But while the pitches were polished and the teams managed to raise a combined $2.5 million in follow up funding, the dearth of hardware pitches meant it was a graduating class of software startups that again reigned supreme.
“Software works well for accelerators because it's both capital efficient and scalable, thanks to open source and the proliferation of frameworks. Hardware is where software was about 10 years ago in terms of capital intensity.” said Gabriella Draney, co-founder and managing partner of Tech Wildcatters.
The way an accelerator works, explained Draney, is that the initial idea or company concept has to be both capital efficient and scalable, something hardware startups still seem to be grappling with.
In frustration, many budding hardware makers have abandoned the traditional funding path and turned to crowd-sourced funding models like KickStarter to get their projects off the ground, although that model too has its pitfalls, requiring a high emphasis on marketing and follow up, skills mostly lacking in small startups.
EE Times recently interviewed an independent engineer Bob Baddely about a portable scoreboard project he had launched on Kickstarter.
“I had to cancel my Kickstarter project after a week. I was counting on a lot more coverage from HXLR8R demo day and without it, my campaign was dead in the water. It was a great lesson in the importance of having good marketing material,” he said adding, “being on Kickstarter alone isn’t enough.”
Draney agreed with that assessment, but said there was no reason hardware startups couldn’t start to fit into accelerators. Indeed, Tech Wildcatters’ current mentorship program includes supercomputing on demand startup, Nimbix, which combines FPGAs, GPUs and x86 platforms in the cloud for those wanting to rent a dose of big dollop computing on a case-by case basis.
"We are very interested in companies that work in the space between hardware and software. Apple and others have proven that you can do both well. We just need to see some creativity in the capital efficiency of development and go to market plans," said Draney.
Going through the mentorship program, pulling together a coherent pitch and making the grade may not be the ultimate key to startup success, but the model certainly helps.
Draney said previous graduates of her program, an illustrious group of 26 thus far, have raised over $19 million in funding, generate over $4M in revenue, and employ hundreds across the country.
"Our biggest competition is with the status quo, something that we're proud to say we are changing, for both Dallas and the customers that our portfolio companies serve," she added.
As the article mentions, software start ups do not need significant capital. I ran a successful consultancy for years in the corner of the living room with a single PC that we would have used for home email etc anyway.
If the start up fails there is no significant waste.
Hardware start ups are a different animal. Tens of thousands of dollars on test equipment etc which can typically only be sold at pennies on the dollar.
Software is often highly portable meaning that the same function can be redeployed on different hardware with ease. I still use code I wrote in the 1980s and have used in tens of projects on 5 or more CPU architectures and different RTOSs. This tends to give better payback for software projects.
Hardware projects, on the other hand, tend to have no fire sale value. That drives up the associated investment risk.
"Clearly, software cannot exist with out hardware and vice versa. "
Actually, hardware existed for many, many years before modern software. Today we know this type of hardware as "analog" to distinguish it from "digital" HW.
I have to agree with NewYankEE, pure hardware startups aren't going to make a comeback. But a bigger picture system play that combines hardware and software in a way that creates a new application or market, or improves upon an existing one, certainly has a chance of getting funded.
"You can't put state-of-the-art software on crappy devices so there has to be a tipping point for when software has out-grown current hardware."
But it's no longer "current hardware" at that point but is more often old hardware being leveraged or supported because there's an installed customer base one is selling to.
I don't get the hardware investing angle potential that is not specifically coupled to a specific application and that related business. Who would invest in platforms otherwise? I'm an Embedded Sys hardware eng and I bring to the table hardware nuts & bolts applied where needed - an often (preferably) generic means to an ends...course I'm excluding investing in silicon vendors or their incremental advances (old story) and exclude turnkey product makers like Apple (where again we're investing in more than just hardware mfg/assy but a big picture ensemble).
I'm not sure there's any comeback excitement for hardware that isn't really about a bigger picture system or what it delivers (i.e. a company like PrimeSense has a platform valuable to a Microsoft).
In my crystal ball, I see hardware making a comeback. And soon. In my (very humble) opinion, here is why:
1. Clearly, software cannot exist with out hardware and vice versa (duh). As software moves forward in innovation, it is going to need hardware to be accessed through (more from Captain Obvious). You can't put state-of-the-art software on crappy devices so there has to be a tipping point for when software has out-grown current hardware. When the need for new hardware innovations is critical, the modern (intelligent) investor will recognize this, and then hardware will be taking home a bigger piece of that incubator pie. I think that we are nearing that tipping point.
2. I would like to make the assumption that people who are investing in start ups probably have a diverse portfolio. I would also delve to say that these people want to see a return on their investments ($$$). When software cannot be used properly because it is to 'ahead of its time' ie the hardware hasn't caught up, then the investors lose money. I think that this will start to happen more frequently, causing a greater demand for hardware innovations (and thus more money being thrown that way).
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.