Arm & NVIDIA – The unthinkable pt. VI.

The risks are demonstrably manageable.

• Herman Hauser has once again criticised the acquisition of Arm citing clear signs of unfair competition but what I think he has done is merely re-iterate the risks of this acquisition which everyone already knows are there.
• The promotion of the idea of a merger between Graphcore and Arm does not help the credibility of his view in my opinion.
• Herman Hauser is an Austrian entrepreneur who co-founded Acorn Computers in 1978 and from which Arm was spun out in 1990.
• In 1997 he co-founded Amadeus Capital Partners which has become one of the most influential investors in the Cambridge Technology scene.
• One of its highest-profile investments at the moment is Graphcore which designs processors to be used in the creation of AI algorithms.
• It most recently raised $222m in a Series E round with a post-money valuation of $2.77bn.
• This is why I think that publicly promoting the idea of a merger between Arm and Graphcore is not a great idea as this would greatly increase the value of Graphcore creating what I see as a conflict of interest.
• In the stockbroking world, this is referred to as reverse broking and happens when a fund manager tries to convince a broker about the merits of a certain stock because the fund manager already has a position.
• There is no ethical issue in this practice that I am aware of, but it highlights that there are incentives outside of the idea being a great investment.
• What Mr. Hauser is referring to as anti-competitive is NVIDIA’s Arm-based server chip Grace where NVIDIA is proposing its proprietary NVLink interface between Grace and the GPU.
• This move effectively means that NVIDIA will be using its own products for the whole architecture rather than using partners as it has in the past.
• I can’t see how this is anti-competitive as NVIDIA is merely throwing down the gauntlet to its competitors daring them to make a better vertically integrated server product.
• However, what Mr. Hauser’s comments do is remind everyone of the risks of this acquisition which are:
  • First, time: With NVIDIA and Arm under the same roof, there is the risk that NVIDIA has access to Arm’s latest innovations before its competitors thereby giving NVIDIA an advantage.
  • Second, roadmap: This refers to the risk that NVIDIA directs the roadmap of Arm to suit its own strategy and chip designs as opposed to those of its competitors.
• These are not new risks and have been raised time and time again whenever any acquisition of Arm has been discussed.
• I also think that the chances of them occurring should NVIDIA successfully acquire Arm are very small.
• Samsung has successfully managed an almost identical problem within its business units for years with no real issue.
• Samsung’s handset competitors have been purchasing memory, storage, and displays from Samsung for years and would have stopped doing so in a heartbeat had there been even a whiff of unfair treatment.
• Hence, I think that these risks are entirely manageable but are not what is causing the acquisition difficulty at the moment.
• These are the problem with Arm China (see here), Arm customer complaints (see here), and the UK regulator.
• NVIDIA has now applied for regulatory clearance in China in what I think is an indication that it thinks that the issue there will be sorted out within a reasonable time frame.
• I still think that the best place for Arm is as an independent public company but the hurdles for its acquisition by NVIDIA do seem to be lessening.