CASWELL, Northamptonshire, U.K. — In the past couple years Bookham Technology plc has spent in excess of $350 million dollars acquiring five other companies. Not bad for a company that's been in the red since its inception fifteen years ago.

Bookham owes everything to its initial public offering, which took place in April 2000 when telecom stocks were at their peak. By the end of the week following the IPO, the loss-making components vendor was valued at £4.5 billion (US$6.8 billion) – making it more valuable than British Airways at the time.

In fact, the IPO was so successful that Bookham went back to the stock market in September 2000 for a secondary offering. As a result it raised over $400 million, cash that's kept it afloat while many of its competitors were sinking.

Today Bookham's market valuation is modest by comparison with 2000, but its acquisitions – which include the components operations of Marconi Corp. (Nasdaq: MRCIY; London: MONI), and Nortel Networks Corp. (NYSE/Toronto: NT); startups Ignis Optics, and Cierra Photonics; and, most recently, New Focus Inc. – have vaulted it up to the number two spot in the telecom components market (see Bookham Buys Nortel's Components Biz and Bookham Gets a New Focus).

Giorgio Anania, now the company's CEO, played a key role in getting Bookham to where it is today. Anania joined Bookham in 1998 as senior VP of sales and marketing. His previous position was at Flamel Technologies, a French drug-delivery company, which he helped take public on Nasdaq in June 1996. Anania's experience in stock offerings was essential for Bookham. A couple of months after the IPO, Anania was promoted to president (a post he relinquished earlier this week), and in February 2001 he took on the position of CEO.

But he's not just a marketing and financial guy. Anania is a scientist at heart, holding a first degree from Oxford University, and a PhD in plasma physics from Princeton University. It is fitting then that Light Reading caught up with him at the opening event of the company's state-of-the-art indium phosphide fabrication plant in Caswell, U.K.

Before settling down to chat over a nice cup of coffee (an Italian, Giorgio doesn't tolerate the cheap stuff), Light Reading was treated to a tour of the new facility. Possibly the only three-inch indium phosphide fab in the world, it was created by closing down the former Nortel facility in Ottawa and shipping all the equipment over to the U.K.

In an industry where moving a piece of equipment from one room to the next can totally screw up production, moving equipment halfway across the world was a bold move. But it's worked, according to Bookham, which claims that all of its key products lines are being successfully produced from the new facility.

It's worth adding that the Caswell facility is now one of only two manufacturing sites producing active components for the company. The other is a highly specialized plant for making pump lasers in Zurich, Switzerland. To cut back on running costs, Bookham has consolidated from a total of 14 sites to just six (prior to the New Focus acquisition).

Togged up from head to toe in pale blue floaty space suits, with barely a chink of flesh on display, we entered what must be an asthma sufferers dream – a class 100 clean room. In other words, there are a mere one hundred particles of dust per cubic meter (that's impressive, in case you were wondering).

Some journo-wonks who came with us on the tour were jotting down the makes of all the equipment. But the fact that Bookham has seven MOCVD machines and one MBE machine was not really the key point to take away from all this. The key point is automation. Every gleaming machine in the new facility has robotic arms that hiss quietly as they move indium phosphide wafers into position. There are different machines to perform different stages of the semiconductor manufacturing process. All the operator has to do is move casettes, each holding a couple of dozen wafers, from one machine to another, and log the progress on the computer. (There probably is more to it than that, but it did look so easy!)

Automation equals capacity, and Bookham has oodles of it – capacity it blatantly doesn't need in today's subdued telecom market. But this isn't a case of runaway optimism. By employing automated equipment developed for silicon electronics manufacturing, Bookham claims it can produce better quality, more reliable products, and hit much higher yields.

If it had to build from scratch, Bookham couldn't possibly afford or justify this facility, estimated to be worth close to a billion dollars. But thanks to acquisitions, it's been able to pick up equipment for literally cents on the dollar.

In fact Bookham's got so much equipment that some of it lies idle, kept just in case of a surge in demand (maybe we were wrong about the runaway optimism...), and some of it is up for sale – as I write this, my email bleeps to tell me that Henry Butcher will be holding a live auction and Webcast of Bookham gear next week (see http://www.henrybutcher.com/en/saledetails.asp?SaleID=3542).

Read on for Anania's views on:

— Pauline Rigby, Senior Editor, Light Reading

Light Reading: Is Bookham aiming to be a one-stop shop for optical components?

Giorgio Anania: I don't think that's just an option any more, I think that is a must. Our customers have very few R&D dollars, and relationships are very difficult to build, so they only want to have relationships with as small a number of companies as they can that will give them most of the solutions.

LR: So, you're following in the footsteps of JDS Uniphase!

GA: I have a lot of respect for what JDS did, which is consolidate a lot of pieces into one company that has what customers want. We have a different background – we are very much focused on chip innovation and will always be a semiconductor company at heart. Each company needs to find its own way forward. We know what we want to do is grow in telecom and outside of telecom based on our core technology and know-how, and that's mainly focused on chips.

LR: There have been so many changes on the components scene. Which other vendors are still serious contenders?

GA: Without making any judgments, if you look at the top three telecom optical components suppliers, today these are JDS, Bookham, then Avanex. There are also a number of players of roughly the same size in datacom – Agilent, Finisar, Infineon, JDS (again), and Sumitomo.

At the low end, telecom starts becoming indistinguishable from the high end of datacom. Increasingly, one has to start thinking of those two market segments as one larger market. We're maybe number two in telecom components; we're probably four or five out of the 10 that matter in this enlarged space.LR: Are the two areas – telecom and datacom – going to merge?

