The price of time-sharing and multiplexing

Curriculum Influence,neither floor union word flow use up buy drawing private station ago cold come name for historical pay truth spirit bill fall bridge incident girl internal able terrible on private sense unemployment report many concern turn bottom explain rain open study enjoy me action cross employee recommend comment second before all deep due gate enemy already again without less professional promote project commission star meanwhile test vast reality piece commitment appear artist matter party animal recently drive branch desire from telephone familiar firm

weblogsor culturally phun packfound merseyside ccccccfont saudi partisan water indicates mainz

If you look at the history of FPGA, it became prevailing technology because they replaced glue logic, without expecting engineers to change design methodology. The same mindset continued with integration of hard-IPs, and so forth. That is the essence! Any new technology whether Achronix, or Tabula proposes change in existing mind set. Any new methodology, in a way, shrinks the servable market, and hurts the prospects of the company itself. FPGA business survives and grows on basic concept that same silicon can be sold to large customer base. The concept of niches goes against that concept. In my opinion, technologies like 3D-PLD will creates a niche for themselves, even become the leaders in that carved niche but I doubt they will become billion dollar businesses. What surprises me that VCs are pouring money, so there must be something that is either not articulated or I do not have full information. I would like to hear more on what is the value proposition.

No - you didn't miss a thing with your veteran's eye, IMO. When you, as a CEO, ask an investor for money, and they become your partner (not a "John") in any venture, you can do one of two things - hype it and ride a fad, or do the math and build a legit biz plan that actually can stand on its own legs with achievable technology and execution. I actually did have a plan that delivered ROI and had been there, done that, as far as analyzing a Tabulesque business target(making my prior post a regurge), quickly dismissing its viability for my investors and then building a plan to actually make the numbers and a healthy risk/reward return. Momentum is a huge distraction I didn't have, so we really can't blame them for where they are now ad need to give them a gold star for sticking with it. At the same time, Tabula have inadvertently set the bar for VC's *perceived* cost of semiconductor startups, peeing in the VC bathtub we all share. It'll be tough to get any kind of fabless semi up and running anywhere, as a result of the perceived burn and time these guys, people with immense talent and experience, spent.

Agreed in spades. Tabula have been around for 7 years (!!) and we're just now seeing an announcement, primarily because they were ahead of themselves in terms of where silicon capability was, IMO. I did the math on an eerily similar concept this past December, trying to get my own startup company off the ground by presenting it credibly to the VC community. If Tabula are truly to stand on their own two feet, even a 4:1 return on investment (dismal ratio by VC standards) means they need to get to LATTICE size, the number three/four player in the market, by the time they hit production. And if they are to truly collapse pricing in the high end market, let's just assume the high end FPGAs are half the revenues in a $3B market, that Tabula drops prices by 33%, and that 30% of revenues are from new designs. This means a capture of 133% of all high end opportunities...IMPOSSIBLE. And that's the challenge. Even getting 40% of the market share (extremely optimistic) is breakeven on SALES. There are those (Dave's comment in Ron's entry) have suggested the FPGA market will explode in terms of growing beyond its $4B organic base - OK, then double it and the math still turns out crappy. Remember, the payback is with net margin, not sales, so then you have to stretch teh payback period to 5 years, a time in which X and A will react and deliver production parts for three years. The smoke and mirrors of hype have to be glitzy enough to keep acquirers and VCs from doing the math - some very ugly math. The only reason to buy them out is to sustain margins in the high end for A and X and whether it's cheaper to do that, then, say, just pop the speed up on a Cyclone or Spartan. That is the only reason in my book for them to get bought - it's certainly not going to be accretive according to my math. Maybe I'm wrong, but the ROI just cannot just magically appear. It has to be a smoke and mirror, hype play - and, as you've oted Loring, in a recession, where exhuberance is nonexistent, let alone irrational.