As FWR shows in a range of interviews with wealth managers, they see artificial intelligence as a driving force for market performance. But there is a hurdle. After New York suspended construction of major data centers due to their high consumption of energy and water, it raises questions about the pace of AI growth in some jurisdictions.
(Editor”s note: While some US states are pushing back against
data centers, it is clear that some investors are fretting about
how realistic valuations of AI-related stocks are. Today, a
selloff in semiconductor stocks extended, amid evidence that
China’s advanced chipmaking capabilities are significant. US
technology-heavy Nasdaq 100 futures fell 0.6 per cent, while in
Asia, a 7.5 per cent slump put a Bloomberg gauge of semiconductor
shares on course for its biggest decline since April 2025,
according to Bloomberg today. One way or the other, the AI trend
is going to involve a bumpy ride. A point to consider is that
while some countries might restrict data centers, others will use
it as an excuse to seize a competitive lead.)
Data center construction has been surging recently due to
demand for processing power from artificial
intelligence firms. But in view of their consumption of
scarce resources and a public backlash, New York governor Kathy
Hochul signed an executive order this month prohibiting the
construction of major data centers for a year. New York was the
first US state to make such a move.
A number of farmers in the US are also raising red flags on the
potential drain on local resources that the data center boom
poses to rural regions, using farmland, electricity and water
needed to raise livestock and grow crops.
David Harrison, fund manager of the Rathbone Greenbank Global
Sustainability Fund, told this news service that he believes that
data centers could become an increasingly political issue in 2026
and beyond. “Some of the largest planned sites will require over
1 gigawatt of power and have significant water usage demands. The
sheer scale of capital involved (over $700 billion invested in
new data centers in the US in 2026 alone, according to Moody’s)
is staggering and creates both opportunities and risks for
investors,” he said. “We think that it only accelerates the need
to invest in grid infrastructure, renewable power sources and
energy storage solutions. We are also seeing strong demand for
companies offering solutions linked to water infrastructure and
HVAC equipment.”
“The move in New York could be replicated across a number of
states, particularly given environmental concerns and rising
energy bills for consumers. Increased regulation or political
intervention appears a strong probability,” Harrison continued.
“This may slow down the roll out of data centers which could
create uncertainty for the speed of AI adoption. On the other
hand, it’s important to remember that new technology and
approaches to improving the efficiency of a data
center (cooling, power usage) are continually evolving,
which could also bring a new angle to the debate.”
Katsunori Ogawa, chief portfolio manager for Sakigake High Alpha,
SuMi TRUST, believes the challenges are unlikely to fundamentally
undermine the growth trajectory of AI-related investment. “The
more important question is not whether AI demand itself will
continue to grow, but whether power supply and infrastructure
development can keep pace with the speed of AI adoption. Looking
ahead, competitive advantage is likely to depend not only on
greater computing power, but also on improvements in energy
efficiency and cooling efficiency,” Ogawa said.
“We believe this changing environment is likely to create new
growth opportunities for Japanese companies,” Ogawa continued.
“As AI adoption accelerates, the importance of energy-saving
technologies, advanced power management systems, next-generation
materials, and power semiconductors is expected to increase
significantly. These are all areas in which Japanese companies
have developed strong expertise over many years, and the benefits
of AI investment are therefore likely to extend beyond
semiconductor chips themselves to the broader supply chain that
supports them.”
Meanwhile, Euan Ker, senior responsible investment analyst at
Aegon Asset Management, thinks that a growing backlash against
the environmental footprint of hyperscale data centers is
starting to materially affect the pace and location of build-out.
“This is not an isolated development,” Ker told this news
service. “Over 70 jurisdictions globally are now considering or
implementing restrictions, ranging from moratoria to tighter
permitting rules and cost-allocation mechanisms,” Ker said.
“Public opposition is also rising, with support for new data
center construction in the US notably low. Combined, these
factors are beginning to reshape project timelines, siting
decisions and economics.”
“While demand for AI-related infrastructure remains strong,
delivery risk is increasing – reflected in recent project
cancellations and expectations that only around half of planned
US capacity may be delivered on schedule in the near term,” Ker
continued. “Against this backdrop, the quality of operators’
energy strategies is becoming more important in securing
approvals, particularly where they address grid impact and
renewable integration.”
“From an ESG perspective, the debate highlights a broader
transition challenge: balancing digital infrastructure growth
with local environmental and social constraints,” Ker added. “For
investors, differentiation is likely to emerge between developers
that can demonstrate credible, system-aligned energy solutions
and those more exposed to regulatory and community pushback.”
Mark Brennan, portfolio manager of
Guinness Real Assets Fund , also highlighted that data
centers are resource hungry and New York will not be the
last to impose a mortarium, which could delay the development of
AI. “There will be pockets where deployment is delayed. But where
I come at it from, as a real assets investor, I am looking at the
listed landlords that manage the data centers. This sort of news
is interesting for them as it puts more scarcity into the market.
As the landlord, it gives you pricing power,” he continued. “This
is actually not bad news for them. We are seeing increased
scrutiny and caution on where data centres will be and on the
environmental impact. But that tilts things in favor of the
existing landlords,” he added. “So there’s two sides of the coin
in terms of what it means for investors. Since it was announced,
we have seen positive share price performance for some of big
data centres REITs that we hold.”
“Also I think the data centers are becoming more environmentally
efficient. AI clearly also has the ability to improve
productivity. The incentives are there for developers to become
as environmentally as possible,” Brennan said. “NY hasn’t had a
direct negative impact on my investments. Growth in AI from a
real asset perspective is a major tailwind and driving increases
in earnings growth. It is a major source of growth for my sector
and it has a long way to run. I remain lent to AI infrastructure
opportunities. It is important,” he added. The firm recently
acquired Foresight Capital Investment where Brennan co-developed
Foresight’s real asset range.
Mikhail Zverev, co-manager of the Amati Global Innovation Fund,
said that AI data center build-out is the biggest capital
investment boom in a generation. “This boom is stretching the
whole supply chain and ecosystem (literally and metaphorically),
from semiconductors to construction capacity to availability of
energy and water. No wonder SpaceX is contemplating data centers
in space,” he said.
“We have found the most compelling asymmetric risk/reward in
electric cables and grid construction. We own Prysmian, Italian
listed global leader in high voltage cables, and MasTec, a
leading US energy network construction specialist. Even if the
pace of AI capex slows, they have years of work to do in
modernizing electric grids to improve reliability and resilience
and serve growing electrification demand across the economy,
beyond data centers.” Zverev added.
For more coverage on the various implications of AI, such as on
client confidentiality within professional services
relationships, see an example from FWR here.