UBS just tripled its price target on Micron Technology. Not doubled. Tripled. Analyst Timothy Arcuri raised his target to $1,625 from a previous $535, maintaining a Buy rating on the memory chip giant. Shares responded by jumping roughly 8.4% in premarket trading on May 26.

What’s behind the massive target hike

Micron makes the DRAM and NAND memory chips that power everything from phones to the massive GPU clusters training the next generation of AI models. High-bandwidth memory, or HBM, has become the bottleneck component in advanced computing workloads. Micron is one of only three companies on earth that can manufacture it at scale.

Arcuri’s thesis centers on the expected rise of long-term supply agreements. Instead of Micron selling memory chips on the spot market where prices swing wildly quarter to quarter, major customers are locking in multi-year deals. That changes the entire earnings profile: predictable revenue, better margins, less cyclical pain.

The new $1,625 target implies more than 100% upside from Micron’s recent trading levels near $798. At that valuation, Arcuri is pricing the stock at roughly 15x next-12-month earnings, which he argues compares favorably to NVIDIA’s multiple.

UBS had a $450 target back in February 2026, then bumped it to $535 in April. The jump to $1,625 suggests something fundamentally changed in the demand picture between April and now.

The memory super-cycle thesis

Micron’s financials tell the story. The company reportedly posted revenue growth of 196% year-over-year in one recent fiscal quarter. Nearly tripling revenue in a single quarter is the kind of number that makes analysts rethink their models from scratch, which appears to be exactly what Arcuri did.

What this means for investors

The premarket surge of 6-12% across various tracking points shows that institutional money moved quickly on this call. If memory companies can secure long-term supply contracts with hyperscalers and AI companies, the entire sector’s risk profile changes. Cyclical stocks start looking more like growth stocks.

The risk is that if AI spending decelerates, or if competitors ramp HBM production faster than expected, the long-term contract thesis could unravel. There’s also the question of whether 15x forward earnings is actually cheap for a memory company, or whether that multiple only makes sense in a world where AI demand grows uninterrupted.