The wonderful thing about the technology industry is that there is always some interesting rumor or piece of news hitting the wires. The latest hot discussion topic is that Amazon (AMZN), the world's leading e-tailer, is in "advanced talks" to purchase Texas Instruments' (TXN) OMAP system-on-chip division, purportedly for "billions" of dollars. Does such a deal even make sense?
A Gross Margin Play?
It is no secret that Amazon is selling its Kindle Fire products essentially at cost and is relying on selling content to ensure that this is eventually a viable venture. Naturally, the company is likely looking for any way to reduce costs, likely in a bid to improve the gross margin profile of the devices. The cost of a smartphone or tablet SoC is roughly $15 - $30, hardly the largest part of a tablet's bill-of-materials, but still nontrivial for a device selling for $150 - $200.
In theory, then, Amazon could be moving to increase its gross margins by having the device design teams in house. However, the improvements may not be so drastic for the following reasons:
- TI's Not Selling Fabs - TI has design teams and patents, not chip manufacturing facilities. This would force Amazon to pay to have any potential designs manufactured at a fab such as Taiwan Semiconductor (TSM), diminishing any potential gross margin advantage.
- The Division Is Not Profitable - As sales of the OMAP chips have declined in light of formidable competition from Nvidia (NVDA), Qualcomm (QCOM), Samsung (SSNLF.PK), and Apple (AAPL), the division has actually begun to produce losses. By reducing volumes further, presumably by only using the chips for in-house designs, the losses will likely widen, especially as these designs become more complex.
- ARM Needs Its Cut - Another beautiful part of the ARM ecosystem is that SoC designers need to pay royalties of ~1.2% of the chip price to ARM for each chip sold. This would also eat into the gross margin profile of the chips and potentially hurt the case for having a team in-house.
- The Products Are Falling Behind - While TI was initially a fairly large player in the smartphone/tablet chip space, the company's products have become increasingly uncompetitive, largely due to the sluggish time-to-market for systems-on-chip sporting new CPU and GPU technology from ARM (ARMH). Qualcomm is the leader with Nvidia and Intel (INTC) aiming for a slice of that pie, and TI has seemed to indicate that it will fall by the wayside in this race.
Would Amazon Be Overpaying?
The next reason that such a deal would not make sense is that it is reportedly for "billions of dollars". First of all, the division generated $342M in revenues in the most recent quarter, representing a 39% year-over-year drop and an 8% sequential drop. The division also lost $51M in the most recent quarter, while it earned a scant $82M in the year-earlier quarter. Why would Amazon pay billions for a little over $1B/yr in revenues at best?
More importantly, TI's OMAP chips are not particularly differentiated. While Qualcomm designs its own CPU and GPU cores along with the SoC, Nvidia designs its own graphics processors while using ARM's core designs, and Intel is using its own custom CPU (with plans to move to an in-house graphics processor down the line), TI has no real "secret sauce" here. It uses standard off-the-shelf ARM cores, off-the-shelf graphics from Imagination Technologies and a number of its own DSPs for specialized functions.
Further, TI divested itself of the cellular baseband business some years ago, so any 3G/4G enabled Amazon phones would need to license additional technology (likely Qualcomm).
In short, paying "billions" may not be such a wise move on Amazon's part.
Great For TI, Though
This would be wonderful for TI, however. It is clear that the mobile chip business is a crowded one with many powerful players. Margins are likely to shrink and R&D costs will go up as these devices become a lot more sophisticated. TI has already announced its plans to scale back its investments in this area in light of the fierce competition. Further, the company has extremely solid core businesses with its analog, embedded, and "other" segments - it doesn't need an unprofitable, commoditized chip business hurting its bottom line.
If TI can actually sell its mobile chip division to Amazon, then this will be positive news for TI and its shareholders. On the flip side, I'm not so sure if the benefits outweigh the costs of this kind of deal for Amazon. The gross margin improvements may be offset by increased operating expenses. I would be very wary if I were an Amazon shareholder - the valuation is already stretched as it is in light of very scant profitability, and this sort of deal may be the final piece of ammo that short sellers need to drive the stock down from its lofty prices.