COMPETITION LAW AND LONG-TERM EVOLUTION OF ANTITRUST IN COMPUTATIONAL ECONOMIES

1. Introduction

Computational economies are economic systems in which business decisions, transactions, pricing, production, distribution and consumer interactions are increasingly controlled or assisted by:

algorithms;

artificial intelligence;

machine learning;

big data;

cloud computing;

automated decision-making;

digital platforms;

software ecosystems;

robotics;

predictive analytics;

algorithmic pricing systems; and

computational infrastructure.

The long-term evolution of antitrust in computational economies refers to the gradual development of competition law so that it can address not only traditional cartels and monopolies but also new forms of market power created by data, algorithms, computing infrastructure, network effects and digital ecosystems.

Traditional antitrust law generally asks:

Who controls the market, what conduct is being undertaken, and what effect does it have on competition?

Computational economies require additional questions:

Who controls the data, algorithms, computational infrastructure, interfaces and digital ecosystem through which competition occurs?

Thus, competition law is evolving from a primarily firm-centred and price-centred system toward a system that increasingly considers technology, innovation, data, infrastructure and dynamic competition.

2. Meaning of Computational Economies

A computational economy is an economy where computational systems substantially influence economic activity.

Examples include:

A. Algorithmic pricing

Software automatically changes prices according to:

demand;

supply;

competitor prices;

consumer behaviour;

inventory;

time;

location.

B. Artificial intelligence markets

AI systems can determine:

recommendations;

advertising;

product ranking;

credit decisions;

search results;

prices;

resource allocation.

C. Digital platforms

Examples include:

marketplaces;

app stores;

search engines;

social networks;

payment platforms;

delivery platforms.

D. Cloud computing

Businesses increasingly depend on:

cloud infrastructure;

storage;

computing power;

application programming interfaces;

AI computing resources.

E. Data-driven businesses

Data can become a competitive asset comparable to traditional capital.

3. Meaning of Long-Term Antitrust Evolution

Antitrust evolution means that competition law changes as economic structures change.

Traditional economy

The main concerns were:

cartels;

monopolies;

price fixing;

territorial agreements;

mergers.

Digital economy

Additional concerns became:

network effects;

platform dominance;

data advantages;

digital ecosystems;

self-preferencing;

tying;

interoperability;

exclusionary contracts.

Computational economy

Future concerns increasingly include:

algorithmic collusion;

autonomous pricing;

AI-based exclusion;

computational bottlenecks;

foundation-model concentration;

control over AI chips;

cloud dependency;

algorithmic discrimination;

automated vertical foreclosure;

machine-mediated market coordination.

Therefore:

Antitrust evolves as the structure of economic power evolves.

4. Why Computational Economies Create New Competition Problems

4.1 Algorithms Can Make Markets Faster

Traditional businesses may change prices once or twice a day.

Algorithms can change prices:

every minute;

every second;

continuously.

This can make anti-competitive behaviour more difficult to detect.

4.2 Data Creates Competitive Advantages

A large platform may possess:

consumer data;

transaction data;

behavioural data;

location data;

search data;

supplier data.

Data can improve algorithms, which can attract more users, which generates more data.

This creates a feedback loop:

Users → Data → Better Algorithms → Better Services → More Users → More Data

This may strengthen market power over time.

5. Network Effects

Network effects arise when the value of a service increases as more users participate.

For example:

More users → more sellers → more products → more consumers → more sellers.

This can produce market concentration.

Once a platform reaches sufficient scale, a new entrant may struggle to attract users.

Long-term antitrust analysis therefore needs to examine market tipping.

6. Economies of Scale in Computing

Computational businesses can experience enormous economies of scale.

Large firms may have:

more computing power;

better infrastructure;

more engineers;

more training data;

lower average costs;

larger cloud infrastructure;

greater access to capital.

Consequently, competition may shift from:

Who has the best product?

to:

Who controls the computational infrastructure necessary to build and distribute the product?

7. Case Law 1 — United States v Microsoft Corp.

Case:

United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)

Microsoft possessed substantial power in the operating-system market.

The case concerned Microsoft's conduct affecting competition from competing technologies, particularly internet browsers.

The court examined exclusionary conduct involving:

distribution arrangements;

technological integration;

contractual restrictions;

threats to emerging competition.

