Should Imagination Technologies buy Altair Semiconductor to bring 4G LTE IP inhouse?

Imagination Technologies, whose acquisition of MIPS was completed in February 2013, announced today that CPU cores from the company would be used in Altair’s two new FourGee-38xx baseband processors.

The press release doesn’t say which cores were used but it is likely that MIPS32 1004K was used rather than the Aptiv range (the statement mentions multi-threading, multi-core, hence our deduction). The two companies have been collaborating for a while; Chinese SoC maker Ingenic showcased a 4G/LTE tablet that used both Altair and MIPs early 2012.

Now, it doesn’t take require a rocket scientist to foresee the advantages Imagination Technologies could derive from acquiring Altair Semiconductor. The Israeli company has an impressive portfolio of baseband processors and uses proprietary Software-Defined Radio technology to keep its hardware updated.

Altair is somewhat similar to Icera, a soft modem chipset maker that Nvidia acquired in May 2011 for $367 million (and had already raised $258 million over nearly a decade) and in an environment where independent pure-play baseband companies are becoming rarer, that could give rise to a ferocious bidding war (Sequans, being the other big LTE player in the market).

Because there’s a trend that appearing in the industry: Qualcomm, Samsung Semi, Intel, Nvidia, ST/Ericsson (even if they split up), Renesas, Huawei, Broadcom, Mediatek and Texas Instruments all have ARM-based application processors and LTE baseband modems and a lot of them have integrated SoC.

Imagination Technologies doesn’t and it is going to be one heck of a weakness going forward as it tries to single-handedly push MIPS as a viable alternative to ARM’s quasi-hegemony.

A baseband acquisition will provide Imgtec with a more complete portfolio than ARM, who cannot possibly follow the same route without potentially damaging relations with some big partners. And some people within the industry have clearly told us that ARM is rapidly becoming big enough to be considered a quasi-monopoly and they would hate to see the Cambridge-based company get the same clout as Intel did in x86.

Altair is smaller than Icera and has raised “only” $94 million in five financing rounds. We suspect that Altair wouldn’t sell for less than $150 million (around £100 million) and that is, of course, pure speculation. MIPS, in comparison, sold to Imagination Technologies for $100 million. So should Altair Semiconductor sell out to Imagination Technologies or are other alternatives possible? Suggestions to be sent on a postcard.