AMD Targets Europe, Acquires Silo AI for $665 Million

Finnish Silo AI is being acquired by AMD for $665 million. As the “largest private AI lab” in Europe, it already provides services to companies like Rolls Royce and Philips. But the acquisition is primarily AMD’s move to enter the AI ​​battle with Nvidia, for which it has increasingly good credentials.

Silo AI has a multifaceted offering and a diverse customer base. Use cases include AI helping diagnose cancer in Philips medical devices, and driving aids in Rolls Royce, Honda and Mitsubishi cars. Co-founder and CEO Peter Sarlin sees a wide range of applications in many other sectors, from retail to finance.

As a European player, it is trying to make AI a key part of businesses – the European way. That is: deploying AI infrastructure on the continent, with support for European languages ​​and an open-source base for building value, as Sarlin outlined. This will be music to the ears of many Brussels decision-makers, especially since Europe is so far behind when it comes to AI infrastructure. Enter AMD, which is spending a similar amount on Silo AI to what Google paid for DeepMind in 2014.

AMD takes over European players

AMD announced Wednesday afternoon that it would acquire Silo AI for $665 million, or about €613 million. The move gives the chipmaker an AI startup that not only has a mature customer base but also 300 researchers and other experts, more than half of whom have PhDs in AI. AMD is effectively gaining an additional AI division, which becomes even more obvious when you consider that Peter Sarlin is staying on board. He will report to Vamsi Boppan, AMD’s vice president of AI, once the deal closes later this year.

The Silo AI acquisition isn’t nearly as much as AMD paid for FPGA maker Xilinx in 2022 ($49 billion). Nor is it an expansion of its own portfolio, as was the case with the Radeon division, which was formed by AMD’s $5.6 billion acquisition of ATI in 2006, which made it not just a CPU maker but also a GPU supplier. Still, expanding its own AI offering is a big step.

AMD is lagging behind in the AI ​​battlefield no matter how you look at it. Nvidia, with nearly 98 percent of the data center market share, is the best when it comes to GPUs. While CPUs certainly play a role in AI acceleration, GPUs are now considered the real workhorse that enables GenAI. There would be no ChatGPT, GitHub Copilot, or Google Vertex AI without the fleet of Nvidia chips that are pumping through IT infrastructures around the world. The fact that nearly every IT vendor worth its salt is now an Nvidia partner speaks volumes about the company’s massive impact on the tech landscape. This goes beyond hardware, as we discussed earlier, and cries out for a mature competitor.

Opportunities in Europe

There’s also work to be done in Europe. Mistral AI complained a little over a month ago about a shortage of data center capacity, and with it AI training capacity, on the continent. It’s much more likely that you’ll find an Nvidia chip in America or Asia than in Europe. While there are many promising AI players in the EU, it seems likely that they’ll eventually be forced to move to America or East Asia if things don’t change over time. But there’s an opportunity in Europe to do AI, “EU style.” That requires a plan in advance, even if it may require the help of a North American company like AMD.

Why is this important in light of AMD’s acquisition of Silo AI? Because it reflects AMD’s clear thinking from which real opportunities in Europe will emerge. The whole vision behind the EU AI Act has its detractors, the most important of which is once again Mistral AI, but it is based on sovereignty, security and due diligence. This is in contrast to the “move fast and break things” a credo that has, for now, taken root in the U.S. approach to AI, where concerns about copyright, privacy, and the impact of AI on society are resolved in the courts rather than defined by law.

AMD is steadily building an ecosystem that supports Europe’s AI vision. This makes it attractive for infrastructure projects to bet heavily on AMD GPUs. In terms of performance, the latest MI300 chip offering falls between Nvidia’s older H100 and the upcoming Blackwell series, but is said to be significantly cheaper. Like Intel’s Gaudi 3, it still counts as an AI alternative when no Nvidia offering is available or affordable. That could change if AMD becomes a prominent player focused on Europe. Given Europe’s need to catch up in order to provide sufficient AI capabilities, the continent could find a notable supplier in AMD with its improved AI Silo technology.

Read also: Intel Feels AMD Epyc 5’s Hot Breath with Xeon 6 Chips