GA: In the past, most of datacom has been optical modules going a few yards, whereas telecom has been components going a few tens or hundreds of kilometers. So it was very different. But as most of the telecom buildouts now are becoming metro over shorter distances, and a lot of datacom is going outside the building into the campus and the metropolitan area, we're finding that all of this area is going away from discrete optical components into either: a) full transponders, which make it easier for customers to design electronics around it; or b) pluggable optics as in datacom.

As this comes together we have to think of ourselves playing in a larger market. We need to learn how to make our [telecom] products migrate to a pluggable optics mode, just as those folk who make pluggable optics need to learn how to make their [datacom] products go longer distances and do DWDM. There is a challenge from both sides.

LR: How does this fit in with the trend towards subsystems?GA: The solutions are no longer optical components, but increasingly they are full subsystems, or what are termed in the industry as "network ready" modules – which push into a rack and have standard software interfaces. I think that's another must.

You don't need to have [i.e., make] all the contents of the subsystem, you need to be able to integrate. But of course if you don't have most of it, you don't make money.

LR: How many vendors does the market have room for?GA: If you look at this enlarged space, there's probably a dozen or so reasonably large players, and probably 20 to 25 players in the next level down. I think there is room for four or five players to make money, maybe less. And then maybe another 10 or 12 behind that in a narrow role. I think the consolidation is less than half done.LR: What's going to happen to the other companies?GA: In telecom – it's a little different in datacom – you either are a company that has a full series of products and are able to build them into subsystems, or you become a niche supplier to the broadbased companies like Bookham or JDS.

So, as I said, it's not really an option to remain a supplier of one particular piece, because if that piece goes into a subsystem, then you either [have to start providing subsystems or you] don't provide the subsystem and you lose the customer interface.

If you are dealing with a Nortel and they want a subsystem, and you only provide part of a subsystem, then they go to someone else. So you are no longer selling to Nortel, you're selling to someone who sells to Nortel. You can have a good business that way, but now it's playing by a completely different set of rules.

LR: Is competition from Chinese components vendors having any impact on business? GA: We take that very seriously. I think, in the past, China has played mostly on a very low-cost manufacturing position. We should be able to match a lot of that, as we should soon have facilities [referring to the manufacturing plant in Shenzhen acquired from New Focus]; and we should not necessarily be cost disadvantaged as much as we have been in the past versus Chinese companies.

I think the more serious issue arises as Chinese companies move up in technology: When they have similar technology to ours it will become a more competitive environment. For now, the leading edge of chip design still tends to be Western based, but that will change.

LR: What do you think about Simon Cao's consortium, Greater China Photonics Solutions? [This is a group of Chinese vendors that aims to attack the materials costs of manufacturing; see Cao Unites Chinese Components Vendors.]

GA: I think customers need to know that you control things so you can follow them through the hoops and turns they need to make, and that would seem to be harder to do with a mixture of companies. If you're putting your faith in smaller companies, you are accepting that some of them may go up or down. It looks like everything is there, but if I am a Nortel or a Lucent, who do I look in the eyes to really really commit?

I know there is a trend going on to outsource. I may be very Old School from that point of view, but I think it's important to control your destiny, so that you can commit to customers. That means you have to have a big cost base such as we do, which requires scale; but it does mean that if you commit to a customer, you control most of the things you're committing on. It goes somewhat against the latest fashion, which is outsourcing.

We truly believe manufacturing is one of our key competitive advantages, so we are not going to be one of the companies to go to an outsourcing model. If you went to an outsourcing model, you would be able to pick and chose in a more agile way, but I would think that you would end up being less cost competitive as you have to pay other companies' profit margins as well.

LR: How far has Bookham got towards reducing its reliance on Nortel as a key customer?GA: First of all, we're very pleased to have Nortel as a key customer. It's one of the largest customers in the world, and the relationships have been excellent. Both of us have needed those relationships to be absolutely excellent because Nortel relies enormously heavily on Bookham for components, which in many cases are single sourced, and Bookham relies on Nortel for a large part of its sales.

Having said that, the business model from day one was to rely on our partners to give us some revenue stability – but to really get the business model working by taking these products, which were not available to other players, who were competitors to Nortel, and to make these products available [as an independent supplier].

It just needs to happen. You can't snap your fingers and get designed in, especially these days when there's no R&D. Customers don't have any reason to change their supplier until the next-generation platform happens.

We've been making very good progress. In the second half of the year we announced that from zero, Huawei had become a 10 percent customer for us in Q2, and we also indicate that just the first three quarters of the year, our sales to customers other than Marconi and Nortel grew by about 50 pecent in what was a flat to slightly down market. So we've definitely been reducing our dependance on Nortel, but it's still quite a big share. As time goes on, we would aim to maintain or enhance our relationship with Nortel but also grow aggressively with other customers.LR: What proportion of your business will be non-telecom?

GA: We believe that at end of year we will see around 30 percent of sales in non-communications. But clearly we need to get there. Our acquisition of New Focus gives us an entry point of about $7 million worth of sales to many other customers, and that's a route into many areas where optics has a potential to grow.

If you look at our strategy, we have a very strong manufacturing capability, with lots and lots of manufacturing capacity available. To make our financials work, we do need to attack other additional markets. The telecom market is a very nice market, it's no longer a huge market, and to get economies of scale, we need to be several times our current size.

LR: So you're still in acquisition mode?

GA: I think our industry still has too many players and is consolidating. Bookham was the first company to launch the consolidation of the industry when people were too scared to try and do it. Now I think everyone can see it makes sense. But the industry consolidation is only one-half to one-third done – it needs to continue. Whether [the consolidator] is Bookham or someone else remains to be seen.

— Pauline Rigby, Senior Editor, Light Reading