Micron nears $1 trillion, while UBS predicts a trajectory towards $1.8 trillion within the next year.
Timothy Arcuri increased his price target for Micron Technology to $1,625 from $535, supported by long-term supply agreements for HBM that he believes stabilize the memory market into a structural duopoly rather than a boom-bust cycle. Following this announcement, Micron’s shares surged by nearly 19% on Tuesday, with Arcuri's new target suggesting a market valuation approaching $1.8 trillion, assuming the prediction holds true over the coming year.
This new target is the highest among the 46 analysts monitoring the stock and marks one of the most significant single-day re-evaluations of a mature semiconductor firm on record. The rationale behind Arcuri’s assessment is particularly compelling; UBS's position is not that AI demand will permanently elevate memory prices, which is the typical argument for cyclical benefits. Instead, UBS argues that long-term agreements between memory manufacturers and major clients—especially hyperscalers and Nvidia—now sufficiently cover the market to change the conventional boom-bust memory cycle into a more stable, utility-like earnings profile.
UBS estimates that as much as 30% of global DDR memory volumes are currently secured under pricing agreements just below current market rates, which trade some immediate gains for extended predictability. If this trend continues, Micron may start resembling an AI-infrastructure annuity rather than a typical cyclical commodity producer. UBS forecasts that Micron's earnings per share will surpass $100 annually through at least 2029. At current valuations, this projection supports a market capitalization closer to $1.8 trillion rather than the approximately $1 trillion at which the company is presently valued; achieving a trillion-dollar valuation was unthinkable just a year ago.
The situation with HBM is more complex than it initially appears. SK Hynix has emerged as a dominant player in the cycle, controlling around 70% of HBM4 supply for Nvidia’s flagship Vera Rubin platform, while Samsung holds most of the remainder. Micron, initially excluded from the premium VR200 NVL72 configuration, has since gained a role as a supplier for HBM4 in the Rubin CPX, a mid-range accelerator, and began volume shipments of its 36GB 12H HBM4 stacks earlier this quarter.
UBS’s outlook hinges on the expectation that Micron will continue transitioning into the higher-tier HBM4 supply segment over the next 18 months, as SK Hynix and Samsung may struggle to scale quickly enough to meet demand. This competitive landscape warrants consideration; SK Hynix reached a market cap of $1 trillion on Wednesday in Seoul and currently accounts for approximately 57% of the global HBM market share, while Samsung holds about 28%, leaving Micron with a smaller portion. UBS’s prediction is essentially that the smaller player can gain from the market's expansion as clients seek a third source for risk diversification, even if the prices within this tier are lower than those of the leader.
This thesis is reasonable but not guaranteed; it relies on Nvidia's Rubin order volume actually utilizing the projected HBM amounts and assumes that SK Hynix does not capture all the potential upside. The broader memory market also reacted on Tuesday alongside Micron, with SanDisk rising 8% and Western Digital climbing 10%, benefiting from the anticipation that long-term agreements are extending into NAND and traditional storage. If true, the trend UBS describes would signify a significant category shift rather than being confined to just one stock.
Micron’s forthcoming earnings report is expected in late June, with capacity utilization guidance and any updates on its Nvidia HBM4 supply expansion being critical metrics to monitor in light of UBS's new target.
Other articles
Micron nears $1 trillion, while UBS predicts a trajectory towards $1.8 trillion within the next year.
On Tuesday, UBS analyst Timothy Arcuri raised his price target for Micron to $1,625, suggesting a valuation of $1.8 trillion, due to the expectation that long-term HBM agreements will shorten the memory cycle.
