- Chip stocks have seen major volatility, briefly tumbling into a bear market this month before rebounding.
- A blurry rate outlook, profit-taking, and jitters about competition sparked the latest rout.
- Yet, analysts at Bank of America and UBS say there are reasons to stick with the AI hardware trade.
July has been volatile for chip makers, but analysts say it's worth holding on and sticking with AI hardware makers through the swings.
The red-hot corner of the AI trade has been rocked by volatility this month. Profit-taking after months of gains, competition from AI models in China, and an unclear interest rate outlook helped push the chip sector into a bear market last week.
But Wall Street analysts say stick with it, because the AI market is still strong despite recent scares.
"Latest releases in open-source models, including China's Kimi K3 launched July 16, reinforce our bullish thesis on memory," wrote Bank of America. "From a high level, we view the aggressive Chinese open-model API pricing (up to 5-350x cheaper than Western models) as reflective of business-model choices, not necessarily reflective of hardware costs."
Moonshoot AI, the Chinese startup behind Kimi K3, has been in focus since it released its model last week. Chip stocks plunged on the news, reminiscent of the release of DeepSeek's R1, another Chinese AI model that triggered a sharp selloff in January 2025.
Despite the initial decline, BofA noted that open-source models like Kimi K3 from fast growing Chinese labs may hurt tech companies with similar products, they'll also create fresh demand for chips.
"While frontier model labs are spending capital to grab share, the market is likely expanding its aggregate compute and memory capacity in the process; and every enterprise that downloads Qwen, GLM, Hunyuan, or MiMo instead of paying for Claude Opus creates a new customer-side memory footprint," a BofA analyst noted.
Even if competition from foreign AI developers weighs on demand from US frontier labs, there's no reason to expect a glut of high-end GPUs to flood the market, UBS said.
Ulrike Hoffmann-Burchardi, CIO of UBS Americas, recently wrote that her team doesn't see the recent bear-market decline in chip stocks reflective of potential oversupply. They're also not worried about hyperscalers scaling back on AI capex if they don't start seeing a profit in the near term, even as AI chip costs continue to rise.
Tech giants view artificial intelligence as a winner-takes-all, existential battle," she wrote in a note to investors. "Even if immediate financial returns are low, companies will continue buying hardware to prevent rivals from gaining a permanent technological advantage."
Similar to BofA, Hoffmann-Burchardi added that UBS sees new AI models like Kimi K3 and Inkling from Thinking Machines, creating more growth opportunities for AI hardware makers, as competition spurs heightened demand for chips.
If you enjoyed this story, be sure to follow Business Insider on MSN.