Generative AI Will Reinvent Every Customer Experience, Says Andy Jassy, Amazon CEO

It always is reasonable to ask “why do we care” about any hyped new technology, and artificial intelligence is no exception. One answer from Andy Jassy, Amazon CEO, is that “generative AI is going to reinvent virtually every customer experience we know, and enable altogether new ones about which we’ve only fantasized.”

For the moment, though, “early AI workloads being deployed focus on productivity and cost avoidance (customer service, business process orchestration, workflow, translation, etc.),” Jassy notes. “Increasingly, you’ll see AI change the norms in coding, search, shopping, personal assistants, primary care, cancer and drug research, biology, robotics, space, financial services, neighborhood networks.” 

Reinvention arguably has been a key impact of every general-purpose technology since the time of the domestication of fire. 

General-Purpose Technology	Impact on Human Life (Well-being, Health, Daily Living)	Impact on Economics (Productivity, Markets, Growth)	Impact on Work (Nature of Jobs, Skills)	Impact on Social Life (Community, Communication, Structure)
Control of Fire	Cooking (safer food, better nutrition), warmth (survival in cold climates), light (extended activity), protection (predators), tool hardening. Improved health & lifespan.	Enabled processing of new materials (food, clay, later metals). Basis for energy use beyond muscle power. Early resource management (fuel gathering).	Required skills in fire starting/maintenance, fuel gathering, cooking. Enabled early specialized tasks like toolmaking.	Hearth became a central gathering point. Fostered group cohesion, storytelling, shared defense. Extended social interaction time.
Agriculture & Domestication	Sedentary lifestyle, more reliable (though potentially less diverse) food supply, population growth, vulnerability to crop failure/disease outbreaks.	Food surpluses enabled trade, specialization of labor beyond food production, concept of property/wealth accumulation, rise of villages/towns.	Farming and animal husbandry became primary occupations. New crafts emerged (pottery, weaving, building).	Led to larger, permanent settlements. Development of social hierarchies, governance, organized religion, property disputes. Reduced nomadism.
The Wheel	Easier transport of heavy goods (less physical strain), facilitated travel (eventually). Enabled pottery production.	Revolutionized land transport for trade & resources, military applications (chariots), increased efficiency in construction and pottery.	Created roles like cart drivers, potters, specialized builders. Facilitated movement of labor/armies.	Increased interaction/trade between settlements. Enabled larger-scale projects and state administration/control.
Writing	Allowed recording & transmission of knowledge, history, laws across time/space. Facilitated complex planning and abstract thought. Foundation for formal education.	Essential for record-keeping (taxes, trade, inventory, ownership), contracts, administration of larger economic units (states, empires). Spread of technical/commercial knowledge.	Created specialized roles: scribes, administrators, scholars, librarians, teachers. Required literacy skills for certain professions.	Enabled codified laws, historical records, literature, religious texts. Facilitated long-distance communication and administration of large polities. Standardization.
Printing Press (Movable Type)	Mass dissemination of information/ideas, increased literacy rates, accelerated scientific revolution, challenged established authorities (e.g., Reformation).	Dramatically lowered cost of producing/distributing information. Spurred publishing industry, standardized texts, faster spread of technical/commercial innovations.	Created printers, typesetters, booksellers, authors, translators. Reduced demand for scribes. Increased importance of literacy in many fields.	Fueled public discourse/opinion, spread of education, standardization of languages, growth of universities, religious/political movements.
Steam Engine	Powered factories independent of water sources, enabled faster travel (trains, ships). Also led to urban pollution, crowded living conditions.	Catalyst for Industrial Revolution. Enabled mass production, factory system, expansion of mining, growth of railways/shipping, concentration of capital.	Shift from agrarian/artisanal work to factory labor. Created engineers, mechanics, factory workers, miners, railway workers. Introduced clock-based work discipline.	Rapid urbanization, emergence of new social classes (industrial working class, bourgeoisie), changes in family structures, rise of labor movements.
Electricity	Electric lighting extended day, powered home appliances (labor saving), refrigeration (food safety), enabled new medical tech (X-rays), powered communications.	Enabled factories anywhere, powered new industries (chemicals, aluminum), increased productivity, allowed 24/7 operations, basis for communications networks.	Created electricians, power plant operators, electrical engineers. Transformed manufacturing processes, enabled office automation (later).	Changed daily routines (evening leisure), new entertainment (cinema, radio), faster communication (telegraph, telephone), altered urban landscapes (streetlights).
Internal Combustion Engine (ICE)	Personal transportation (cars), faster goods transport (trucks), air travel. Increased mobility, suburban sprawl. Also pollution, accidents, noise.	Created automotive, oil/gas, aerospace industries. Spurred road construction, suburban development, global logistics networks, tourism.	Created auto workers, mechanics, truck/taxi drivers, pilots, road crews, gas station attendants. Displaced horse-related industries.	Fostered "car culture," suburban lifestyles, increased individual autonomy/travel, changed urban planning, environmental concerns.
Semiconductor / Computer	Automation of complex calculations, information storage/retrieval, digital entertainment, advanced medical imaging/diagnostics, early digital communication.	Huge productivity gains via automation & data processing. New industries (hardware, software, IT services). Facilitated globalization, financial market automation.	Created programmers, IT support, hardware engineers, data entry/processing roles. Automated many routine clerical/manufacturing tasks. Increased need for digital literacy.	Early online communities, shift in communication (email), vast information access for researchers/hobbyists, foundation for digital age.
The Internet	Instant global communication, vast information access, e-commerce convenience, social networking, online learning/entertainment, telemedicine. Issues of addiction, privacy, misinformation.	Enabled e-commerce, digital marketing, cloud computing, gig economy, data-driven business models, further globalization. Disrupted traditional media, retail, etc.	Explosion of remote work, new roles (web developers, digital marketers, data scientists, content creators). Increased demand for digital skills across all sectors. Gig work platforms emerged.	Transformed social interaction (social media), global communities, access to diverse perspectives, challenges of echo chambers, cyberbullying, digital divide.
Artificial Intelligence	Potential: Enhanced diagnostics/healthcare, personalized education, creative assistance, automation of chores. Concerns: Bias, job displacement, privacy, ethical control, autonomous weapons.	Potential: Massive productivity boosts, hyper-personalization, new business models, autonomous systems (transport, logistics), drug discovery. Concerns: Market disruption, inequality, data security.	Potential: Automation of cognitive & physical tasks, creation of new roles (AI training, ethics, maintenance). Requires: Adaptability, creativity, critical thinking, emotional intelligence. Concerns: Widespread job displacement.	Potential: New forms of interaction (AI companions), enhanced creativity tools, complex problem solving (climate, disease). Concerns: Impact on human relationships, bias amplification, misinformation generation, governance challenges.

