Intel wanted to Acquire Nvidia years ago for $20 billion

Former Intel boss Paul Otellini apparently had a big idea in 2005. According to a report in the New York Times, he wanted to make a takeover bid for graphics specialist Nvidia and pay up to $20 billion.

Ex-Intel boss saw opportunity, board rejected

As the New York Times (NYT) reports, citing sources close to Intel who are said to be familiar with the discussions in Intel’s top management at the time, it was assumed at the time that Nvidia chips would one day play a major role Data centers could play.

Although this has now become true as a result of the boom in artificial intelligence that has been going on for some time, the board was against Otellini’s proposal almost 20 years ago. At the time, it would have been the most expensive takeover in Intel’s history, especially since there were various concerns about the possible integration of Nvidia into the existing Intel company.

Larabee flop and more missed opportunities

So instead of investing up to $20 billion in the takeover of Nvidia, as Otellini had planned – and thus, from today’s perspective, catching a cash cow – Intel decided to invest in the development of its own graphics solutions.

The result was “Larabee”, a kind of hybrid of CPU and GPU that used Intel x86 technology. As is well known, Intel later discontinued the project, but today it is again offering its own graphics solutions with the Xe and Arc GPUs. Today, Nvidia is worth significantly more than Intel, whose market value is less than $100 billion. Although Intel is now integrating its own Neural Processing Units (NPU) into its platforms to carry out AI tasks, which are certainly impressive in terms of performance, they are clearly losing out compared to the No. 1 AI chip supplier, Nvidia pulled.

Nvidia isn’t the only case where Intel appears to have missed an opportunity to benefit greatly from the emerging AI boom. You already had the opportunity to invest in OpenAI in 2017 and 2018. But there was no deal with OpenAI either because the then Intel boss Bob Swan rejected the project. He simply didn’t believe that AI models could soon become a large market.