Funds are operating in an extraordinary social and economic environment, with Scott Chan, chief investment officer of CalSTRS, saying he has never witnessed so many “large shifts stacked on top of the other” in his investment career. Amid the change, investors are increasingly shifting to a scenario and regime-based approach to asset allocation.  

At the Top1000funds.com Fiduciary Investors Symposium, Chan – who took over the top investment job at the $374 billion CalSTRS a little over a year ago – said the fund had moved to a more defensive position ahead of Liberation Day in April, expecting a higher probability of a recession in the US than what was priced in.  

“That meant increasing the more diversifying elements of our portfolio – fixed income and hedge funds – and raised cash. We felt pretty good as the market was going down. 

“I tell the team that this was actually a highlight of the year, because one of the things that’s unique to CalSTRS is we have about 44 per cent of our assets allocated to private markets and alternatives. So liquidity management is becoming very key to this environment. 

“To ballpark, we probably have to have something like $35 billion in liquidity if we go into a correction… we actually were in position in front of what we thought could have been a correction with enough liquidity.” 

However, Chan acknowledged the approach has its risks: the defensive position cost around one basis point of performance during the subsequent market recovery and the team needed to “trade up quickly”, he said. “But I’ll take that, because I think we were ready again in front of what we thought could be something way different, like a crisis might have ensued, or a recession.” 

Sentiment shifts 

Anna Langs, managing director of asset allocation, risk management and innovative solutions at San Francisco Employees’ Retirement System, said the fund has similarly been strengthening its liquidity management. Around half of the fund’s assets are allocated to private markets, but it also has close to a 10 per cent allocation to hedge funds, which Langs said provided both return and liquidity during COVID when the two traditionally liquid asset classes – equity and fixed income – failed to support the liquidity needs.  

Matilde Segarra, president and US chief executive of the $722 billion APG Asset Management, said the Dutch fund is also focusing more on regime analysis to inform more dynamic and nimble asset allocation.  

One big question the fund is grappling with now is how it approaches geographical diversification. Investors are rethinking their US exposure due to nervousness around debt and deficit, but Segarra pointed out that the problem is not unique to the world’s largest economy: “there are a lot of indebted governments around the world”, she said.  

What she is really worried about is whether investors’ sentiment towards a certain country is increasingly dominated by short-term events.  

“I am very worried about sentiment at the moment. What I mean by that is if I see the conversations that are going on in the company – again, I work for a firm that has offices across the world – I really see that the Europeans look at the United States right now with angst and scepticism,” she said. 

“And American colleagues that feel constantly attacked at the moment by the questions that they get from their European counterparts, are constantly pointing out that Europe isn’t growing enough and that Europe has an innovation problem. 

“I see almost that the polarisation that you see in society and in political systems across the world is also happening in my organisation on a small scale.  

“We’re supposed to be long-term investors, and I see that the conversations at even the highest level of the company, are conversations driven by sentiment and with a lot of short-term emotions dominating the dialogue…  at some point that is going to affect the quality of decision making.” 

Amid geopolitical and economic uncertainty, CalSTRS’ Chans said investors should always be ready to pivot their strategy as a “bulletproof” investment now could unravel very quickly should capital or ownership control policies get enacted in different countries. To do that, the fund is applying a total portfolio approach in its investment process so that it can “map the total risk exposures” better.  

“With the geopolitical risks and the issues around governments becoming more involved, as well as all the other risks we describe, we could be in for a lot more volatility.  

“So we have to have a team in a design where we can get more diversified or dynamic, but also be able to trade up quickly too, which I think is something that, if I looked over the past couple decades, wasn’t as necessary. 

“We have been exposed to growth investments, and that’s worked rather well. I don’t think you want to be out of growth… but I’m similarly worried about how quickly that might disrupt other industries. We’re looking through our portfolio and asking ourselves the question of what do we need to sell here? Alongside the question of what we’re buying?” 