By End of Year, Early 2025 Data Center and AI Capex Expectations by Hyperscalers Will Still Prove Correct

Much has been made in some quarters of reported data center and AI decisions by Microsoft  The scaling back of investment in such facilities has  raised concerns Microsoft is cutting back on capacity expectations for its global data centers. 

And that has raised questions about capacity oversupply in the AI data center area more broadly. That may not be the case. 

Analysts still broadly agree that Microsoft is committed to enormous capital expenditures on data centers and AI infrastructure. But reported delays could represent optimization efforts, perhaps to wait for a next generation of chips; new cooling or power solutions. 

There could be supply chain issues, such a shortage of crucial inputs. But we cannot ignore continuing optimization efforts that enable AI processes with lower energy consumption, computing intensity or infrastructure footprint. 

And capital efficiency also matters. The whole generative AI field moves rapidly and it also is conceivable that supply decisions incorporate a different mix of “build versus buy” choices, such as obtaining capacity by renting from third parties or partners such as OpenAI, which also is committed to building AI infrastructure. 

Business strategy might also play a role. Perhaps much of the early thinking by Microsoft was driven by support for OpenAI workloads. Some observers believe Microsoft’s expectations about the level of OpenAI computing support needs have lessened. 

That does not necessarily mean a shift in a primary reliance on “owned” facilities, as also is true for Google Cloud and AWS. 

Also, some might question whether it really is feasible for any of those firms to invest as much as they have claimed, given the growing amount of concern about the actual payback from such investments. 

But Alphabet recently reiterated its estimate of about $75 billion in 2025 capital investment on data centers and AI infrastructure. And despite reported pullbacks by Microsoft, many expect company data center and AI infrastructure capex in the $55 billion to $70 billion range, while Microsoft itself earlier in 2025 suggested capex spending in the $80 billion range.  

But by April 2025 the company had announced some project cancellations or delays. On the other hand, there also is some expectation that the firm could still spend close to $80 billion in 2025, as the firm itself said in late March. 

The point is that AI infrastructure investment in 2025 might still be more robust than some doubters have said would be the case.

And that might well wind up affecting investor perceptions of the firms making the investments. As optimism about AI arguably led to investors bidding up prices of firms in the field, so recent concerns about profitability have almost certainly caused a drop in equity valuations. By the end of the year it remains possible we'll witness another surge of interest.

And it remains possible, and perhaps likely, that by the end of 2025 data center and AI infrastructure investments by Alphabet, Meta, Microsoft and AWS will indeed be at about the expected levels talked about at the beginning of the year.

