UBS analysts believe the artificial intelligence investment cycle is still in its early stages, with the next wave of growth likely driven by enterprise adoption rather than just infrastructure buildout. In a note to clients, the firm highlighted 14 "highest-conviction calls" across technology, media, and telecommunications (TMT), positioning investors for what it sees as a broadening AI opportunity.
The analysts noted that only 8% of companies surveyed have deployed agentic AI at scale, suggesting significant room for growth. As businesses move from experimentation to deployment, spending is expected to shift toward cloud services, networking, data management, and enterprise software. This transition, they argue, will create a second phase of the AI investment cycle.
Amazon leads large-cap internet picks
Among large-cap internet companies, Amazon is UBS's top pick. The firm's estimates for Amazon Web Services (AWS) revenue exceed consensus, and it believes Wall Street will catch up as major customers adopt Amazon's Trainium 2 and Trainium 3 chips from late this year through 2027. UBS set a $318 price target on Amazon shares, implying roughly 25% upside from current levels.
The analysts also noted that open-source AI models are gaining traction due to cost advantages, but enterprises often deploy them through major cloud platforms like AWS and Microsoft Azure. This dynamic supports continued spending by the largest cloud providers as demand for computing capacity grows.
Texas Instruments and data-center beneficiaries
In the semiconductor space, UBS selected Texas Instruments as its top pick. The analog chip cycle, driven by automotive and industrial demand, is beginning to recover, and the company is gaining market share. UBS expects data-center growth to add further tailwinds, with total revenue potentially rising 20% even if data-center growth decelerates. The firm set a $380 price target, representing 43% upside.
Celestica and TD Synnex were highlighted as key beneficiaries of data-center infrastructure spending. UBS expects OpenAI's custom chip rollout next year to support at least 70% revenue growth for Celestica, which already supplies Ethernet switching technology to hyperscalers like Google and Meta. The firm set a $430 price target for Celestica, implying 29% upside. For TD Synnex, whose Hyve Solutions unit manufactures custom hardware, UBS sees a $352 target, or 32% upside, as data-center spending grows at least 30% annually over the next three years.
Palantir at the intersection of AI and data
Palantir Technologies is another major conviction call, described by UBS as "at the nexus of the two most powerful spending trends in software - AI and data." The firm expects revenue to increase 88% and margins to reach the low-60% range, citing strong demand for its AI and automation software and limited competition. UBS set a $250 price target, compared to a recent market price of $177.64.
Other top picks include AT&T, EchoStar, Pinterest, Spotify, Mastercard, Global-e Online, Procore Technologies, and JFrog. The breadth of the list reflects UBS's view that AI spending is extending beyond chipmakers and model developers to a wider range of technology services.
As enterprises deploy AI more widely, UBS expects spending to spread across cloud infrastructure, networking, data management, and software. This could mark a shift from the massive infrastructure buildout that has driven the sector so far, opening opportunities for investors in companies that enable enterprise AI adoption.
This article is for informational purposes only and does not constitute financial advice.