Mark Horowitz, a leading computer scientist and electrical engineer at Stanford University, has declared that Moore’s Law is “basically over”, which will have significant ramifications for artificial intelligence investors who are counting on more computing power to feed into more complex models.  

It is a view shared by Nvidia CEO Jensen Huang, who pronounced the law “dead” in 2022 while justifying a hike in chip prices.  

The law refers to the observation that the number of transistors on an integrated circuit doubles every two years while the costs decrease. That’s an economic factor that has driven the semiconductor industry, said Horowitz, who is chair of the department of electrical engineering and professor of computer science at Stanford University. 

“If I have a product and I’m selling it in high volume, when I move to the next generation technology, it’ll become cheaper to produce, therefore I will make more money, or it will prevent my competition from underselling me. We can then create either the same parts we have now cheaper, or we can build even better parts at the same price point,” he told the Top1000funds.com Fiduciary Investors Symposium.  

“We still are scaling technology. We’re still building more advanced processing nodes. We’re cramming more transistors per square millimetre, but unfortunately, the transistors are not getting cheaper.” 

Machine learning scaling rules state that bigger models tend to mean better performance, which is why the dismantling of Moore’s Law and sluggish computing power growth could be a roadblock for AI development, Horowitz said. 

“This is a major disruption, because our expectation is that we can do more computing and we’re pushing more stuff to the cloud,” he said. 

“Now Moore’s law is actually flat, so all that [AI model] complexity is going to have to be done through sort of algorithms or applications… if we think about the economics, where is money going to be made?” 

Horowitz is of the view that most companies are losing money on their AI projects. “All the hyperscalers in the world [like Google, Meta, Amazon and Alibaba] are spending enormous amount of money trying to protect some of the business they have because they’re worried about losing it,” he said.  

He suggested the profitability of an AI application hinges on two things: the service cost and the liability cost. The latter would be increasingly pronounced as the world moves towards agentic AI applications as service providers will need to start taking responsibility for decisions and suggestions of their AI agents.  

Taking these costs into account, Horowitz said it is highly likely that profitable AI models or applications in the future will not be large, but small. 

“Both of those, to me, indicate that what you’re going to try to do is reduce the scope of the model to something in a particular area, so you can make it cheaper to serve and less likely to make a bad thing,” he explained. 

“If that’s true, then the people who are going to start making money are not the people who are using the really big models. It’s the people who have used the big models to create smaller models or more domain specific models.” 

These views are not pessimistic predictions for the AI space, Horowitz said, adding that he is optimistic about models which can find useful and specific applications.  

“I do think there’s going to be some big craters, because there’s been billions of dollars invested in a lot of companies, and I’m not sure they’re going to be the people who survive,” he said. 

“I have no doubt that machine learning is going to change the world. The world’s going to be very different. Who’s going to be the player there, I think, is still to be decided.” 

The size of the current infrastructure investment gap and the speed at which it is widening mean there is both a desire and a need for more public-private partnerships to unlock funding. Investors say that collaboration with local governments and raising public awareness of private investment benefits are crucial. 

From roads and airports to energy and water, the US is seeing a rapid deterioration in the condition of its critical infrastructure. Industry estimates suggest there is a $2.6 trillion investment gap to modernise these features before 2030, per the American Society of Civil Engineers.  

At the Top1000funds.com Fiduciary Investors Symposium, Janet Cowell, mayor of North Carolina’s capital city Raleigh and former state treasurer, said infrastructure investment is entering a “very fluid and dynamic” era but that it will take some time for some officials to realise private money needs to play a bigger part.  

“[When] people realise the money’s not flowing, they’re going to scramble for existing revenues,” she told the symposium at Stanford University.  

Raleigh is grappling with an explosion in not only the cost of building materials but also that of land, which went up as much as 70 per cent as the city became an attractive real estate investment destination post-COVID, Cowell said. The city had to abandon certain infrastructure projects as a result. 

“For example, we had a State General Assembly member who said, ‘you’re not funding our transportation enough, so I’m going to try to grab your food and beverage money and reallocate that and try to strong-arm the locals’,” she said.  