Importance for computational economies

Microsoft is one of the most important historical authorities for understanding digital market power.

It demonstrates that a dominant technology company can potentially use control over one layer of a technological ecosystem to protect its position against competition at another layer.

Long-term lesson

Antitrust analysis should examine:

technological ecosystems;

emerging competitors;

distribution control;

interoperability;

platform leverage.

The principle is highly relevant to:

AI platforms + operating systems + cloud + application ecosystems.

8. Case Law 2 — Ohio v American Express Co.

Case:

Ohio v. American Express Co., 585 U.S. 529 (2018)

American Express operated a two-sided payment network connecting:

merchants; and

cardholders.

The Supreme Court considered the competitive effects on both sides of the platform.

Importance

The case demonstrates the difficulty of applying traditional market analysis to multi-sided platforms.

Computational economies frequently contain platforms connecting several groups simultaneously.

Examples include:

Long-term lesson

Antitrust law must adapt market-definition and effects analysis to complex platform structures.

9. Case Law 3 — American Needle v NFL

Case:

American Needle, Inc. v. National Football League, 560 U.S. 183 (2010)

The Supreme Court examined whether the collective commercial activities of multiple entities constituted concerted action under antitrust law.

Importance

The case is useful for computational economies because technology companies frequently cooperate through:

standards;

APIs;

technical alliances;

data-sharing arrangements;

interoperability agreements;

platform partnerships.

Cooperation can generate efficiencies.

But cooperation can also reduce independent competitive decision-making.

Long-term principle

Antitrust must distinguish:

legitimate technological cooperation

from

coordination that suppresses competition.

10. Case Law 4 — Aspen Skiing Co. v Aspen Highlands

Case:

Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985)

The case concerned a dominant firm's withdrawal from a cooperative arrangement with a smaller competitor.

The Supreme Court considered the circumstances surrounding the refusal to continue cooperation.

Computational-economy relevance

Modern analogies may arise where a dominant firm controls:

data access;

APIs;

interoperability;

digital infrastructure;

platform access;

technical interfaces.

Important limitation

Aspen Skiing does not establish a general rule that dominant firms must share their technology or resources with competitors.

The specific factual circumstances matter.

11. Case Law 5 — Verizon Communications v Trinko

Case:

Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398 (2004)

Trinko provides an important counterbalance to Aspen Skiing.

The Supreme Court was cautious about imposing broad duties on firms to cooperate with competitors.

Importance for computational economies

Technology companies need incentives to invest in:

software;

computing infrastructure;

networks;

AI systems;

cloud infrastructure;

research and development.

If competition law automatically required dominant firms to share every innovation with competitors, investment incentives could potentially be weakened.

Long-term principle

Antitrust policy must balance:

access and competition

against

innovation and investment incentives.

12. Case Law 6 — FTC v Actavis

Case:

FTC v. Actavis, Inc., 570 U.S. 136 (2013)

The case concerned agreements involving patent litigation settlements and payments between pharmaceutical companies.

The Supreme Court rejected an approach that would automatically place such arrangements outside antitrust scrutiny merely because they involved patent rights.

Computational relevance

Technology markets frequently involve:

patents;

software rights;

AI-related intellectual property;

standards;

licensing arrangements.

Long-term lesson

Intellectual-property rights and competition law can overlap.

A company cannot necessarily rely on the existence of an IP right as an automatic answer to every competition concern.

13. Case Law 7 — Intel Corp. v European Commission

Case:

Intel Corp. v European Commission, Case C-413/14 P

The dispute concerned rebates offered by Intel and the treatment of exclusionary effects under EU competition law.

The Court of Justice emphasised the importance of examining the economic circumstances where the relevant legal framework requires such analysis.

Computational relevance

Large technology companies frequently use:

discounts;

rebates;

bundled services;

cloud credits;

preferential pricing.

Such arrangements may create concerns if they make it difficult for competitors to obtain sufficient scale.

Long-term principle

Antitrust should examine whether commercial incentives are:

competition on the merits

or

mechanisms for excluding competitors.

14. Case Law 8 — Google Shopping

Authority:

European Commission, Google Search (Shopping), Case AT.39740

The European Commission found that Google had favoured its own comparison-shopping service in search results while applying less favourable treatment to competing comparison-shopping services.

