Morgan Stanley offered some of its best ways to play with artificial intelligence as the technology enters a “golden age.” Enthusiasm for AI contributed a large part of the market’s gains in the first half, with giants like Nvidia and Meta Platforms more than doubling. But even after those significant increases, Morgan Stanley believes the tailwinds are far from over. “We expect the benefits of AI to come in stages over the next decade: hardware and semiconductors have substantial AI exposure today, but leadership will change once the diffusion in infrastructure/devices changes that value with the time,” the bank said, expecting a third wave driven by companies creating “usable” applications to match the infrastructure. Given this setup, Morgan Stanley analysts singled out 37 stocks as global AI winners with a combined market capitalization of $2.8 trillion. By 2027, the bank forecasts that the AI ​​technology industry will roughly triple its total addressable market to $275 billion, driven primarily by semiconductor names expected to hit a $500 billion market. More broadly, AI should become a roughly $3 trillion industry by the end of the decade. Of the bank’s picks, 50% are based in the US, while 21%, 14%, and 10% operate in Taiwan, Korea, and Europe, respectively. To compile the list, Morgan Stanley looked for areas of the technology supply chain situated for the strongest AI revenue exposure. They also looked at earnings quality, future growth and upside potential. “We encourage investors to ‘open the net’ by positioning themselves for AI and considering factors such as growth, quality, regional trends, and specific technology niches, respectively,” the team wrote. A handful of popular semiconductor stocks made the cut, given their integral role in creating graphics processing units and the power of many large language models. Many of these names play an integral role in the early stage of AI investment, the firm said. Some of the best US-based options include Nvidia and Advanced Micro Devices. Nvidia’s stock has nearly tripled this year as Wall Street labels the stock a clear leader in artificial intelligence. Shares jumped 24% one day in May after the company shared guidance for success, fueled by strong demand for AI. Morgan Stanley expects the company’s dominance to continue for the foreseeable future, calling it the “cleanest story” in AI hardware. This deserves investors’ attention even after the rise in shares and the increase in valuation in the first half. NVDA YTD mountains Nvidia shares in 2023. Despite Nvidia’s dominance, strong execution, and ability to aggressively “defend its turf,” Morgan Stanley also remains bullish on Advanced Micro Devices, highlighting its low-cost inference segment that it depends on the launch of its MI300 product later this year. . “AMD’s involvement in that opportunity looks more tangible and enduring than our initial expectations, with strong revenue forecasts from two of the four US hyperscalers, with prospects for a third, at drive levels that materially surprised us.” said the bank. Other areas poised to benefit include custom chip design providers. Morgan Stanley sees Marvell Technology as one of the top-ranked names, with shares up nearly 63% this year and poised for a 15% gain from Friday’s close. MRVL YTD mountain Marvell Technology shares this year Taiwan Semiconductor Manufacturing remains another dominant leader in AI, with US-listed shares up more than 30% this year. Morgan Stanley forecasts that 13% of revenue by 2027 could come from AI chip manufacturing. “We see TSMC as a key enabler of future AI semichips given its technology leadership, which should enable it to capture >90% market share in AI semichip manufacturing in the next 5 years, we estimate,” the firm wrote. . Many of the top picks within the second and third wave of AI advancement operate predominantly outside of the US Some secondary beneficiaries highlighted in the report include ASML, poised to benefit from growing demand for cutting-edge chips, and Dell Technologies. . — CNBC’s Michael Bloom contributed reporting.