Coatue Management 13F Portfolio · Philippe Laffont

The standout move is Micron. Coatue took Micron from a rounding error to a 7.46% position, boosting share count by +1793.7% and adding an estimated $3.44B. That is an explicit call that high-bandwidth memory and DRAM sit at the center of AI’s next leg, not just GPUs.

Intel arrives as a new 3.47% stake worth $1.69B, signaling belief that foundry catch-up and x86 relevance in AI PCs and servers are mispriced. Pair that with a fresh, smaller Advanced Micro Devices position and a modest add to Broadcom, and you get a bet that the second wave of AI silicon — CPUs, interconnect, accelerators — still has room.

Coatue is also arming the physical stack around all this compute:

• GE Vernova at 6.18% and essentially unchanged: a huge, long-running bet on grid and generation build-out.
• Eaton nudged up to 4.35%: electrical gear and power distribution for data centers and electrification.
• Equinix lifted by +16.7% to 2.73%: the neutral data center landlord that monetizes rising rack density.

Beyond core infra, they leaned further into hyperscaler demand with Amazon (+49.2% to 5.8%) and modestly added to Alphabet and Microsoft. Together, this is a stacked trade on AI workloads from chip to rack to cloud to consumer.

If Micron and Intel are the new heroes, Lam Research, Applied Materials, Taiwan Semiconductor, and ASML are the piggy banks. Coatue trimmed Applied Materials by -19.7% (shedding about $748.1M), cut ASML by -40.5% (about $399.5M), and reduced Lam Research and Taiwan Semiconductor modestly. Gains vs cost are huge — Lam Research is up 387.6% versus their buy price and Taiwan Semiconductor 183.2% — so this looks like risk recycling, not a thesis reversal.

The software-heavy side of AI and digital ads is where conviction is clearly cooling. They slashed Synopsys by -58.3% and trimmed Meta by -10.2%, while keeping Microsoft and Alphabet as the core platform plays. The message: keep the hyperscalers, but don’t overpay for design tools and adtech as infrastructure becomes the scarcer asset.

Within consumer internet, Netflix was cut hard (-31.7%), and smaller software and engagement names were bled down: AppLovin (-9.2%), Pinterest (-28.7%), DoorDash (-4.9%). These trims sit alongside tiny, underwater positions like The Trade Desk, Zillow, and some health-tech and fintech names where Coatue chose to sit tight rather than add or capitulate.

Tech’s share of the book actually ticked down from 70.01% to 67.21%, but the quality of that exposure changed. Coatue is swapping some software and the most extended equipment winners into a denser cluster of AI-critical hardware: memory, CPUs, legacy nodes, and networking.

The most notable net move outside tech is into energy and quasi-infrastructure finance. Energy went from 0% to 2.61% via a new $1.15B position in Forgent Power Solutions, essentially an industrial power play aligned with data center and grid demand. Finance more than doubled from 1.93% to 4.19%, driven by new Hut 8 exposure plus existing Nubank and First Citizens.

Consumer Discretionary crept up to 9.55%, but that masks a barbell. At one end, they scaled Amazon and held Generator maker Generac; at the other, they cut back on more cyclical or execution-sensitive names like Netflix, Carvana, and Mastec. Health Care, Utilities, and Real Estate all edged lower in aggregate — they’re still present, but clearly not where Coatue wants incremental risk in an AI and infrastructure-driven tape.

This 13F says Coatue believes the easy money in AI narratives has been made, and the next phase will reward owning the scarce inputs: compute cycles, memory bits, megawatts, and racks. Concentrating nearly 32% of the book in four semiconductor names and adding Intel and Advanced Micro Devices is a statement that capacity, not app-layer growth, will drive outperformance.

The extension into Forgent Power Solutions, Equinix, GE Vernova, and Eaton suggests they see AI as an electricity and real estate problem as much as a software one. Hut 8 layers on a high-beta, crypto-tied way to express that same power-plus-compute thesis through the lens of digital assets and mining economics.

Looking forward, expect Coatue to keep harvesting gains from richly valued equipment and consumer software to feed this AI plumbing complex. If shortages in memory, power, or data center space deepen, this portfolio is built to benefit; if AI demand disappoints or capital floods into new capacity too fast, the same positions will be where they feel the most pain. Either way, the bet is clear: own the bottlenecks, not the buzzwords.