Two giant investment managers.
Two very different approaches to the AI revolution.
Cathie Wood's ARK increased its Nvidia holdings by 85% and SpaceX holdings by 51%.
Meanwhile, Bridgewater reduced Nvidia by approximately 30%, doubled its Micron position and dramatically increased its Sandisk exposure.
Comrades, here's what caught my attention.
Both funds are positioning for AI growth. But they're choosing different parts of the value chain.
ARK is betting more heavily on computing platforms and disruptive innovation.
Bridgewater appears increasingly interested in the memory infrastructure required to power that computing.
And after Nvidia delivered blockbuster earnings but its shares still struggled, I think this divergence deserves serious attention.
π₯ 1. What the Institutional Filings Actually Reveal
Let's look at the Q3 2026 Form 13F disclosures, comparing holdings on June 30 and September 30.
ARK Investment Management
π Nvidia (NVDA): 1.38M to 2.56M shares, up 85.4%
π SpaceX (SPCX): 4.48M to 6.75M shares, up 50.7%
π° Nvidia holding value: approximately $585M
π° SpaceX holding value: approximately $1.02B
ARK also reported approximately $563M invested in Circle (CRCL), making it another substantial position in its disclosed portfolio.
Now look at Bridgewater.
Bridgewater Associates
π Nvidia (NVDA): reduced 30.5%, to approximately 2.69M shares
π Micron (MU): increased 103.2%, to approximately 237,119 shares
π Sandisk (SNDK): increased from just 317 to 66,756 shares
π° Micron holding value: approximately $252.6M
π° Sandisk holding value: approximately $116.1M
Notice the difference.
ARK increased its Nvidia exposure considerably.
Bridgewater trimmed Nvidia but still maintained a substantial position.
That's important.
Bridgewater wasn't abandoning AI. It was changing its exposure to different parts of the industry.
And I think that distinction matters more than the headline percentages.
π 2. Why Micron May Be the More Interesting Signal
Here's my interpretation.
The next stage of AI growth isn't simply about producing more powerful GPUs.
It's about supplying the infrastructure those GPUs require.
That includes high-bandwidth memory, advanced packaging, networking and storage.
A powerful accelerator without sufficient memory bandwidth cannot operate at its full potential.
And as AI systems become larger, memory requirements increase dramatically.
This is where Micron becomes interesting.
Its latest fiscal results showed record performance, with management expecting an even stronger fiscal 2027.
Nvidia's latest earnings also demonstrated that demand for AI computing remains enormous.
But there's a complication.
Nvidia reported $115.8B in quarterly revenue and guided towards $130B next quarter, yet its shares remained under pressure after earnings.
Why?
Because investors are questioning valuation, future margins and the sustainability of AI spending.
Meanwhile, Micron may benefit from growing memory demand even as GPU manufacturers face higher component costs.
My interpretation: Bridgewater may be positioning for the possibility that memory suppliers capture a greater share of AI infrastructure spending.
That is a hypothesis supported by its portfolio changes, not a confirmed explanation of Bridgewater's investment decisions.
βοΈ 3. Is ARK Wrong About Nvidia and SpaceX?
Not necessarily.
ARK's strategy could still prove extremely rewarding.
Nvidia retains major advantages through CUDA, its integrated computing platforms and deep customer relationships.
Even if competitors gain market share, Nvidia can continue growing as the AI market expands.
SpaceX offers a different exposure altogether.
Its launch infrastructure, satellite operations and longer-term space technology ambitions create opportunities beyond conventional semiconductor demand.
But here's the risk.
These are businesses where investors may already be paying substantial premiums for future growth.
Excellent companies can produce disappointing investment returns when expectations become too demanding.
There's another intriguing detail.
Alphabet reported that its SpaceX holding represented approximately 95% of the securities value disclosed in its Q3 13F.
That sounds extraordinary.
But please don't confuse this with 95% of Alphabet's entire assets.
The percentage refers only to the securities included in that particular regulatory filing.
Likewise, changes in institutional share counts are not the same as investment returns.
A 51% increase in shares doesn't mean a 51% gain.
And these filings reflect September 30 positions, not transactions made after that date.
π― 4. My Prediction: Who Performs Better Next?
