High-tech

Due Diligence: Celoxica

By Bipin Parmar

Introduction

Celoxica is a post Chilli R2 electronic design automation (EDA) company. The company was founded in 1996 and was previously known as Embedded Solutions Limited. The company originated as a spinout from Oxford University's research on algorithmic compilation, which was led by Ian Page (now at Seven Spires Investments and a visiting professor at Imperial College), along with Roger Gook. This research resulted in a new way of describing both hardware and software. The new language, which was a variation of the popular C language, was called Handel C. Jon Treanor, now CEO at Conduit Partners, was appointed CEO.

The company went through a challenging period, when trying to establish a new systems language. The company has now repositioned itself as an EDA company, with a clear focus on its target market and audience of system level architects - providing designers the ability to perform a top-down design from system to chip. This is a higher level of abstraction than normal and would potentially allow the earliest possible integration of software and hardware after partitioning at the system level.

Funding history

Standard EDA Model

Value proposition

The Chilli perspective

Funding history

Celoxica kick started in 1999 when it attracted approximately $30m investment from group of private investors via Beeson Gregory in three post seed rounds up to May 2000. Those 'bubble' funds were used to fund a global expansion, with new offices, staff and support in Singapore, Japan and Silicon Valley, in addition to a bigger UK headquarters with an eye to a quick initial public offering (IPO). The management team strategy, headed by then CEO Jon Treanor, was to appoint regional general managers, a move which turned the company into a slow and cumbersome operation, rather than a nimble startup. This was compounded by the fact that some of the management team had little or no domain expertise in either EDA or startups. In hindsight, and to be fair with the then management, the pressure from VCs and investors in the bubble days was to spend, spend, and spend, so when the next funding round came, current investors could carve out for themselves a hefty chunk of a potential pre-IPO company.

By March 2001, as the company launched it's first software product, headcount had ballooned to 150 staff, with expenses run rate running amok at approximately $21m on revenues of less than $2m. Most of the business was from telecommunications systems houses in Europe, including Ericsson and Marconi, which went through some massive metamorphoses, dragging their suppliers with them.

By early 2001, the company was running low on fuel and needed to raise a new round. It worked very hard and managed to close a new round of $30m from Advent Partners, Cazenove, Quester, Intel Capital and Xilinx in September 2001. This time, the funds were based on specific milestones. As of today the company is still living off this round.

In January 2002, the board decided that the next phase of the company's growth would be better served by a new refreshed management team and CFO Bernard Morgan was appointed as interim CEO, while the board initiated a search for a new CEO.

Phil Bishop was appointed as the new CEO and joined the company in July 2002. Bishop was formerly a VP and general manager of worldwide consulting at Mentor Graphics, a leading EDA company and prior to that he led the ASIC teams at Motorola Semiconductor and Boeing Aerospace. Clearly Bishop had the domain expertise of EDA, semiconductors and systems design, to start managing the drastic changes needed to turnaround the company and its direction. He started with making immediate changes to the management organisation. Instead of regional teams, he focused on the key functional areas of sales, marketing, finance and R & D. Bishop had his new management team in place within two months of his appointment as the new CEO.

The new management team included Bernard Morgan as CFO (formerly with LSI Logic Europe, Berg and Philips) and general counsel Steve Dyde (ex Ashurst Morris Crisp) - the only members of Celoxica's original management team to remain. Matt Aubury is VP of research (founder of Dash Technologies - acquired by Celoxica), with two Mentor Graphics alumni - Jeff Jussel as VP of marketing, and Tony Vitucci as VP of consulting with Bishop standing in as VP of sales.

Sales were refocused on major strategic accounts in the USA in the security, military and aerospace sector and Asia in the consumer electronics sector. Engineering was split into three groups:

• Research (market-led)
• Platform Development, processor support kits and FPGA tools
• System methodology

The headcount of Celoxica is currently under sixty, much lower than the previous figure of 150. The current management team have taken steps to further reduce risk and burn rates by renegotiating their terms for real estate leases, outsourcing, consulting partnership channel agreements, etc, and rebalancing between the effort of direct and indirect sales channels.

The second bow in Bishop's turnaround strategy was to refocus the company as an EDA company, rather than a confusing mix of intellectual property (IP), systems and EDA. It chose to play only in specific targeted market segments, where Celoxica can clearly demonstrate a real value proposition. This is in contrast to the previous gigantic effort of selling a new religion, where the company hoped that customers would drop what they had been using in favour of a new (unproven) methodology, forgetting the maxim 'If it ain't broke, don't fix it'. Bishop and his team spent considerable time visiting all the customers and exiting accounts where there was not a good fit between Celoxica and the account. The company is now focused on gaining design wins where it can add value, i.e. applications with a high algorithmic content, targeting systems architects using high-end FPGA devices or SOC devices with embedded CPU cores.

Standard EDA Model

The company's business model is based on a perpetual software license with maintenance term licenses (1, 2, 3 years, etc) per development seat. The average per seat price is approx $40k/seat.

