ARK Sells $85M in Nvidia Shares Before Q3 Earnings
ARK sold about $84.8 million in Nvidia shares while buying crypto-linked stocks. See what the trades may signal—and what they cannot tell you before earnings.

What ARK's $85M Nvidia Sale Says About Its Crypto Bets
ARK Investment Management made a notable shuffle ahead of Nvidia's third-quarter earnings: it sold about $84.8 million in Nvidia shares and bought nearly $40 million of Block, Circle, and Coinbase. It also sold units of Solana and Ether staking exchange-traded funds, according to a Monday trade disclosure. The timing raises two practical questions: how much weight investors should give the "AI trade" before Nvidia reports, and what ARK's crypto-adjacent buys suggest about its comfort with crypto-linked businesses.
That's worth digging into - but it's not a clean read on where Nvidia or crypto markets go next. An asset manager can sell for lots of reasons, including portfolio sizing, risk controls, or client flows, not just because its view has changed. And buying a crypto-exposed company isn't the same thing as buying the underlying tokens.
This article walks through the reported transactions, the limits of reading intent from a trade list, what to watch in Nvidia's results, and a more practical way to interpret ARK's move. It also flags details in the disclosure that need confirmation before treating them as established facts.
What did ARK buy and sell?
Based on the reported Monday disclosure, ARK sold 543,092 Nvidia shares, valued at about $84.8 million using the closing price. It also sold roughly 30,659 Bullish shares (about $605,000) and 350,000 units of the 3iQ Solana Staking ETF plus 300,000 units of the 3iQ Ether Staking ETF. The disclosure does not provide closing values for the staking ETF sales, so those unit counts alone don't tell you how big the dollar reductions were.
On the buy side, ARK acquired 277,523 shares of Block, 180,534 shares of Circle, and 53,760 shares of Coinbase. Using Monday's closing prices, those purchases were valued at about $23 million, $10.4 million, and $6.4 million, respectively. Together, that puts the three buys at roughly $39.9 million.
The report also lists a purchase of 378,515 SpaceX shares valued at about $52 million. That part needs more context. SpaceX is private, and the disclosure as provided doesn't explain the instrument or structure behind the "SpaceX shares" figure. Investors shouldn't assume it's the same kind of publicly traded "stock" holding; ARK's full filing and position records would be needed to clarify what the fund actually owns.
There's another caveat: the transaction values are estimates based on closing prices, not necessarily the prices ARK received. Trades happen throughout the day, and disclosures can report quantities without execution prices. And you don't learn the size of each holding before and after from a single day's activity. In other words, selling $84.8 million of Nvidia sounds significant on its own, but its real meaning depends on ARK's overall Nvidia exposure and the fund's scale.
The most defensible takeaway from the disclosed activity is simple: ARK reduced some Nvidia exposure, added to several crypto- and fintech-linked names, and trimmed staking ETF positions. The trade list alone doesn't prove a complete change in the firm's view on AI, crypto, or any single company.
Why sell Nvidia before earnings?
The timing gets attention because Nvidia's third-quarter results were due after Tuesday's market close, making the sale a pre-earnings trade. Earnings can move stocks sharply when expectations are high. Even a strong report can fall short if guidance, margins, or outlook don't beat what investors already have in their heads.
So the key question isn't just "is Nvidia growing?" It's whether growth - especially in data center revenue - can keep pace in a way that supports Nvidia's valuation. The report referenced an analyst estimate of roughly $109.3 billion in third-quarter revenue, including about $101.8 billion from data center. Those numbers reflect analyst expectations, not Nvidia's own guidance or reported results. They should be checked against Nvidia's outlook and the consensus being formed around the release.
The report also said Nvidia shares fell 4% on Monday to close at $156 and were down more than 16% year to date, citing Yahoo Finance data. Those figures describe a specific snapshot in time. A short-term decline doesn't, by itself, prove the AI investment cycle is over.
A portfolio manager might sell ahead of earnings for several ordinary reasons: reducing event risk, taking gains, rebalancing a position that has grown too large, or freeing cash for other holdings. Those explanations can all be true at once. Public trade data generally won't tell you which one applies to a particular order.
The disclosure also refers to an unusual claim about a change in Nvidia's leadership. It doesn't include enough reliable detail to treat that claim as established, so it shouldn't be used to explain the trade. The analysis here should stick to what's actually in the disclosure and the upcoming earnings event - not to unverified personal or management allegations.
That distinction matters. A trade made before an earnings release might reflect uncertainty about the outcome, but it doesn't prove the seller expects disappointment. Treating timing as motive turns a fact (the sale happened before earnings) into a conclusion the disclosure doesn't support.
