ARM loses its lustre as new licensees falter

ARM loses its lustre as new licensees falter
By Luke Collins, EE Times
October 24, 2002 (9:48 a.m. EST)
URL: http://www.eetimes.com/story/OEG20021024S0001

So what next for ARM Holdings, the Cambridge processor licensing company that defied market gravity until last week when it announced its first quarter of declining sales and sacked 10% of its workforce?

Publicly-quoted ARM had turned in 18 successive quarters of improving revenue until its business prospects unwound in its most recent 19th quarter. The company warned three weeks ago that its profit would be down on expectations - an announcement that carved 60% off the company's share price in a single day.Now its Q3 results show sales of £33.3m, down 23% on Q2 (£43.2m) and 11% down on a year ago. ARM is still profitable, having made £8m in the third quarter, down from £12.9m for the corresponding period last year. But the company will take a £2m charge next quarter and lay off about 100 people in a bid to save £5m a year.

So what happened? In the second week of September, ARM was still making reasonably bullish comments about its prospects. Bu t it appears that, as the quarter closed, several licensing deals which ARM had expected to sign with customers were put on hold. And licences are a substantial chunk of the company's revenue.

The company was careful to sweeten its results with news of three licensing deals for the ARM11 architecture, signing up LSI Logic, Texas Instruments and Qualcomm.

ARM's business has always been built on tomorrow's promise, with much more of its revenue coming from licences, development tools and services sold to clients adopting the architecture than from royalties on devices actually shipped.

In the most recent quarter for which ARM has figures, the company's partners shipped 124 million chips using its core, up from 95 million in the previous quarter. Royalty revenue was £6.2m (19% of revenues) compared with £6.5m (15% of revenues) in the second quarter of 2002.

ARM now has 101 partners comprising 57 multi-use licensees (34 of whom are shipping) and 44 per-use licensees (three of whom are shipping). A decline in per-unit royalty rates in the quarter to 30 June, 2002, was mainly due to a change in end product mix, particularly the introduction of lower-priced smartcards.

ARM may have a large client base equipped with its processor cores, tools and the know-how to use them, yet What it does not have is a majority of users seeing the benefits of their licences.

And as the semiconductor industry downturn has continued, the market for new licences has fallen away. There is no point in taking on a new licence if you cannot see the end market for the product that would use it.

This bubble in ARM's revenue pipeline must be of concern, not least for what it says about the company's ability to predict revenues.

Its half-year outlook, published in July, expected growth in the remainder of the year to match that of the first half, based on the company's sales backlog. But, instead, business turned down as partners deferred buying licences. So ARM is predicting "flattish" revenues based on its Q 3 run rate of £33m "for the foreseeable future".

So what are the longer term prospects for ARM now? Jim Tully, a research director at Gartner Dataquest who tracks the semiconductor intellectual property (IP) industry, thinks the company will emerge stronger.

"ARM will gain as a result of the downturn - not necessarily because of actions of its own but because, when times get tight, things become more conservative and ARM wins because people play it safe," he said. "You go for what is stable, the market leader, and that means ARM.

"It is in a position where it is sitting on the world's favourite architecture and has become a de facto standard in many application areas."

ARM was at pains to use the Q3 statement to claim that its competitive position had strengthened through the downturn as new and existing partners committed to ARM's technology roadmap.

But becoming an industry standard can create problems as users rebel against the provider's terms. ARM has already used the courts to see off attempts to create low-cost clones of the its architecture in response to demand in the Far East.

Malcolm Penn, founder of semiconductor industry analysts Future Horizons, sees this as a potential problem for ARM: "There will be growing resentment - that's the problem with the royalty model, especially in a tight market when you're not making any money and you start to feel peeved that there's another royalty bill to pay."

The fact that such customers would not have a product without ARM's contribution is usually overlooked in these situations. But Tully says ARM's position may not be so problematic: "Because the IP market is much smaller than the device market, ARM does not feel like a powerful company - it holds a lot of power but it doesn't behave as if it does."

ARM faces other problems as it grows, particularly what it does next to broaden its franchise.

Tully said: "It's difficult - ARM has a processor and has done all the right things for the core, with development tools and perip herals. It has also been good at forming relationships with realtime operating system and software vendors.

"So it has a processor in the market in a range of applications. But what then? That's the big question for ARM."

The company's response has been to move into 'platform-based design', in which it wraps up a core, some peripherals, development tool support and basic software as a standard product that will give its users a head-start in a specific market such as wireless handsets.

But Tully says ARM's approach to platforms, such as its PrimeXsys wireless offering, gives much less integration than its rivals. The reason? ARM cannot afford to compete with its processor core customers.

Tully said: "How rich a platform ARM can ever genuinely offer without upsetting its customers is a very delicate balancing act.

"New companies will emerge that are not constrained by customer relationships, and they will be able to create high-integration platforms which allow their customers to enter ne w markets quickly, competing with ARM's less integrated offerings," he said.

Despite these doubts, ARM's key market for processor cores remains a good long-term bet. Almost every embedded device needs a processor, and more of those processors are being provided by ARM as time goes on, creating new sources of royalty revenues.

Tully says that, despite possible constraints on ARM's business development created by its need to avoid competing with its customers, it still has a bright future.

"The question becomes 'how long will the pure processor market be a good enough market for ARM?'," he said. "How far can ARM carry on? It's a question of when processor development is going to stop - and it probably isn't."

Copyright © 2003 CMP Media, LLC | Privacy Statement