AI In 2025: Regulations And How To Buy The Dip — UBS Lists Five Key Factors

Artificial intelligence, which emerged as a key theme in powering the US stock market's performance in 2024, is likely to be a sector to watch out for in 2025, according to UBS Financial Services Inc.

"Overall, with a strong near-term visibility, we remain bullish on the AI theme by maintaining our positive view on AI semiconductors and leading cloud platforms," it said in a note dated Jan. 2.

The observations come in the backdrop of AI semiconductor suppliers like Nvidia Corp. and Taiwan Semiconductor Manufacturing Co. surging 190% and 88% on the bourses respectively over the last 12 months.

UBS suggested that investors could take advantage of any spike in volatility through structured strategies or by "buying the dip" in quality AI stocks. Those investing in AI stocks need to be "more nimble in 2025" amid rising regulations, potential product transitions and tariff-related uncertainties, it added.

The note, which has been authored by analysts of UBS AG Singapore Branch, UBS Switzerland AG and UBS Financial Services, explored five key factors that could affect the AI stocks in 2025.

The capital expenditure towards AI integration by the Big 4 — Alphabet Inc., Microsoft Corp., Amazon.com Inc. and Meta Platforms Inc. — was a key factor in fuelling the rally of AI-themed stocks in 2024. "We raise our industry forecast for AI capex by increasing our estimate for the Big 4's spending in the new year by 5%, thanks to greater visibility into their investment plans," UBS said.

Their combined capex is seen rising from $224 billion in 2024 to $280 billion in 2025, marking a 25% increase. "Upward capex revisions is an additional boost for AI’s investment case. We believe this is positive for AI semiconductors in the near term," it added.

UBS believes that AI monetization "should indeed improve sharply in 2025", even as it pointed towards the ongoing debate in the industry whether AI investments could be significantly monetised at this stage.

UBS pointed towards the acceleration in cloud growth, which reflected in the 25% year-on-year jump logged by Amazon Web Services, Microsoft Azure and Google Cloud Platform in the third quarter of 2024. "US Census Bureau data shows that AI adoption is picking up across industries and is set to broaden in 2025, paving the way for increased monetisation," it said.

While AI revenue will continue to lag behind capex in 2025, the gap "should increasingly narrow over the years ahead". This should "prove positive for software and internet" over the medium term, the note added.

There are both positive and negative drivers that can change valuations of AI stocks in the period ahead. However, the "strong" underlying earnings-per-share growth should be enough to support solid near-term share price performance, according to UBS.

Contrary to the general perception that AI valuations are stretched, most of the rally in AI-related stocks during the last two years was supported by solid earnings growth and less driven by an expansion in price-to-earnings multiples, the brokerage added.

UBS is of the view that investors should remain prepared for "volatility spikes" in AI stocks in this year, adding that this period of volatilities would offer buying opportunities.

"For long-term growth investors, we would refrain from excessive trading because long-term compounding returns are generated by structurally positioning in themes like AI," the note stated. "But investors can take advantage of heightened volatility through structured strategies and buying the dip in quality AI stocks, as we believe focus will eventually return to robust fundamentals."

AI regulations have caught up in 2024, and more rules are set to be introduced in 2025. Regulations have always been, and will always be, a "risk" for the tech sector, more so for AI, according to UBS.

The regulations will not hit the medium-term trajectory of AI stocks, but lead to a temporary decline that can provide opportunities for investors to buy at a dip, it said.

"Be prepared for more AI regulations in 2025, but because excess regulations can stifle innovation, we expect more balanced measures," the note said. "Large corrections due to regulations and geopolitics would be a buying opportunity, in our view."