What should investors watch in Nvidia's results?
A single quarterly revenue headline won't tell you whether AI infrastructure spending can sustain Nvidia's expectations. Data center revenue matters - but so does the quality and durability of the growth. Are orders spread across many customers, or concentrated among a few large buyers? Are customers actually deploying the chips they buy? Do their signals suggest spending is likely to continue into future quarters?
Margins and supply constraints also matter. Strong demand doesn't automatically translate into unlimited earnings growth if manufacturing constraints, product transitions, or rising costs start to affect deliveries and profitability. Guidance is especially important because markets trade on what investors expect next, not just what the company already shipped.
A useful way to review the earnings is to separate four questions:
- Demand: Is data center demand still expanding, and what evidence supports management's view?
- Conversion: Can Nvidia turn demand into shipments and revenue on the expected schedule?
- Profitability: Are gross margins and operating costs consistent with continued earnings growth?
- Customer concentration: How dependent is the growth on a small group of large buyers?
Underneath all of that is a bigger question: does AI spending actually generate enough economic value for customers to keep investing? Cloud providers and other large buyers can fund heavy spending while they build capacity, but it still needs a business case. If AI services generate revenue, improve productivity, or deliver strategic value, infrastructure investment is more likely to persist. If returns are slower or harder to justify than expected, buyers can pull back - even if overall demand for computing stays high.
This isn't a one-quarter "pass/fail" test for whether AI is a bubble or a durable market. Infrastructure cycles take time. One earnings report can reset expectations, especially when positioning is crowded, but it can't settle questions about long-term adoption or returns on capital.
It also helps to separate company performance from stock performance. Nvidia can report strong results and still see the shares slide if investors expected more. Conversely, "solid" results might hold up the stock if guidance lands better than feared. The comparison that matters is not only "this quarter vs last quarter," but "this quarter and outlook vs what the market already priced in."
How should you read an asset manager's trade disclosure?
A trade disclosure is evidence that activity happened - not a full explanation of why. The common mistake is to treat a sale as a direct signal. ARK sold Nvidia, therefore Nvidia is overvalued. ARK bought Coinbase, therefore crypto is about to rally. You can't get from the transaction alone to those conclusions.
Instead, try a five-step approach:
- Verify the filing and date. Make sure the trade appears in the manager's own disclosure and identify when it was executed. News summaries sometimes leave out important detail.
- Compare position sizes. A dollar figure by itself doesn't tell you much without the fund's total assets and what the position looked like before. Look at the change as a percentage of the position, not just the headline amount.
- Separate exposure types. Nvidia shares, equity positions in crypto-related companies, token-linked ETFs, and direct token exposure all behave differently. Don't lump them into one risk bucket.
- Check the event context. Earnings timing can explain why the trade drew attention, but it doesn't prove earnings drove the decision. Rebalancing, risk limits, and fund flows can look similar from the outside.
- Look for persistence. One day's activity is a weak signal. Repeated changes across disclosures - especially when they materially change a position - are more informative.
This framework also helps when a report includes unusual or incomplete details. The SpaceX entry, for example, likely needs clarification because the disclosure doesn't explain how the exposure is held. A careful read flags that uncertainty rather than smuggling in an assumption. The same goes for the ETF sales: without sale values and remaining balances, unit counts don't tell you the scale of ARK's shift.
Key takeaways:
- ARK reportedly sold about $84.8 million of Nvidia shares ahead of the earnings release.
- It bought about $39.9 million of Block, Circle, and Coinbase shares, but those businesses come with different models and risks.
- The disclosure lists Solana and Ether staking ETF sales without dollar values, limiting direct comparisons.
- Trade timing isn't proof of motive, and one disclosure isn't a stand-alone market forecast.
- Nvidia's guidance, customer spending, and data center economics matter more than a single revenue headline.
The point isn't to ignore manager disclosures. They can show where capital is moving and help identify what you should investigate. The point is to treat them as one input in a broader process - not as a substitute for understanding the businesses, the positions, and the risks.
Conclusion
ARK's reported trades show a useful contrast: it reduced Nvidia exposure ahead of a closely watched earnings event while adding to companies tied to payments and crypto-related markets. This looks like a portfolio move, not a definitive verdict on AI spending or digital assets. The disclosure also leaves key gaps, including how large ARK's remaining ETF positions are and what the "SpaceX shares" exposure actually represents.
Read the activity as a prompt to investigate, not as a shortcut to a market call. Verify the filings, compare position sizes, separate stock exposure from token exposure, and evaluate Nvidia against the expectations already embedded in its shares. That approach is more reliable than trying to infer a full investment thesis from a single day's trades.