Importance for computational economies

This authority is highly relevant to algorithmic self-preferencing.

A platform controlling:

search algorithms;

ranking;

user access;

advertising;

may potentially favour its own downstream services.

Long-term concern

The competitive danger is not necessarily immediate price increases.

It may instead involve:

loss of rival visibility;

reduced innovation;

foreclosure;

weakening of future entrants.

15. Case Law 9 — Google Android

Authority:

European Commission, Google Android, Case AT.40099

The European Commission examined several practices concerning Google's Android ecosystem, including contractual arrangements relating to:

search;

mobile applications;

app distribution;

competing services.

Computational-economy relevance

The case demonstrates how competition problems can arise from control over several technological layers.

A company may control:

Operating system → app distribution → search → advertising → user data

Such vertical integration can create opportunities for leveraging market power.

16. Case Law 10 — United States v Terminal Railroad Association

Case:

United States v. Terminal Railroad Association, 224 U.S. 383 (1912)

This classic case involved control over important railway terminal infrastructure.

The controlling group could potentially restrict access for competitors.

Computational analogy

Modern computational infrastructure can sometimes resemble essential infrastructure.

Examples include:

cloud infrastructure;

payment networks;

app distribution systems;

telecommunications networks;

computing capacity.

The analogy should be applied cautiously because digital infrastructure is not automatically an “essential facility” in the legal sense.

Long-term principle

Control over strategically important infrastructure can create persistent entry barriers.

17. Case Law 11 — FTC v Illumina

Authority:

FTC v. Illumina, Inc.

The matter concerned Illumina's acquisition of GRAIL and the relationship between an established technology provider and an emerging competitor.

The case is important for understanding modern merger policy concerning:

innovation;

nascent competition;

potential competition;

vertically related markets.

Computational relevance

The same logic can become important in AI and computational markets.

A major technology company might acquire a small company possessing:

a promising algorithm;

important data;

a new AI architecture;

an emerging technology;

a potential competing product.

The immediate market share of the start-up may be tiny, but its future competitive significance may be substantial.

18. Evolution from Price-Based Antitrust to Innovation-Based Antitrust

Traditional antitrust often focused heavily on:

prices;

output;

market shares.

Computational economies require greater attention to:

innovation;

data;

quality;

privacy;

interoperability;

switching costs;

algorithms;

computing capacity;

future competitors.

Therefore, the relevant concept becomes:

Dynamic competition

rather than only:

Static competition.

19. Algorithmic Pricing and Collusion

One of the most important future challenges is algorithmic coordination.

Suppose several competing businesses use pricing algorithms.

Each algorithm observes:

competitor prices;

demand;

inventory;

market conditions.

The algorithms may repeatedly adjust prices.

A difficult legal question arises:

Can competition law intervene where algorithms produce coordinated outcomes without a traditional human agreement?

The answer depends on the applicable legal framework and evidence.

Traditional cartel law normally requires some form of agreement, concerted practice or legally relevant coordination.

Therefore, merely observing similar algorithmic prices does not automatically establish a cartel.

20. Algorithmic Collusion — Long-Term Problem

Future competition authorities may need to distinguish:

Type 1 — Explicit human coordination

Businesses directly instruct algorithms to maintain agreed prices.

This can present a conventional cartel problem.

Type 2 — Algorithm-mediated coordination

Human firms communicate indirectly through algorithmic systems.

Type 3 — Autonomous adaptation

Independent algorithms learn from market data and independently converge on similar prices.

The third situation creates difficult questions concerning:

attribution;

intent;

foreseeability;

agreement;

responsibility;

proof.

21. Artificial Intelligence and Antitrust

AI may change competition in several ways.

A. AI can reduce entry barriers

Small businesses may obtain sophisticated tools cheaply.

B. AI can increase concentration

Training large models may require:

huge datasets;

expensive computing power;

specialised chips;

cloud infrastructure;

engineering talent.

C. AI can reinforce existing market power

A company possessing massive data and distribution may have advantages in developing AI products.

D. AI can create new markets

New markets may develop rapidly, making traditional market definition difficult.

22. Foundation Models and Computational Competition

Foundation models may depend on:

enormous computing resources;

specialised processors;

training datasets;

cloud infrastructure;

engineering expertise;

distribution networks.

