Google’s cloud is an interesting example of how information flips the supply chain upside down.
As pointed out in vertical integration, the physical supply chain gets flipped upside down, when we look at it from a data/information standpoint.
A classic supply chain moves from upstream to downstream, where the raw material is transformed into products, moved through logistics and distribution to final customers. A data supply chain moves in the opposite direction. The raw data is “sourced” from the customer/user. As it moves downstream, it gets processed and refined by proprietary algorithms and stored in data centers.
To truly appreciate those differences, let’s start with one of the most interesting of the physical supply chains: coffee.
From coffee beans to data farms
The coffee supply chain is among the most interesting, as it’s quite complex and it goes through a cycle of growing beans, harvesting, drying, packing, bulking, blending, roasting, labeling, packaging, distributing, and selling.
Throughout this process, there are many players involved globally, at each stage. Perhaps wherein the growing and harvesting of coffee beans, countries like Brazil and Colombia play a leading role.
Yet, as the bean goes through a process of refinement, it goes through several parts of the chain, and only toward the end, there is the process of roasting, labeling, packaging, and distribution.
Most of the economic value of this supply chain is skewed toward the end. As we get closer to roasting, labeling, distributing, and retailing, that is where most of the economic output is captured.
In other words, of the overall price paid by customers in the shops for their nice double espresso, most of the revenues go toward covering up for the expenses to run the shop/rent, the staff, tax, and profits.
To gain a bit of context, as the Financial Times evaluated on a 2.50-pound cup of coffee, 25 pennies go to the shop as a profit, and only 10 pennies go to the overall coffee chain, and about a penny goes toward the farmer.
If we use this analogy for the world of AI supply chains, the farmer is no longer in a plantation in South America. But anywhere tapping with her/his finger on a 4.5 inches smartphone.
Raw materials are sourced by consumers
In an AI supply chain it all starts from consumers. They are the growers of the raw materials (raw data) that will serve as the basis for the whole supply chain.
It’s worth pointing out, that, as in a coffee supply chain, where the beans are grown by farmers, which are the ones capturing less of the individual economic value from the supply chain.
In an AI supply chain, consumers are like farmers, and they also are the ones that gain the least in terms of economic value from the overall supply chain.
Hardware devices become the harvesting facilities
The raw data gets gathered, harvested and collected through physical devices, which are the most proximate object to the consumer.
As the raw data becomes available, hardware becomes the harvesting facility in the AI supply chain.
Software and operating systems become the harvesting machines
Software, operating systems and everything else that is in between the physical device and the company’s algorithms become the harvesting machines, ready to sort the data, that will go through the industrial machineries of the AI supply chain.
Algorithms are the industrial machinery for data
As this data, partly filtered by the software side will go through a process of industrial refinement, algorithms will play a key role in refining, processing, and packaging of the data for several scopes.
In that sense, the data moves in two directions. On the one hand, it will move toward consumers to improve the services they get for free. On the other hand, it will move into the proprietary technology stack of the company, ingrained in its monetization machinery to generate profits.
Before it can move in those two directions, though, it will need to be stored within its data centers.
Data centers as fulfillment facilities
As the data goes through the data centers, it gets stored, and it moves in many directions. Back to consumers in the form of free services and toward the monetization machinery, where the data processed, refined, and continuously reprocessed becomes the core service the company offers on the market.
Perhaps, as Google highlights “Our data centers keep all of Google’s products and services up and running around the clock and around the world. Whenever you access Gmail, edit your documents, or search for information on Google, you’re using one of our data centers and have the power of a supercomputer at your fingertips.“
Key takeaways
- The consumer as the farmer sources the raw data, and it gets back only a fraction of the overall economic value.
- Most costs go back to data centers, power sourcing, profits, and organizational costs.
- Algorithms work in two directions, by refining data to offer free services to consumers. And by creating the premises for the monetization machinery to work at the best.
- Hardware, as the most proximate thing to the consumer, becomes the harvesting facility.
- Data centers become the fulfillment facilities moving refined and repackaged data in two directions.
- Toward consumers in the form of free services and toward businesses in the form of advertising or premium services.
Is this a permanent design for AI supply chains? Not necessarily. Yet that has become the predominant design for new dominant media companies (Google, Facebook).
Originally published on FourWeekMBA.