Book-to-build ratio was 1.0 and hovered around 1.1 in the last two quarters of 2003. The first half of 2003 was a continuation of the recovery from 2001-02, observes Rob S. Chandra, a general partner with Bessemer Venture Partners. “There are probably four reasons for this,” says Chandra. Inventory in the supply chain depleted drastically, placing a demand in the chain.

Secondly, says Chandra, there has been growth in geographic markets driven by unique users. “I think, across the board, expansions in China and India has also been a good influence in tweaking the downward curve,” opines the investor, “and the industry itself has found optimism.”

The cyclical industry has found momentum for the next phase.”If you look back at the last 40 years, the chip business has been growing at 12 percent, doubling every 6 years in effect,” Chandra says. There are not many industries that show such persistant growth over so many decades.

“George Gilder said it best, ‘Semiconductors invade new markets,’” remarks Chandra. Automotive is a good example. Till the last decade, the steel was the largest component in the vehicle, and now, the electronics have taken the spot. “The overall growth in the automotive industry has barely crossed four percent, yet the automotive chip sector has doubled,” points out the general partner. “The other phase change is happening in the retail market, as RFID gains widespread applications.”

“This is common across industries that show single digit growth, yet the chip sector related to the industry has grown robustly. It is tough not to be optimistic now.”

TCO driving the Market
“Data centers are being re-architected across enterprises to afford a lower Total Cost of Ownership (TCO),” says Chandra. “Look at the Oracle applications running off an IBM 32-bit, 1.7 Ghz PowerPC. At $44,000 for the machine, each 1Ghz of computing power carries a tag of $27,000. The new architecture is now designed grids of server blades that come at about $900 for every 1 Ghz.”

This grid comes with fault tolerance and flexible performance. “The shift in data center architecture has thrown up a new challenge—of how to interconnect these blades, the I/O battle if you will,” says Chandra, which he finds very interesting. The aggregation of computing and the new methods of distributing the computing power is an exciting driver.

Technology Moves out of the Closet
Over the last two decades, adds the general partner, technology has been focused on information processing—digital information, and the end geographical location of the technology has been the server closet.

“Today, technology is in the consumer’s everyday life—in the living rooms, in automobiles, in their pockets. This invasion is fantastic,” exclaims Chandra. The demands from these consumer devices have funneled into one platform—integration. “The integration has to happen at the silicon level,” remarks Chandra.
Even as features are added to a palm-sized device, the device itself is being shrunk to smaller dimensions, reducing power consumption and increasing processing speed.

The digital delivery has engendered superb sophistication in the associated tools—RTL, verification, place and route all have fantastic tools. “This has been a great movement, and has thrown up another field of unexplored opportunities,” observes Chandra.

Analog, anyone?
Digital processing has been successful in the data center worlds, says Chandra, while real-life applications like voice and video—consumer applications, if you will—are not very well developed. Simulation tools, verification or place and route tools are still immature in the analog design world. “One of our portfolio companies, Berkeley Design Automation, is developing some tools,” reveals Chandra.

In much the same way as the digital world depended on reconfigurable processors—FGPAs—the analog world is yet to see any such help in improving the design process. “There is virtually no one in this space and the entrepreneur world can really explode this opportunity,” says Chandra.

Strategy Relook
“In the last six years, technology was focused on communications—wireline, wireless, cellular and so on,” recollects the Bessemer investor. “Communications was the end market that was driving the ASICs investments. But now, the vertical applications are not as clear.”

Numerous companies are involved in every vertical. “Look at the players in the 802.11 space, or the GSM-GPRS-CDMA integration space,” comments Chandra. “We are moving away from vertical application investments to investing more in the horizontal enabling technologies.” For example, an analog tool company that could enable many of these vertical application startups to produce a new set of ASICs would interest investments. “Memories are now seeing an upward peak,” Chandra opines. Novel approaches to lower cost memories, driven by new tools are enabling new products into the market.

“Our strategy has changed now,” says Chandra. “There have been innovations in storage in the last few quarters, but how many clients will this space find? Enterprises are the slowest to change vendors or legacy, and time is the enemy of the entrepreneur. This is not the same anxiety that the market had six years ago,” laughs Chandra. “The consumer market has become sexy again.”

India Power Plan
“There is no denying the talent pool of the Indian market,” says Chandra. “We are now able to design and deliver chips at significantly lower costs.” The chip capabilities out of India has been growing at phenomenol levels. “The fact that Texas Instruments, Intel, Synopsys and other companies are investing seriously in Indian talent should be a good signal to the startup world too,” observes Chandra.

“In that context, now is just about the best time to start a company. The Valley is full of talent, costs are low and the markets are looking up.”

Venture funding, says Chandra, is about investing around milestones. “Market, technology and people are the three risks that influence our investing strategies,” he comments. “The risks we feel comfortable taking are with people and technology. We are not comfortable with markets.” Conversations with customers and markets have shown that this risk cannot really change, especially for the startup world.

The fab-world itself is changing, says Chandra, as companies are exploring new methods of fabrication. “New, viable sources of wafers are now available and will drive further inovations.”