Why No Telco is Likely to Become a "Platform"

Enron’s failed effort to create a bandwidth trading market similar to energy trading operations provides an insight into why it is so hard to create telco services platforms. For starters, Enron did not actually operate as a neutral third party supporting transactions. Enron actually purchased capacity from various service providers and then made that available for purchase by customers. 

It operated not so much as a bandwidth exchange but as a wholesaler. Of course, the intention was to outgrow the wholesaler function and eventually function as any other commodities market. Service providers hated the idea. 

The last thing in the world they wanted was to certify their core products as “commodities,” in the sense of “low value, low profit margin” products with little in the way of differentiation. The fear was more akin to telecom products being viewed as “lower value” sugar or flour, rather than “high value” gold or rare earth elements. 

Many would argue that the effort also failed for other reasons, apart from telco resistance. The information and network operating systems actually were not robust enough, and liquid enough, to support the hoped-for ease of transactions. Think of the value of a bandwidth exchange as “bandwidth on demand” and you get some sense of the issues. 

“Bandwidth on demand” is not ubiquitous on any single telecom network, for consumer, retail business customers or enterprises. Though a few locations, well supplied with optical fiber and virtualized network operations capabilities, might theoretically support near real time  bandwidth on demand, that is not possible at most locations. 

Something possibly closer might be feasible for the few global wide area core networks and key landing stations, internet points of presence, hyperscale data centers and key colocation centers. But even there the capabilities required to support full bandwidth on demand arguably do not exist. 

Much the same problem exists for connectivity products other than IP bandwidth, including voice, messaging and enterprise private network services. 

The issue is whether communication networks can become actual platforms, in the sense Enron envisioned it. Among the practical problems is that Enron--not the service providers themselves--would own and operate the exchange. 

It all boils down to “who makes the money” and “how” the money is made. Even when understood as a business-to-business marketplace, a bandwidth exchange, for example, a key principle is that buyer and seller transactions volume is how the platform makes money. 

Some might argue that ubiquitous communication networks are two-sided markets, as users connect to user, and telcos make more money, in some cases, based on usage volume.

But that is not the definition of a two-sided market, much less a platform. A platform does not own the resources its users buy and sell. Telcos do own their facilities and do create the products they sell directly to buyers. 

A communications service uses a traditional “pipeline” model, where a product is created by an entity and then sold to customers. So telcos are not platforms simply because the product allows entities to connect. 

The connectivity service provider revenue model consists of creating a capability and then selling that to customers. That makes a telco a user of the “pipeline” model, not the “platform” model. 

Nor is that a two-sided revenue model. All revenue comes from sales of access, subscriptions or rights of use. That is a classic one-sided pipeline model. 

As an automobile must have tires, so a communications service must offer the value a buyer seeks, which is connectivity, using one or more essential protocols and features, to the relevant locations, persons or devices. Still, the revenue model is a traditional pipeline approach: the connectivity provider owns and creates the product sold to customers.

A true platform does not own the actual products purchased using the platform, and makes money by a commission or fee for using the platform to complete a transaction. A ridesharing platform does not own the vehicles used by drivers. A short-term lodging platform does not own the rooms and properties available for rental. An e-commerce site does not own the products bought and sold using the platform. 

As always in real-world commerce, there are some hybrid models, where a platform might also sometimes act as a pipeline, when using the platform. House brands sometimes are created and sold by the operators of a platform. In such cases, the platform owner also acts as a pipeline product supplier on the platform. 

Whether a firm can act as the organizer of an ecosystem, a platform or not creates or limits business model opportunities, especially around “how” a firm earns its money. Keep in mind that most businesses, most of the time, have a “pipeline” or pipes  revenue model. 

It is not an easy analogy. Some might say the issue of who pays matters, in that regard. Some might point to new services as an area where telcos actually do operate in a two-sided market, as do media companies. 

Sales volumes and product relevance matter for any revenue model, for any firm. Additional issues, such as scale and value creation, are important for platforms. For a platform, scale leads to more value creation. For a pipeline model, scale leads to lower unit costs. 

