Morgan Stanley says a cyclical downturn in global tech is almost certain next year. “Our base case is that a cyclical downturn is a near certainty in 2025,” it said in an Aug. 20 report, adding that it will be caused by a “likely reversal” in the rate of change of revenue growth and tight supply-demand conditions. “Certain segments of tech – such as semiconductor materials and the AI supply chain – could be hit the hardest in view of higher volatility in pricing and operating leverage,” said the bank. However, it added that the effects are likely to be felt across all sectors of tech. Playbook for a potential peak cycle Morgan Stanley says the semiconductor industry is in the late cycle, moving “from optimism to euphoria,” with the risk-reward not being as attractive anymore. Chips have been riding the artificial intelligence boom but the next leg of the AI trade “may not be quite so simple,” the bank said. “The key question is the ‘pull forward’ on AI demand going into 2025-26, as no one would debate the obvious near-term demand for [graphics processing units],” it added. Over time, applications for the use cases of AI will have more value than infrastructure, it said. “Despite the current shortage of AI computing chips, we have to remember that this is not a train that will run forever – eventually AI chips will catch up to demand, become cyclical and more difficult to sustain the current pace of gains,” Morgan Stanley wrote, adding that it expects topline growth will start to moderate into 2025. “In our view, it is always better to avoid overpaying for stocks late in the cycle, even if this does not allow us to precisely time the top. Once we move past the peak, it becomes the same as any other cycle,” it concluded. To position for such tech downturns, Morgan Stanley said, go for quality names with strong free cash flows, as well as stocks with cheap valuations. It says companies with revenues that have proven to hold up better during recessions have been more resilient than the tech sector and broader market. “This also includes companies where the demand for their products is steady and repeatable,” the bank said. Stocks Several areas of tech, as well as individual stocks, have a combination of defensive, countercyclical and growth opportunities that should help them outperform in the event of a correction, it added. “Eventually the storm will pass and the traditional leaders – Samsung, TSMC and Apple, for example – will have pole position for the next tech growth curve,” it said. Here are U.S. stocks Morgan Stanley still likes. U.S. semiconductors Nvidia has been the go-to stock to play the AI trade, but Morgan Stanley says it might be too early to call the top. AI spending continues to be resilient, and Nvidia is well positioned to monetize those investments at an even higher rate than before, said the bank. “And should the AI cycle be coming to an end we would favor NVDA over its suppliers and competitors,” it said, explaining that an inventory correction could be more severe for secondary components. U.S. IT hardware Historically, IT hardware upcycles last four to six quarters and the segment strongly outperforms the S & P 500, the bank said. Right now, it’s only two to three quarters into an upcycle, which could last close to six quarters. Stocks that the bank likes in this space are Apple , Seagate and Dell . U.S. networking Morgan Stanley says ethernet names with exposure to AI can outperform, referring to a type of networking standard. Networking in normal tech terms refers to a network of devices that can transmit and share information over physical or wireless communications. In AI, however, the requirements are higher because of large language models and other AI applications that require very high bandwidth and low latency. It sees Arista Networks as best positioned to capture this opportunity, calling it the “highest performance networking name.” Here are some global names Morgan Stanley listed as its “fundamental top picks.” Samsung Electronics : The “quality rotation” that occurs during downcycles will begin to favor Samsung’s more defensive properties, such as earnings stability, rather than growth, said the bank. TSMC : Morgan Stanley likes TSMC’s “quality and defensive” nature during a downcycle for semiconductors. Quanta : The bank says the laptop maker has a strong balance sheet and generates strong cash flow, with growing exposure to data center AI infrastructure. — CNBC’s Michael Bloom contributed to this report.