Even with industry consolidation, the robustness of the ecosystem is worth more than any single company.
The massive consolidation that has been underway over the past couple years is about to slow down. Interest rates are expected to increase in the very near future—the Fed has been dropping hints about this for months—ending the era of cheap capital. The cost of borrowing already is creeping up in anticipation of this, and it’s happening globally because money lending is a global industry.
That doesn’t mean consolidation will end entirely, but it will be much harder to justify to shareholders why a company should be bought if it’s just to reconfigure a pool of assets. The price will be higher, and the ability to spin off unwanted assets will be much tougher. Companies will be forced to work with what they have, and reach outside for partners who can fill in the gaps for what they don’t have.
This may look like a change in direction, but it really isn’t. The string of acquisitions that has occurred to date was the result of two very distinct trends. One is that Moore’s Law will likely end over the next node or two. Building chips beyond 5nm (the real 5nm node, that is) will be possible, but it probably won’t be profitable without massive volume, new lithography, 450mm wafers and an investment of tens of billions of dollars. And even then, there will probably not be enough companies moving to those nodes to ever pay back the investment. Physics has its limits, and companies are repositioning themselves in the face of these limits.
The second trend is that the number of companies that can develop a chip that will sell billions of units will continue to shrink. As smartphone sales level off, and the focus is on even less expensive models with similar features, even the juggernauts of consumer electronics will begin having earnings blips.
Still, those trends are coincidental to the bigger shift that has been underway for some time. The recent spate of aggregation won’t stop the bigger trend toward disaggregation because specialization is always more cost effective in growing markets. It drives innovation and competitive cost structures. And only part of the semiconductor industry is consolidating. The IoE opens new doors everywhere, including more sophisticated chips in some markets that in aggregate are expected to dwarf the number of complex SoCs sold in the past.
The IoE and connected devices are the next big market. However, the economics are different. No one company can do it all, and the economies of scale are achieved only with multiple companies working together rather than one company trying to do everything. Put simply, the market battles of the future will be fought by tightly knit ecosystems, not giant companies. And while there may be big companies anchoring those ecosystems, it’s also possible to build ecosystems with much smaller companies based upon well-defined processes for integrating IP and industry-defined rules for putting those pieces together.
This plays well for EDA and IP companies, at least in the long run. The transition may be a bit bumpy, as are all changes. But once that transition is over, tools will still be required to make sure all of this stuff works, and that it works well enough over time. And IP will become even more valuable because the emphasis will be on time to market, as well as maximum flexibility with as many options as possible to tweak power, performance and area for a particular need.
ARM has been pitching an ecosystem approach for years, and it has amassed a huge ring of partners across a broad swath of existing and emerging markets. Imagination Technologies has been working on a similar strategy. Even Intel, which has staunchly maintained its IDM model, is beginning to show off to potential customers its business arrangements with makers of boards and other components that can integrate its menu of chips. And Synopsys, Cadence, Mentor Graphics, Rambus, Arteris and Sonics, all of which develop their own IP, have been carving places for themselves horizontally across all of these ecosystems.
There has been a lot of talk about consolidation, but it’s ecosystems that will fight the next battles. While that reality may never bubble to the surface in consumer advertising, it certainly will determine which products ultimately win, which tools and standards are used to build them, which process is used to manufacture them, and what the semiconductor industry will look like over the next decade.