Invezz

How SpaceX’s earnings call turned bullish for Micron, SK Hynix, other memory stocks

How SpaceX’s earnings call turned bullish for Micron, SK Hynix, other memory stocks
Vatsala Gaur
Aug 05, 2026, 05:37 AM

powered by

Invezz
Micron (MU)

Buy Micron. SpaceX’s CEO basically confirmed the core trade: memory is the bottleneck—demand is growing far faster than supply, so pricing stays firm for years even as capex ramps. Micron is the cleanest way to own that tightness and the AI/HBM buildout, with upside reinforced by Bank of America calling the sell-off a buying opportunity.

Key Risk: AI/cloud capex slows sharply and memory demand growth falls back toward supply growth, forcing prices and margins down.

SK Hynix (HXSY)

Buy SK Hynix. It’s the other direct beneficiary of sustained HBM demand and the “memory bottleneck” narrative. If elevated memory prices persist through 2029–2030, SK Hynix should capture outsized earnings power versus more diversified peers.

Key Risk: HBM demand weakens (or customers redesign around lower-memory requirements), causing a faster-than-expected pricing normalization.

  • Elon Musk says AI memory demand is rising about 10X faster than supply growth.
  • Deloitte says memory supply tightness, high prices to persist through 2029-30.
  • Bank of America says Micron's recent sell-off is a buying opportunity.

SpaceX's earnings may have sent its shares sharply lower as investors fretted over the rise in AI spending, but CEO Elon Musk delivered a bullish message for one of the semiconductor industry's most volatile segments: memory chips.

Speaking during the company's earnings call, Musk argued that demand for memory chips is rising at a pace far exceeding global production, suggesting prices could remain elevated for years despite aggressive capacity expansion by manufacturers.

His comments reinforce the long-term investment case for companies such as Micron Technology, SK Hynix, and Samsung Electronics, all of which are racing to meet surging demand from artificial intelligence infrastructure.

The remarks also come as investors debate whether the current AI spending boom can sustain strong pricing for memory products, particularly high-bandwidth memory (HBM), which is essential for training and running advanced AI models.

Musk says memory remains the industry's biggest bottleneck

Musk made the comments while responding to a question from JP Morgan analyst Doug Anmuth, who asked about supply-demand dynamics and whether SpaceX would be able to maintain its premium pricing.

"Look at the rate at which logic and memory is being produced. One must always consider the limiting factor here. The limiting factor currently is memory," Musk said.

"The memory output is increasing by around 20% per year. Now, normally, that would be fantastically fast and amazing for any large mature industry."

He contrasted that with the pace of demand growth.

"Ask yourself, is the demand increasing by 20% a year? No, the demand is increasing by 200% a year, maybe higher. If you have got demand increasing much faster than supply, Economics 101 would suggest that the price increases. It does not decrease."

The comments come as AI companies continue pouring billions of dollars into data centers equipped with graphics processors and advanced memory chips, creating sustained demand across the semiconductor supply chain.

Memory stocks remain volatile despite improving outlook

Musk's statement comes at a time when memory stocks have experienced significant volatility in recent weeks.

Micron shares were down around 2% in premarket trading on Wednesday after gaining ground over the previous two sessions, while SK Hynix slipped roughly 3%.

Investor concerns have centred on whether hyperscale cloud providers could eventually slow their AI spending and whether Chinese memory producer CXMT could emerge as a stronger competitor.

Sentiment, however, has improved this week following stronger-than-expected earnings from Amazon, Microsoft, and Google, whose cloud businesses continued to report accelerating AI demand.

Broader market optimism surrounding easing tensions in the Middle East has also supported semiconductor stocks.

Deloitte expects shortages to persist for years

Industry forecasts continue to suggest that memory supply will struggle to keep pace with demand despite record investment.

According to Deloitte, the world's three largest memory manufacturers- Micron, Samsung, and SK Hynix- are expected to increase their combined capital expenditure by nearly 340% between 2024 and 2027 to expand production.

Memory-related investment could account for roughly half of total semiconductor industry capital expenditure by 2026.

Even so, Deloitte noted that additional capacity will take years to come online because new fabrication plants typically require three to five years to build and ramp up production.

The consulting firm expects hyperscale cloud providers to allocate around 30% of their 2026 data-centre investments to memory, with that share projected to rise to 36% in 2027.

Memory components also account for roughly one-quarter of the bill of materials for high-end AI server racks.

As a result, Deloitte forecasts global memory sales could exceed $1 trillion in 2027, compared with approximately $230 billion in 2025.

"The current memory supply tightness and elevated prices may persist until 2029 or even 2030, assuming continued demand among hyperscalers for memory chips. Other customers that need memory for devices such as PCs, smartphones, and other consumer electronics, as well as for non-AI data centres, will likely also need to contend with high memory prices," Deloitte said.

Bank of America sees Micron sell-off as buying opportunity

Musk's comments closely align with a bullish note issued by Bank of America this week, in which the brokerage reiterated its Buy rating on Micron and maintained a price target of $1,550, implying roughly 72% upside from Tuesday's closing price.

Analyst Vivek Arya acknowledged that memory pricing and margins would eventually normalise as new capacity enters the market between mid-2027 and 2028.

However, he argued that investors have become overly focused on a future downturn despite continued improvement in current industry fundamentals.

"Hyperscaler spending continues to rise despite higher component costs, suggesting semis/memory pricing power," Arya wrote.

Bank of America also noted that GPU rental rates remain close to record highs and that none of the major cloud providers has indicated memory shortages are constraining AI deployments.

The brokerage further dismissed concerns over Chinese manufacturer CXMT, arguing the company remains focused on commodity DRAM and does not currently pose a significant competitive threat in high-bandwidth memory used for AI workloads.