Cathie Wood's Latest Moves: Buying CoreWeave and Circle Ahead of the Test

ARK's trades look more like rotation than broad enthusiasm

The clearest read on ARK's latest activity is rotation. The fund put roughly $22.04 million into CoreWeave and about $7.01 million into Circle-about $29 million together-just before a tight earnings window: Circle reports on August 5 and CoreWeave reports on August 11. At the same time, ARK sold 66,940 Shopify shares worth about $8.19 million. That points to a preference for AI compute and stablecoin infrastructure over some software exposure ARK no longer wants to overweight.

The daily trades matter more here than the 13F snapshot

The Q1 2026 13F still matters for context. ARK disclosed a roughly $12.9 billion portfolio and active reweighting across the book. But the more immediate signal is the daily ETF trade tape, because it shows what ARK wanted to own right before earnings, not just what it held at quarter-end.

There is also a useful caution here. ARK's earlier 10X Genomics accumulation was later followed by a 13G/A showing a reduced stake. That is a reminder that an ETF buy report does not always mean clean, lasting conviction.

CoreWeave and Circle look like redeployment bets, not blind optimism

CoreWeave: financing flexibility before a fundamental check

The starting point is why the stock weakened. CoreWeave highlighted better terms on a $2.6 billion loan, which management said would help expand computing capacity for major customers. ARK then bought 298,243 shares worth about $22.04 million as the name traded lower. That looks like a bet that extra financing flexibility could matter if demand for compute stays tight.

That thesis fits ARK's broader AI framing. AI inference costs are collapsing, while demand for processed tokens and data-center system spending is rising. If usage expands from training into inference and agents, access to compute remains a key bottleneck. CoreWeave is directly exposed to that demand side, so the setup is less about generic AI excitement and more about whether the company can turn capacity and customer need into durable results.

Circle: broader ETF participation, but a tougher case

Circle is the less obvious add, but the pattern is similar. ARK spread its purchase across three of its ETFs, buying 109,129 Circle shares for about $6.7 million on Friday and adding through the broader ARKK/ARKW/ARKF complex. That looks more like portfolio-level confirmation than a one-fund experiment.

The bullish angle is that stablecoin infrastructure may be becoming more than a crypto-native story. Even after Bernstein cut Circle's price target to $140 from $190, the analyst said Open USD poses less of a threat than feared and said major partners, including Visa, still work with Circle. So the debate is less about whether the category matters and more about whether Circle can defend its franchise.

Why this basket still carries execution risk

The bearish case is straightforward. CoreWeave and Circle both look more narrative-sensitive than established software companies, and both need proof points around demand, partnerships, and monetization. The trade-offs ARK made matter too: the fund trimmed Shopify, CrowdStrike, and Snowflake to fund part of this rotation.

What the basket suggests: - Bull case: compute access and dollar-pegged token infrastructure could prove more durable than the market currently credits. - Bear case: both names still look sensitive to sentiment and execution misses. - Decision point: the coming earnings reports should show whether demand and partnerships are translating into measurable business progress.

10x Genomics still does not look like a clean high-conviction hold

Small ETF buys do not erase the bigger reduction

ARK's earlier spring accumulation in 10x Genomics was real: it bought 98,722 shares worth about $1.8 million. But that signal is harder to read against the larger picture. ARK later disclosed 11,887,396 shares, down from 12,845,713 shares-a 7.46% reduction in disclosed ownership, even though it still held 10.39%.

That mix does not read like straightforward escalation of conviction. It looks more like a position ARK keeps dabbling in rather than one it is adding to cleanly across all reporting layers.

Why that matters more in biotech

In AI and software, repeated small buys can make sense while revenue ramps and investors figure out how to value the business. In biotech, timing and execution matter quickly, and a stock can lose valuation support fast if milestones slip.

The issue is not one purchase. It is the combination of small ongoing ETF accumulation and a lower overall disclosed stake. That is a weaker signal than a full conviction add, and it argues for more care before treating 10x Genomics as a leadership position.

What to watch in the next 10 days

What matters now is confirmation.

The near-term scorecard

• CoreWeave: the next hard test is its second-quarter results due on August 11. The key question is whether the loan translates into visible capacity gains and durable demand.
• Circle: the earlier test comes on August 5. Investors need evidence that partnerships and franchise value still matter more than the pricing and competition debate.

What would count as confirmation

• Good: later filings show ARK still holds, or adds to, its pre-earnings stakes. That matters more when viewed against the backdrop of a roughly $12.9 billion portfolio, where position changes usually reflect portfolio-level decisions.
• Better: earnings validate why ARK bought before the print.
• Less convincing: another round of small daily ETF buys while the underlying ownership filing gets softer. The 10x case still applies after ARK later reported 11,887,396 shares in a 13G/A filing that showed a lower stake than before.

Daily ETF trades are worth watching, but filing-level confirmation is what separates durable conviction from short-term headline chasing.