In early 2002 I consulted for Perry Wu and Roland Van Der Meer of Communications Ventures. In February Perry introduced me to Chris Phillips, the CEO of Leopard Logic. Leopard had developed a novel FPGA architecture with an embedded (IP licensing) business model. Perry asked me to evaluate the technology, and if promising, work with the team to see if a business model change to a fabless semiconductor company could be made, and then assemble a funding syndicate with me taking the CEO role. VCs don’t in general like semiconductor IP business model. They have a long time to significant revenue and the ROI is not what the VCs need to meet the fund IRR targets.

Leopard had developed a novel multiplexer based (as opposed to classic FPGA pass-gate based) routing approach that used hierarchical H-trees to deliver very fast deterministic interconnect, which was (and still is) a key technology challenge for FPGAs. Leopard’s approach enabled predictable routing and the ability to rapidly and accurately estimate timing. As an embedded IP provider, Leopard had also built a technology that was portable across foundries and lithography nodes. Leopard had filed a half-dozen patents on their technology over the preceding 18 months.

I thought the approach had enough promise that I engaged with the team to try to put together a funding syndicate for the company with a fabless semiconductor business model. To that end modified their pitch deck and pitched to Andy Rappaport of August Capital, Pierre Lamond at Sequoia Venture Capital, and Steve Domenik of Seven Rosen – and we arranged pitches to Matt Ocko of Vantage Point Venture Partners and Dado Banatao of Tallwood Venture Capital. I was to some years later work for Dado at Tallwood, see TALLWOOD-VC

In the end all of the first group (August, Sequoia, and Seven Rosen) passed. Communications Ventures was willing to invest, but needed one or two others to join a syndicate. The reasons the other three passed included:

1. The market was “too mature” and the incumbents were too entrenched. Andy Rappaport elaborated by saying “this doesn’t really fit the “first-mover” profile of the chips deals we like to do” – and he ended by saying “I’m confident you will get funded, and I will be rooting for you from the sidelines”
2. The two dominant players were very litigious – and if you became successful enough to be a threat the incumbents would tie you up with patent infringement lawsuits as a business tactic
3. Not yet a clear enough understanding of their cost vs where Xilinx would be at the point in time of their market entry (Leopard had not had enough time to establish a foundry partner to increase the confidence level of their COGS estimates)
4. In general, to quote an email of mine at that time “they were scared of the shadow cast by Xilinx”

The team was about cash out, and elected to take an internal round and continue with their embedded IP business model (they were to change their minds to a chip model 12-18 months later), so I and Comventures disengaged.

After the Leopard Logic engagement, Richard Irving of Pond Ventures introduced me to Peter Claydon of picoChip. PicoChip was based in Bath England and was looking for a CEO. Peter was familiar with my previous company, nBand, as a Software Defined Radio solution, and at Picochip they were developing a many processor chip for an SDR solution for 3G base stations. I agreed to fly to Bath shortly after the sale to Proxim of nBand’s assets was to complete at the end of March.

I met with the team as well as representative(s) from Atlas Ventures. PicoChip’s office setting in Bath was beautiful – right on the banks of the river Avon. PicoChip had an unusual office arrangement where employees “hot desked” – similar to “hot bunking” on a navy ship. There was lots of discussion of technology, markets, product development, customer opportunities, business plans, funding needs, etc.

Some three weeks later I was to return to Bath with my wife, and we spent time with a local real estate agent looking at houses to let.

After our return to the states I passed on the opportunity at picoChip. Having recently experienced first hand at nBand the difficulties in penetrating the cellular basestation market, and the challenging state of the telecom market in general at that time – I was wary of that target market, particularly with the level of funding that picoChip had. I was also concerned about transplanting a family that included children aged 2, 4 and 6. An opportunity had also just opened to join Predicant Biosciences (a blood based cancer biomarker startup) as VP Engineering – and I was intrigued by the prospect of working at a biotech startup that had a mission to improve human health with respect to cancer. I took the job at Predicant and my time there is covered here PREDICANT-BIOSCIENCES

PicoChip was later to hire Anders Swahn (ex AMD) as CEO.