RockstarMarkets
All news
Markets · Narrative··Updated 8m ago
Part of: Semiconductor Cycle

Memory Chip Stocks Rally On Record Volumes Yet Valuations Compress: NVDA, MU at 78 RSI

Despite semiconductor and memory stocks reaching all-time highs, valuation multiples have compressed significantly as earnings growth lags stock price momentum. Trading volumes in chipmakers have surged to record levels, with RSI indicators flashing overbought conditions, signalling potential profit-taking risk despite AI capex tailwinds.

R
Rocky · RockstarMarkets desk
Synthesised from 8 wires · 44 mentions in the last 24h
Sentiment
+40
Momentum
70
Mentions · 24h
44
Articles · 24h
79
Affected sectors
Related markets

Key facts

  • NVDA, AMD, MU RSI indicators flashing 70-78 overbought levels
  • Forward P/E multiples compressed despite stock prices at all-time highs
  • Meta-CoreWeave $21B partnership signals AI inference capex acceleration
  • JPMorgan raised Taiwan Taiex target to 50,000 on AI buildout thesis
  • ONDS briefly most-traded US stock despite 1,000x smaller market cap than NVDA

What's happening

A curious disconnect is emerging in the semiconductor complex: stock prices are soaring to record highs, but price-to-earnings multiples are actually tightening. Nvidia, AMD, Broadcom, and Micron have all rallied sharply on the back of relentless AI capex demand, yet forward earnings multiples have compressed meaningfully from their peaks earlier this year. The reason is stark: the market is pricing in such aggressive earnings growth that even with stock prices up 20-40% in recent weeks, the valuation is becoming less stretched on a forward basis.

Trading volumes have reached record levels, particularly in speculative vehicles. A single stock, ONDS, briefly became the most-traded in the US market this week, surpassing Nvidia by dollar volume despite being 1,000 times smaller by market cap. This retail euphoria is mirrored in social media mentions of chip stocks, which hit their highest levels since the 2021 bubble. RSI indicators on NVDA, AMD, and memory names are flashing overbought conditions above 70-78, classic signatures of potential pullback risk.

The bull case rests on unprecedented demand: Meta's $21 billion CoreWeave partnership signals that AI inference capacity, not just training, is becoming a bottleneck. JPMorgan raised its Taiwan-linked targets to 50,000 on the Taiex, framing Taiwanese chipmakers as "the most pure-play exposure to global AI buildout." Goldman Sachs upgraded multiple semiconductor names, citing secular demand tailwinds. Yet analyst notes increasingly acknowledge that valuations are 'stretched' and that any miss on margin expectations could trigger sharp mean-reversion trades.

The debate hinges on whether AI capex growth can sustain 30-40% CAGR for the next two years or if it normalises faster than consensus assumes. Some bears point to Samsung's recent sell-off and subsequent tech-sector weakness in Asia, citing rotation risk into value and defensive names. Others note that Nvidia's guidance and Q1 results did beat, but by narrower margins than recent quarters, raising questions about whether the growth hockey stick can persist. The lack of significant corrections in chip stocks for months may have masked execution risk.

What to watch next

  • 01Nvidia Q2 earnings and guidance: late May
  • 02AMD earnings call and margin guidance: next week
  • 03Taiwan semiconductor capex guidance: Q2 earnings season
Mention velocity · last 24 hours
Coverage from these sources
Previously on this story

Related coverage

More about $NVDA

Topic hub
Semiconductor Cycle: AI Capex, Memory and the SOX Trade

Live coverage of the AI semiconductor cycle — NVDA, AVGO, AMD, ASML, memory demand, capex run rates and overbought signals.