Skip NavigationMarketsBusinessInvestingTechPolitics & PolicyVideoWatchlistInvesting ClubPRO
LivestreamMenuTechnology analysts and traders are not buying the flood of warnings about the dangers of artificial intelligence and the need to slow down development that have recently come from frontier model platforms. Instead, they see political motivation behind the doomspeak, and a messaging playbook in which safety hype spurs government regulation that doubles as a competitive moat. “Mostly what’s going on is that [the frontier models] are pulling up the ladder. They have a big lead, and they don’t want anybody to catch up to them, so they’re trying to scare us and our politicians into creating so much regulation that would slow down their competition,” Gil Luria, head of technology research at DA Davidson, told CNBC Monday. Investor Michael Burry, famous for shorting the U.S. housing market before the 2008 financial crisis, also dismissed the warnings, pointing to competitive threats to frontier models Anthropic and OpenAI, both of which are expected to go public before the end of next year. “IPOs need hype & puffery; ‘We are so awesome it could become dangerous’ is hype & puffery,” he wrote in a Monday social media post . “Competition is coming up fast” and slowing development “benefits incumbents.” Traders at investment bank Jefferies said Monday that frontier model makers have been “leaning into this increased regulatory regime around AI for quite some time,” leading to “swift agreement” among senior leaders. The warnings from Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman and an Anthropic researcher who quit his job last week have contributed to a pullback in tech stocks, with the Nasdaq-100 at one point tumbling as much as 1.7% early Monday. Shrugging off red flags But traders are still keen on the sector and see the latest share price decline as an opportunity. “We are still bullish on tech and like the AI theme, semis and Mag7 as the way to express this view. Given today’s price drop on the ‘AI Pullback’ story, we would be dip buyers as we think it is unlikely to materialize into revenue declines [or] margin compression,” the JPMorgan trading desk wrote Monday. Tech analysts at Bank of America called recent appeals for restraint in AI development “noise,” in the context of still accelerating capital expenditure. “Washington [is] unlikely to handicap domestic champions, even as policymakers weigh guardrails,” Vivek Arya at Bank of America wrote to clients Monday. He expects the industry will eventually regulate itself. Arya recommended chipmakers and networking equipment manufacturers as ways to play the trend, though he cautioned it could remain volatile through the November midterm elections. “We expect Compute ( NVDA , AMD ), Networking ( MRVL ), and Analog ( ADI , ON ) to demonstrate greater resilience,” Arya said. “Memory ( MU ) and semicaps ( LRCX ) could rebound strongly as momentum resumes.” Cybersecurity stocks gained thanks to the safety warnings emanating from frontier large language model (LLM) developers. First Trust Nasdaq Cybersecurity ETF (CIBR) added nearly 6% in early trading Monday, Amplify Cybersecurity ETF (HACK) climbed more than 7% and Global X Cybersecurity ETF (BUG) soared 10%. Cybersecurity providers Palo Alto Networks jumped 13% and Zscaler soared 15% Monday. PANW 1D mountain PANW 1 day Frontier LLMs are at the forefront of an increasingly competitive industry as algorithms with publicly available parameters and source code tempt companies away from proprietary software. Open models from both Chinese and U.S. companies can perform nearly as well as the frontier models, often at far less cost. The skepticism on Wall Street Monday toward safety warnings was echoed by the Chinese Ministry of Foreign Affairs, which described the appeals for a slowdown as a form of “fear mongering.” “Fear-mongering, confrontation, competition will just disrupt [the] process of global AI governance,” Guo Jiakun, a spokesperson for China’s Foreign Ministry, said Monday via an English translation published by Reuters.Read More