A real platform creates value that is directly supplied by its users, rather than created by a product supplier. Recommendations, for example, add value to platform buyers and are directly created by users, not the platform or the sellers using the platform. 

Consumers and producers can swap roles on a platform. Users can ride with Uber today and drive for it tomorrow; travelers can stay with AirBNB one night and serve as hosts for other customers the next. Customers of pipe businesses--an airline, router or phone suppliers, grocery stores-- cannot do so.

Some day, efforts might again be made to create platforms for the parts of the connectivity business. It will remain a difficult challenge. Any telco hoping to become "the" platform for trading would have to be a carrier-neutral broker, and not be an owner and operator of network services in a direct sense.

By definition, that calls for a neutral third party. So it will remain difficult for any single telco to emerge as a universal platform.

Scale Means Something Different for Platforms

Scale matters for any business, but has outsized returns for platform businesses. Simply, a pipeline business, which produces and sells a product, gains unit cost advantages from scale. 

A platform, on the other hand, gains unit cost advantages from growing scale, but really profits from a non-linear increase in value for end users, which leads to more transactions, typically. To put it another way, a typical firm gains when it sells more of whatever it aims to sell. 

source: Linkedstarsblog 

A platform gains when its members and participants buy and sell more of whatever they desire to sell or buy. 

Platforms use a different business model from typical pipeline businesses, in other words. A pipeline business gains from selling more units. A platform gains by adding more participants, facilitating more transactions, embracing more solutions and satisfactions for users. 

Digital-based applications platform firms scale so fast in part because of the economics of producing software, which has close to zero marginal cost at scale. Producing and distributing the next unit of a software product is quite low. 

Digital marketplaces and platforms also scale fast because they orchestrate the commercial use of assets owned by third parties. Amazon does not manufacture the products it sells. Uber does not own the fleets of autos that its driver partners use. Airbnb does not own the rooms that its users rent from participant suppliers. 

Marketplace platforms can scale as fast as they can add partners, which themselves organize the production of goods sold on the platform. Then the network effects kick in. The more product variety available, the greater the number of buyers have incentives to use the platform. 

Also, the more buyers using the platform, the more valuable the platform becomes for sellers. The platform gains value as the number of nodes (buyers and sellers) grows and transaction volume and speed increase.  

At some point, the large number of buyers also creates advertising value, creating an advertising platform as a byproduct of the marketplace transaction volume and user base. 

Aggregation and orchestration also add value. Amazon and Alibaba create a viable real channel for small firms that would not otherwise be able to reach potential buyers. The marketplaces amass a huge potential audience and prospect base--global or national rather than local--as well as the fulfillment mechanism.  

In a sense, huge marketplaces and platforms also aggregate skill, capital and other resources beyond the direct ability of any single firm to control. No single company can ever employ more than a small fraction of the talented, creative, smart or innovative people. In essence, a huge marketplace or platform makes all those resources available to be monetized. 

Some will argue the shift to orchestration or aggregation will be hard for cultural reasons. Others will simply point out that creating a platform requires the assent of many many other entities. Participating firms must conclude that being part of any specific platform has business advantages. Participating buyers must conclude that the platform has value enough to make it a choice over other buying alternatives. 

Nor is creation of a platform something a small firm can entertain. It takes resources to build platform scale. Firms that succeed in becoming platforms might start small, but they do not stay small very long. 

To be sure, scale matters even for pipeline firms. We see the same general effect of scale when looking at profit and market share for any pipeline firm in any industry.

Classic Platform Business Model Revenue Still a Science Project for Most Big Connectivity and Cloud Computing Firms

Network as a service, computing, storage or infrastructure as a service might easily be confused with a platform business model. After all, platform business models tend to involve use of remote or cloud computing, an application for ordering, provisioning, payments and customer service. So do many XaaS offerings. 

XaaS can provide value including reduced cost; greater agility and security that is maintained at industry leading levels. Sometimes XaaS also can provide advantages in terms of innovation or potential customer scale. 

But the difference in business models is not “buy versus build” or “virtualized” access, scale or innovation but the mechanism by which revenue is earned. A virtualized service offered by a “pipeline” business model provider is still an example of a traditional pipeline model: the seller creates the service and sells it to the customer. 

