①The chip sector has been volatile since July, but Wall Street analysts are calling on investors to hold onto AI hardware stocks despite market fluctuations;
②After several days of consideration, Wall Street analysts now unanimously believe that the recently released Chinese AI model Kimi K3 will further boost demand for memory chips;
③This was also one of the drivers behind the sharp rebound in U.S. memory stocks overnight.
Cailian Press, July 22 (Editor Bian Chun) The chip sector has experienced volatile movements since July, but Wall Street analysts are urging that, despite the market swings, investors should continue to hold AI hardware stocks. After several days of deliberation, Wall Street analysts now widely agree that the recently launched Chinese AI model Kimi K3 will further drive up demand for memory chips.
This was also one of the catalysts for the sharp overnight rebound in U.S. memory stocks. On Tuesday, U.S. memory concept stocks staged a strong rebound, with the Roundhill Storage ETF (DRAM) closing up 10.91%, SanDisk up 14.27%, SK Hynix up 13.75%, Western Digital up 12.51%, Micron Technology up 12.17%, and Seagate Technology up 11.14%.
Over the past few weeks, the hardware stock sector, a previously red-hot area of AI trading, has experienced significant volatility. After months of gains, a combination of factors including investor profit-taking, competitive pressure from Chinese AI models, and uncertainty over interest rate prospects pushed the chip sector into bear market territory last week.
However, Wall Street analysts are advising investors to maintain their holdings in chip stocks, as the fundamentals of the AI market remain robust despite recent investor jitters.
Bank of America: Kimi K3 to Generate New Chip Demand
Bank of America recently noted that the latest open-source model releases, including the Chinese model Kimi K3 launched on July 16, further validate our bullish thesis on memory chips.
Analysts at Bank of America, led by Vivek Arya, stated that Chinese open-source large models offer highly competitive API pricing, with Western model prices being 5 to 350 times higher by comparison. However, this reflects more a choice of business model rather than a significant reduction in hardware costs.
Last Thursday, Chinese AI developer Moonshot AI released its latest Kimi K3 model, an open-weight model that ranks comparably to leading U.S. models in certain benchmarks.
This has reignited questions about whether other model developers are over-investing in AI hardware and computing power. Affected by this, U.S. chip stocks plunged, reminiscent of the market reaction triggered by the release of another Chinese AI model, DeepSeek R1, in January 2025.
Nevertheless, Bank of America points out that rapidly growing Chinese open-source models like Kimi K3, while potentially challenging overseas tech companies with similar products, will also generate new demand for chips.
Bank of America analysts noted that while frontier model labs are spending to compete for market share, the overall market's computing power and memory capacity may expand in the process. Each download of an open-source model implies that customers need to deploy the model themselves, thereby creating new demand for memory such as HBM, DRAM, and NAND, which is not the case with closed-source models.
Wall Street Collectively Bullish on Chip Stocks
Ulrike Hoffmann-Burchardi, Chief Investment Officer for the Americas at UBS, stated that her team views the recent bear market in chip stocks as not reflective of underlying oversupply. They are also not concerned about hyperscalers cutting AI capital expenditures, even if profits are not immediately visible and AI chip costs continue to rise.
She wrote in a note to investors: "Tech giants view AI as a 'winner-takes-all' battle for survival. Even with low short-term financial returns, these companies will continue to procure hardware to prevent competitors from gaining a permanent technological advantage."
Echoing Bank of America's view, Hoffmann-Burchardi added that UBS believes new AI models like Kimi K3 will create more growth opportunities for AI hardware manufacturers, as competition is stimulating higher demand for chips.
Furthermore, Morgan Stanley told investors in a report on Monday that it sees the recent sell-off in U.S. memory chip stocks as providing a good entry point, citing the ongoing and intensifying shortage of data center memory chips despite mixed signals from other parts of the market.
UBS Group's trading desk also recently stated that the rapid sell-off in momentum stocks may be nearing its end, creating an opportunity for investors to rebuild positions in AI and semiconductor stocks.
JPMorgan also believes the recent decline in chip stocks is not the start of a prolonged downturn but rather a consolidation before the next leg up. The bank recently noted that the divergence between chip stock prices and fundamentals is widening, while technical indicators also show "oversold" conditions, recommending increasing holdings over the summer. The bank also expects AI-driven tightness in DRAM supply and demand to persist until 2028, with the Q2 earnings season acting as a catalyst for the next rally.
Nancy Tengler, CEO of Laffer Tengler Investments, compared the market reaction to the Kimi K3 model release to the sell-off triggered by the Chinese open-source DeepSeek model last year, noting that the previous sell-off created a "tremendous opportunity" for bargain-hunting investors.
Hendi Susanto, a portfolio manager at Gabelli Funds, stated that despite significant share price gains for memory chip companies in recent months, the "investment thesis remains intact" as demand is expected to continue exceeding supply next year. He said this implies "favorable pricing and strong earnings" can persist.
