Microsoft’s Inflection AI transaction from an Allocator’s Perspective

On March 21st, The Information reported that Microsoft hired two of Inflection AI’s co-founders, most of its staff, and entered into a licensing deal that in effect made venture investors in Inflection AI whole on their investment.

As per the report, Microsoft entered into an unconventional arrangement, presumably to avoid regulatory review, where they hired Mustafa Suleyman, Inflection’s co-founder and CEO and most of the Inflection staff as employees. Since this would destroy value for Inflection’s investors, Microsoft appeared to sweeten the deal for early venture investors in Inflection AI, including Greylock (a venture firm run by Reid Hoffman who is also a Microsoft board member), Bill Gates, Eric Schmidt, Nvidia and Microsoft as well.

The payout from the licensing arrangement would be about 1.5x for early investors and about 1.1x for later investors. Decent, but hardly the kind of expected returns from venture investments, which seek power law type exits.

I’ve been thinking about this from the perspective of an allocator, and the challenges of placing overlapping exposures into separate asset allocation buckets. In this case, let's say a pension fund hypothetically invested in Greylock through its Venture Capital bucket while it owns Microsoft in its Public Equity bucket. In both buckets the allocator is exposed to the same factors, namely the ‘innovation’ and ‘early stage’ economics of Inflection AI. Yet the returns differ greatly, and one could speculate that in this case the return to Microsoft may be greater than the 1.5x less fees that an LP in Greylock would earn. In other deals it might be the other way.

So where am I going with this? I believe that the neat buckets that we like putting investments into for asset allocation purposes may not be as neat as the narratives and assumptions that drive those buckets. For example, the narrative that innovation and private opportunities can be accessed through venture may not be strictly accurate since you can access those same opportunities through your equity investment in a Microsoft or Google or NVDA for example (not investment advice). Furthermore, as public and private companies begin to operate and transact in similar realms (e.g. data centers can be private or public, a building can be owned by a public REIT or a private REIT), the distinction between the underlying economics of public and private begin to blur.

The broad factor bets in public and private markets are similar, and a private allocation and public allocation access similar factors. 

Therefore, the bet that a long-term investor makes in private markets is more nuanced. First, it's a bet on the management team - in the example above, the indirect bet is that the management team at Greylock is more capable at identifying opportunities than Satya Nadella. The second is a concentration bet - the sector/opportunity presented by your venture portfolio may be a different basket of opportunities and weights than your public portfolio.

For such bets to be effective, an allocator needs to break the buckets that have kept private and public investments separate. Instead, the private opportunities that an investor selects to invest in need to be evaluated in terms of the value they bring over and above their equivalent public market factors.