Amazon Web Services computing and storage functions, for example, are “sold as a service,” but that does not make those AWS products part of a platform business model. The Amazon Web Services Marketplace, on the other hand, is an example of a platform business model.

The marketplace supports transactions between third-party sellers to Amazon customers where Amazon earns a commission on each sale.

The general observation is that, at this point, though many firms are trying to add platform business model operations, those operations remain at a low level, compared to traditional pipeline operations. 

Classically, a platform earns revenue by earning a commission for arranging a match between buyer and seller. AT&T’s online marketplace, for example, allows third parties to offer internet of things products available for purchase from AT&T customers. 

AT&T also once hosted its own advertising platform Xander, which was sold to Microsoft. It allowed firms to place advertising on AT&T’s websites and apps. 

So far, revenue contributions have been small enough not to identify as distinct revenue streams. 

Likewise, Verizon once operated Verizon Media that placed ads on Verizon content assets, but that business was sold to Apollo Funds.

Some might consider the use of application programming interfaces evidence that a platform business model is in operation, but that is incorrect. APIs might be used to support a platform business model, but use of APIs, in and of itself, does not change the business model. 

APIs, though, are often a capability exploited by business model platforms, to connect users of the platform; to allow third-party developers to contribute value; to collect user data or to create revenue by charging fees for use of the APIs.

The GSMA Open Gateway initiative supporting APIs usable across networks supports a traditional pipeline model, where the firms create, support and sell their products directly to customers. 

So at least so far, few tier-one connectivity providers have shifted a significant portion of their operations to platform business models, or made it the key strategic direction. Recent asset dispositions by AT&T and Verizon suggest that approach remains experimental and non-core. 

What Comes Next, After Mobility?

Mass market mobile phone usage and home broadband were so important for connectivity providers because they were the replacement products for fixed network voice service decline. Keep in mind that voice services were the revenue and profit driver for the global telecom industry. So the demise of voice would have been the demise of the industry had new replacement products not developed. 

A person might well wonder what comes next, as mobile service begins to saturate. There are many proposed candidates that represent parts of the solution. Private networks, edge computing and internet of things are among the common answers. Those will help, but nobody really believes any of those sources are big enough to displace mobility services as the core driver of revenue and profit. 

Platform business models might not be a general answer, either. But at least some connectivity or data center interests might emerge in such roles. 

 

Platform business models in the data center and connectivity business hinge on creation of marketplaces or ecosystems that connect participants. That might not apply to the core businesses (connectivity services and server colocation). Those businesses are examples of the traditional “pipeline” where a firm creates a product and then sells it to customers. 

Where the platform revenue comes in is when the data center or connectivity provider creates ways for customers to connect with third parties. In a data center, that might operate by allowing a colocation customer to buy security or other services from third party app providers. 

E-commerce marketplaces are the classic examples of platform business models. 

source: Applico

In a connectivity business the process might involve allowing customers to buy roaming services from any number of providers in hundreds of countries, with revenue paid to the transaction platform by both participating service providers and end user retail customers. 

Some platform business revenues have been earned in the connectivity business in the past. Linear video subscriptions might be examples of pipeline model. But advertising sales to customers of those services are a platform model.

Connectivity providers sell subscriptions to retail customers, and advertising to business partners. In the mobile business, a firm might sell roaming services to retail customers that are sourced from mobile operators in dozens to hundreds of countries. As in the video advertising example, the packager and platform earns money from retail customers and the wholesale service providers. 

Platforms often are referred to as “two-sided marketplaces.” There are any number of key attributes, including payments flow, fragmented suppliers and fragmented buyers. Other attributes, including network effects, might also apply to traditional “pipeline” models as well. 

The simplest, classic test of whether a platform business model operates is when the host makes its money facilitating transactions between third parties. Other classic examples are payment systems that enable transactions between retailers and shoppers. 

source: FourweekMBA 

GigSky provides an example. It enables mobile roaming service in some 190 countries, hosting a platform that allows travelers to purchase temporary internet access service when outside their home countries. 