Over the next two to four weeks, I see three possible outcomes.
π’ Bull Case: 30%
AI sentiment improves and semiconductor stocks rebound broadly.
Nvidia recovers above $165β$170.
Micron reclaims $720 and challenges $750.
Both strategies benefit, with momentum potentially favouring previously beaten-down AI leaders.
π‘ Base Case: 45%
AI stocks remain volatile and selective.
Nvidia struggles to sustain its earnings recovery, while Micron trades around $670β$730.
Investors begin differentiating between companies facing margin pressure and suppliers that may benefit from tighter memory availability.
This is my most likely scenario.
π΄ Bear Case: 25%
The broader AI trade weakens further.
Nvidia loses $150 decisively.
Micron breaks below $670 and risks testing $630β$650.
Even strong memory fundamentals cannot necessarily overcome widespread institutional selling.
These are judgment-based scenarios, not statistically calculated probabilities or guaranteed targets.
My prediction? I favour selective exposure to memory over aggressively chasing the most crowded AI names, but only when price action confirms that buyers are returning.
And I would not assume that Micron is immune to a broader semiconductor correction.
π° 5. How Would I Position Myself?
Personally, I find Micron the more interesting name to monitor right now.
But that doesn't mean buying blindly.
Here's my trading framework.
Micron (MU)
With MU recently trading around $700, I would watch $680β$690 as a potential stabilisation area.
A convincing reclaim of $720, followed by $750, would strengthen the recovery case.
A sustained breakdown below $670 would make me more cautious.
I would also monitor HBM demand, memory pricing and whether higher revenue is translating into stronger margins and cash flow.
Nvidia (NVDA)
I want to see the stock recover above $160β$165 and sustain buying interest.
A move towards $170 would become more credible if the broader semiconductor sector also stabilises.
If $150 fails, I would avoid assuming the recent earnings beat guarantees a recovery.
SpaceX (SPCX)
I would monitor whether the stock can defend the $135β$140 region and regain momentum towards $150.
However, I would evaluate its launch, satellite and financing economics separately from the AI chip trade.
One important point.
I wouldn't buy Micron just because Bridgewater increased its position.
The fund could have purchased those shares at very different prices, and its allocation may have changed since September.
π§ 6. The Biggest Lesson From These Billion-Dollar Trades
Here's what I think investors should remember.
ARK isn't necessarily betting against memory.
Bridgewater isn't necessarily betting against Nvidia.
And neither fund has a monopoly on understanding the AI revolution.
What their filings reveal is different portfolio positioning.
ARK is emphasising growth platforms and disruptive innovation.
Bridgewater's disclosed changes indicate increasing exposure to memory and storage.
But here's my contrarian thought.
The most profitable part of the next AI cycle may not be the company receiving the most attention. It could be the supplier solving the industry's most expensive bottleneck.
That's why I'm watching Micron so closely.
However, memory is cyclical.
If producers eventually expand supply faster than demand, today's attractive pricing environment could reverse.
Investors must distinguish a temporary shortage from a durable improvement in earnings power.
π‘ My Final Take: Which Whale Would I Follow?
If I were choosing a theme rather than blindly copying a fund manager, I currently lean towards the memory-infrastructure opportunity.
I particularly like Micron's exposure to HBM and advanced data centre memory.
But I want confirmation from both business fundamentals and price action.
Nvidia remains an extraordinary business.
SpaceX offers a different long-term growth opportunity.
And ARK's continued exposure to Circle reminds us that its broader strategy remains heavily tied to disruptive technologies.
My strongest conviction is that AI infrastructure spending will continue, but the stock-market winners may rotate as profitability shifts across the supply chain.
That's where I think investors should focus.
Not simply on which billionaire bought the most shares.
But on which business can translate rising demand into durable earnings and attractive shareholder returns.
Comrades, which strategy appeals to you more?
Cathie Wood's Nvidia and SpaceX exposure, or Bridgewater's increasing Micron and Sandisk positions?
Personally, Micron is the name I'm watching most closely.
I am not a financial advisor. Trade wisely, Comrades!
If you find this post interesting, give it wings! πͺ½ Repost and share the insights. Do consider 'Follow me' and get firsthand read of my daily new post. Thank you.
Comments