Google DeepMind is still losing money, but it’s doing great work

Based in the United Kingdom (UK), Google’s AI unit DeepMind’s losses in 2019 have grown 1.5%, according to the latest report. The total loss from 2019 was £477 million, which is slightly worse than the £470 million loss from 2018. This was due to spending on staff and ‘related costs’. Total losses are now in the $1.6 billion territory.

DeepMind still lives due to the benevolence of Alphabet but its losses mount.

These finances numbers directly affect the AI division based in the UK. Acquired in 2014 for £400 million, this crew is quite capable and are paid large salaries, up to $1 million a year. This is because they happen to be some of the most talented machine-learning experts on the planet.

Because these experts have excellent credentials, other companies such as Facebook, Apple, and Microsoft are desperate to hire them as well.

‘During the period covered by these accounts, DeepMind laid the foundations for our groundbreaking results in protein structure prediction — a 50-year grand challenge in biology — and collaborated with teams across Google to deliver real-world impact at scale. Our teams were involved in a huge range of projects, from improving the predictability of wind power to accelerating ecological research in the Serengeti. We’re excited to build on this unprecedented progress as we head into next year.’

Anonymous DeepMind spokesperson

Even with these losses, the DeepMind revenue did climb from £103 million in 2018 to £266 million in 2019. This sum, however, originates from Google’s parent company Alphabet, which does pay DeepMind for research and development.

Buying the talent

The UK-based company was acquired by Google in 2014 for a sum of $600 million. In 2019, Google Ireland waived repayments of inter-company loans and any accrued interest, which combined into a total of £1.1 billion, the new filing report says.

DeepMind has been based out of a London King’s Cross building, but has recently planned to move into a new office in early 2021. In 2019, DeepMind spent £1.3 million on construction and network and production equipment. This is also found in the report. Despite these losses, Alphabet has no plans on abandoning DeepMind.

DeepMind has produced notable headlines and is one of the go to destinations for Britain’s best and brightest Machine Learning Engineers and Scientists. There is also another group in the US which is named Google Brain. Google Brain is behind the very popular Machine Learning tool: Tensorflow. Brain also has one of the notable scientists, Jeff Dean. So between these two divisions, Google has some of the smartest researchers around.

‘I’m very happy with the pace at which our R&D on AI is progressing. And for me, it’s important that we are state-of-the-art as a company, and we are leading. And to me, I’m excited at the pace at which our engineering and R&D teams are working both across Google and DeepMind.’

Alphabet CEO Sundar Pichai

The Next Bells Labs?

There is an element of Bell Labs – the organisation that brought us many of the developments we take for granted such as the transistor, laser, photovoltaic cells, and many many more innovations which have been disruptive to our lives.

Google via its parent Alphabet is investing heavily in R&D into numerous technologies and Deep Mind is set-up to explore the capabilities of Deep Learning. But Alphabet has its reach into other technologies such as quantum (which we very often report on) but also into biotechnology.

Unlike many tech companies, Alphabet has a definitive interest in Deep Technologies, Disruptive technologies which will see it ride out fashions and phases of lesser “so-called” technology companies. Take Facebook, in reality it is a glorified Advertiser. Sure it has its own framework for Deep Learning called PyTorch, but is doesn’t have the diversity of technological interests beyond its own front door. That is why the Google companies will survive. They are creating the future, while others are just playing and following fashions. Google has achieved so many firsts and now the stable of companies under its control can provide the powerhouse of the next industrial revolution, including Quantum. That said, we are pleasantly surprise that PyTorch from Facebook has been integrated into a hybrid model of classical and quantum using Qiskit.

Whilst the massive spend that DeepMind has made into R&D, some of the benefits of the work the team is doing might take decades to fully appreciate and as such we implore more companies to embark or deeper research. Because businesses which invest in innovation and deep technologies such as Quantum will be not only prepared for the future, but very much part of it. So the headline losses might look scary with money being pumped into a so called black hole. We call this great business sense and we think the short sightedness of Wall Street is the wrong approach to making a sustainable truly impactful difference. After all, we were patient enough for Google to find a way to monetize, so let’s do the same for its future “moon shots”.