Micron and SK Hynix Cross $1 Trillion; Fireworks AI Seeks $15B Valuation

Memory chips enter the trillion-dollar club on back-to-back days as Fireworks AI targets a $15 billion funding round.

📊 Micron and SK Hynix Both Cross $1 Trillion in Market Value

Decoded: Micron Technology crossed $1 trillion in market capitalization for the first time on May 26, with shares rising 17.4% to $881.60 after UBS raised its price target to $1,625 — the highest among 46 brokerages covering the stock, per Reuters. Bloomberg reported on May 27 that South Korea's SK Hynix also crossed the $1 trillion threshold, making both the U.S. and Korean memory giants trillion-dollar companies within 24 hours. UBS analysts described Micron as sitting at the center of AI's memory demand surge, driven by HBM (high bandwidth memory) requirements for large language model inference and training. Micron's shares have risen more than eightfold over the prior 12 months. Samsung, the world's largest memory chipmaker, had previously crossed $1 trillion, meaning three of the four dominant global memory suppliers now carry trillion-dollar market capitalizations. (Reuters, May 26, 2026; Bloomberg, May 27, 2026)

Why it matters: Memory chips are no longer a cyclical afterthought to the AI trade — they are now central infrastructure with valuations to match. Micron (MU) and SK Hynix's simultaneous crossing of the $1 trillion mark signals that investors have structurally re-rated memory as a durable AI infrastructure play, not a boom-bust commodity cycle. The shift is driven by HBM3E adoption in Nvidia (NVDA) Blackwell GPUs, where memory bandwidth is as performance-limiting as compute. For investors, the memory bifurcation is now clear: companies supplying HBM and AI-grade DRAM — Micron, SK Hynix, and Samsung — are commanding growth multiples previously reserved for fabless chip designers. The remaining question is whether supply can keep pace with AI-driven demand before margins compress.

💰 Fireworks AI in Talks to Raise at $15 Billion Valuation

Decoded: Fireworks AI, a startup that provides infrastructure enabling companies to run and deploy AI models at scale, is in talks to raise a new funding round at a $15 billion valuation, Bloomberg reported on May 27. Fireworks AI operates in the AI model serving layer — providing the API and runtime infrastructure that allows enterprises to call inference endpoints across multiple AI models, including open-source and proprietary options, without building custom deployment pipelines. The company competes with Together AI and Anyscale in the inference infrastructure segment, a layer that has attracted significant capital in 2026 as enterprise AI adoption moves from proof-of-concept to production deployment. OpenRouter, a competing multi-model routing platform, separately raised $113 million this week. (Bloomberg, May 27, 2026)

Why it matters: A $15 billion valuation for an AI inference infrastructure company — one layer below the model developers and one layer above the cloud providers — signals that investors are pricing in the inference economy as a durable revenue layer, not a transitional utility. As frontier models proliferate and enterprises adopt multi-model routing strategies, the infrastructure to serve, switch between, and monitor AI endpoints becomes a critical middleware category. Fireworks AI's funding round, alongside OpenRouter's $113 million raise this week, suggests capital is consolidating around inference infrastructure at scale. For investors, the inference layer represents a higher-margin, more defensible position than commodity GPU compute rental — and the $15 billion valuation reflects that.

Stay decoded. See you tomorrow.

— The Get AI Decoded Team