“People are going to find other goofy ways to try and do this [invest in infrastructure], and ultimately they’re going to have to come to more of a public-private [model]. I think we’re going to need some enabling legislation.” 

Mikael Limpalaer, head of Americas at AustralianSuper, the A$387 billion ($254 billion) Australian pension fund which has over A$40 billion ($26 billion) in global real assets, said people are “much more amenable” to private or semi-private infrastructure funding when they can see that it is directly linked to the growth of their retirement savings.  

There also needs to be more awareness from the public that, in the absence of a large private infrastructure investor, the alternative funding methods could be much less desirable.  

“Is it going to be more taxes? More debt for the state or the municipality? That’s a very large part of the educational work local and state politicians need to do with their constituents to make that [public-private] mix work,” he said.  

“Then you’ve got the existing, more traditional road networks, airports, water system and energy grids, which are in need of maintenance and upgrade, and particularly in the US, the governance and the ownership structures are not designed or built to accommodate, in most cases, public-private partnerships. 

“We’re meeting with lots of governors and municipal authorities that are keen to engage with investors in the space.” 

There are around 500 passenger airports in the US and only one is run by a private operator – the Luis Muñoz Marín International Airport in San Juan, Puerto Rico. Andrea Mody, IFM Investors head of North America clients and strategy, said that number is out of pace with other countries.  

“In Australia alone, we’ve invested in eight airports. We’ve invested in four airports in Europe. And that is not an asset class that’s even really available,” she said. 

Regulatory reasons are a part of that, Mody said.

“When airports were built in America, they were built with public money, and there were federal grants that states and municipalities took out, and some of the requirements for taking those grants were that you could never privatise your airport.” 

“The US also has a very well-developed municipal financing market… but obviously states and municipalities are capped at what additional debt they can issue in 2025 and beyond.” 

The point with infrastructure investments is not to build cheaply, but build well, Mody said. It’s a delicate balancing act between capital expenditure and affordability for the infrastructure’s end consumers, which Mody said the public sector could control via mechanisms like pricing caps and maintenance standards.  

Many critical moments of US growth in its history have been driven by traditional infrastructure – periods like the Gilded Age, where the railroads and energy contributed to economic prosperity, and the establishment of the Interstate Highway System post-World War II. 

“We’re thinking a lot about data and digital and AI, and that’s a really important part of our growth trajectory, but the movements of people and stuff are still going to need to be there in a large way if we’re going to continue to grow,” Mody said. 

“And it’s going to take a real [public-private] partnership and a different way of thinking about how we did it in the past versus how we’re going to do it in the future.” 

Stanford president Jonathan Levin said the university’s top priority is maintaining the partnership with the federal government while safeguarding its operational freedom, as the institution balances financial reliance on Washington and political scrutiny from the Trump administration. 

The university has been swept up in US president Donald Trump’s crackdown on the nation’s higher education institutions, which he has repeatedly criticised as “liberal” and “woke” with an “anti-American” or antisemitic bias, using the reduction of federal funding as powerful leverage. This August, Stanford made 363 staff redundant, citing budget reductions.  

But at the Top1000funds.com Fiduciary Investors Symposium hosted on the Stanford campus, Levin said some recent criticisms directed at American universities are not new and are, in fact, fair. 

“The US is a country that has become more and more politically divided over the last however many years, and university faculty and student populations are, by and large, on one side of that political divide,” said Levin, who is also an economist and Bing Presidential Professor at the university.  

“We should learn a lesson from that and make sure that, going forward, universities are not so enmeshed institutionally in politics. The students and the faculty can be involved and that’s their freedom to do that, but the institutions shouldn’t.” 

Levin said that at the heart of American universities’ excellence is their ability to operate freely and independently, including in setting curriculum and pursuing research projects. That freedom stems from the post-World War II framework for federal support of science established by Vannevar Bush, an engineer and the head of the US Office of Scientific Research and Development at the time. 

The framework established the federal government’s responsibility to fund research and development as industry support would lack appropriate scale, and emphasised a peer-reviewed, merit-based allocation of grants. 

