An Emerging Policy Dilemma for Developing Countries: Water for People vs Water for Machines – Dr Anthony Turton
The world is a dynamic place. The prime minister of Canada – Mark Carney – has recognized one of the major drivers of change being the “rupture” caused by a fundamental shift in the role played by America as either a protector or a predator. This Rupture is likely to have a cascading effect across all markets. A central feature of this Rupture is the emergence of technology providers, most notably in the field of artificial intelligence (AI).
This is best described as a collision course between an emerging technology that is pivotal to the future wellbeing of all nations, with the reality of changing perceptions about sovereign control.
Technology versus sovereignty is a central component of the Rupture, so we are likely to see a lot more of this clash coming through our news feed.
We are already seeing it in changing threat perceptions within the regulatory environment of governments regarding dependence on American software such as Microsoft and Palantir. We are also seeing it play out in the collapse of the Petrodollar system, and the emergence of alternative platforms for financial transfers between sovereign nations wanting to wean themselves off the increasingly unhealthy dependence on the US Dollar. This will become a growing challenge to SWIFT and the hegemony of Visa and Mastercard.
The petrodollar system ensures dollar dominance, but this is changing because of the “rupture” defined by Canadian Prime Minister Mark Carney.
All of this requires datacentres that need to be set up in locations other than America.
Datacentres are the new pathway to the reclamation of sovereign control as these big changes take place. They are the new battlespace where policy choices today will have long-term implications.
Datacentres are therefore a necessary element of sovereign control, with three new issues that are emerging.
The first big issue is the policy environment needed to enable the establishment of datacentres, as critical elements of infrastructure, in developing countries where poverty is often an underlying factor that governments need to deal with. Seen in the context of the Rupture, the government of every developing country will be forced to make a choice. This will probably become a binary option between two alternatives. Either accept that the Rupture is happening and then deciding where their own economy is going to fit into the new but ill-defined world order; or resist this change by failing to respond and thereby running the risk of failing to optimize any future opportunities.
The second big issue is control of technology. This is a wide field indeed, so I can only touch on it in the most cursory way in an article of this nature, but in essence the axes of future comparative advantage are likely to spin around the key areas of microchip manufacture, software platforms and cooling technologies for datacentres. There will be others, but these three issues are likely to dominate the decision-making agenda at strategic level in both corporations and governments.
The third big issue is dealing with the public perception that governments are favouring big corporations by allocating water rights to datacentres at the expense of local communities who see water as a fundamental human right. In this regard we need to see datacentres as new water users that have not previously been granted water allocations. In political economies where finite limitations to existing water supply are not a factor, this will be an easy task for government and regulatory authorities, because water is available surplus of demand. The opposite is likely to become true for developing countries where water resources are approaching closure.
Let us expand on this notion of resource closure, because this is likely to be central to the emerging issues around water for people versus water for machines. Closure refers to that moment in time where all available water resources have already been allocated to legal users, so any additional user will become a constraint on existing rights. This is a policy problem confronting all governments going through the Rupture, but it is more acutely felt by developing countries where poverty is still a factor.
In policy terms, this poses a dilemma for governments. Do they build new infrastructure to meet the needs of a water consuming sector that is not yet fully understood; or do they simply ignore the public outcry driven by environmental and social justice NGO’s concerned with the human rights aspect of water for people versus water for machines? A dilemma is a problem for which all apparent solutions are equally undesirable. In this case, the government of the developing country must make a tough choice between providing water for human survival versus water for economic growth.
Sectoral Water Efficiency (SWE) and Allocative Efficiency are two policy tools for governments in water constrained economies. Both have a role for AI platforms, so the question becomes how technology providers choose to position themselves vis-à-vis government regulatory bodies making decisions about sovereign risk.
Here the policymaker is confronted with a new set of tools. Sectoral water efficiency (SWE) refers to the conversion of a unit of water into a unit of production within a defined sector. This is about incentivizing efficiency, so we can think of it as “more crop per drop” in the agricultural sector, or “more efficient cooling per drop” in the datacentre sector. Both are necessary, but each sector has a different gearing ratio, so the benefit to society is different within each sector. This exposes the policymaker to another tool in their tool-box – allocative efficiency. This refers to the relative benefit of improved SWE when compared across sectors. SWE is about efficiency within a given sector, whereas allocative efficiency is about quantifying the relative benefit derived across different sectors. Stated differently, what is the nett benefit arising from a 5% improvement in efficiency in the agricultural sector versus a 5% improvement in efficiency in the technology sector seen within the context of a national economy?
This requires data to be measured that is currently not readily available, which is where digital systems start to open new policy doors. By accurately measuring changes to core parameters, digital artificial intelligence systems can provide the policymaker with defendable options around the reallocation of scarce water away from people to machines. Let us look at a simple example by asking a series of questions to which binary answers create a logic cascade.
Is surplus water available within an existing economy? If yes, then simply allocate some water to datacentres. If no, then proceed to the next step.
Do datacentres represent a nett benefit to society in terms of becoming an economic enabler? If yes, then proceed with the quantification of the cost-benefit by comparing the SWE with the Allocative Efficiency when making decisions about new water users in the technology sector. If no, then the status quo prevails, and the developing economy remains unable to leverage advantage from the changing circumstances arising from the Rupture.
When deciding to elevate the technology sector as a legitimate water user, is it necessary to provide potable quality water for cooling purposes? If yes, then prepare for the public perception that government is prioritizing water for machines over water for people. If no, then determine what alternate sources of water are available.
When deciding what alternate sources of water are available, what are the critical parameters that need to be measured and monitored to ensure efficient process cooling while providing the most defendable policy position for government.
Logic cascade for decision-makers in water-constrained national economies.
Seen through the lens of changing comparative advantage to any developing economy, the Rupture becomes a useful policy tool. When the government concerned is interested in regaining sovereign control over technology, then the way that water is allocated to cooling functions within datacentres is an emerging area of solution thinking. When all parties realize that the core problem is in fact the risk of public perception of government choices in allocating water to people versus water for machines, then technology providers can choose to align their solutions with the stated policy objective of the governments in whose jurisdiction they operate.
When a technology provider can give sovereign control back to their own government around water allocation decisions, that becomes a win-win-win outcome for policymakers, technology-developers and society. This is the new frontier. We will explore it further in future articles where we will delve into the actual numbers.