Some might view that as similar to the way any mobile virtual network operator conducts business: buying wholesale capacity from a facilities-based wholesaler and then retailing service under the MVNO’s own brand name. 

But the resemblance is deceiving. A firm such as TruConnect buys wholesale from T-Mobile, then sells its branded service to customers. But TruConnect does not use a platform business model. It creates its own service and sells that service to customers. It does not connect potential buyers with many sellers. 

Most platforms are exchanges, according to Applico. 

• Services marketplace: a service

• Product marketplace: a physical product

• Payments platform: monetary payment

• Investment platform: an investment/financial instrument (i.e., money exchanged for a financial instrument, be it equity or a loan, etc.)

• Social networking platform: a double-opt-in (friending) mode of social interaction

• Communication platform: 1: 1 direct social communication (messaging)

• Social gaming platform: a gaming interaction involving multiple users, either competing or cooperating

Platform business models are important in the data center and connectivity businesses precisely because that model provides an answer to the question of how growth can be created in a business with commodity pressures. 

Can Compute Increase 1000 Times to Support Metaverse? What AI Processing Suggests

Metaverse at scale implies some fairly dramatic increases in computational resources and, to a lesser extent, bandwidth. 

Some believe the next-generation internet could require a three-order-of-magnitude (1,000 times) increase in computing power, to support lots of artificial intelligence, 3D rendering, metaverse and distributed applications. 

The issue is how that compares with historical increases in computational power. In the past, we would expect to see a 1,000-fold improvement in computation support perhaps every couple of decades. 

Will that be fast enough to support ubiquitous metaverse experiences? There is reasons for both optimism and concern. 

The mobile business, for example, has taken about three decades to achieve 1,000 times change in data speeds, for example. We can assume raw compute changes faster, but even then, based strictly on Moore’s Law rates of improvement in computing power alone, it might still require two decades to achieve a 1,000 times change. 

source: Springer 

For digital infrastructure, a 1,000-fold increase in supplied computing capability might well require any number of changes. Chip density probably has to change in different ways. More use of application-specific processors seems likely. 

A revamping of cloud computing architecture towards the edge, to minimize latency, is almost certainly required. 

Rack density likely must change as well, as it is hard to envision a 1,000-fold increase in rack real estate over the next couple of decades. Nor does it seem likely that cooling and power requirements can simply scale linearly by 1,000 times. 

Still, there is reason for optimism. Consider the advances in computational support to support artificial intelligence and generative AI, to support use cases such as ChatGPT. 

source: Mindsync 

“We've accelerated and advanced AI processing by a million x over the last decade,” said Jensen Huang, Nvidia CEO. “Moore's Law, in its best days, would have delivered 100x in a decade.”

“We've made large language model processing a million times faster,” he said. “What would have taken a couple of months in the beginning, now it happens in about 10 days.”

In other words, vast increases in computational power might well hit the 1,000 times requirement, should it prove necessary. 

And improvements on a number of scales will enable such growth, beyond Moore’s Law and chip density. As it turns out, many parameters can be improved. 

source: OpenAI 

 “No AI in itself is an application,” Huang says. Preprocessing and  post-processing often represents half or two-thirds of the overall workload, he pointed out. 

By accelerating the entire end-to-end pipeline, from preprocessing, from data ingestion, data processing, all the way to the preprocessing all the way to post processing, “we're able to accelerate the entire pipeline versus just accelerating half of the pipeline,” said Huang. 

The point is that metaverse requirements--even assuming a 1,000-fold increase in computational support within a decade or so--seem feasible, given what is happening with artificial intelligence processing gains.

AWS Wavelengths Does Not Create a Platform Opportunity for Telcos

One of the most-common suggestions for connectivity service providers selling to consumers is the notion that the business model has to evolve from “connectivity” (dumb pipe) to something else, up to and including “becoming platforms.”

So look at what telcos have been doing in the edge computing business so far. You might argue the approach is not “becoming a platform” but supplying dumb pipe (hosting, in this case). Amazon Web Services, for example, is partnering with several tie-one telcos to create edge computing as a service nodes. 

