I have spent years on both sides of this. In 2017, as a cybersecurity undergrad in Jakarta, I was manually auditing ICO whitepapers — more than fifty of them — and I found a re-entrancy flaw in a high-profile token contract hours before mainnet. The vesting schedule in that project's docs looked immaculate. Four-year linear, twelve-month cliff, team allocation clearly labeled. What the docs never said was who held the keys to the vesting contract, or whether those keys could be re-pointed to a different beneficiary. That information was not disclosed. It was not legally required to be disclosed. It was just gone.
That gap has not closed in eight years. It has widened.
Here is the side-by-side, and I want you to feel how lopsided it is.
TradFi gives an insider a cooling-off window. DeFi gives a cliff unlock a fixed block number. TradFi compels a personal certification of clean knowledge. DeFi compels nothing — the vesting contract executes whether the team is buying or dumping. TradFi bans overlapping arrangements. DeFi routinely stacks a cliff, a linear vest, a protocol-owned liquidity release, and an airdrop claim window inside the same 30-day stretch, and calls it a "growth phase." TradFi requires a Form 4 within two business days of every single sale. DeFi requires you to know the team's wallet cluster, follow the multisig, and hope the OTC desk they use routes through an address you've already labeled.
TradFi's insider selling has a paper trail. DeFi's has a block explorer and a prayer.
And before anyone tells me the Arbitrum comparison is unfair — that ARB's unlock went to a DAO treasury and ecosystem funds, not to insiders — I've heard it. It's true and it's irrelevant. The float is what prices the asset. The recipient's intent changes the destination of the coins, not the supply shock. Liquidity is the only religion in the DeFi temple, and liquidity does not read press releases about ecosystem grants.
Let me put numbers on the violence. ARB's March 2024 cliff added 1.1 billion tokens against roughly 1.27 billion circulating — call it an 87% immediate float expansion. Compare that with TIA, APT, SUI, and the rest of the 2024–2025 unlock cohort, all running monthly linear releases that compound into double-digit annual float growth. These are not edge cases. They are the design standard.
Now the Ellison number again. $8 billion. Sounds apocalyptic. It is roughly 2.5% of float, and it will not land as a block — it will be drip-fed through a plan with fixed parameters, administered by a broker, on a schedule that was set weeks or months in advance. Scaled to the same measuring stick, it's the equivalent of a token team announcing a twelve-month, low-single-digit float expansion — the kind of thing that barely moves a mid-cap.
The $8 billion headline is the least informative number in the story. As always: data lies, but volume never cheats.
The angle nobody is writing: the alpha isn't the sale, it's the hedge
Here's where my 2025 work becomes uncomfortable.
Earlier this year I built and ran a small detection prototype to hunt AI-driven manipulation in DEX volume. It flagged a bot network controlling roughly 15% of trading activity on a niche layer-2 — algorithmic market makers cycling inventory across a handful of pools to farm incentive emissions. That was the stated project. But the tool surfaced something else at the edges, and that's the finding that should be in every unlock note written this cycle.
Around a scheduled cliff, the spot volume is theater. The real positioning shows up in perpetual basis and open interest before the unlock date, and it shows up from wallets that never appear on the unlock list. What I mean, concretely: a recipient of an unlocking allocation does not have to sell. They can short the perpetual at three times the notional weeks in advance, let the unlock arrive, spot-fade the release, and close the perp for a profit that never touches the foundation's address. The coins never move on-chain. The block explorer shows you a transfer into a cold wallet and a tidy "HODL" narrative. The P&L was already made elsewhere.
This is the exact envelope that 10b5-1 was built to seal. TradFi cannot fully stop an insider from hedging — the rules on pledging and hedging company stock are real but imperfect — but TradFi at least forces the adoption date, the plan terms, the per-transaction execution, and the quarterly disclosure into the public record, so an analyst can reconstruct the sequence after the fact.
