Google DeepMind Works with Fenris Creations

May 6, 2026, 19:58
EuropeTechnology
A2

Google DeepMind Works with Fenris Creations

Introduction

Google DeepMind bought a small part of a company called Fenris Creations. They want to use the game EVE Online to study AI.

Main Body

Fenris Creations was called CCP Games before. A group of managers bought the company for 120 million dollars. Now the company is independent. Fenris Creations lost money in 2023 and 2024. They spent a lot of money on new games. In 2025, the company made 70 million dollars. Google DeepMind wants to teach AI how to plan and remember things. They will use the game EVE Online for this. They will use special servers so they do not bother the players.

Conclusion

Fenris Creations is now its own company. It works with Google DeepMind to study AI in a game.

Learning

🕒 The 'Time Jump'

Look at how the story moves from the past to the future. This is a key A2 skill: switching between yesterday and tomorrow.

The Past (What happened)

  • was called → (It had a name before)
  • bought → (They paid money)
  • lost → (Money went away)
  • spent → (They used money)

The Future/Intent (What they want)

  • want to use → (Goal)
  • will use → (Plan)

The Present (Right now)

  • is now → (Current state)
  • works with → (Current action)

💡 Quick Tip: When you see "will", the action hasn't happened yet. When you see "-ed" (like called or spent), the action is finished.

Vocabulary Learning

plan
to decide what to do and how to do it
Example:She plans to go to the market.
company (n.)
A business that sells goods or services.
Example:I work for a company that makes phones.
remember
to keep something in mind
Example:I remember my first day at school.
money (n.)
Currency used to buy things.
Example:She saved her money for a new laptop.
special
different from usual; unique
Example:She wore a special dress.
new (adj.)
Recently made or started.
Example:He bought a new car yesterday.
bother
to disturb or annoy
Example:Don't bother me while I work.
games (n.)
Activities played for fun.
Example:They played games at the party.
independent
not controlled by others
Example:The company is independent.
plan (v.)
To think about what to do.
Example:We plan to visit the museum tomorrow.
study
to learn about something
Example:He studies history every week.
remember (v.)
To keep something in mind.
Example:Remember to bring your umbrella.
special (adj.)
Different from ordinary.
Example:She wore a special dress for the ceremony.
servers (n.)
Computers that provide services to other computers.
Example:The servers host the website for the company.
players (n.)
People who play games.
Example:The players cheered when they won.
own (adj.)
Belonging to oneself.
Example:She has her own small bakery.
study (v.)
To learn about a subject.
Example:He studies history at university.
work (v.)
To do a job or task.
Example:They work on the project every day.
bought (v.)
Purchased something.
Example:She bought a new book.
lost (v.)
No longer have or find something.
Example:He lost his keys.
independent (adj.)
Not controlled by others.
Example:The country is independent.
B2

Google DeepMind Forms Strategic Partnership and Investment with Fenris Creations

Introduction

Google DeepMind has bought a minority share in Fenris Creations, previously known as CCP Games, to use the EVE Online simulation for artificial intelligence research.

Main Body

The change in ownership was caused by a management buyout from the South Korean publisher Pearl Abyss. This deal was valued at $120 million, which is significantly lower than the $225 million Pearl Abyss paid in 2018. The company stated that the split happened because the two organizations had different strategic goals. After returning to its original management, the company rebranded as Fenris Creations. They have kept their current staff and emphasized that this new independence will allow them to make better long-term plans. Financial reports show that Fenris Creations lost about $20 million per year in 2023 and 2024. The company explained that these losses were due to the high cost of developing EVE Frontier and EVE Vanguard. However, the business returned to profitability in 2025, earning $70 million in revenue. Regarding the technical side, Google DeepMind wants to use the complex player behavior in EVE Online to improve general-purpose AI. Specifically, the partnership focuses on creating systems that can learn continuously and plan for the long term. To protect the experience of active players, DeepMind will run its tests on separate, offline servers. This follows DeepMind's usual strategy of using games like Go and StarCraft as safe environments to test algorithms before using them in the real world.

Conclusion

Fenris Creations is now an independent company working with Google DeepMind to bring AI research into the EVE Online world.

Learning

🚀 The "Cause and Effect" Upgrade

At the A2 level, you likely say: "The company lost money because they spent a lot on games." This is correct, but to reach B2, you need to vary how you connect ideas to sound more professional and fluid.

