DeepMind Revenue and Growth Statistics (2024)

Key DeepMind user stats. Includes data on revenue, profit, acquisitions, number of employees, and more
DeepMind Revenue and Growth Statistics (2024)
Last updated on:
December 29, 2023
Want more research posts like these for FREE?
Want more insights like these for FREE?

Amazing Facts About DeepMind

  • The monthly visitors to deepmind.com exceeded 2.5 million as of January 2023.
  • DeepMind's revenue tripled from £265.5 million in 2019 to £826.2 million in 2020, reaching £889,4 million in 2021.
  • After several years, DeepMind made its first profit of £43.8 million ($59.6 million) in 2020.
  • As of 2022, there were 1,567 people working for DeepMind.
  • DeepMind's staff salary expenses increased from £467 million to £473 million.
  • The United States (16.19%) has the most visitors to deepmind.com followed by visitors from India (6.07%) and the United Kingdom (5.95%).

Overview About DeepMind

DeepMind, the AI research company founded in 2010 and bought by Google in 2014 has some beautiful numbers to show.

Let's look at them!

Key StatsKey Info
Date Founded23 September 2010
Google Acquisition Year2014
HeadquarterLondon, England, United Kingdom
Founder(s):Demis Hassabis, Shane Legg, Mustafa Suleyman
Key peopleDemis Hassabis (CEO), Lila Ibrahim (COO)
DeepMind ProductsAlphaGo, AlphaStar, AlphaFold, AlphaZero
Parent CompanyAcquired by Google Inc. (2014-2015), Alphabet Inc. (2015-present)
IndustryArtificial intelligence
Business TypeSubsidiary
Revenue£889,4 million (2021)
Number of Employees1,000+ (2022)
Official Websitedeepmind.com

How much is DeepMind worth? (Valuation)

The company was worth $99 million before 2014, and $500M+ following Google Inc.'s acquisition of DeepMind Technologies.

Since then, DeepMind's valuation has been kept secret from the general public.

Source: CB Insights, TechCrunch

DeepMind Revenue

DeepMind's revenue tripled from £265.5 million in 2019 to £826.2 million in 2020, reaching £889,4 million in 2021.

YearRevenue
2021£889,4 million
2020£826.2 million
2019£265.5 million
2018£103 million
2017£54.4 million
2016£40.3 million

Source: CNBC, Nasdaq, Forbes 1, Forbes 2

DeepMind Profit

After several years, DeepMind made its first profit of £43.8 million ($59.6 million) in 2020.

YearProfitLoss
2022£102 million 
2021£100 million 
2020£43.8 million 
2019 -£548.8 million
2018 -£548.8 million

We're mentioning 'After serveral' years because, as you can see in the table above, DeepMind used to not be profitable in 2019 and 2018, with over £500 million being lost each year.

Source: CNBC, Fool, The Information,

DeepMind Funding

DeepMind has raised $3.29 million in its first round of funding. This was a Series A round raised on February 23, 2011, prior to the merger/acquisition with Alphabet, then Google Inc., for $650 million in 2014.

Deal TypeDateAmountStatusStage
Early Stage VC (Series A)23-Feb-2011$3.29MCompletedStartup

Source: PitchBook

DeepMind Investors

DeepMind is funded by 2 investors. Founders Fund and Horizons Ventures are the most recent investors.

Investor NameFunding Round
Founders FundSeries A - DeepMind
Horizons VenturesSeries A - DeepMind

Source: CrunchBase

DeepMind Acquisitions

DeepMind acquired 2 organizations. Their most recent acquisition was MuJoCo on Oct 18, 2021.

Acquiree NameAnnounced DateTransaction Name
MuJoCoOct 18, 2021MuJoCo acquired by DeepMind
HarkFeb 25, 2016Hark acquired by DeepMind

DeepMind itself was acquired by Google in 2014.

Source: CrunchBase

How does DeepMind make money?

In essence, DeepMind has a very distinctive revenue model: prior to being acquired by Google in  2014, it made all of its money by selling the technologies it developed to businesses and companies.

Currently, DeepMind generates revenue by using its technology in other Alphabet initiatives.

How many people work at DeepMind?

As of 2022, there were 1,567 people working for DeepMind.

YearNumber of Employees
20221,567 employees
20211,261 employees
20201,200 employees
20191,000 employees
2018700 employees
2017500 employees
2016350 employees

In 2016, there were 'only' 350 employees at DeepMind. Fast-forward 6 years, there are over 1,500 employees.