DeFi offers none of it. The unlock date is public. The exposure is invisible. Chaos is where the institutional money hides — and this particular chaos is manufactured, scheduled months in advance, and sold to retail as a growth milestone.
So when I see an $8 billion insider plan from a man who owns 41.5% of a company, I do not read it as a top signal. I read it as the mature version of a primitive DeFi has not yet learned to operate. Ellison gets a cooling-off period. Ellison gets a personal certification. Ellison gets a checkbox on a federal form. Ellison gets an Item 408 disclosure four times a year. Ellison's broker reports every fill within two business days.
A token team gets a blog post and a vibes-based Twitter space.
There is one more layer, and it's the part that makes governance-token holders uncomfortable. A governance token pays no dividend. There is no buyback tied to the token, no revenue share, no claim on cash flow. Structurally it is a non-dividend equity with no covenant. Which means the only realization path is selling to a later buyer at a higher price — and the only variable that reliably determines whether that is achievable is the unlock cadence. The vesting schedule is the entire investment thesis. Everything else is narrative.
Which is exactly why the missing disclosure layer is not a minor compliance footnote. It is the whole game. If a holder cannot see when supply arrives, who receives it, and whether the recipients are already short, they are not pricing an asset. They are pricing a rumor with a chart attached.
Contrarian, stated plainly
The consensus read when this news broke was that a founder dumping $8 billion is a warning on the AI trade. That is a lazy read, and I'll tell you why in mechanical terms rather than philosophical ones.
First, base rates. Ellison has sold ORCL repeatedly across a decade in which the stock compounded violently. Founder sales are dominated by diversification, tax planning, philanthropy, and estate structures. The plan's existence tells you almost nothing about the insider's view on the next four quarters. The plan's size relative to the holder's position tells you a little more. $8 billion against a ~$240 billion position, at pre-drawdown prices, is roughly 1.4%. This is a trim, not a thesis reversal.
Second, the float math. 2.5% of float, drip-fed over a plan horizon, against an average daily volume that has been running in the tens of millions of shares. The options market absorbs this without breaking a sweat.
Third, and this is the contrarian hinge — the reason a trad-fi insider filing matters to a DeFi reader has nothing to do with Oracle's stock. It matters because it shows the industry the finished version of a process it is running in beta. Every token that launches with a cliff and no disclosure is running the 1990s version of insider selling, the pre-Sarbanes-Oxley version, the era when executives could move size in the dark and nobody had standing to ask. That era ended in equities for a reason. It cost retail trillions.
DeFi's cliff unlocks are a rerun, and the audience is paying for the original sins a second time.
Don't read Ellison's plan as a crypto signal. Read it as a specification. It is the document that shows precisely which protective layers the token unlock industry never installed: the cooling-off, the certification, the overlap ban, the two-day reporting, the quarterly Item 408 disclosure. Five layers. DeFi has zero of them.
What I'm watching next
Not the ORCL price. The cadence.
Watch the Form 144 filings and Form 4 footnotes over the next two quarters. A one-off plan is noise. A second plan adopted inside six months, or a cluster of senior Oracle insiders adopting in parallel, changes the base rate and would be the first genuine internal signal in this cycle.
On the crypto side, watch for the first DAO that voluntarily publishes unlock-exposure disclosure — not just the schedule, but the hedging positions of recipients, the OTC routing of foundation allocations, and a per-transaction log of foundation wallet outflows. A handful of teams have started gesturing at it. Nobody has shipped it. The first team that does turns an unpriceable risk into a priceable one, and that is worth more than any airdrop they could run in its place.
And watch the basis, not the spot, into the next cluster of large cliffs. The spot is marketing. The basis is intent.
One question to sit with. If an $8 billion sale by a man who owns 41.5% of a company requires a 90-day cooling-off period, a personal certification, a checkbox on a federal filing, and quarterly public disclosure — why does a $2.3 billion token cliff require nothing but a blog post and a promise?
Speed isn't the entire product. Disclosure is the product we never shipped.