“Stanford – probably more than any other university – was a beneficiary of that. Stanford was founded in 1891 and coming out of World War II, we were a regional university of fairly modest calibre,” Levin said. 

“We were in financial difficulty at the time too, because the endowment hadn’t been that well managed, and folks leading Stanford at the time saw there would be an opportunity with federal funding for science to build up a research enterprise.” 

The university built up its engineering and science departments, and later complemented them with humanities, history and other arts departments. It also lent land to technology companies such as Fairchild and Varian, which cemented Stanford’s leadership in the early development of Silicon Valley. 

“That really all came out of the federal-university partnership that led to the US having this extraordinary innovation economy. It’s a great story, and it’s still the best recipe that this country has to be the leader in innovation and to be at the frontiers of knowledge,” Levin said. 

“So a big focus for us right now is how do we sustain that great partnership that universities have enjoyed with the federal government, and get back to a place where the American people support us enough to fund us, and give us the freedom that allows our universities to be exceptional, respected and contributing to the country and to the world.” 

Produced in partnership with Blue Owl

Global data centre spending is on track to exceed $1 trillion by 2029, according to Dell’Oro Group, driven by an insatiable demand for hyperscalers to fuel AI and machine learning technologies.

Robert Hartog, who is a senior member of Blue Owl Capital’s digital infrastructure team, believes the hype is justified, with global data centre capacity up more than 200 per cent in the past decade and McKinsey predicting that usage could climb from 60 gigawatts today to up to 298 gigawatts by 2030.

Meeting this demand will require more land, capital and expertise to build more data centres and ancillary digital infrastructure, with the cost estimated to be north of $6.7 trillion globally, based on analysis by McKinsey.

Hartog describes the demand and supply dynamic as a “generational market opportunity.”

“Multi-trillion-dollar investment is needed for real estate, construction, IT equipment and power so the opportunity set is just massive,” he says.

“If you focus on cloud-computing alone, the growth in that market has been 20-30 per cent year on year, which is enough to build a very strong investment thesis, but then add to that AI and it significantly increases the scale of the opportunity set.”

“Hyperscaler facilities provide critical infrastructure for the roll out of cloud and AI, which is essential for society and communities to function.”

With Blue Owl’s digital infrastructure team having close to 10 years of data centre investing experience, the group is well-positioned to take advantage of the scale and capital required, and with asset owners to develop and build the digital infrastructure for the future. Blue Owl’s digital infrastructure strategy has $14 billion in assets under management across the entire digital infrastructure opportunity set from real estate and infrastructure development to credit and stabilised assets. Critically, Hartog says the ability to execute is crucial, as projects continue getting bigger, more complex and more expensive.

“If you look at the capex profiles of hyperscalers, including Microsoft, Amazon, Oracle, Meta and Google, they just keep expanding globally,” he says.

“There is a massive roll out of new and ongoing projects right now and they’re just getting larger and larger.”

“Data centre development is not easy and the larger these projects get, the bigger the constraints.”

Hartog says there are constraints on capital, power, land, supply chains and construction with other considerations including backlogged utility queues, delayed transmission upgrades, and complex regulatory conditions, all of which require an experienced partner with subject matter expertise, global scale, and local presence.

 

From asset owner to asset manager

An M&A lawyer by background, Hartog joined Blue Owl in 2024 through the firm’s acquisition of global alternatives manager, IPI Partners.

During his time as managing director of IPI Partners, the group grew to become one of the largest private data centre investors.

Prior to that, Hartog was the head of communication infrastructure investments at the €251 billion Dutch pension fund, PGGM and he describes the transition from asset owner to asset manager as an “extension of activities.”

“At PGGM my role was to find suitable digital infrastructure assets around the world, including telecommunications towers, data centres and fibre networks, and my work at Blue Owl builds on that,” he says.

“The data centre industry, especially the hyperscale data centre industry, is very interesting for institutional asset owners because of the long-term contracts and stable cashflows.”