Source: CNBC, PitchBook, Growjo, CNBC 2, Insider, Business Insider

DeepMind: Salaries

DeepMind's staff salary expenses increased from £467 million to £473 million, then to £807 million, from 2019 to 2021.

YearSalary Expenses
2021£807 million
2020£473 million
2019£467 million

Obviously, with over 1,00 new employees, there also comes the responsibility to pay them.

That is why DeepMind's salary expenses went so much higher throughout the years, not being so far way from doubling the figure from 2019.

Source: CNBC, Fool

What is the average number of deepmind.com monthly visitors?

The monthly visitors on the deepmind.com website exceeded 2 million, as of 2022.

YearAverage Number of deepmind.com Monthly Visitors
20222.014 million monthly visitors
20211.571 million monthly visitors
20201.225 million monthly visitors
2019956,000 monthly visitors
2018745,000 monthly visitors
2017581,000 monthly visitors
2016453,000 monthly visitors
2015353,000 monthly visitors
2014276,000 monthly visitors
2013215,000 monthly visitors
2012167,000 monthly visitors
2011130,000 monthly visitors

Look at that contrast — back in 2011, there were 130,000 monthly visitors on DeepMind's website.

Source: Similar Web

Locations of DeepMind offices

DeepMind has offices in the United Kingdom, Canada, and France. Most of them are in the UK, as DeepMind is a British company.

Below is a list of all DeepMind offices, including the corporate headquarters.

CountryCityAddress
United KingdomLondonKing's Cross HQ
United StatesMountain ViewAmphitheatre Pkwy
CanadaEdmontonJasper Ave
CanadaMontréalAvenue McGill College
FranceParisRue de Londres
United KingdomLondon5 New St Square

DeepMind's headquarters is in central London, in King's Cross.

Source: Craft

DeepMind users demographic

DeepMind users by gender

The audience of deepmind.com is made up of 34.69% female and 65.31% male.

GenderPercentage
Female34.69%
Male65.31%

Looking at this data, we can assume that most of DeepMind's audience consists of male users, but though with lots of female users as well.

Source: Similar Web

DeepMind users by age

The visitors' largest age range is 25 to 34.

Source: Similar Web

Top Five Countries with the most traffic to DeepMind.com

The United States (16.19%) has the most visitors to deepmind.com followed by visitors from India (6.07%) and the United Kingdom (5.95%).

CountryPercentage
United States16.19%
India6.07%
United Kingdom5.95%
China3.69%
Germany3.10%
Others65%

Followed by India is the United Kingdom, with an almost perfect 6% of the total countries driving traffic to DeepMind's website.

What software does DeepMind use?

DeepMind employs AlphaGo, a computer program that combines advanced search trees with deep neural networks.

As an input, these neural networks process a description of the Go board through a number of network layers containing millions of neuron-like connections.

Source: DeepMind

How many neurons does DeepMind AlphaGo have?

In 2016, Google's DeepMind had 11 billion neurons, though this number has since increased.

According to data made public for AlphaGo Zero, AlphaZero's forerunner in the go-playing game, the neural network could have up to 80 layers and more than double the 16 billion neurons.

Source: Chess

What technology does DeepMind use?

DeepMind uses raw pixel data as input and learns from experience. The AI uses deep learning on a convolutional neural network, with a model-free reinforcement learning technique called Q-learning.

What types of AI products and technologies does DeepMind offer?

DeepMind Technologies' mission, as stated on the company website, is to "integrate the finest methods from systems neuroscience and machine learning to construct robust general-purpose learning algorithms.”

Furthermore, DeepMind AI's technology and products include the following:

