Jul 10, 2008This article was originally published by RFID Update.
July 10, 2008—RFID chip and reader manufacturer Impinj today announced a bold strategic move that could elevate the company from its current status as one of a small handful of industry leaders, to a new level that is unique in the RFID space. Impinj has acquired the UHF RFID reader chip operation from semiconductor manufacturer Intel. There are many implications and interpretations of this deal, but first the facts.
Intel's RFID operation was a unit within the vast company's New Business Initiatives, which incubates new technologies and business opportunities. The operation was working on RFID reader chip technology, the first fruits of which were introduced last March with the commercial release of the R1000 (see Intel Announces Gen2 RFID Reader Chip). The R1000 consolidates 90 percent of the components that comprise a standard reader onto a single chip, resulting in lower cost and dramatically smaller form factors. This move from a componentized reader device to a single chip -- called "silicon integration" in chipmaker parlance -- is a major step in the evolution of reader technology. As evidence of that fact, when it released the R1000 Intel also announced a handful of clients (that is, RFID reader manufacturers) that were going to develop reader products based on the technology (see RFID Reader Vendors Rally Around Intel's New Chip). To date, about 20 companies have developed R1000-based readers, and 20 or so more have products in development.
Impinj has purchased essentially the whole kit and caboodle. It now owns the R1000 and all associated products, rights, and intellectual property, according to vice president of finance Evan Fein. As for employees, most of the R1000 sales and marketing people will go to work for Impinj, including Kerry Krause who had been Intel's RFID marketing director and spokesman. The R1000 engineering team was split between in-house Intel employees and outsourced engineers. The Intel employees will stay with Intel, while Impinj will continue the relationship with the outsourced engineers but may or may not use them for future product development.
The R1000 has been re-christened the Indy R1000, to go with Impinj's racetrack branding theme.
There are many facets to this acquisition, but in a nutshell the three takeaways appear to be:
- This is a major win for Impinj.
- Intel's exit from the industry is not welcome, but not necessarily a bad sign.
- The dynamics of the Gen2 reader space will change.
Each of these are explored in more detail below.
Impinj: Less David, More Goliath
The benefits accrued to Impinj are myriad. First, the R1000 is reportedly a very good piece of technology in its own right. It was developed by the world's leading semiconductor company and has been deployed in the products of 20 companies, with 20 more in the offing. While it is not perfectly suited to every single application out there (sources say that it is not enough of a workhorse for very demanding applications), it is a solid "all-purpose" chip.
Perhaps more importantly, it is essentially the only game in town. Competition against the R1000 from other Gen2 reader chips is negligible to non-existent. "The R1000 has 100 percent market share," said Fein. WJ Communications announced a Gen2 reader chip at some point, but the company struggled badly and was acquired earlier this year. Starport Systems announced a Gen2 reader chip last year, but the company is reportedly no longer developing UHF technology. The only competition on the horizon comes from austriamicrosystems, which offers the AS3990. That product appears to have very little traction in the market.
"For the scope of functionality that the R1000 offers, no other reader chip does it," said Impinj's Fein. "No competing product has the same level of silicon integration."
The R1000's near-100 percent market share is reminiscent of the Monza chip's monopolistic position circa 2006. Recall that Monza, Impinj's Gen2 tag chip, was the only commercialized, production-quantity Gen2 chip on the market from its introduction in late 2005 into 2007. During that time literally all of the Gen2 tags sold commercially were based on Impinj's chip. This feat was all the more impressive given Impinj's small size relative to semiconductor giants like NXP, Texas Instruments, and STMicroelectronics, all of whom had designs on the Gen2 chip market. To this day, only Alien and NXP have gained traction with competing Gen2 chips, and both are still playing catch-up.
The R1000's dominance also means that Impinj will enjoy business from a wide range of RFID applications. The company already had a Gen2 reader product in Speedway, which is a higher-end fixed reader designed for supply chain and retail applications. (Understand that the R1000 is not a reader itself, but a chip that drives a reader -- much like the Intel or AMD processor in a laptop is not the computer itself, but drives the computer's inner workings.) With the R1000, Impinj's reader technology will be present in mobile readers, printers, USB storage sticks, and whatever other mundane or exotic RFID reader devices are ultimately developed with the chip. The potential applications of the R1000 chip are more diverse and numerous than the supply chain applications for which the Speedway was designed.
