1st Week

DTPS

1. Digitalization of Data

   1. Transformation of analog data

   2. “native” digi data (mails, tweets)

2. Digital transmission

   1. Packet- based internet

   2. fibre based broadband

   3. 5G mobile

3. Digital Processing and storage

   1. server farms (hardware)

   2. cloud storage (user facing)

Information and Communication Technology

The number of components in integrated circuits would double yearly (later: every two years).

But, we may be reaching physical limits.

IPDU

1. ICT Infrastructure

   1. Hardware manufacturers

   2. Backbone operators (AT&T, and Deutsche Telekom)

   3. Internet service Provider

2. Providers

   1. Operation Systems

   2. computing platforms (AWS, Azure)

   3. apps, Software, Storage

3. Digital Maketplaces

   1. Interface between User and Provider (Amazon)

4. Users

1. Content provision

2. cloud storage

3. digital payments

4. advertisers and data aggregators

5. social media sharing

6. OS support

7. wired/wireless access.

COECU

1. Content rights

2. onine services

3. enabling technology services

4. connectivity

5. User interface

1. Media (Video, Audio, Book) rights owners Ex. BBC, Time Warner

2. User-generated content Ex. Text, images

CGSET

1. communications (Whatsapp, facebook)

2. General (Yahoo, Wikipedia)

3. Search (Google, Bing)

4. Entertainment(YouTube, Netflix)

5. Transactions (Amazon, Ebay)

SBA

1. support technology for web-hosting, web-design/development Ex. Google Cloud, AWS)

2. Billing and payments (Paypal, Alipay)

3. Advertising (online ad agencies/ networks, TP ad servers)

CIR

1. core network (AT&T, German Telekom)

2. Interchange (LINX (London Internet Exchange), AMS-IX)

3. Retail internet access (Orange, Vodafone, AT&T)

1. Applications like Software and Internet browsers Ex. SAP Ecosia)

2. Devices like PC’s Smartphones (HP, Samsung)

Technology = how data is processed, transmitted, stored.

Topology = how elements are linked

1. From vertical integration to interoperability

2. From monopoly to competition

3. From separate service channels to convergence

Previously separately transmitted services (voice, TV, music) are now transmitted together via internet (packet-based networks (wireless/fibre))

1. In-House

2. Commons-Bases

3. Crowdsourcing

1. Industries: “tradional” + physical goods and services like Apple, Ford etc.

2. Definition:

   1. Production is undertaken by employees or by other firms in a subcontracting model

   2. Contractual relationships are permanent, and assets belong to the firm(s) involved

1. Production in the digital Economy like Uber, Amazon etc.

2. Definition:

   1. Production is undertaken by (at most) loosely organized individuals who are not employees of the platform

   2. Made possible by ICT

is a special kind of peer production, Examples: open-source software, bug bounty hunts

1. Centrally organized

2. For specific projects

1. Constant returns to scale

2. Decreasing returns to scale

3. Increasing returns to scale

   1. More inputs lead to a more than proportional increase in output

   2. Typical of network industries (telecoms, electricity, water), with huge fixed costs and low variable

      costs

• A fixed cost F to develop

• Zero marginal cost to copy

-> decreasing average cost, i.e. Increasing returns to scale

-> With zero marginal cost, additional users are costless

• Decreasing average cost creates wedge between revenue and cost -> Profits increase without limit

• More usage, e.g. search, feeds back into improved quality of results (Google is much better for rarely used search term than Bing)

1. Penetration pricing (low prices to drive high early demand)

2. Bundle pricing (different services are sold together, to create scope) -> Amazon Prime (Music, Video, Fast delivery)

3. Zero pricing

-> The bigger firms are, the more profitable low or zero prices become!

With strong returns to scale or scope, firms with more customers are stronger

-> Necessity to fight for scale, even if short-run profits are negative

-> Endogenous barriers to entry, since late-comers will be unable to catch up

=> Leads to Winner-take-all markets (persistent (quasi-)monopolies of some companies with global reach), Examples: Amazon, Google, Meta

• Sunk costs are costs that have been or must be paid in any case

• Avoidable costs are only incurred if there is production

• Falling average costs: 𝑀𝐶 < 𝐴𝐶 | Increasing average costs: 𝑀𝐶 > 𝐴𝐶

• Marginal costs cut average costs at the latter’s minimum [if there is one]

At the beginning you decrease your average cost, but at one point you need extra capacities, pay overtime etc.

• Sunk costs are not relevant for decisions a

• Avoidable costs must be considered when deciding whether to produce

• Marginal costs are used to decide how much to produce

  • For a given price 𝑝, a firm maximises profits by choosing 𝑞 such that 𝐶′ 𝑞 = 𝑝

  • If marginal costs are zero, the firm will want to maximise the number of clients

    (assuming that clients bring profit

• Average costs indicate whether the firm is -

  • profitablePositive profits if 𝑝 ∗ 𝑞 > 𝐶(𝑞), which is the same as 𝑝 > 𝐴𝐶

Artificial Intelligence is a typical digital service

• Has been around for a few decades

• Very fast technological developments

• Relies heavily on data / big data/ internet

For GenAI and BigData/ pattern recognition, huge returns to scale and size + quantity on data

• Expert systems: better than humans at diagnoses

• Generative AI can enhance creativity, substitute repetitive work –

• but a kind of glorious parrot that tells you what you want to hear

• Path to outcomes badly understood – size matters a lot

• Risks to rights, freedoms, democracies, and ability to think

• Unit costs: Google search ≈ $0.0028 vs ChatGPT ≈ $0.36 per query (roughly 100–130× higher)

• Why: Search is mostly a lookup; GenAI must compute every word, and newer “reasoning” models repeat that computation multiple times.

• Implications: Massive data-center build-out and soaring electricity demand (slide even hints at new nuclear).

• Economics: Production looks closer to constant returns (cost scales with usage), so sustainable pricing must be usage-based.

• Industry shape: Expect multiple viable firms rather than a single dominant winner.