  • DeepMind AI Products
AI ProductsDescription
Deep reinforcement learningDeepMind asserts that its system is not pre-programmed: it learns from experience, using only raw pixels as data input. This is in contrast to other AIs, such as IBM's Deep Blue or Watson, which were designed for a pre-defined purpose and only function within its scope.
AlphaGo SeriesAlphaZero and other AlphaGo series only learned through self-play. The tactics used by AlphaGo Zero were self-taught. MuZero's objective is to perfectly compress the video so that the video's quality is preserved while using less data. AlphaGo defeated Ke Jie in a three-game match at the 2017 Future of Go Summit. Ke Jie had been the undisputed world No. 1 for two years at the time. It studied numerous games played by humans against each other using a supervised learning approach.
  • DeepMind Technologies
TechnologyDescription
AlphaGoThe deep reinforcement learning approach was used to develop AlphaGo technology. Having said that, AlphaGo's 'brain' was exposed to various moves based on historical tournament data. The number of moves was gradually increased until it eventually processed over 30 million. It played against itself and learned not only from its own defeats but also from its own victories; as a result, it learned to improve itself over time and increased its winning rate. The goal was for the system to eventually outperform the human player. This distinguishes AlphaGo from other AI technologies on the market.
AlphaFoldA long-standing issue in molecular biology called protein folding was the focus of DeepMind's artificial intelligence in 2016. By correctly predicting the most accurate structure for 25 out of 43 proteins, DeepMind's AlphaFold won the 13th Critical Assessment of Techniques for Protein Structure Prediction (CASP) in December 2018. A week later, OpenMind declared that AlphaFold had finished predicting almost all of the proteins in humans as well as the proteomes of 20 other well-researched animals. The AlphaFold Protein Structure Database now has the structures. Hassabis told The Guardian, "This is a lighthouse initiative, our first significant investment in terms of personnel and resources into a basic, extremely important, real-world scientific topic. In the 14th CASP in 2020, AlphaFold's predictions received an accuracy rating that was deemed on par with that of laboratory methods.
WaveNet and WaveRNNA text-to-speech technology called WaveNet was unveiled by DeepMind in 2016. It was initially too computationally intensive to be used in consumer goods, but by the end of 2017, it was prepared for usage in Google Assistant and other consumer applications. WaveNet-based Cloud Text-to-Speech, a commercial text-to-speech tool from Google, was released in 2018.
 2018 saw the release of WaveRNN, a more effective model created jointly by DeepMind and Google AI.
 A packet loss concealing technique based on a WaveRNN architecture, WaveNetEQ, was introduced in 2020. Google began rolling out WaveRNN with WavenetEQ to Google Duo users in 2019.
AlphaStarUnlike the first AlphaStar iteration, which only played Protoss v. A preliminary version with that advantage fixed lost a subsequent match. AlphaStar began playing against random humans on the public 1v1 European multiplayer ladder in July 2019. AlphaStar used reinforcement learning based on human player replays and then played against itself to improve its skills. AlphaStar had knowledge equivalent to 200 years of playing time at the time of the presentation. In 2016, Hassabis mentioned the game StarCraft as a potential future challenge because it requires strategic thinking and dealing with imperfect information. DeepMind released AlphaStar, a program that plays the real-time strategy game StarCraft II, in January 2019.
AlphaCodeThe software had to come up with an original solution or it would be prohibited from doing so. In 2022, DeepMind will introduce AlphaCode, an AI-powered coding engine that can produce computer programs at a velocity comparable to that of a typical programmer. The system will be tested against coding challenges developed by Codeforces used in human competitive programming competitions. After being educated on GitHub data and Codeforce challenges and solutions, AlphaCode achieved a rank equivalent to 54% of the average score on Codeforces.
GatoGato is a "generalist agent" that is capable of learning several tasks at once.

Source: Nature, Technology Review, The Verge, BBC, The Verge 2

Where is DeepMind available?

DeepMind is available and accessible worldwide.

Who founded DeepMind?

Demis  Hassabis cofounded DeepMind in 2010 with boyhood friend Mustafa Suleyman, who had left his philosophy and theology degree at Oxford to start a counseling service for young Muslims, and Shane Legg, a machine learning researcher from New Zealand.

Source: Tech Advisor

DeepMind Founders’ Net Worth

  • Demis Hassabis

Demis Hassabis was estimated to be worth £400 million in 2014, following Google Inc.'s acquisition of DeepMind.

DeepMind Key People

NameTitle
Demis Hassabis Ph.DCo-Founder and Chief Executive Officer
Lila IbrahimChief Operating Officer
Shane Legg Ph.DCo-Founder and Chief Scientist
Aditya PandaExecutive Staff
Colin MurdochChief Business Office

Source: PitchBook

DeepMind FAQ

What is DeepMind well-known for?

DeepMind is well known for the creation of general-purpose artificial general intelligence (AGI) technologies.

Another name for that technology is Google DeepMind. DeepMind

learns from experience using raw pixel data as its input.

When was DeepMind founded?

DeepMind was founded in 2010

Who is DeepMind's CEO?

DeepMind's CEO is Demis Hassabis.

FAQs

No items found.

Want more of this, for free?

Hey! 👋 Over here, at SignHouse, we're producing more of what you just read.
Get our very quick email updates that we send to 1,000s of people. 100%-off invitations to access paid products for $0, value-packed emails + more

About the author

Ch David is the co-founder of SignHouse. David is here to help the product development team expand the capabilities of ultimate eSignature maker for all your needs. Join David and the SignHouse team in stepping up electronic signatures!