An intriguing report hit our inboxes earlier this morning, one which mentions Amazon as being a potential suitor for Texas Instruments’ OMAP system-on-chip business, a deal that could potentially be worth “billions of dollars” according to the source, Assaf Gilad from Israeli financial website, Calcalist.co.il.
Sadly though, it doesn’t make much sense. Amazon’s cash and short term investments stand at around $10 billion with its gross profit standing at around $11 billion in 2012. Even a deal of a couple of billions of dollars is likely to have a significant impact on the company’s bottom line and more importantly, on the company’s stock price. Coincidentally, TI has today published a PDF guide called ARM Processor Selection Guide, that succinctly shows what's on offer should Amazon go ahead with the purchase.
In comparison, Apple spent around $500 million to develop the A6, a figure that includes a string of acquisitions, acquiring the architecture license and developing a new architecture, almost from ground up.
Also historically, Amazon never purchased a hardware business. The US retailer is inclined to either invest strategically in content websites (DPreview.com, IMBD), platforms (Audible, Lovefilm, Mobipocket) or technology (Engine Yard, Snaptell).
There are indeed a fair few reasons why Amazon may shy away from such a wieldy acquisition: acquisition costs, integrating a grown-up, mature business into a retailer that has little experience about dealing with ODM and OEM customers (at least from a provider’s perspective) plus an ultra-competitive landscape that would discourage such a move.
After all, TI is only a partner of ARM and doesn’t own the technology and given that there are dozens of ARM partners globally, picking one with the right expertise will cost less even if, as rumour has it, Amazon Web Services and its tens of thousands of servers, are what the deal would be all about.
If servers are what it is all about, AMCC, Calxeda and even Chinese-based Nufront may be better and more affordable acquisition targets, should Amazon try to emulate what Apple did, but in the data centers.
There's at least one major potential target worth considering; MIPS. At less than $400m, the Sunnyvale-based company is similar to ARM but on a much smaller scale and with a much smaller price tag as well.
So rather than being one player amongst dozens of others, Amazon would own the technology and the IP behind one of the major chip architectures on the market.
The company has been rumoured to be on sale since April 2012 and back in May 2012, the company announced a range of new microprocessor cores, Aptiv, aimed at giving ARM’s Cortex-A15, a run for its money by challenging on performance, performance per unit area and power consumption.