AI safety is in the spotlight with increasing calls from the industry and beyond for paced development. However, private banks believe the investment story remains intact, especially with regard to the infrastructure buildout.
There have been increasing calls for a slowdown in artificial intelligence (AI) development across the board, including from industry leaders, with a focus on ensuring that safety standards are met. Anthropic CEO Dario Amodei led the pack with a published essay on September 12 proposing three steps, including assessment by independent evaluators, industry-wide coordination, and global coordination.
What does this mean for markets and investors?
No expected effect on capex and model training
“Pacing does not necessarily imply lower capex,” said UBS in an investment note, retaining its 2027 AI industry capex forecast at USD 1.2 trillion, up 33% from its estimate of USD 900 billion earlier this year. “The safety proposal also explicitly distinguishes pacing from halting model training, and xAI’s Elon Musk confirmed on Sunday that further training was under way for its leading-edge Grok 4.8 model. OpenAI also recently reiterated its expectations for continued growth in memory demand.”
“Taking a step back, most importantly, pacing the frontier does not mean halting model training or AI progress,” added Enrico Chinello, Next Generation Research, Julius Baer, in a commentary. “For model development, the practical effect is that evaluation moves from a release checkbox to a gating step, with potentially longer pre-release windows, more capability withholding, and competitive differentiation shifting from raw frontier capability towards cost, reliability, and safety standards.”
Continued optimism in infrastructure outlook
On the investment outlook, private banks remain positive on AI, especially regarding infrastructure.
UBS continues to favor a diversified approach across the AI value chain, combining infrastructure beneficiaries, including semiconductors, networking, power, and cloud, with larger platforms and software companies positioned to monetise adoption. Julius Baer is also still constructive on cloud computing and AI, driven by strong demand for AI solutions and an accelerating buildout of AI infrastructure.
“The stated intention of slowing down LLM development between fierce competitors (and not too much, to avoid Chinese models catching up) should not lead to a major decline in investor’s demand assessment for chips and data centres and key components. The shortage of supply and bottle neck in these areas will not disappear. As a result, any dip related to the announcement should be shallow and temporary,” said Willem Sels, global CIO, HSBC Private Bank and Premier Wealth, highlighting the bank’s continued preference for chips, the cloud and AI-enabled applications.
