Cathie Wood's Latest AI Stock Picks: Buying the Dip on AI Stocks

When AI stocks sold off hard in late June 2026, Cathie Wood did what she is famous for: she bought. Her ARK Invest funds went on a multi-hundred-million-dollar spree during the dip. But the story is more nuanced than the “buying the dip” headline suggests — some names had pulled back sharply (Tesla, Palantir, CoreWeave, Cerebras), others were near record highs (Snowflake, Nvidia), and ARK was a seller of one of the biggest AI names, AMD. This is a factual look at what ARK bought and sold, with each stock’s recent price action shown through simple technical indicators (MACD and RSI) — not a view on whether any of it is a good idea. This article is for information and education only. It is not investment advice, and not a recommendation to buy, sell, or hold any security.

What is Cathie Wood buying?

Between roughly June 18 and July 8, 2026, ARK’s disclosed daily trades show it adding to a cluster of AI and AI-adjacent names as the market fell: Tesla, Palantir, CoreWeave, Cerebras, Snowflake, Amazon and Alphabet, plus AI-power and AI-defence adjacencies like X-Energy and Kratos. ARK publishes its trades every day, which is why we can see the buys down to the share.

The framing to keep in mind: ARK was rotating, not just buying. It funded the shopping spree by selling older or “crowded” positions — most notably trimming AMD, liquidating Alibaba, and dumping Roku. And the single biggest surprise is what it did in chips.

The AI dip: what actually sold off, and why

The backdrop was a real, two-stage pullback. The Nasdaq set a record on June 2, 2026, then rolled over. First came a semiconductor slide on June 4–5 after Broadcom’s soft AI-chip guidance (~$16B vs ~$17.2B expected). Then a broader “AI-capex bubble” leg hit during the week of June 22–26, dragging the Nasdaq more than 6% off its high. The worries were familiar: Big Tech is on track for roughly $725 billion of AI infrastructure spending in 2026 with uncertain near-term payback, and a hot jobs report kept the 10-year Treasury yield up around 4.4–4.5%. Chip stocks alone shed more than $1.3 trillion in market value before a rebound began around July 6. That is the “dip” Wood was buying.

Tesla (TSLA): ARK’s largest holding

Tesla is ARK’s single largest holding. On July 2, ARK bought 96,935 shares (~$38.1M) around the $393 level after Tesla fell ~7.5% despite a Q2 delivery beat — a classic “sell-the-news” drop. That followed earlier June adds. ARK’s Tesla case is an AI case: it values robotaxi and full self-driving as the bulk of Tesla’s future worth.

Palantir (PLTR): down about 39% from its highs

Palantir was one of the hardest-hit AI-software names, falling to a 52-week low near $107 during the selloff — roughly 48% below its late-2025 peak. ARK bought it repeatedly on the way down (81,254 shares (~$9.7M) on June 23, and more near the low on June 26). Worth noting for honesty: ARK had been trimming Palantir earlier in June before flipping to buying it back — a reminder that these are active, changing positions.

CoreWeave (CRWV): an AI-infrastructure name well off its highs

CoreWeave — the Nvidia-backed AI GPU-cloud provider — has been one of the most volatile AI names. The stock is down about 40% from its 12-month high — knocked back by light guidance at its May earnings, and again in early July when Meta said it would sell its own excess AI compute capacity. ARK added in the June spree and again on July 7 near $90.

Cerebras (CBRS): the new AI-chip name

ARK also bought Cerebras Systems, the wafer-scale AI chipmaker and Nvidia rival, on its post-IPO dip. Cerebras only listed on May 14, 2026 (priced at $185, opened at $350) and has since fallen roughly 50% from its post-IPO high toward ~$184. Because it has traded for under two months, there is not yet enough history to draw a meaningful MACD/RSI chart — the indicators need more data — so we have left it off the chart set for now.

The chip picture: Nvidia bought, AMD sold

Here is the part that surprises people. For years, ARK largely avoided Nvidia. In 2026 it reversed course: on June 2 it bought about 300,000 Nvidia shares (~$67M) across all five of its biggest ETFs. But note the timing — that buy was near Nvidia’s highs during an “AI PC” rally, not a dip. Since then Nvidia has pulled back about 15% from its 12-month high.

At the same time, ARK was selling AMD — its long-time AI-chip holding — trimming roughly $57M in early June, ~$72M on June 15, and more in July. AMD, unlike most of ARK’s buys, is sitting near its highs, which is exactly the kind of “crowded” position ARK was rotating out of. So the accurate read on chips is a rotation: Nvidia in, AMD out — not a retreat from AI silicon.

