AI

Micron forecasts $50B quarter on AI memory demand

The chipmaker's outlook beat estimates by $7 billion as strategic customer deals reshape the volatile memory market.

Omega Editorial· June 25, 2026· 2 min read

Micron's AI-driven forecast exceeds expectations

Micron Technology issued a fiscal fourth-quarter revenue forecast of approximately $50 billion, substantially above the $43.2 billion Wall Street had expected. The company also projected earnings of roughly $31 per share, compared to analyst estimates of $25.31. Shares rose in premarket trading following the announcement.

The outlook signals continued strength in AI-driven demand for memory chips, even as the broader semiconductor sector has faced recent pressure. Micron is the largest U.S. manufacturer of computer memory chips.

Strategic agreements aim to stabilize pricing cycles

Micron disclosed that it has secured 16 strategic customer agreements averaging three years in length. According to Bloomberg Intelligence analyst Jake Silverman, these deals could help dampen the boom-and-bust cycles that have historically characterized the memory chip industry.

Silverman noted the agreements should support upward pricing revisions through at least 2027, though likely at a slower pace than recent increases. The deals may also reduce pricing volatility as supply begins to catch up with demand around 2029.

Data center spending tied to artificial intelligence has driven demand for both conventional memory and high-bandwidth memory (HBM). Shortages have affected multiple sectors, including computers, phones, and automobiles.

Supply constraints expected through 2027

For the fiscal third quarter ended May 28, Micron reported sales of $41.5 billion and earnings of $25.11 per share, both ahead of analyst projections. Adjusted gross margin more than doubled to 84.9%, exceeding the 81.9% analysts anticipated.

CEO Sanjay Mehrotra told analysts the company sees no near-term resolution to supply constraints. He said tightness is expected to persist beyond calendar 2027, with availability potentially improving gradually in 2028.

Micron also confirmed its HBM4 memory will be used in Nvidia's next-generation Vera Rubin platform alongside products from competing suppliers. Nvidia CEO Jensen Huang verified this partnership.

Why it matters

Micron's multi-year customer agreements represent a structural shift for an industry known for extreme price volatility. If these deals succeed in stabilizing pricing and reducing cyclicality, they could make memory chip investments more predictable for both manufacturers and customers. The extended supply shortage timeline also suggests AI infrastructure buildout will remain constrained well into the second half of the decade, potentially limiting the pace at which companies can deploy new AI capabilities.

These details were first reported by AI Watch, citing Bloomberg Intelligence analysis.

#micron technology#memory chips#ai infrastructure#semiconductor industry#high-bandwidth memory#nvidia

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in AI

AI· 4 min read

General Intuition raises $320M to train AI agents in video games

The startup believes gameplay data with embedded player actions offers a faster path to real-world robotics than traditional training methods.

Via AI Watch · Jun 25, 2026
AI· 3 min read

Chinese AI Model Z.ai Gains Traction After U.S. Restricts Anthropic

Z.ai's GLM-5.2 offers performance nearly matching top American models at a fraction of the cost, capturing Silicon Valley attention as six Chinese models now rank among the world's top ten.

Via AI Watch · Jun 25, 2026
AI· 2 min read

Qualcomm targets $15B data center revenue by 2029 with AI chips

The mobile chipmaker is pivoting to rack-scale infrastructure with custom silicon for hyperscalers including Meta and Microsoft.

Via AI Watch · Jun 25, 2026