NEW YORK – After warnings last week of falling quarterly sales stemming from weak PC demand, Marvell Technology Group sought to change perceptions about the company’s prospects that one analyst called “less than inspiring.”
Countering several financial analysts who have downgraded Marvell, co-founder Weili Dai stressed in an interview that Marvell is ready for the coming “super-integration era” of storage (controllers for solid state drives), connectivity, video, 3-D graphics and modem technologies for Ultrabooks, tablets, smartphones and smart TVs. According to Dai, not even rival Broadcom has the “ingredients” necessary for broad range of future smart devices.
Dai stopped short of providing specifics as to when she expects the so-called “super-integration” era to start contributing to Marvell’s bottom line.
[Get a 10% discount on ARM TechCon 2012 conference passes by using promo code EDIT. Click here to learn about the show and register.]
Sticking mostly to company talking points, Dai (right) cautioned against betting against Marvell – especially the company’s staying power. “We don’t live by quarters. We are here to stay [for the] long term.”
She noted that Marvell has rebounded several times in the past, proving naysayers wrong. Among the examples: Marvell’s entry into the cut-throat HDD controller business 17 years ago; its decision to buy Intel’s XScale processor division in 2006; and development of its own TD-SCDMA modem chip.
Marvell also stresses its leading position in SSD controllers which will become an important ingredient in a range of products beyond Ultrabooks. The company is hoping other key elements including connectivity, multicore CPUs, modem, graphics and video will also play a critical role in smart devices, including smart TVs, smartphones and tablets.
Marvell - RIM connection
The key issue for financial analysts is that Marvell seems unable to find the right customers.
Gary Mobley of Benchmark Co. called Marvell’s quarterly results and underlying guidance “less than inspiring.” “Some large diversified chip companies such as Broadcom seem to have an uncanny ability to find the right growth end markets (e.g. comm ICs) and the right customers (e.g. Apple)," he said in a research note. "Marvell, on the other hand, continues to struggle with too much exposure to no-growth markets (e.g. HDDs; 47 percent of revenue) and has suffered along with struggling customers (e.g. RIM). We reiterate our Hold rating.”
Indeed, Marvell owes its initial success in mobile phones to design wins in RIM’s Blackberry. As RIM tanked in the smartphone market, so have Marvell's fortunes. Asked about company plans to make up for losses associated with RIM in the global market, Dai stood by her customer, noting, “You should know that I go to the war for my customers. RIM is going through some tough times. Sure, as a part of the ecosystem, we are affected. But I am confident that RIM will be leading the pack with its enterprise software capability.”
Marvell is meanwhile betting big on its multimode, multifrequency band TD-LTE modem chip set for TD-SCDMA and LTE markets. The TD-LTE modem must operate on TD-SCDMA and GSM in the China Mobile network, while offering a roaming capability on 4G and FDD. Marvell’s PXA1802, announced last February is designed to meet those requirements. “This will become a foundation for our global leadership,” claimed Dai.
Let me clarify. The woman in my first comment should be gender neutral. What I was saying is: Marvell is ruled by a family, WeiLi and her husband. Given what I knew of how the company was governed, the future may not hold well. Indeed, similar opinions were held by many analysts and people of long history in the semiconductor industry.
Much like the consolidation that happened in Enterprise servers which dwindled down to 4 players (now 3), Telco to 5 (2 from china), the ASSP based semi companies will dwindle down to 4. The papttern in this consolidation is there will be high value player and a cost player. INTC, BRCM, QCOM. These are innovators. There is room for 1 or 2 value players. Mediatek has taken that stake. Perhaps MRVL can fit in the latter category since it has IP and channel, but high cost structure. If MRVL operates like Netgear - which is a US Company with primarily Taiwan/China operations and focuses on the value part of ASSP it has a chance. As the ASSP market consolidates as even $1B run rate in revenue is not sustainable in this winner-takes-all game, MRVL needs to be part of somebody (very difficult given personalities and product mix) or need to become the value player. So it has outsmart Mediatek at its own game to be a player. Increasingly value based SoC will be Taiwan/China based as it has happened with TV chips last year.
Someone's comment that seems to imply that a woman driving a company is bad, is a totally foolish comment, from a close-minded thinker, not a world wide one.
However, most of what I got from the article sounded like someone desperately fighting for survival, and while that may get you food for the next day, to me it does not point to a carefully-considered strategy for long-term growth of a company. Passion for the products you build and the markets you chase is one thing; viciously defending everything you've gotten yourself into, sounds problematic. You can't force people to buy the end product. When the market moves, you can get left behind.
Marvell, kind of missed out on the smart phone bandwagon in their focus on storage ICs but even now it's not too late since the market for quadcore based tablets and smart phones is seeing tremendous growth.
I will answer what I know. It was my bad that I didn't mention this in the story, but Forward Concepts baseband figures are based on units, not by revenue.
The numbers include both types: baseband-only chips and baseband/apps combo chips.
Dear Junko, thank you for another interesting article. And -- one certainly should wish Marvell and Weile to have a great success with their 3G/4G cell phone IC platform.
I do have a question for Bill Straus and his Forward Concepts since this is not the first time that EETimes uses his/their market share data.
By what accounting approach he arrived to his smartphone baseband market shares - whether they are in units or revenues?
Smartphones typically have either:
-- a standalone baseband (BB) - that goes together with a standalone application processor (AP)-- typically in high end smartphones a la Apple/ASIC or Samsung's Exynos processors
-- an integrated processor (BB+AP)- typically used in all low-end smartphones
According to Forward Concepts Marvell is about two times larger than either MediaTek or Broadcom in 1H12 smartphone baseband shipments...!? Could you, Junko, kindly ask Forward Concepts to explain - how was that arrived to - based on all known factual company data, including Broadcom success at Samsung and MediaTek's success in China's smartphones?
Of course, in feature phones there are only BBs -- just MediaTek sells there more than 0.5B BB units there but that was not the subject of the article and Forward Concepts graph.
Many thanks, once again and my regards
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.