Nvidia shares can extend their 2023 rally once the chip maker’s growing artificial intelligence business is fully appreciated by Wall Street, according to HSBC. Frank Lee, the firm’s head of technology research in Asia, double-upgraded the chip maker’s stock to buy from reduce. Lee also more than doubled his price target to $355 from $175. His new price target implies an upside of 31.5% over Monday’s close. “We believe there’s more earnings upside vs market expectations in FY24e and beyond,” he said in a note to clients Tuesday. “Overall we believe we were too cautious on Nvidia.” Shares rose 2.2% in the premarket. The stock has surged 85% this year, outperforming the technology heavy Nasdaq Composite ‘s 16.2% advance. NVDA .IXIC,.SPX YTD mountain Nvidia That rally has made some previously question whether the stock has further upside. But the analyst said there’s much more to go. “Nvidia’s incredible AI pricing power (is) not fully priced in,” Lee said. The analyst noted that artificial intelligence will provide a “significant boost” prices as AI chips are between 10 and 20 times more expensive than the average price of a typical graphics processing unit for gaming. Because these AI chips are so much more expensive, Lee said the company does not need to see much of a volume increase to meet revenue growth expectations. A potential downside risk to earnings in 2024 from a slowdown in datacenter momentum drove earlier pessimism, according to Lee, given that 60% of the company’s sales comes from this business. The company has seen downside to server growth expectations, while he said investors now debate whether there could be future revisions to the business outlook going into the rest of 2023. Lee also noted concern about rising inventory. Slowing demand growth from wafer orders related to semiconductor maker TMSC was also an earlier cause for concern, Lee said. But he said the changes between company revenue and wafer orders should instead be taken as a sign of the company’s business model changing to meet the moment within AI. He said the AI business should be able to more than offset weakening datacenter demand and Wall Street was “too focused” on the later before. Lee also pointed to the fact that the majority of the inventory buildout is related to AI chips, which should help mitigate any issues related to oversupply. Lee noted there has not been any significant upward revisions to consensus revenue or earnings expectations over the past three months. But he raised his earnings expectations in the 2024 and 2025 fiscal years. “We acknowledge we underestimated the potential of Nvidia’s AI business and speed of ramp-up post Microsoft ‘s ChatGPT launch,” he said. “We believe the motivation to adopt more generative AI, especially by both China and US cloud service providers, has exceeded what we initially thought.” — CNBC’s Michael Bloom contributed to this report.