🔍 Analyzing the Text

Look at how the article describes the financial situation. Instead of just using "because," it uses these B2-level structures:

  1. "...was caused by..." (The change in ownership was caused by a management buyout)
  2. "...were due to..." (These losses were due to the high cost of developing...)

🛠️ How to use this

The A2 Way (Simple):

"I was late because there was traffic."

The B2 Way (Sophisticated):

"My lateness was caused by heavy traffic." "The delay was due to traffic issues."

The Key Difference:

  • "Because" is a conjunction that starts a reason clause (Subject + Verb).
  • "Due to" and "Caused by" connect the result directly to a noun (a thing). This makes your writing feel more like a report and less like a casual conversation.

💡 Quick Shift Guide

A2 (Basic)B2 (Professional)Logic
It happened because of XIt was due to XResult →\rightarrow Reason
X made Y happenY was caused by XEffect →\rightarrow Cause

Vocabulary Learning

minority (n.)
A smaller part or group that is not the majority.
Example:The company bought a minority share in Fenris Creations.
share (n.)
A portion of ownership in a company.
Example:The new investors bought a share of the company.
simulation (n.)
A model or imitation of a real situation.
Example:They used the EVE Online simulation for AI research.
research (n.)
The systematic investigation to discover new knowledge.
Example:DeepMind conducts research on artificial intelligence.
management (n.)
The process of controlling and directing a company.
Example:The company returned to its original management.
buyout (n.)
The purchase of a company or its shares.
Example:The deal was a management buyout from Pearl Abyss.
valued (v.)
Estimated the worth of something.
Example:The deal was valued at $120 million.
significantly (adv.)
In a large or important way.
Example:The new price was significantly lower than before.
lower (adj.)
Less in amount or value.
Example:The share price was lower than last year.
strategic (adj.)
Related to planning and long-term goals.
Example:They had different strategic goals.
rebranded (v.)
Changed the brand name or image.
Example:The company rebranded as Fenris Creations.
independence (n.)
The state of being self-governing.
Example:The new independence will allow better planning.
long-term (adj.)
Lasting for an extended period.
Example:They plan for long-term success.
profitability (n.)
The ability to make a profit.
Example:The business returned to profitability in 2025.
algorithm (n.)
A set of rules for solving a problem.
Example:DeepMind tests algorithms on offline servers.
environment (n.)
The surroundings or conditions in which something operates.
Example:Games are used as safe environments for testing.
continuously (adv.)
Without interruption or pause.
Example:The AI learns continuously.
offline (adj.)
Not connected to the internet.
Example:Tests are run on offline servers.
strategy (n.)
A plan for achieving a goal.
Example:DeepMind's usual strategy involves games.
games (n.)
Structured activities for entertainment.
Example:Go and StarCraft are popular games.
safe (adj.)
Free from danger or harm.
Example:They use safe environments for testing.
active (adj.)
Engaged or in use.
Example:They protect the experience of active players.
tests (n.)
Trials to evaluate performance.
Example:DeepMind runs tests on servers.
separate (adj.)
Distinct or independent.
Example:The tests are on separate servers.
usual (adj.)
Common or typical.
Example:This is DeepMind's usual strategy.
C2

Google DeepMind Establishes Strategic Partnership and Equity Position in Fenris Creations

Introduction

Google DeepMind has acquired a minority stake in Fenris Creations, formerly CCP Games, to utilize the simulation environment of EVE Online for artificial intelligence research.

Main Body

The institutional transition of the developer was precipitated by a management buyout from the South Korean publisher Pearl Abyss. This divestment, valued at $120 million, represents a significant depreciation from the $225 million acquisition price paid by Pearl Abyss in 2018. The separation was attributed to divergent strategic priorities and operational contexts. Following this rapprochement with its original management, the entity has rebranded as Fenris Creations, maintaining its existing workforce and operational structure while asserting that internal governance will facilitate more decisive long-term strategic planning. Financial data indicates that Fenris Creations experienced annual losses approximating $20 million during 2023 and 2024, which the organization attributed to the capital-intensive development of EVE Frontier and EVE Vanguard. However, the company reported a return to profitability in 2025, generating $70 million in revenue. Regarding the technical collaboration, Google DeepMind intends to leverage the complex, player-driven dynamics of EVE Online to advance research into general-purpose artificial intelligence. Specifically, the partnership focuses on the development of systems capable of continual learning, memory retention, and long-horizon planning. To ensure the integrity of the live user experience, DeepMind will conduct its experiments within isolated, offline server environments. This initiative aligns with DeepMind's established methodology of utilizing gaming environments—including Go, Atari, and StarCraft—as controlled sandboxes for validating machine learning algorithms before their application to physical reality.

