Custom Search

Monday, October 19, 2009

Bluetooth Low Energy Hype 2: Semiconductor Price is Below $1.00

"Bluetooth Low Energy will add almost zero cost to Bluetooth chip." "The price of Bluetooth Low Energy chip will be sub $1.00." If you have ever attended any Bluetooth Low Energy presentation, you probably have heard these messages. This concept has been well accepted and used as a price benchmark for technology evaluation in companies and in organizations. The reason is because 1) who said it. For me, I heard this message from Robin Heydon, the co-chair of BTLE working group. 2) how much people like or expect it. People believe the volume of Bluetooth chip will be several hundred million a year, and this volume, as well as multiple vendors in competition, should drive the price down. Sub $1.00 cost sounds just so reasonable.

But for most of the manufacturers, sub $1.00 price is irrelevant!!!

In my last post (hype 1: cell phones will offer the critical mass), I already pointed out the complex product matrix of Bluetooth. So you first question should be "is the sub $1 price for dual mode, single mode or both?" This answer is very straight forward, this price can only be possibly true on dual mode device. BTLE single mode chip will be more expensive than dual mode. Ironically, this message was also delivered in a Q&A session of a Bluetooth conference that I attended, but it seems did not go far.

Dual mode chip sounds more complex than single mode, so why the price of dual mode Bluetooth is even cheaper. Because dual mode chip and single mode chips are targeting to two different design environment, one is mobile platform and the other is ultra low power embedded systmes.

Dual mode chips are supposed to be used in cell phone or PC, where a much powerful host computer is in presence. So Bluetooth protocol implementation is usually cut at the Host Controller Interface(HCI), leaving other high level applications and profiles out to the host. These dual mode IC do not need to have a complex IO ports to look after analog and digital inputs or outputs, except for the interface with the host.

Semiconductor die size and volume are big factors for the low price too. The more transistors able to be integrated, the smaller the size and the cheaper is the semiconductor. The most advanced and mature semiconductor manufacturing technology today is 45nm, and foundry companies are working towards the next generation 28nm today. Any generation forward is a huge investment that only high volume device is capable to bear and amortize the huge investment. And low volume devices will just be more ecnomically use technology often several generations back.

Remember nowadays, Bluetooth is coexisting with other radios (usually Wi-Fi, FM and GPS) on one chip. These chips adopts multi-core architecture often with reserved memory or MCU core for new functions to fit the market dynamics. Adding BTLE to this kind of architecture could basically zero hardware cost. 

Other than dual mode for mobile platform, single mode chips mainly dealing with embedded system design, require a much complex digital and analog mixed and System on Chip architecture to include a complete protocol stack, an API for applications, a comprehensive I/O ports, A2D and serial interface and a ultran low power friendly timing and power management. It requires a much different semiconductor expertise to design, but with a 10x lower volume in production. I have not seen any SoC (802.15.4 or preprietary) is at 45nm, most of them at 180nm. 

So any idea of the target price for BTLE single mode chip? I guess at around $2.00 when it is launched at relative high volume say 100k. My guess is based on the product announcement from Nordic Semiconductor, which is the leading vendor of ultra low power wireless in 2.4GHz, including BTLE, ANT and its own proprietary. In its recent ANT part nRF24AP2 press release, "Pricing is USD $1.8 and USD $2.1 for 50k shipments of nRF24AP2-1CH and nRF24AP2-8CH respectively." I believe the Nordic BTLE part will be at the similar range and even higher. ANT and BTLE will compete neck to neck to the same market. In order to survive, ANT has to offer some price advantage to compensate its disadvantage (perceived but may not be true), such as lack to cell phone accessibility and not from an standard organization.    

So sub $1.00, if you are top five cell phone manufacturers, is your Bluetooth chip price; if you are a sensor manufacturer, forget about sub $1, and budget your products with at least $2.00 for the BTLE chip, possibly $3-4 at low volume to start.

As I publish this post, Bluetooth SIG will hold the First International Bluetooth Low Energy Technology Conference. If you attend, why not raise your hand and get the confirmation from Bluetooth SIG directly?

1 comment:

  1. This comment has been removed by a blog administrator.

    ReplyDelete