Wavelength is a physical deployment of AWS services in data centers operated by telecommunication providers to provide low-latency services over 5G networks. Operators signed up so far include Verizon, Vodafone Business, KDDI and SK Telecom.

Keep in mind, in this arrangement, it is AWS that becomes the platform. The telco participates as a supplier of rack space and related services, and benefits indirectly to the extent that its connectivity service adds value. 

Taxonomically, the telco acts as a “pipeline” business, creating a capability (server hosting) and selling it direct to a customer (AWS). Most businesses historically have been pipelines, creating products and selling to customers, so that is not unusual. 

The important fact to note is that, for this particular opportunity, telcos are not seeking to create a platform. AWS is the platform. Telcos sell a pipeline service, which, by definition, is sold to a single type of customer. 

A platform, also by definition, involves becoming a marketplace where services or products are sold to at least two different groups of constituents, and where the platform enables transactions. 

Ridesharing services, for example, are platforms, linking drivers and riders, but not owning or creating the resources used for fulfillment. 

The optimistic view on creating a platform is that any product can become a platform if information or community can create new value. The unstated corollary is that the “platform” activities must generate incremental revenue. 

And there is no shortage of recommendations that telcos become platforms. “Operators will have to shift from traffic monetization (relying mainly on connections) to traffic value monetization (inclusive of rate, latency, and slicing),” say HKT, GSA and Huawei. “In other words, operator business models must provide both intelligent platforms and services instead of merely traffic pipes.”

If you have been in the communications business long enough, you have heard that suggestion almost all the while you have been in business. “Value, not price” is the way forward, one hears. 

That is correct, up to a point. Any pipeline business can add, augment or replace the actual products it creates and sells to customers. It matters not what the product is that is created and sold to customers. Generally speaking, this is the meaning of the advice to “move up the stack,” supplying value beyond connections, bandwidth or minutes of use. 

The more-challenging notion is “become a platform.” The Wavelengths deal is not the only way telcos can participate in the edge computing business. But it is unlikely to create a platform for telcos, because, by definition, the sale of hosting to AWS is not a platform business model. 

Quite to the contrary, Wavelengths is both traditional “hosting” and also seems a direct outgrowth of the way AWS has in the past sourced computing infrastructure, including a mix of owned and leased facilities. Along the way, AWS has had to create and get comfortable with the idea of its servers operating in somebody else’s facilities.

Up to this point, the “somebody else” has been third party data centers. But edge computing requires even more decentralized facilities. Hence, Outposts, a rack of servers managed by AWS but physically on-premises. 

The customer provides the power and network connection, but everything else is done for them. If there is a fault, such as a server failure, AWS will supply a replacement that is configured automatically. Outposts runs a subset of AWS services, including EC2 (VMs), EBS (block storage), container services, relational databases and analytics. S3 storage is promised for some time in 2020. 

Local Zone, currently only available in Los Angeles, is an extension of an AWS Region, running in close proximity to the customers that require it for low latency. Unlike Outposts, Local Zone is multi-tenant. AWS deploys only when there is a critical mass of customers unable to take advantage of an established AWS region. 

All three services essentially are built on Outposts and local server facilities using third party sites. As AWS had to get comfortable with third party data centers hosting its servers, so Outposts extends that hosting to enterprises. 

A Local Zone is effectively a large group of outposts. 

Wavelength involves Outposts located inside a telco facility of some kind, likely often a central office. AWS is early to move, but the other hyperscale computing-as-a-service providers also are expected to make big moves toward edge computing facilities as well. 

By 2023, by some accounts, the hyperscale cloud computing firms will be spending $23 billion in a single year to create edge computing facilities, about half of total capex in that year. 

All interesting. And telcos are likely to experiment with other initiatives in edge computing. But Wavelengths does not achieve the objective of creating a platform.

"Platform" is Overused, Misapplied and Difficult

Almost no word gets tossed around as a business “strategy” as “platform.” Becoming a platform often is touted as the key to success and a way to create brand value and escape commodity pricing. Compounding the problem are the likely different definitions people seem to use. 

In computing, a platform is any combination of hardware and software used as a foundation upon which applications, services, processes, or other technologies are built, hosted or run.