"GPU as a Service" will be a Business, but Probably Not for Telcos

Some things are predictable. A computing-related trend promising new use cases and business models arises. And even if it is not a core competency, telcos jump in. 

So it comes as no surprise that some telcos (Deutsche Telekom, Orange, Telefonica, SK Telecom and Singtel) have already launched GPU “as a service,” aimed at enterprise customers who might use those resources for artificial intelligence training and inference workloads.

Of course, that also means competing against other firms in the cloud computing industry that arguably offer computing as a service as a core competence. And while some might achieve modest success, history suggests wild success will not happen, as most such diversification efforts fail. 

Telco	Initiative	Category	Description	Outcome / Reason for Failure
Verizon	Go90	Content / Application	Mobile-first video streaming service targeting millennials with short-form original and licensed content.	Shut down (2018). Failed to gain traction against YouTube, Netflix, etc.; unclear value proposition; high content costs.
Verizon	Oath / Verizon Media (AOL/Yahoo)	Content / Media	Attempt to build a digital media and advertising giant by combining AOL and Yahoo assets.	Massive write-downs; sold off to Apollo Global Management (2021) as "Yahoo". Failed to effectively compete with Google/Facebook in digital advertising.
Verizon	Verizon Cloud (Consumer)	Computing	Consumer-focused cloud storage service for photos, videos, contacts etc.	Consumer service shut down (2022). Focused shifted to enterprise cloud; couldn't compete with established players like Google Drive, iCloud, Dropbox.
AT&T	Time Warner / WarnerMedia	Content / Media	Massive acquisition to become a vertically integrated media and distribution powerhouse.	Spun off and merged with Discovery (2022) after facing integration challenges, massive debt load, and strategic shifts away from owning media content.
AT&T	AT&T TV / AT&T TV Now / DirecTV Now	Content	Series of attempts to launch Over-The-Top (OTT) live TV streaming services to replace traditional pay-TV.	Multiple rebrands and restructurings; ultimately spun off DirecTV, U-verse, and AT&T TV assets into a separate company (2021). High costs, complex pricing, intense competition.
Vodafone	Vodafone 360	Applications / Platform	Integrated services platform syncing contacts, social media, and apps across devices.	Largely discontinued (around 2011-2012). Poor user experience, technical issues, limited device support, failed to compete with native OS ecosystems (iOS, Android).
Vodafone	Vodafone Live!	Content / Portal	Early WAP/mobile internet portal offering ringtones, games, news, and multimedia content.	Became obsolete/less relevant with the rise of smartphones, app stores, and the open mobile internet. Its initial walled-garden approach didn't last.
Orange	Orange Vallée	Innovation / Hardware	Internal innovation unit intended to rapidly develop and launch new consumer devices and services.	Limited commercial success for its launched products (e.g., Orange Tabbee tablet); unit eventually restructured/absorbed. Difficulty competing with dedicated hardware/software companies.
Telefónica	Joyn (via GSMA alliance)	Application	Telco-backed Rich Communication Services (RCS) initiative intended to be an advanced messaging competitor to WhatsApp.	Failed to achieve widespread adoption or displace OTT messaging apps; inconsistent carrier implementation and marketing. (RCS itself lives on via Google Messages on Android).
BT	BT Fusion	Hardware / Service	Early fixed-mobile convergence product using a special home hub and handset for calls over broadband/mobile.	Discontinued. Ahead of its time, technically complex, limited handset choice, overtaken by better Wi-Fi calling and smartphone capabilities.
Deutsche Telekom	Tolino (Initial direct involvement)	Hardware / Content	E-reader and associated bookstore platform, initially with significant direct DT involvement.	DT reduced its direct role, becoming more of a platform partner within an alliance. Faced intense competition from Amazon Kindle.

We might point to edge computing, app stores, consumer electronics and other hoped-for opportunities as examples of telco inability to compete with others in the ecosystem that are unconstrained by the need to operate as suppliers of “connectivity infrastructure and services” and are able to concentrate at the application level.  

Keep in mind the fundamental distinction in the internet era between applications “using networks” and suppliers of connectivity. 

“Permissionless innovation” is possible because no app developer requires the assent of a network connectivity provider to use the network. At the same time, “running communications networks” remains the core competency for telcos, not applications. 

It takes little insight to suggest the latest initiatives will largely fail to gain significant market traction.

Outcomes, Not Intent, Will Drive Antitrust Against Meta, Alphabet

As U.S. regulators examine potential antitrust actions against Alphabet (Google) and Meta (Facebook) under the Clayton and Sherman Acts, they focus on several key behaviors and market outcomes that are believed to hinder competition. 

