Micron Technology just posted the kind of quarter that makes analysts look conservative. The company reported fiscal Q3 2026 revenue of $41.46 billion on June 24, a figure that didn’t just beat the consensus estimate of roughly $35 billion. It obliterated it by more than $6 billion.
To put the growth in perspective: Micron pulled in $9.3 billion during the same quarter last year. That means revenue grew more than fourfold in twelve months.
The numbers behind the blowout
DRAM, the workhorse memory used in everything from servers to smartphones, accounted for $31.3 billion of the quarter’s revenue. NAND flash memory contributed another $9.9 billion.
Adjusted earnings per share came in at $25.11. Wall Street had penciled in around $20.39.
Perhaps the most striking number in the entire report: adjusted gross margins of approximately 85%. In the semiconductor world, margins like that signal extreme pricing power in a supply-constrained market. For context, Micron’s gross margins have historically hovered much lower, often in the 30-50% range during typical cycles.
CEO Sanjay Mehrotra framed the results around a larger thesis.
The results reflect the strategic value of memory in the AI era.
A supply deal with Anthropic and bullish Q4 guidance
Micron didn’t just report backward-looking numbers. The company announced a strategic supply agreement with Anthropic, the AI lab behind the Claude family of models. The deal underscores how memory manufacturers are forging direct relationships with frontier AI companies, bypassing the traditional supply chain where cloud providers served as the primary intermediary.
Looking ahead, Micron guided Q4 revenue between $49 billion and $51 billion. If the company hits the midpoint, that would represent yet another sequential leap, roughly 20% higher than the Q3 figure that already shocked the market.
The stock reacted accordingly, with Micron shares posting sharp gains following the report.
The memory supercycle is real, but what about crypto?
Micron’s entire earnings narrative revolved around AI. There was no mention of cryptocurrency, blockchain, or any adjacent technology in the company’s results or commentary.
That’s a notable departure from previous hardware cycles. During the 2017-2018 crypto boom and again in 2021, memory demand was partially driven by mining operations and crypto-adjacent hardware needs. GPU shortages during those periods rippled into memory pricing.
This time around, the demand driver is singular: AI. Companies like Anthropic, OpenAI, Google DeepMind, and Meta are consuming memory at a pace that dwarfs what crypto mining ever demanded.
For crypto investors, the implication is subtle but important. The semiconductor supply chain that once allocated capacity partly based on crypto demand is now almost entirely oriented toward AI workloads. If crypto mining or blockchain infrastructure needs surge again, those customers may find themselves competing for chips against AI labs with multi-billion-dollar budgets and long-term supply agreements already locked in.