Unlock Editor’s Digest for free
FT editor Roula Khalaf picks her favorite stories in this week’s newsletter.
AMD is to buy Finnish artificial intelligence start-up Silo AI for $665 million in one of the biggest such takeovers in Europe, as the US chipmaker looks to expand its AI services to compete with market leader Nvidia.
California-based AMD said Silo’s 300-person team will use its software tools to build its own large language models (LLM), a type of artificial intelligence technology that underpins chatbots such as OpenAI’s ChatGPT and Google’s Gemini. The all-cash acquisition is expected to close in the second half of this year, subject to regulatory approval.
“This agreement helps us accelerate our customers’ engagement and deployment, while also helping us accelerate our own AI technology stack,” Vamsi Boppana, AMD’s group senior vice president for artificial intelligence, told the Financial Times.
According to data from Dealroom, the acquisition is the largest privately owned AI start-up in Europe since Google bought UK-based DeepMind for around £400m in 2014.
The deal comes as buyouts by Silicon Valley companies have come under tighter scrutiny from regulators in Brussels and the United Kingdom. European AI start-ups including Mistral, DeepL and Helsing have raised hundreds of millions of dollars this year as investors look for a local champion to compete with the likes of US-based OpenAI and Anthropic.
Among the largest private AI labs in Europe, Silo AI, based in Helsinki, offers customized AI models and platforms to enterprise customers. The Finnish company launched an initiative last year to build an LLM in European languages ​​including Swedish, Icelandic and Danish.
AMD’s AI technology competes with Nvidia’s, which has taken the lion’s share of the high-performance chip market. Nvidia’s success pushed its valuation past $3 billion this year as tech companies push to build the computing infrastructure needed to power the biggest AI models. AMD began introducing its MI300 chips late last year in direct competition with Nvidia’s “Hopper” chip series.
Peter Sarlin, co-founder and CEO of Silo AI, called the acquisition “a logical next step” as the Finnish group seeks to become the “flagship” AI company.
Silo AI is committed to “open source” artificial intelligence models that are freely available and customizable by anyone. This sets it apart from OpenAI and Google, which prefer their own proprietary or “closed” models.
The start-up previously described its family of open models, called “Poro”, as an important step towards “strengthening European digital sovereignty” and democratizing access to LLMs.
Meanwhile, the concentration of the most powerful LLMs in the hands of a few US-based Big Tech companies is attracting the attention of antitrust regulators in Washington and Brussels.
The Silo deal shows that AMD is looking to rapidly expand its business and increase customer engagement with its own offering. AMD sees Silo, which creates custom models for clients, as the link between its “core” AI software and real-world applications of the technology.
Software has become the new battleground for semiconductor companies as they seek to tie customers to their hardware and generate more predictable revenue outside of the boom-and-bust chip sales cycle.
Nvidia’s success in the AI ​​market stems from its multibillion-dollar investment in Cuda, its proprietary software that enables chips originally designed for processing computer graphics and video games to run a wider range of applications.
Since Cuda development began in 2006, Nvidia has expanded its software platform to include a variety of applications and services, largely aimed at enterprise customers who lack the in-house resources and skills that large technology companies have to build on their technology.
Nvidia now offers more than 600 “pre-trained” models, meaning they are easier for customers to deploy. The Santa Clara, Calif.-based group last month began rolling out a “microservices” platform called NIM that promises developers to quickly build AI chatbots and “co-pilot” services.
Historically, Nvidia has offered its software for free to buyers of its chips, but this year it said it plans to charge for products like NIM.
AMD is among several companies contributing to the development of an OpenAI-led rival to Cuda, called Triton, which would allow AI developers to more easily switch between chip providers. Meta, Microsoft and Intel also worked on Triton.