What Investors Want
By Ed Sperling
What exactly do EDA startups need to get going? Early-stage investors Jim Hogan and Paul McLellan tag teamed an audience filled with would-be startup executives at the Design Automation Conference with a summary of what works, what doesn’t, and how investors view EDA startups.
For one thing, startup capital as close to zero as possible is essential. Hogan said that at Altos Design, which was bought last month by Cadence, no one received a salary for the first three years. He said the average price for cubicle rentals and electricity was $350 per month during that time.
The first step is proof of technical proficiency, which is round A funding. From there the company has to start producing something, which is where the real investors come in. The challenge for startups is to retain as much ownership of the company—as much as 75%—so the exit strategy is easier. That way it’s easier to agree on a sale price and the price is more reasonable.
“We’re now starting to see companies in the C round,” said Hogan. “Apache now has support for broad-based customers, and both Apache and Atrenta have to worry now about sustainable returns. They have to show predictable revenue and profit.”
Hogan said that the initial seed round, which lasts a maximum of 18 months, is run by technologists. From there they acquire more capital in round A, which lasts from 12 to 24 months. During that time there is a CEO, a business development person and two or three FAEs. In Round B , which lasts from 20 to 30 months, the company adds a CFO, a marketing vice president, and five to eight FAEs. In round C, which lasts from 20 to 40 months, there is also a vice president of sales, three to four salespeople, and 10 to 20 FAEs. At IPO, which takes 40 to 60 months, an entire G&A team and customer support are added.
The key to making this process work is breaking even by three years. Otherwise companies end up at asset sales, which generally bring in no more than $2 million to $10 million. A company sale, in contast, can bring in up to $50 million in round A, and up to $200 million in round C.
There are plenty of reasons why companies fail, though. McLellan said the most common is by being too early to market rather than too late. Another is getting the market wrong.
McLellan said the top opportunities for EDA startups involve dealing with second-order effects, the ends of the spectrum including system-level design and software, DFM and yield optimization, and those that focus on IP and assembled chips rather than those chips that are designed.
He said customers are willing to pay for IP because it’s in the chip and therefore mission-critical. They also will pay for optimization, especially in the area of power, and verification, where ROI is easily measured. What they will not pay for a cheaper tool because it raises the risk.
“When you look at faster, better, and cheaper, cheaper is bad,” Hogan said. “And when you’re dealing with faster in commoditized markets, it can’t be just two tims faster. It has to be orders of magnitude faster.”
Tags: Altos, Apache Design Solutions, Atrenta, Cadence, VC funding