Operating systems are platforms, allowing software and applications to be run. Devices are platforms. Cloud computing might be said to be a platform, as systems are said to be platforms. 

Standards likely are thought of as platforms by some. 

In other cases components such as central processing units, physical or software interfaces (Ethernet, Wi-Fi, 5G, application programming interfaces) are referred to as platforms. Browsers might be termed platforms by some. Social media apps are seen as platforms as well. 

Platforms, in this sense, are a foundation upon which other things are built or created. That is true enough, but arguably too broad an interpretation to be useful when used with reference to business strategy.  

For purposes of business strategy, a platform earns revenue in a different way than most traditional products have done. Traditional products are sold with a “pipeline” business model, where one firm creates a product and then sells it. 

The platform business model requires creation of a marketplace or exchange that connects different participants: users with suppliers; sellers with buyers. A platform functions as a matchmaker, bringing buyers and sellers together, but classically not owning the products sold on the exchange. 

Perhaps the best models are multi-product e-tail firms such as Amazon, Alibaba or eBay; ride hailing companies such as Uber or Lyft; content exchanges such as YouTube; payment services such as PayPal; lodging exchanges such as airBNB; food delivery services such as GrubHub;  messaging platforms such as WhatsApp or social networks such as Facebook. 

source: Innovation Tactics 

A pipe business creates and then sells a product directly to customers. Amazon is a platform; telcos and infrastructure suppliers are pipes. So you can sell the enormity of the challenge. A connectivity provider would be a platform if it enabled a huge ecosystem of suppliers creating and delivering apps over its platform, in nearly all cases without owning any of those apps, or even earning direct revenue, except in the form of a commission or fee for each transaction. 

Platform creation is not especially easy for a connectivity services provider. If you think about every business as either a “pipe” or a “platform,” then most businesses are “pipes.” They create a specific set of products and sell them to customers. That is a classic “one-sided market.”

This taxonomy illustrates why it is a challenge for any connectivity services provider to become a “platform.” Not only must “connectivity” (the current function) be provided, but also all the other functions that allow third parties to build applications on the platform. Note that the “networking” function is the foundation, but only that. 

source: Adnet 

A platform often creates value because of the scale and scope of the interactions between members of the ecosystem, so the range and depth of interactions might be a better metric for a platform. In other words, the platform is easy to join, easy for participants to use and easy to federate. 

Effective application program interfaces are one aspect. But effective logistics, settlements, data exchange, payments and information on ecosystem participant behavior might be other important aspects for ecosystem transactions and interactions. 

All that suggests any substantial connectivity provider platform would necessarily be built by some entity other than a single telco, no matter how large its operations. Among the reasons: creation of the ecosystem and platform would necessarily require adding roles and functions far beyond connectivity. 

In fact, the obvious paradigm already exists. The internet ecosystem functions with connectivity as an abstraction. It is assumed to exist. In that sense, the internet is the platform. 

What might remain to be created are industry platforms for apps and services with specific communications requirements that possibly are dynamic. Platforms for industrial use cases; healthcare or automated vehicles might provide examples. Or so many hope.

If Generative AI is "Winner Take All," That Dictates Investment Bets

Contestants in the generative artificial intelligence model business are following a “winner takes all” approach to the market, investing at a pace that gets criticized by financial analysts (and rightly so, in some respects) for exceeding what seems immediate and tangible financial payback.

But the “winner takes all” strategy has worked often in the internet era, across many different segments of the market. So we already can predict some outcomes. 

“Wild” levels of spending are going to pay off for at least one and perhaps two providers of GenAI models, in terms of leadership of the market (ecosystems built on them). Some of the contestants will eventually “break even” for their investors, neither gaining or losing equity value. 

But most will fail, losing their investors most to all of the invested capital. That is simply what has happened in winner-take-all markets. 

As we saw early on, even pre-internet, for operating systems, and later saws for search engines, mobile operating systems, e-commerce, social media, ridesharing, peer-to-peer lodging and other app segments, market leadership is highly concentrated with one or sometimes two providers dominating. 

