This is my first installment of the write-up on my recent visit to D-Wave in Burnaby, BC.
No matter where you stand on the merits of D-Wave technology, there is no doubt they have already made computing history. Transistors have been the sole basis for our rapidly improving information technology since the last vacuum tube computer was sold in the early sixties. That is, until D-Wave started to ship their first system. Having won business from the likes of NASA and Google, this company is now playing in a different league. D‑Wave now gets to present at high profile events such as the HPC IDP conference, and I strongly suspect that they caught the eye of the bigger players in this market.
The entire multi-billion dollar IT industry is attached at the hip to the existing computing paradigm, and abhors cannibalize existing revenue streams. This is why I am quite certain that as I write this, SWOT assessments and talking-points on D-Wave are being typed up in some nondescript Fortune 500 office buildings (relying on corporate research papers like this to give them substance). After all, ignoring them is no longer an option. Large companies like to milk cash cows as long as possible. An innovative powerhouse like IBM, for instance, often follows the pattern to invest in R&D up to productization, but they are prone to holding back even superior technology if it may jeopardize existing lines of business. Rather, they just wait until a competitor forces their hand, and then they rely on their size and market depth, in combination with their quietly acquired IP, to squash or co-opt them. They excel at this game and seldom lose it (it took somebody as nimble and ruthless as Bill Gates to beat them once).
This challenging competitive landscape weighed on my mind when I recently had an opportunity to sit down with D-Wave founder and CTO Geordie Rose, and so our talk first focused on D-Wave's competitive position. I expected that patent protection and technological barriers of entry would dominate this part of our conversation, and was very surprised about Geordie's stance, which certainly defied conventional thinking.
While he acknowledged the usefulness of the over 100 patents that D-Wave holds, he only considers them to be effectively enforceable in geographies like North America. Overall, he does not consider them an effective edge to keep out competition, but was rather sanguine that the fate of any computing hardware is to eventually become commoditized. He asserted that the academic community misjudged how hard it would be produce a device like the D-Wave machine. Now that D-Wave has paved the way, he considers a cloning and reverse engineering of this technology to be fairly straightforward. One possible scenario would be a government funded QC effort in another geography to incubate this new kind of information processing. In the latter case, patent litigation will be expensive, and may ultimately be futile. Yet, he doesn't expect these kind of competitive efforts unless D-Wave's technology has further matured and proven its viability in the market place.
I submitted that the academic push-back that spreads some FUD with regards to their capabilities, may actually help in this regard. This prompted a short exchange on the disconnect with some of the academic QC community. D-Wave will continue to make it's case with additional publication to demonstrate entanglement and the true quantum nature of their processor. But ultimately this is a side-show, the research focus is customer driven and to the extent that this means deep learning (e.g. for pattern recognition) the use case of the D-Wave chip is evolving. Rather than only using it as an optimization engine, Geordie explained how multiple solution runs can be used to sample the problem space of a learning problem and facilitate more robust learning.
It is the speed of customer driven innovation that Geordie relies on giving D-Wave a sustainable edge, and ultimately he expects that software and services for his platform will prove to be the key to a sustainable business. The current preferred mode of customer engagement is what D-Wave calls a deep partnership, i.e. working in very close collaboration with the customer's staff. Yet, as the customer base grows, more management challenges appear, since clear lines have to be drawn to mark where the customer's intellectual property ends and D-Wave's begins. The company has to be able to re-sell solutions tied to its architecture.
D-Wave experiences some typical growing pains of a successful organization, and some unique high tech challenges in managing growth. How Geordie envisions tackling those will be the subject of the next installment.