The SoC market is consolidating fast.
That poses an interesting challenge for the new, larger, companies – especially as they continue to grow AFTER the step-change of a "consolidation event".
They have an expanding portfolio of different architectures and techniques which are used internally. These often come with a lot of legacy; are the result of a lot of investment; which in turn has produced a lot of trusted, proven code which is of great value and is familar to users. But it is no easy matter to simply swap out.
For example, one large SoC company “boasts” that it has built up a range of eight unique DSP architectures from its acquisitions; another has four different host architectures, all being used for new designs in different groups.
That poses two related problems.
First, with every acquisition, the portfolio grows even larger. More architectures, more tools, more legacy software, more complexity in hardware, software and “wetware” (humans).
Second, the consequent desire to rationalize is not for free. How do you plan a progression? What is a smooth trajectory that uses the assets you have just expensively bought, but gets you to a simpler, more supportable future?
The existence of licensed IP certainly simplifies this. The world is consolidating around a number of platforms, and there are fewer processor architectures or in-house CPUs than there might have been ten years ago.
But “simpler” is only relative; this landscape is far from simple.
And that is especially true when it comes to platform IP: interconnect fabrics, development tools and the like clearly support the convergence we see.
Take, for example debug & development support. This has usually been approached on a company-by-company basis, with a hybrid of JTAG (industry standard, universal but limited), vendor-specific tools (eg CoreSight for ARM – more powerful but in a silo) and in-house technology (everything else).
As “in-house” gets complicated and messy, there is a growing necessity for a “neutral platform” that can help debug all the different architectures, work with all the legacy blocks & subsystems. This can then act as a “wrapper" to help integration; it can be used as a bridge to help combine the different subsystems; it provides a route and a roadmap to rationalization in the future. A platform debug solution can combine the existing blocks, with all legacy and code, into a single coherent fabric, and help debug and development.
Then, over time, the same debug fabric stays as a a wrapper or bridge, ensuring trust and a “golden reference” as blocks get upgraded and swapped for a new, more unified portfolio.
There will be similarities in other segments, where licensed IP delivers an immediate benefit, and can also be part of a strategy.
The semiconductor world is consolidating. IP can be an important part of this, and both acquirer and acquiree should plan their IP roadmap accordingly, as part of their M&A integration plans.