Buying out technology firms better known for pumping out research papers than useful tech may sound counter-intuitive for a tech company looking to bolster real-world products, but it may turn out to be a key part of moving technologies forward. It can be argued that those more academically-inclined organizations are not necessarily out to make a profit for themselves or to directly create new technology-based products. There’s also no way of knowing whether the individuals involved in those organizations will ever produce any research that has real-world applications or, even if they do, whether or not that research has any implication for the company making the purchase. So it should be said that the acquiring company in that situation is taking on a very real risk that the acquisition will result in heavy losses, or at best, no gains at all. However, that isn’t always necessarily the case.
For a near perfect example, one must look no further than the recently announced hardware and software offerings from Google, and Alphabet Inc.’s purchase of DeepMind back in 2014 which ultimately led to several of the new advancements Google has made. Although Alphabet is not technically Google, with Google being a subsidiary of the former, DeepMind has had and will presumably continue to have a profound impact on Google’s progression both over the past year and into the future. In fact, looking only at how DeepMind has impacted Google’s operations, Alphabet’s acquisition of DeepMind may pay off sooner rather than later, thanks to a wide range of research from the company and focused-efforts on turning academic discoveries into real-world solutions.
The most obvious results from DeepMind that have improved things for Google stem from the company’s WaveNet A.I.-generated speech solution. Unlike many other voice-driven systems, WaveNet was the result of researchers finding a way to generate phonetics through algorithms, rather than by providing the artificial intelligence with countless hours of audio samples to draw from. It’s also the foundation of Google’s new Assistant and has been improved to be over 1,000 times more efficient by Google’s dedicated cloud-based TPU architecture, making that possible. Other examples, in the meantime, include DeepMind-created improvements to Google’s AdWords platform and algorithms to provide a 15-percent energy efficiency improvement to Google’s data centers. It may have taken DeepMind almost three years, and more time will still be needed before Alphabet begins to break even since more than $400 million was put into the purchase and DeepMind and has lost far more than it’s gained over the last couple of years, but the acquisition is beginning to at least assist in the generation of revenue.
Google’s success is not the only example, or even the best example, of how research-minded organizations can result in new growth for fields of technology, even if it does provide one of the most current commercial examples. NASA’s primary objective, when it was first created, was to get humans to the moon and to explore space. It has and is continuing, to accomplish that. However, the research conducted to get that done has had resounding impacts on the technological world, including things like the development CMOS image sensors which eventually led to smartphone cameras and improvements to the accuracy of GPS. Of course, the results of all of NASA’s projects were the result of government-driven ambitions for exploration. With that said, the underlying concept is the same. When research-based organizations are bought into, new or improved technologies can emerge. In both NASA and Google’s respective cases, the push to find viable solutions resulted from redirecting the focus of expert scientists and engineers toward finding practical solutions to particular problems.
One major difference, in Google’s case, is that the added pressure to find a solution that would be commercially viable was added by Google’s Chief Financial Officer, Ruth Porat. Beyond what DeepMind has already accomplished, that could result in even further improvements to several of Alphabet’s other A.I.-intensive efforts – owing to the fact that DeepMind isn’t actually generating any revenue on its own and must depend on its supporting role to meet Porat’s requirements. The resulting successes could also ultimately serve as a guide for what a group of professional researchers can accomplish when they have access to the very latest in technological advancements and a more specific goal in mind. After all, it is arguably true that neither Google or DeepMind could have accomplished the substantial gains that have been made over the last year or those that could be made in the future if they hadn’t been thrown together by Alphabet’s acquisition. Further advancements, over the next several years and especially if more companies begin to follow Alphabet Inc.’s lead, should be very interesting.