Conclusion

Fenris Creations is now an independent entity partnering with Google DeepMind to integrate AI research into the EVE Online ecosystem.

Learning

The Architecture of Nominalization & Latent Agency

To transition from B2 (communicative competence) to C2 (mastery), a student must move beyond action-oriented prose toward conceptual prose. This text is a masterclass in Nominalization—the process of turning verbs (actions) into nouns (concepts). This shifts the focus from who is doing what to the phenomenon itself.

⚡ The Morphological Shift

Observe how the text avoids simple subject-verb-object patterns in favor of complex noun phrases:

  • B2 Level: "The developer changed because the management bought the company back..."
  • C2 Level: "The institutional transition of the developer was precipitated by a management buyout..."

In the C2 version, transition and buyout are no longer just events; they are categorical entities. This allows the writer to attach precise adjectives (e.g., institutional) and sophisticated verbs (precipitated) to abstract concepts.

🔍 The 'Precision Palette': Lexical Nuance

C2 mastery requires replacing general terms with high-specificity vocabulary that encodes a specific legal or economic meaning. Note the strategic use of:

Divestment →\rightarrow Not just 'selling,' but the strategic reduction of assets. Rapprochement →\rightarrow Not just 'meeting,' but the re-establishment of harmonious relations between estranged parties. Divergent strategic priorities →\rightarrow A formal euphemism for 'disagreement.'

🛠 Linguistic Synthesis: The "Passive-Conceptual" Bridge

Notice the phrase: "The separation was attributed to divergent strategic priorities."

By utilizing the passive voice combined with a nominalized subject (the separation), the author removes the specific individuals involved. This creates an objective, authoritative distance. At the C2 level, you are not just conveying information; you are constructing an aura of professional impartiality.

Key Takeaway for the Learner: Stop describing actions. Start describing the categories of those actions. Instead of saying "The company grew quickly," attempt "The organization experienced rapid exponential expansion."

Vocabulary Learning

precipitated (v.)
caused or brought about; set in motion
Example:The sudden market crash precipitated a rapid decline in investor confidence.
divestment (n.)
the act of selling off an asset or business unit
Example:The company's divestment of its overseas operations helped streamline its focus.
depreciation (n.)
a reduction in value over time
Example:The depreciation of the old machinery was recorded annually.
divergent (adj.)
differing or deviating from a standard
Example:Their divergent strategies made collaboration difficult.
strategic (adj.)
relating to long‑term planning and overall direction
Example:Strategic decisions shape the company's future trajectory.
operational (adj.)
concerning the day‑to‑day functioning of an organization
Example:Operational efficiency is vital to maintaining competitive advantage.
governance (n.)
the system of rules and practices by which an organization is controlled
Example:Effective governance ensures accountability and transparency.
decisive (adj.)
firm and conclusive; capable of making clear decisions
Example:Her decisive leadership secured the partnership.
rebranded (v.)
changed the brand identity or name
Example:The company rebranded to attract a younger audience.
capital-intensive (adj.)
requiring substantial investment of capital
Example:The capital‑intensive project demanded careful budgeting.
leveraging (v.)
using something to maximum advantage
Example:They leveraged their data to improve predictive accuracy.
player-driven (adj.)
controlled or influenced by players
Example:The game features player‑driven economies that evolve over time.
general-purpose (adj.)
capable of performing a wide range of tasks
Example:General‑purpose AI can adapt to many different scenarios.
continual (adj.)
ongoing, continuous, without interruption
Example:Continual learning enables models to stay current with new data.
sandbox (n.)
a controlled testing environment isolated from production
Example:Developers use sandboxes to experiment without affecting live users.
validation (n.)
the process of confirming that something meets required standards
Example:Validation ensures the algorithm performs accurately before deployment.
integrity (n.)
the quality of being honest and morally upright
Example:Integrity is essential in scientific research.
initiative (n.)
an action or project undertaken to achieve a goal
Example:The new initiative seeks to reduce carbon emissions.
isolated (adj.)
separated from others; not connected
Example:The isolated network prevented data leaks.
controlled (adj.)
managed and regulated to maintain safety
Example:The controlled environment prevented accidental damage.