Memory Chips, Optical Communication Sectors Rise Together, US AI Recovery Gradient Verified; Is It Still a Buy?

TradingKey - On June 8, ET, the memory chip and optical communication sectors were the first to rebound.

This validated the viewpoint previously published by TradingKey in "S&P 500 Total Market Value Is About to Surpass $70 Trillion, Is There a Bubble? When to Buy to Ensure Relative Safety?" regarding the pattern that "upstream sectors of the AI industry chain exhibit the greatest recovery elasticity."

Following the systemic sell-off on June 5, which saw the Philadelphia Semiconductor Index plunge over 10% in a single day, these two sub-sectors have become the vanguard segments leading the rebound in the AI valuation recovery.

On June 8, ET, the Philadelphia Semiconductor Index's gains expanded to 7%, with Nvidia ( NVDA) rising 1.77%, and TSMC ( TSM) up 4.18%.

Why Memory Chips and Optical Communications Were the First to Rebound

The decline in the AI supply chain during this round has exhibited distinct tiered characteristics. Leading storage chipmakers, including South Korea's SK Hynix and Samsung Electronics, as well as U.S.-listed Micron Technology and SanDisk, plummeted by over 10% during the crash but subsequently emerged as the stocks with the strongest rebounds.

Memory chips represent the sector within the AI supply chain with the most certain supply-demand gap.

SK Hynix has officially announced that the HBM capacity gap will persist until 2030, while Samsung Electronics and Micron Technology have simultaneously reached a state where their production capacity is fully sold out. Goldman Sachs projects that the HBM market size will grow from $56 billion in 2026 to $116 billion in 2027, further expanding to $168 billion by 2028. With the supply-demand gap expected to last until 2028, the visibility provided by "full order books and sold-out capacity" serves as an earnings anchor that other AI segments find difficult to match.

The investment logic for the optical communications sector is equally certain.

Previously, Lumentum ( LITE) released news that garnered market attention: driven by accelerating demand from large-scale AI data centers in the United States, the company's orders are expected to be fully booked through 2028.

In June, while sharing the stage with Marvell Technology's ( MRVL) Matt Murphy at COMPUTEX, Jensen Huang stated that optical communications leader Marvell Technology will become the next trillion-dollar market cap company. This statement directly ignited the global optical communications sector, with Marvell Technology skyrocketing by over 32% that day, Coherent surged over 17%, and Corning ( GLW) rose more than 13%. Communications represent the essential utilities of AI data centers; as long as the AI computing power arms race continues, optical communications demand will not diminish as clusters expand by orders of magnitude.

Institutions raise price targets en masse

Memory chip leaders have become the focal point for concentrated institutional price target upgrades.

Morgan Stanley doubled its price target for Micron from $520 to $1,050, maintaining a Buy rating; analysts warned that DRAM chips have become a major bottleneck in AI infrastructure development, yet tech giants remain willing to pay premium prices.

Susquehanna significantly raised its price target for Micron from $600 to $1,750; Wells Fargo raised it to $1,220, and Citi to $840. Among the 41 analysts covering Micron, the vast majority assign Buy or Strong Buy ratings.

Citi significantly raised its price target for SanDisk from $1,300 to $2,025, while Morgan Stanley hiked its target from $1,100 to $1,750. Goldman Sachs explicitly supported the P/E valuation logic for memory, raising price targets for the three giants—South Korea's SK Hynix and Samsung, and Japan's Kioxia—projecting that supply shortages will persist until 2028.

Optical communications have also entered an institutional revaluation phase. Marvell Technology surged over 32% in a single day following Jensen Huang's public endorsement, with U.S. optical communication leaders rallying collectively as Innolight, Eoptolink, and TFC hit successive all-time highs. Huatai Securities expects a significant supply-demand gap in global optical fiber demand between 2026 and 2027, with leading firms poised to fully benefit from the industry upcycle. This chain-wide wave of intensive upward revisions, from chips to infrastructure, signifies that the capital market is re-pricing the earnings certainty of the AI upstream sector.

Recovery does not imply a reversal.

Although storage chips and optical communications have shown the strongest rebound, the nature of this recovery rally remains an oversold rebound rather than a confirmed trend reversal. The following risk factors warrant a cautious outlook:

The trigger for the rebound is a recovery in expectations rather than accelerating fundamentals. Broadcom's Q2 revenue surged 48% year-over-year, and AI semiconductor revenue skyrocketed nearly 80%, yet the stock plummeted after failing to raise its 2027 AI sales guidance, later rebounding about 2.8%. This indicates the market remains in a state of "extreme sensitivity to expectations." UBS clearly pointed out that Broadcom's failure to raise its 2027 AI sales guidance disappointed investors; the market's threshold for "surprises" has become extremely high. Against a backdrop of crowded positioning, any data that "meets expectations but is not surprising enough" could trigger a new round of volatility.

Valuations for some underlying assets remain in historically extreme territory. Before the recent crash, Micron had accumulated a gain of nearly 90% in May, with year-to-date gains still as high as 232%. Given such massive prior gains, any signal of a marginal deceleration in AI capital expenditure could trigger a more violent sell-off than the current one, as valuation safety cushions remain thin.

There are concerns regarding highly uncertain capital structures. The driver of this rebound remains the reallocation of existing capital rather than the entry of incremental funds. During the Asian trading session on June 9, the KOSPI staged a sharp rally of nearly 8%, but the preceding crash stemmed from large-scale foreign capital outflows and excessive retail leverage accumulation; market depth remains fragile. Any unexpected macro or industry-level news could trigger a second round of large-scale selling under the current positioning structure. The short-term rebound window is not a safe zone.

Structural suppression from the macroeconomic environment has not been lifted. The 10-year U.S. Treasury yield continues to run near 4.5%, and the probability of a Fed rate hike this year is close to 50%. High interest rates inherently exert systemic valuation pressure on high-valuation growth stocks. Even if corporate fundamentals do not deteriorate, the increase in external discount rates is depressing valuation medians. The decision-making weight of such exogenous variables on "whether to buy now" cannot be ignored.

Overall, the recovery gradient for the storage chip and optical communication sectors has been verified; they are the two segments with the highest certainty in the AI industry chain. However, entering with a heavy position all at once at this stage is not recommended.

A more suitable strategy is to use storage chip and optical communication ETFs as core base components of AI upstream holdings, building positions in tranches to average costs while controlling position concentration in the AI sector within the overall portfolio.

More importantly, every sharp correction in the AI sector is accompanied by intensive upward revisions of price targets by institutions, which in itself is a confirmation of the sector's long-term value. For investors already holding positions, the current price is not the time for full liquidation. For investors with no current holdings, waiting for a new round of expectation signals to converge before positioning will be superior to rushing in with heavy positions while volatility has yet to stabilize.