So the companies developing GenAI frontier models, such as OpenAI, Google DeepMind, and Anthropic, are assuming the market  will eventually shake out with a “winner takes all” structure, with market leadership highly concentrated and creating a new ecosystem of value around the leading models, with the hundreds of other would-be contestants relegated to history.

Market/Product Category	Dominant Players
Search Engines	Google
Social Media	Facebook, Instagram, TikTok
E-commerce	Amazon
Mobile Operating Systems	Apple iOS, Google Android
Desktop Operating Systems	Microsoft Windows, macOS
Web Browsers	Google Chrome, Safari
Streaming Video	Netflix, YouTube
Cloud Computing	Amazon AWS, Microsoft Azure, Google Cloud
Online Advertising	Google, Meta (Facebook)
Smartphones	Apple (iPhone)
Productivity Software	Microsoft Office, Google Workspace
Mobile Payment Systems	Apple Pay, Google Pay, PayPal
Digital Maps	Google Maps
Food Delivery Apps	DoorDash, Uber Eats
Ride-Hailing Apps	Uber, Lyft (US), Didi (China)
Video Conferencing	Zoom, Microsoft Teams

By way of contrast, end users of GenAI--especially enterprises--will have to answer for their choices, as they would for any other capital investments made to support their businesses. And that goes for suppliers fo hardware and software incorporating AI functionality. 

GenAI model leadership will undoubtedly shake out into a “winner takes all” pattern. But that also means hardware or software suppliers might “bet on the wrong horse,” as might their customers. That generally means high danger for model suppliers and those firms that are direct parts of those ecosystems. 

Hardware and software participants might face moderate to high impact if they are in the “wrong” ecosystem or platform. End users--whether consumers or businesses--might face low repercussions if their chosen GenAI supplier does not emerge as the eventual market leader (either provider one or two). 

Basic or firm-specific functionality might not be the big problem, as niche solutions might be perfectly viable, especially if a model emerges as providing superior industry-specific functionality. 

Value Chain Segment	Example Participants	Degree of Danger if Supplier/Partner is Not a Winner	Reason for Risk
Foundation Model Providers	OpenAI, Google DeepMind, Anthropic, Meta	High	If the chosen provider loses, end-users face high costs in switching models due to integration dependencies.
Model Fine-Tuners & Adaptation	Hugging Face, Scale AI, Stability AI	High	Fine-tuning investments and adaptations may not be transferrable to new, winning models.
Hardware Providers	Nvidia, AMD, Intel, custom chip startups	Moderate to High	Dependence on specialized chips may limit flexibility; non-dominant players may face R&D cutbacks.
AI Infrastructure Platforms	AWS, Microsoft Azure, Google Cloud	High	Non-winner platforms may lack interoperability, leading to high switching costs and stranded data assets.
Middleware Providers	Databricks, Snowflake, API integration services	Moderate to High	Middleware built for a specific provider may need reconfiguration, reducing compatibility with other models.
Application & Service Developers	Jasper, ChatGPT plugins, custom AI tools	Moderate to High	Apps tightly integrated with a non-winning model may struggle with compatibility and require costly rewrites.
Consulting & Implementation	Deloitte, Accenture, AI-specific consultancies	Moderate	Models selected may lose support, limiting longevity of solutions delivered to clients.
AI Tooling & DevOps	MLOps tools, AI pipeline solutions (e.g., Weights & Biases)	Moderate	DevOps tools tied to a specific model may lose relevance if clients pivot to dominant providers.
Enterprise & Custom Solutions	SAP AI, Oracle AI, Salesforce Einstein	Moderate	Custom solutions based on non-winning models may have compatibility and feature limitations.
Data Providers	Web-scraping firms, domain-specific datasets	Low to Moderate	Dependence on any one model is less critical, though dataset compatibilities may shift.
Consumer-facing Applications	Grammarly, Canva, AI-powered image and video tools	Low to Moderate	User-facing tools may need updates for compatibility but can adapt faster than core infrastructure layers.
End Users & Businesses	Large enterprises, SMEs, individual users	Low to Moderate	Users may face inconvenience, but model-agnostic interfaces could minimize direct switching costs.

Still, longer term, the specific GenAI ecosystem should matter, as network effects should eventually emerge.