With their goal of 50/50 handset/non-handset revenue split by 2030, and their recent acquisitions pointing in the same direction, it stands to reason that they will do a lot of high capex investments into things like chiplet/chiplet communication for datacenters, automation/automotive, as well as edge AI. We can also observe they're baking in a lot of fpga-style configurability into a lot of these product lines - the connectivity fabric they acquired along with alphawave semi, their hexagon dsp, nuvia(oryon which they won the legal case for recently), etc,. which is another hint for the type of market they're targeting.

My opinion is that they should productize ESP [1] (no, not that one) which will be super harmonious with their goals.

Arduino acquisition, IMO, is putting one foot into manufacturing automation/automotive/sensors field. They have done similar in the past, arriver was an ADAS compute thing.

Personally I don't believe they will take the execution risk and scale up on all of these things. They will probably wait for the right time and chop off a few of these things and focus on whatever looks like it's going to be a cash cow.

Finance wise, there will be near term margin pressure but long term (IMO) they will execute superbly on a portion of their bets.

The main problem is the clock is ticking, handsets becoming commodified leading to vertical integration, licensing losing value, etc. Apple modem agreement running out soon too, and 6G modems too will not be as high margin due to diminishing improvements in telecom tech, even operator uptake at this point is looking unlikely after the 5G... debacle.

Which explains the very diverse bets they have made.

Will be interesting to see what they execute in this limited timeframe.

[1] https://www.esp.cs.columbia.edu/