Eye On AI: AMD’s Big Deal Is Just Their Latest In AI

This column is a look back at the week that was in AI. Read the previous one here.

While we often focus on the machinations of the likes of Microsoft, Google and Nvidia in the ever-increasing world of AI, other large tech players sometimes escape our thoughts.

Such is the case with AMD — which caught everyone’s attention this week when it announced it would buy server maker ZT Systems in a cash and stock transaction valued at $4.9 billion as it looks to challenge chip giant Nvidia for data center tech in the AI space.

It wasn’t a deal completely out of nowhere — AMD’s market cap is more than $250 billion and it is the second-biggest provider of the graphics processors needed for AI — but it did highlight its ambitions in a space it has been active in as an investor for a while now.

AMD itself has made a quartet of investments — per Crunchbase — in the startup world in just the past year that are heavily AI focused:

The company took part in AI infrastructure startup Fireworks AI’s $52 million Series B just last month, as did others in the AI space such as Nvidia and Databricks Ventures.
Late last year, AMD participated in Essential AI’s $56.5 million in funding. The startup is attempting to use AI to automate monotonous corporate tasks. Nvidia and Google also took part in the round.
AMD co-led a $22 million Series B last October for Moreh, which builds an AI software tool that optimizes and creates AI models.
Finally, last August it took part in the huge $235 million round for Hugging Face at a $4.5 billion valuation. The New York-based startup allows companies to store and use AI software, and hosts hundreds of thousands of open-source AI models that developers can use for AI applications. That mega-sized, Salesforce Ventures-led 1 round included a who’s who of tech, with Qualcomm, Nvidia, Intel, IBM, Google and Amazon all participating.

AMD’s venture arm — AMD Ventures — has been just as busy in the AI space, making a flurry of mostly large deals this year alone, per Crunchbase:

In March, it took part in the massive $175 million Series C for optical interconnectivity startup Celestial AI — whose photonic fabric platform helps separate compute and memory, making processing extensive AI faster and computing more energy-efficient.
In April, AMD Ventures took part in enterprise AI platform Lamini’s seed and Series A. The startup enables enterprises to safely and cost-effectively build their own AI.
It also was part of a massive collective that took part in Scale AI’s huge $1 billion round that valued the data labeling and evaluation startup at a stunning $13.8 billion in May. Other Big Tech-related investors in that round included Nvidia, Meta, Amazon, Cisco Investments, Intel Capital and ServiceNow Ventures.
Finally, just last month it took part in another megadeal — participating in large language model builder Cohere’s $500 million Series D which values the startup at $5.5 billion. The round also included the likes of Cisco, Nvidia and Salesforce Ventures.

While it certainly doesn’t come as a shock that AMD is investing heavily in the generative AI space, those deals — some massive — along with the huge acquisition of ZT Systems clearly show the semiconductor company plans to take a back seat to no one when it comes to AI and, more importantly, the huge amount of infrastructure that is needed to support it.

Although Nvidia has certainly grabbed the majority of headlines when it comes to that in the early going for AI dominance, it certainly has its share of deep-pocketed competitors who show no signs of leaving the AI race.

Things that caught our eye and other stuff:

Defense tech has been in the news a lot recently (like here and here). AI certainly has played a role in why and this week we got another example. Defcon AI, which developed an AI-powered decision tool for military logistics planning, raised a $44 million seed round led by Bessemer Venture Partners. The Air Force’s Air Mobility Command already has deployed the platform. The company claims the Air Force can now build airlift networks for moving personnel and equipment around the world during crises in less than 10 minutes.