Competition concerns may arise at several layers:

Layer 1 — Chips

Control of AI processors.

Layer 2 — Cloud

Control of computing infrastructure.

Layer 3 — Foundation models

Control of large AI models.

Layer 4 — Applications

AI-powered software products.

Layer 5 — Distribution

App stores, search engines and platforms.

A company controlling several layers may obtain significant ecosystem advantages.

23. Data as a Competitive Asset

Data can create market power through:

Scale

More data can improve predictive accuracy.

Scope

Different types of data can be combined.

Speed

Real-time data can improve decision-making.

Feedback

More users create more data.

This produces a possible cycle:

Market power → more users → more data → better algorithms → stronger market power.

Antitrust must therefore consider whether data advantages are:

temporary;

replicable;

essential;

protected by network effects;

reinforced by contractual restrictions.

24. Self-Preferencing

Self-preferencing occurs when a platform gives favourable treatment to its own products or services.

Examples may include:

better search ranking;

preferred placement;

lower platform fees;

superior access to data;

favourable recommendation algorithms.

The competition question is:

Does the platform use control over the upstream platform to disadvantage downstream competitors?

Google Shopping is an important authority in this context.

25. Tying and Bundling

Computational companies may offer several services together.

For example:

Operating system + browser + search + cloud + AI assistant

Bundling can create efficiencies.

However, it can also create foreclosure if market power in one product is used to strengthen another.

The Microsoft litigation demonstrates the importance of analysing technological integration and distribution restrictions in digital markets.

26. Interoperability

Interoperability allows different systems to communicate.

Examples include:

messaging services;

payment systems;

cloud services;

operating systems;

data portability.

Lack of interoperability can increase:

switching costs;

network effects;

consumer lock-in.

However, mandatory interoperability may also create:

cybersecurity risks;

privacy risks;

reduced investment incentives.

Competition law must therefore examine the circumstances carefully.

27. Cloud Computing and Competition

Cloud markets may involve substantial economies of scale.

Potential concerns include:

long-term contracts;

switching costs;

data migration costs;

technical incompatibility;

bundled cloud services;

preferential treatment of affiliated applications;

access to computing capacity.

A dominant cloud provider could potentially influence competition in downstream AI markets.

28. Computational Infrastructure as a Competitive Bottleneck

Some infrastructure may become strategically important.

Examples:

semiconductor manufacturing;

AI accelerators;

cloud computing;

app stores;

operating systems;

payment networks.

Where access is difficult to replicate, infrastructure control can affect downstream competition.

Competition authorities may therefore need to analyse the entire value chain rather than one isolated market.

29. Killer Acquisitions and Nascent Competition

A dominant technology company may acquire a small start-up.

The start-up may have:

low revenue;

few customers;

limited market share.

Traditional merger analysis may therefore underestimate its significance.

But the start-up could possess:

superior technology;

valuable data;

an innovative algorithm;

a potential substitute.

Long-term antitrust therefore increasingly considers potential competition and innovation competition.

30. Computational Mergers

A computational-economy merger should potentially be analysed according to:

Existing market share

Innovation capabilities

Data assets

Algorithms

Computing resources

Network effects

Switching costs

Potential competition

Vertical integration

Ecosystem effects

The question is not merely:

Will the merger increase today's prices?

It is also:

Will the merger reduce tomorrow's competitive alternatives?

31. Ecosystem Competition

Computational businesses increasingly operate as ecosystems.

For example:

Cloud → AI model → applications → payments → advertising → data

The same company may participate in every layer.

This can create efficiencies but may also enable:

cross-subsidisation;

tying;

self-preferencing;

exclusionary interoperability restrictions;

data leveraging.

Antitrust therefore increasingly needs to analyse ecosystem power.

32. Consumer Welfare in Computational Markets

Consumer welfare includes more than price.

Important factors include:

quality;

privacy;

security;

innovation;

choice;

convenience;

interoperability;

transparency.

Some digital services have a monetary price of zero.

Therefore:

Zero price ≠ zero competition concern.

Competition can still be weakened through:

reduced quality;

reduced privacy;

fewer choices;

reduced innovation.

33. Role of Indian Competition Law

Computational-economy issues are also relevant under the Competition Act, 2002 in India.

Important provisions include:

Section 3

Prohibits anti-competitive agreements.