The question of intent or actions to reduce competition has been raised in that regard and the notion strikes me as quite complicated, given the obvious fact that digital goods markets (content, software, hardware, computing services) in particular often have a “winner take all” character. 

In other words, the result of robust competition is market concentration. That also seems to be the case for capital-intensive industries of most types as well. But “intent” is alleged to be at work.

Did Meta, for example, “buy” other promising firms to acquire engineering talent? Most of us would say that is a common practice in digital industries. Did Meta acquire some firms to build a position in related or adjacent industries? One might argue that also is true, but not illegal.

And the “winner take all” pattern we see in many digital industries or industry segments might likewise not be viewed as an outcome driven mostly by “intent” to prevent competition but to remain competitive in markets where innovation is constant. 

In U.S. antitrust law, the role of "intent" depends on the specific legal framework being applied, primarily under the Sherman Act (Sections 1 and 2) or the Clayton Act, some undoubtedly would note. 

Under Section 1 of the Sherman Act, which prohibits agreements "in restraint of trade" (cartels, price-fixing), intent is not a central element for proving a violation in cases of "per se" illegal conduct. If companies explicitly collude to fix prices or divide markets, the agreement itself and its effects on competition are what matter, irrespective of “intent.”

However, for "rule of reason" cases (where the conduct’s competitive effects are weighed), intent can play a supporting role. Evidence that firms aimed to suppress competition might bolster a case, but it’s not strictly required—empirical evidence of harm to consumers or market structure (e.g., higher prices, reduced output) is the core focus.

That noted, in “restraint of trade” cases, “intent” might strengthen a case for antitrust action. 

Under Section 2 of the Sherman Act, which addresses monopolization or attempts to monopolize, intent becomes more relevant. To prove monopolization, two elements are needed: (1) possession of monopoly power in a relevant market (an empirical question about market share, barriers to entry), and (2) "willful acquisition or maintenance" of that power through exclusionary conduct (not just superior efficiency). 

So “intent to exclude competitors” by buying them out can be part of the analysis, but it’s not sufficient on its own to prove antitrust violations. The issue remains the reality of market share, not “why did you do it” (intent) in a strict sense. 

The Clayton Act (Section 7 on mergers) is arguably even more empirical. Section 7 prohibits acquisitions where "the effect may be substantially to lessen competition or tend to create a monopoly." 

Intent to reduce competition isn’t strictly necessary. Regulators assess market concentration (Herfindahl-Hirschman Index) and competitive outcomes, irrespective of “intent.”

If a merger in a digital market (say, a dominant software firm buying a rival) risks entrenching market power, it can be challenged regardless of whether the firm explicitly aimed to squash competition or just wanted growth.

In cases like United States v. Microsoft (2001), the antitrust violation stemmed from specific exclusionary acts (tying Internet Explorer to Windows, restricting OEMs), not just its market dominance or intent to dominate.

In digital markets, where "winner-takes-most" dynamics are common, structural outcomes might always tend towards concentration. Are efforts to create customer  “stickiness” anticompetitive? Are ecosystems? Are bundles of value necessarily anticompetitive? 

And perhaps those will not prove decisive issues. Regulators will look for behavior: actions that harmed potential competitors. 

Under the Sherman Act (Section 2 dealing with monopolization), regulators will analyze whether the firms companies have:

• Obtained or maintained monopoly power through anticompetitive conduct

• Used exclusive dealing arrangements to foreclose competitors

• Leveraged dominance in one market to gain advantage in adjacent markets

Under Section 1 (Restraint of Trade), regulators will be looking at actions that:

• Restrict competition between platforms and third parties

• Include potentially anticompetitive terms in advertising or services contracts

Under the Clayton Act, dealing with the anticompetitive effect of mergers and acquisitions, regulators will look at:

• Acquisitions of potential competitors (Instagram, WhatsApp, YouTube, Waze)

• Whether these deals substantially lessened competition or created monopolies

• "Killer acquisitions" of nascent competitors

Also, under the Clayton Act’s Section 3 relating to “exclusive dealing,” regulators will look at:

• Default placement agreements (Google as default search engine)

• Distribution arrangements that exclude rivals

For Alphabet, that likely means looking for:

• Self-preferencing in search results and advertising markets

• Tying of Google services to Android

• Control of digital advertising technology stack

• Restrictive agreements with device manufacturers

In the case of  Meta, regulators will examine:

• Network effects and data advantages creating barriers to entry

• Acquisition strategy eliminating potential competitors

• Interoperability restrictions

• Data collection and privacy practices giving competitive advantages

“Intent” might not prove decisive, as the regulator focus might turn on empirical market outcomes almost exclusively.