"We can offer the silicon reader in almost every application," said Fein. "Before we saw ourselves operating in a segment of the reader space. Now we see ourselves as the reader company."
Yet another benefit for Impinj is that it now controls both sides of the air-to-air interface. That is, it sells the chips that power the tags on one side of the RF "handshake", and the chips that power the readers on the other. That allows the company to improve and optimize this all-important link which is at the core of any RFID system's performance. It can especially optimize the link between an Impinj tag chip (Monza) and an Impinj reader chip (Indy R1000), and thereby make a credible assertion to potential customers that their RFID system will perform best if they pair Monza-based tags with Indy-based readers.
Impinj has always been high on the industry short list of companies that might eventually go public. This acquisition puts Impinj even higher -- highest, perhaps -- on that list. Just last week, the company divested its non-volatile memory (NVM) business, which had little to do with RFID directly, and announced it would move forward as a pure-play RFID firm (see Impinj Sells Non-RFID Assets, Moves Ahead as Pure Play). While management stresses that the NVM divestiture and the Intel acquisition are unrelated, they have a combined, complementary effect of ratcheting up the company's position in the industry. "We think this enhances the value of our company," Fein deadpanned. Intel itself seems to agree; it sold the RFID operation in exchange for Impinj stock. Impinj would not comment as to how large the stake was or its valuation, but did say the deal was pure equity, no cash.
It really is hard to see a downside in this for Impinj. The industry analysts and observers consulted by RFID Update concurred. There is always the chance that despite all the benefits Impinj paid too much, but that is impossible to evaluate without knowing the purchase price. (Note that Intel is scheduled to release its quarterly results this coming Tuesday, which may include further details about the transaction.)
Intel Exits RFID
Clearly it never seems good when a large global player pulls out of an industry, as Intel appears to have done here. Impinj counters that Intel hasn't fully exited, and will remain active through its board observer status with Impinj. But this is hardly the same level of participation as having a devoted team of business people and engineers developing the RFID venture for Intel's own immediate benefit.
The point was also made that this is standard practice for Intel's incubator division, which routinely invests in and develops new technology only to later sell it off. And in fact, Intel maintained since the beginning that it wasn't interested in selling R1000 chips as an end in itself, but as a means to spur RFID adoption. The idea is that when RFID is ubiquitous, the vast amount of data generated will drive demand for Intel's core chip products.
Such a long-term bet on RFID might seem farfetched from the perspective of an impatient RFID industry that tends to think in terms of one to two years out. But Intel is a 40-year-old, $114 billion behemoth whose investment time horizons are measured in half-decades and decades, not single years. Given that, it does seem plausible that the R1000 project was simply a technique to give RFID adoption a much-needed nudge.
Collective Groan from Impinj Competitors
It is one thing to have a strong competitor. It is quite another to have your products depend on your competitor. That is the position some RFID reader manufacturers now find themselves in, as their R1000-based readers depend on an Impinj product. Examples include Alien and ThingMagic. Will they continue basing their products on the Indy R1000, now that Impinj owns it? Or, will Impinj try to handicap them by refusing to sell them any more R1000s?
Probably neither. To the first question, recall that the Indy R1000 is the only real option on the market, so reader manufacturers that want the many benefits of a Gen2 reader chip have no other choice than to keep buying it to put in their products (for probably at least another year or two). To the second question, such aggressive tactics do not appear to be Impinj's style. The company has been very active in EPC standards development and other initiatives, which suggests it wants to see RFID blossom into a huge market, rather than be over-protective of its current position in what is a relatively small market.
"We don't have any plans in the near term to make changes that would affect the R1000 customer base [reader manufacturers]," assured Fein. "On the contrary, we've been on the phone with the largest customers to explain exactly that to them."
Whatever Impinj and its competitors in the reader space do, one thing is for certain: this move dramatically strengthens the company's hand.
Conclusion
Ultimately the most significant question does not have to do with why Intel sold its RFID operation or with the competitive machinations of the reader space. It is whether Impinj will prove a good steward of the R1000 technology. As mentioned earlier, silicon integration is crucial to bringing reader prices down and making the form factors smaller, which is in turn crucial to growing RFID adoption generally. By all accounts, Impinj did a great job in the realization and commercialization of Gen2 tag chips. Hopefully it will be as successful with the Indy R1000. The entire industry stands to benefit if it is.
