The recent 13F filing from Renaissance Technologies, the quant powerhouse founded by Jim Simons, offers a clear signal regarding capital allocation in the current tech landscape. A substantial increase in its Micron Technology (MU) stake, adding nearly 1.81 million shares for approximately $520 million in new exposure, stands in sharp contrast to a significant reduction in its Nvidia (NVDA) holdings. This isn't merely a portfolio rebalance; it’s a directional bet on where the next phase of value creation lies within the AI infrastructure.
Renaissance’s total Micron position now approaches $859 million, elevating it to a top-tier holding. This move is particularly notable given that it occurred in a quarter where the fund also reduced its stakes in other AI bellwethers, including Google-parent Alphabet. The implication is that while the broader AI narrative remains compelling, the firm sees more compelling, perhaps less priced-in, upside in the foundational components enabling that narrative.
The timing of this Micron bet is critical. The memory market is transitioning from a period of oversupply to one of tightening supply, a shift that is already translating into climbing prices. Micron’s own guidance suggests a trajectory of margin expansion through 2026. This positive outlook is intrinsically linked to the relentless demand for High-Bandwidth Memory (HBM), which has emerged as the most supply-constrained and strategically vital segment of the chip ecosystem, directly fueled by AI workloads.
Micron’s financial performance underscores this shift. Fiscal 2025 saw data-center sales surge to 56% of total sales, accompanied by 52% gross margins. HBM sales in fiscal Q4 reached approximately $2 billion, indicating an annualized run rate of nearly $8 billion. Looking ahead, the forecasts are even more aggressive: DRAM contract prices are projected to rise 90-95% quarter over quarter in Q1 2026, with NAND prices potentially skyrocketing by 55-60%, and PC DRAM pricing possibly jumping over 100% quarter over quarter. Micron is also committing substantial capital expenditure, with a baseline of approximately $4.5 billion per quarter for 2026, without a significant impact on inventories. This aggressive capex, coupled with demand-driven pricing power, suggests a conviction in the longevity and profitability of this memory cycle. Renaissance’s increased exposure suggests a belief that the market may still be underestimating the magnitude and duration of this pricing power, positioning itself for returns that could exceed broader expectations.
This wasn’t about chasing growth. It was about identifying where the market was still mispricing the underlying dynamics.
Conversely, the decision to trim Nvidia is equally telling. Renaissance decreased its Nvidia stake by nearly $887 million, reducing its position from $1.05 billion at the end of Q3 to a mere $162.5 million by year-end. Nvidia has been the undisputed poster child of the AI trade, experiencing a phenomenal run over the past three years, reaching an all-time closing high of $207.03 in late October 2025. By mid-February 2026, it still commanded an eye-popping $4.5 trillion market capitalization, trading at more than 45 times non-GAAP earnings (trailing-12-months).
This is not a bearish call on AI itself, but rather a pragmatic assessment of valuation and relative opportunity. It appears to be a classic move of portfolio hygiene: harvesting substantial gains from a position that has seen an extraordinary appreciation and redeploying that capital into an area where the pricing power cycle is still in its earlier stages. The fund is essentially rotating out of a mature, highly valued AI leader into a foundational component supplier that is just beginning to realize the full financial benefits of the AI build-out.
The underlying message is clear: while the AI narrative continues to drive market sentiment, the most attractive investment opportunities may be shifting from the obvious beneficiaries to the less heralded, but equally critical, enablers. The capital is moving to where the fundamental supply-demand imbalance is most acute and where the pricing power is still accelerating, rather than merely sustaining.
This is a calculated move, reflecting an understanding that market leadership rotates, and that even the most dominant trends create secondary and tertiary opportunities. The focus is on the structural shifts in component pricing and supply, rather than simply riding the momentum of established giants. It suggests a preference for the early-cycle leverage of memory producers over the later-cycle, premium valuations of the chip designers who have already captured much of the initial AI euphoria.
The market is always about relative value. And sometimes, that value is found in the less glamorous, but fundamentally essential, parts of the supply chain.
Renaissance’s latest 13F activity snapshot reveals a market value of $64.5 billion, down from $75.8 billion, indicating a -14.9% outflow as a percentage of total market value. The firm made new purchases in 466 stocks, added to 1,030 stocks, sold out of 738 stocks, and reduced holdings in 1,680 stocks. The top 10 holdings concentration remains relatively modest at 12.1%, suggesting a diversified quantitative approach.
Top buys included Costco ($0.59 billion), Micron Technology ($0.52 billion), Tesla ($422 million), Netflix ($419 million), and Procter & Gamble ($313 million). Top sells, by value, were Nvidia ($0.89 billion), Alphabet Class A ($0.72 billion), Alphabet Class C ($0.51 billion), AppLovin ($460 million), and Gilead Sciences ($328 million).
The shift is evident. It’s a re-allocation, not a retreat.
