It's a no-brainer these days to say AI is driving tech mergers. As I wrote earlier this month, big AI deals and a deal-friendly presidential administration have pushed the value of U.S. announced tech mergers to more than $543 billion this year, the highest level since 2021. 

To Morgan Stanley bankers, the deals this year and over the next several years will be driven by one fundamental thesis: Companies are building or rebuilding the underlying technology infrastructure—software, databases, cloud services and chips—for the era of AI.

The main areas of dealmaking are data companies and semiconductor makers. A string of data-related deals happened in the past three months. Morgan Stanley advised companies on three of them: Confluent on its $11 billion proposed sale to IBM, the merger of data startups Fivetran and dbt Labs, and data management company Veeam Software's $1.73 billion purchase of Securiti. 

Companies are focusing on using and managing data to get an advantage in the AI race. "Every single one of the players in and around the data ecosystem is trying to establish their vision for what the underlying data architecture needs are for AI," said Rohan Mehra, managing director and global co-head of AI investment banking at Morgan Stanley.

Take IBM's bid for Confluent, which was founded in 2014 and went public in 2021. Confluent, a leader in the niche category of data streaming, sells software that analyzes data in real time. IBM is betting that future AI systems will depend on constantly flowing data.

Other companies certainly agreed with IBM. At least one other public company and a private equity firm have shown an interest in buying Confluent, according to people with knowledge of the discussions. Bankers said there's often a domino effect across industries in mergers and acquisitions, where one deal pushes other companies to do similar transactions. 

AI is also driving big deals in semiconductors, according to Mark Edelstone, chair of global semiconductor investment banking at Morgan Stanley, who recently advised Celestial AI on its sale to chipmaker Marvell Technology for up to $5.5 billion.

Edelstone, who covered Nvidia and other semiconductor companies, first as a research analyst in the 1980s and then as a banker, has seen waves of consolidation. He said the industry is gearing up for a transition from AI development focusing on training large language models to running them. 

That shift requires chips to be more energy efficient and to have less delay in transferring data between processors and memory chips, he said. 

Marvell aimed to address the speed issue with its acquisition of Celestial AI. Celestial's technology uses light to move data, which is faster than moving it as electrical signals on copper wires. That could potentially boost Marvell's data center offering. Other companies, including d-Matrix and Mythic, are racing to design chips to shorten the distance between memory storage and processing. 

The boom in AI kicked off by ChatGPT is upending the old order in tech, and the reordering will take some time. "That's kind of led to the haves and have-nots across the tech sector and the broader economy, and we're still early days," Edelstone said.

New From Our Reporters

Exclusive OpenAI Has Discussed Raising Tens of Billions at Valuation Around $750 Billion By Sri Muppidi and Katie Roof

Exclusive OpenAI, Anthropic Discuss Data Deals With Biotech and Other Companies By Valida Pau and Stephanie Palazzolo

Exclusive Banks Pitching Mega IPOs Like SpaceX Seek to Limit Mass Selling By Cory Weinberg and Valida Pau

Exclusive OpenAI in Talks to Raise At Least $10 Billion From Amazon and Use Its AI Chips By Anissa Gardizy, Sri Muppidi, Cory Weinberg and Amir Efrati