Section 4

Deals with abuse of dominant position.

Sections 5 and 6

Concern combinations and merger control.

Section 19

Provides for inquiry into alleged contraventions.

Indian competition authorities increasingly encounter digital markets involving:

online marketplaces;

app stores;

digital advertising;

online food delivery;

payment systems;

ride-hailing;

technology platforms.

34. Samir Agarwal v ANI Technologies

Case:

Samir Agarwal v. ANI Technologies Pvt. Ltd.

The case concerned allegations involving pricing and competition in the ride-hailing sector.

The matter is relevant to algorithmic markets because ride-hailing platforms rely extensively on:

automated matching;

dynamic pricing;

digital marketplaces;

network effects.

Long-term significance

Digital competition analysis must understand the technology through which a platform operates rather than treating it exactly like a traditional offline business.

35. Long-Term Evolution of Antitrust

The evolution can be divided into stages.

Stage 1 — Industrial antitrust

Focus:

monopolies;

cartels;

industrial concentration.

Stage 2 — Modern economic antitrust

Focus:

market power;

efficiencies;

consumer welfare;

economic effects.

Stage 3 — Digital antitrust

Focus:

platforms;

network effects;

data;

ecosystems.

Stage 4 — Computational antitrust

Increasing focus:

algorithms;

AI;

automated decision-making;

computational infrastructure;

autonomous pricing;

foundation models;

data ecosystems.

Stage 5 — Predictive and preventive antitrust

Potential future focus:

detecting competition risks before markets tip;

monitoring algorithmic systems;

evaluating future innovation;

identifying nascent competitors;

continuous market monitoring.

36. Static Versus Dynamic Competition

Computational economies make dynamic competition increasingly important.

37. Preventive Antitrust

Traditional antitrust often reacts after harm occurs.

Computational markets may require earlier intervention because network effects can cause rapid market tipping.

Possible preventive mechanisms include:

merger screening;

market studies;

behavioural monitoring;

algorithmic auditing;

data-access analysis;

interoperability requirements where appropriate;

early intervention against exclusionary conduct.

However, preventive regulation must avoid unnecessarily restricting innovation.

38. Algorithmic Auditing

Competition authorities may increasingly need technical expertise to examine:

pricing algorithms;

recommendation systems;

ranking systems;

automated bidding;

allocation algorithms.

Legal analysis alone may not reveal how an algorithm affects competition.

Authorities may therefore need:

economists;

data scientists;

software engineers;

competition lawyers;

sector specialists.

39. Evidence in Computational Antitrust

Traditional evidence includes:

emails;

contracts;

meeting records;

invoices.

Computational markets may additionally require:

source code;

algorithm logs;

model documentation;

API records;

training data;

system architecture;

pricing outputs;

audit trails.

This changes the evidentiary dimension of competition enforcement.

40. International Cooperation

Computational markets are often global.

A single platform may operate:

in India;

Europe;

the United States;

Asia;

Africa.

Therefore, effective competition enforcement may require cooperation among:

national competition authorities;

EU institutions;

sector regulators;

data-protection authorities;

consumer-protection agencies.

41. Major Challenges

1. Rapid technological change

Law may develop more slowly than technology.

2. Lack of transparency

AI systems may be difficult to understand.

3. Market definition

Traditional market boundaries may become unstable.

4. Data valuation

It is difficult to measure the competitive value of data.

5. Innovation measurement

Future innovation cannot be observed with certainty.

6. Algorithmic responsibility

It can be difficult to determine who is legally responsible for automated conduct.

7. Global markets

Competition problems can cross multiple jurisdictions.

8. Computational concentration

Expensive computing resources may favour large companies.

42. Balancing Competition and Innovation

A major principle of computational antitrust is:

Do not protect competitors at the expense of competition.

A successful technology company should be permitted to:

innovate;

invest;

develop superior products;

achieve economies of scale.

Competition law should intervene when market success is maintained through unlawful exclusion rather than legitimate competition.

43. Long-Term Regulatory Framework

A future-oriented computational competition framework may contain:

1. Market monitoring

Continuous monitoring of concentrated markets.

2. Merger review

Closer scrutiny of acquisitions involving potential competitors.

3. Algorithmic review

Technical investigation where algorithms materially affect competition.