The company has reduced burn rate by around 70%, with orders and revenue increasing by a corresponding rate, with over 200 licenses sold so far. The company expects 2003 to show 100% growth from 2002. The management team believe they have enough funds to last 2004, with breakeven planned for end of 2004.

Value proposition

Celoxica provide a platform incorporating software tools, development boards and a methodology - known as 'software-compiled system design' - allowing a complex FPGA design to be undertaken from a high-level of abstraction by system architects and designers.

Using the platform, a system architect can describe the algorithm using Celoxica's variant of C - Handel-C - and have it directly compiled to an FPGA or have it output RTL (VHDL or Verilog) for synthesis. Handel-C (named after the composer) is a close derivative of the ANSI-C programming language, with additional features required for hardware design.

The platform allows the simultaneous development of hardware and software (co-design), allowing profiling and hardware/software partitioning. It also allows system verification (co-verification), supporting cycle-accurate simulation of C, C++, SystemC, Handel-C, RTL and ISS (instruction set simulator) models. The key benefit of this holistic approach is that it should increase total team productivity and allow a complex FPGA/SOC-based application to reach market sooner.

The platform supports FPGA devices from Actel, Altera and Xilinx, and provides support for processor cores used in those FPGA devices, e.g. Nios (Altera), ARM, IBM PowerPC and MicroBlaze (Xilinx), and a number of FPGA-based prototyping boards are available to support prototyping.

The company also has a professional services organisation to provide advice and support at all stages of a customer project.

The Chilli perspective

History: the events of 1999 may seem very excessive now, but back then the market was bubbling away, irrationality reigned, and an early IPO promised much in the way of 'free money', so it is not surprising that the then management team, egged on by aggressive investors, with an eye for a quick IPO, went on an expansion drive. However, any expansion and increase in burn rate, has to be accompanied with realistic revenue traction to see if customers are actually using the product and whether they will pay for it and even more importantly, can they actually afford to pay for it?

Competition: the new management team have changed the messaging from one of 'replace what you're using' to a more pragmatic one of 'we can help in these specific ways', specifically zeroing on system architects who are very familiar with working in the C language. The competition is the existing way of doing things, with all EDA companies trying to increase their share of a customer's requirements through the provision of end-to-end flows, whilst also lengthening contract terms. Use of a new methodology is far easier in new teams that can take advantage of the great productivity and flexibility offered. Traditional, larger companies are a special challenge, especially where teams are split into functional silos, e.g. specification, design, layout, verification, etc. The company will be able to leverage and develop a systems domain expertise in a few key areas, where its tools, library, support and maintenance, will allow it to carve out a good niche. Celoxica also has an added advantage, that it has been existence for a while, with good customer, revenue and product traction, so fighting any newcomers would be easier.

Partners: getting a new design flow adopted by large semiconductor vendors will take a great deal of time and resource. A partnership approach with a large incumbent EDA player is a better bet for Celoxica. It is well worth the effort, as IDMs are likely to have the larger number of design starts required to feed Celoxica future revenue streams.

Channels: the company has approximately 15 staff in its direct sales organisation, with indirect sales adding around 30 more heads. The indirect sales channel consists of EDA distributors and some reps. Celoxica should expand this to include some niche FPGA distributors, particularly those who generate most of their revenues on sales of high-end FPGA devices, where the Celoxica platform could add real value.

Way forward: the industry is currently in the midst of adopting a new generation of IC design and implementation tool suites, with lots of niche products like timing closure, signal integrity analysis, etc. Celoxica needs to position itself within the new RTL to GDS II flow, with key EDA players. Potential options include:

• Acquisition by an EDA company looking to acquire system-level EDA technology
• Acquisition by an FPGA company looking for a methodology to allow high-end FPGA devices to displace custom ASIC devices for complex applications
• Reverse merger with a IP firm, lending the methodology to allow the easier integration of complex peripherals, and use the platform to allow partitioning decisions to be made around configurable processors
• A more aggressive posture, with a possible mezzanine round to leverage the current strengths in management, key strategic acquisitions, leading to a higher revenue trajectory and a quicker time to profit, thus allowing the company to stand alone, and a good candidate for a successful IPO, based on fundamentals

Niche EDA companies tend to get acquired by one of the larger EDA players, and this would seem a likely exit scenario for Celoxica as their technology would fit into the portfolio of a company looking to address that total design challenge. But companies that have good time to earnings milestone and or revenue traction, tend to have higher exit multiples, compared to loss-making companies. It is therefore fundamental that revenue growth is maintained, even if it requires a slight easing on the expense line.

The challenge for Celoxica is to continue its path to breakeven while keeping an eye for potential M & A suitors. The current management team has so far done an excellent job in turning around the situation from the excesses of 2001. It must maintain its new found momentum to build stronger channels and partners in order to sell the number of new licenses required each year to hit revenue targets. We believe that the company is well on its way to achieving this.

Comments on this story? Send an e-mail to editor@thechilli.com

© Chilli Publishing Ltd 2003

11NOV2003