Snowflake (SNOW): bought near its highs, not on a dip

Snowflake shows why “buying the dip” does not describe the whole spree. ARK bought 223,690 shares (~$52M) of the AI data-cloud company on June 18 — but Snowflake was surging, up around 41% in a month after a strong earnings report. On the technicals, its RSI has been near overbought territory rather than oversold — a purchase into strength, not a pullback.

ARK also topped up mega-caps Amazon ($9.6M) and Alphabet ($8.2M) in the same June 23 spree — both near record highs — so, again, “dip” only describes part of the picture.

What Cathie Wood is selling

The buys were funded by sells, and the sell list is telling. Alongside trimming AMD, ARK liquidated Alibaba — selling well over $70M across June 25–26 amid a controversy over AI model “distillation” — a clear sign ARK is not chasing Chinese AI names. It also dumped roughly $77M of Roku (a streaming name, not AI) and trimmed several biotech holdings. In short: out with the old and the crowded, in with the AI-innovation dips.

Cathie Wood’s thesis: AI as a deflation engine

Why buy into the fear? Wood’s argument is macro. In a June 5, 2026 post she argued the bond market is “discounting something much more powerful: the deflationary impact of technological innovation, particularly artificial intelligence, which is beginning to increase productivity across broad swaths of the economy.” She added that “the next phase of this cycle could be characterized by accelerating growth, declining inflation, falling interest rates, and a strengthening U.S. dollar” — a backdrop she thinks favours exactly the high-growth innovation stocks ARK owns.

The catch: ARK’s record

It is only fair to note the other side. ARK’s flagship ARK Innovation ETF (ARKK) has been a serial underperformer: it is up only about 7% in 2026 versus roughly 20% for the Nasdaq-100, and over five years it has lost around 35% while the Nasdaq-100 more than doubled. Its 0.75% expense ratio is also well above a plain index fund’s. Cathie Wood’s conviction is real, but her track record this cycle is a reason to treat these picks as ideas to research — not gospel.

What AI stocks is Cathie Wood buying?

In the June–July 2026 dip, ARK Invest bought Tesla (TSLA), Palantir (PLTR), CoreWeave (CRWV), Cerebras (CBRS), Snowflake (SNOW), Amazon (AMZN) and Alphabet (GOOGL), plus AI-adjacent names like X-Energy and Kratos. Tesla was the largest single add.

Is Cathie Wood buying Nvidia?

Yes — and that is a change. After years of avoiding Nvidia, ARK bought about $67M of Nvidia (NVDA) on June 2, 2026 across its funds, while selling its long-held AMD position. Note the Nvidia buy came near the highs, not on a dip.

Is Cathie Wood really “buying the dip”?

Partly. The buys made after sharp pullbacks were Tesla, Palantir, CoreWeave and the newly-listed Cerebras. Others — Snowflake, Nvidia, Amazon, Alphabet — came while those stocks were near their highs, so “buying the dip” describes only part of the spree.

How has ARK’s ARKK performed?

Poorly relative to the market: ARKK is up ~7% in 2026 against ~20% for the Nasdaq-100, and it is down ~35% over five years while the Nasdaq-100 has more than doubled.

Track more market moves with our explainers on Jim Cramer’s latest stock picks and Samsung’s AI-token strategy, or browse the full Drawpie blog.

Important information

This article is for informational and educational purposes only. It is not investment, financial, legal, or tax advice, and is not a recommendation, offer, or solicitation to buy, sell, or hold any security. Nothing here is tailored to your individual circumstances.

The technical analysis shown — candlesticks, MACD (12, 26, 9), RSI, and any “tops” or “bottoms” — is one interpretation of historical price data. These indicators are heuristic and lagging, frequently produce false signals, and do not predict future prices. Past performance does not guarantee future results.

Prices and indicator values are as of July 9, 2026 and change continuously; this content may be out of date by the time you read it. Details of ARK Invest’s trades are drawn from its public daily disclosures and financial-media reporting and may differ slightly from the official filings. All investing involves risk, including the possible loss of principal.

The author is not a licensed financial adviser or broker-dealer. Do your own research and consult a qualified, licensed financial professional before making any investment decision. Drawpie and the author accept no liability for any loss arising from the use of this content. Price data via Yahoo Finance.