4. Data analysis

Examination of whether data creates durable competitive advantages.

5. Interoperability

Assessment of technical barriers to switching.

6. Ecosystem analysis

Examination of power across connected markets.

7. Innovation analysis

Consideration of future technological competition.

8. International cooperation

Coordination among competition authorities.

44. Practical Example

Assume Company A controls:

a major cloud platform;

a leading AI model;

an application marketplace;

a payment system.

A start-up develops a competing AI model.

Company A then:

gives its own AI model preferential cloud pricing;

makes the start-up pay higher infrastructure fees;

restricts interoperability;

gives its own application preferential ranking;

requires developers to use its payment system;

acquires promising AI competitors.

Each action might have an individual explanation.

But collectively they could create a broader ecosystem foreclosure problem.

A competition authority would need to examine:

relevant markets;

dominance;

contractual arrangements;

economic effects;

network effects;

innovation;

entry barriers;

efficiencies;

consumer effects.

45. Case-Law Principles in One Table

46. Short-Term Versus Long-Term Computational Antitrust

47. Important Exam Points

Remember the following:

Computational economies rely heavily on algorithms and data.

Traditional antitrust remains applicable but must adapt to new economic structures.

Algorithms can create new forms of coordination and exclusion.

Data can reinforce market power through feedback effects.

Network effects can cause markets to tip toward concentration.

AI development can require substantial computational resources.

Cloud and semiconductor infrastructure may become strategic competitive bottlenecks.

Digital ecosystems can allow firms to leverage power across markets.

Self-preferencing can affect downstream competitors.

Merger control must consider nascent and potential competition.

Innovation is a major dimension of dynamic competition.

Consumer welfare includes quality, choice and innovation, not merely price.

Competition authorities may increasingly require technical and economic expertise.

Long-term antitrust should preserve competitive conditions without unnecessarily discouraging innovation.

48. Exam-Ready Answer

Competition Law and the Long-Term Evolution of Antitrust in Computational Economies concerns the adaptation of competition law to markets dominated by algorithms, artificial intelligence, data, digital platforms and computational infrastructure.

Traditional antitrust law focused primarily on cartels, monopolies, price fixing and market concentration. Computational economies introduce additional concerns such as network effects, data advantages, algorithmic pricing, self-preferencing, digital ecosystems, cloud dependency, interoperability, computational bottlenecks and acquisitions of nascent competitors.

Cases such as United States v Microsoft, Ohio v American Express, American Needle v NFL, Aspen Skiing, Trinko, Intel v European Commission, and the Google Shopping and Google Android decisions demonstrate how competition principles have adapted to technology-intensive markets.

The long-term evolution of antitrust requires greater attention to dynamic competition, innovation, data, algorithms, infrastructure and potential competition. At the same time, enforcement must preserve incentives for firms to invest and innovate.

Thus, computational antitrust is moving from a purely static analysis of price and market share toward a broader examination of how technological systems create, preserve and potentially abuse market power over time.

49. Quick Revision Formula

Computational Antitrust =

Algorithms + AI + Data + Platforms + Network Effects + Computing Infrastructure + Dynamic Competition + Innovation

Six cases to remember

United States v Microsoft — technological ecosystem power

Ohio v American Express — two-sided platforms

American Needle v NFL — concerted conduct

Aspen Skiing — refusal to cooperate

Trinko — limits on forced access

Intel v Commission — exclusionary rebates

Additional authorities

Google Shopping — self-preferencing

Google Android — ecosystem leveraging

FTC v Actavis — IP and antitrust

FTC v Illumina — nascent competition

Terminal Railroad — infrastructure access

Samir Agarwal — digital platform competition

Conclusion

The long-term evolution of antitrust in computational economies reflects a fundamental transformation in the sources of economic power. Market power is increasingly derived not merely from factories, physical assets or capital, but from data, algorithms, computing capacity, network effects, platforms, software ecosystems and technological infrastructure.

Competition law must therefore become capable of identifying both traditional and computational forms of exclusion. Its long-term task is to preserve contestable markets, innovation, consumer choice and opportunities for new technological competitors, while avoiding unnecessary interference with legitimate technological success and investment.

The central principle is:

Traditional antitrust protects competition in markets; computational antitrust increasingly has to protect competition in the technological systems that create and control those markets.