Institutional clients have different participant pools, liabilities and time horizons, making it difficult to make a blanket statement about the ideal allocation to digital infrastructure, Hartog says.

“For long-term institutional investors, digital infrastructure and hyperscalers, in particular, can offer the benefits of risk-adjusted returns and capital preservation. We focus on resilient, long-lived facilities with long-term contracts and high credit rated tenants, which underpins stable, predictable cashflows,” he says.

 

Not all data centre developers are created equal

But no matter how favourable the macro thematic tailwinds, hyperscaler investments – like any investment – carry risks, including counterparty risk and development risk.

To help manage risk, investors can work with a manager with a track record of delivering large-scale projects.

“Ultimately, we are developing hyperscale facilities together with our hyperscale tenants, and they are relying on our capability to deliver on schedule, on budget, as per the specs,” he says.

“That means having the required land and permits and ensuring our supply chains can deliver all the necessary equipment and power. Scale is critical to be able to purchase equipment and items ahead of time, and engage large contractors, because there is a limited amount of time to deliver these projects.”

For established players with capital and experience, these challenges spell opportunity, according to Hartog.

“Projects are flowing to new markets that can offer energy resources and infrastructure advantages, which creates jobs and spreads economic development,” he says.

“Over the last decade, we’ve built a business focused on meeting the needs of hyperscalers. Our team is dedicated to solving problems for our tenants and providing them with the infrastructure and real estate they need for the roll-out of their product to the end-users. We’ve built a vertically integrated operating platform with specialised resources in local markets across the world to help address the challenges they are facing.”

These value-added services include site selection, securing alternative power solutions, coordinating land and power permits to drive developments, and engaging with the communities and regulators where we are developing.

“This doesn’t happen overnight. It has taken the better part of a decade to be this specialized and to build and acquire inhouse capabilities to address the specific challenges of delivering purpose-built hyperscale facilities.”

Blue Owl employs a range of specialists in areas like power, land acquisition and property development.

“We continue to be a strategic investor in the space and in order to deliver we need to make sure that we have the right capability sets inhouse, which is why we have power specialists who really understand the grid and the transmission and distribution challenges,” Hartog says.

“We have people who specialise in land acquisition strategies and developing sites into data centres. These are skill sets that are needed to continue to deploy assets in this space.”

While the opportunity for data centres and digital infrastructure may indeed be generational, Hartog believes that managers with the right skill set, expertise, hyperscaler relationships, community presence, scale, and proven ability to execute will be positioned to deliver.

Source: Bloomberg New Energy Finance (BNEF), “Data Center Market Overview”, June 2025.

Red-hot demand for data centres and a lack of infrastructure capital is creating a rare market dynamic in which lower-risk credit can yield more than equity investments in the sector, according to investors speaking at the Fiduciary Investors Symposium at Stanford University.

“Given the lack of debt capital relative to equity capital, we can generate some nice premiums,” said Jennifer Hartviksen, managing director and head of global credit at Canada’s Investment Management Corporation of Ontario (IMCO), citing high single-digit to low double-digit returns for single-B to BB- risk.

“Twelve months ago, when spreads were wider, we were seeing mid-teens for those types of paper that have really great structural downside protections with very nice returns relative to what we get in other markets and, in some cases, relative to what the equity below us can be getting.”

She said there were probably only 20 infrastructure credit funds targeting below investment grade that were lending to data centres, with the sector requiring a completely different skill set than investing in the far more popular direct middle market loan space.

“There’s just not a lot of folks doing that type of lending, and there’s really only two or three that do big funds that are multi-billion funds, so there’s a lot less capital there. And as a result of that, the spreads have come in less.”

Patrick Lawler, portfolio manager and head of core acquisitions on Blue Owl Capital’s digital infrastructure team, said some equity deals were closing with negative leverage, where the year-one cap rate is below the cost of debt. Those investors were betting on mark-to-market rent increases over the next few years.

“It can still be an attractive risk-adjusted return over a seven- to 10-year hold… given the immense rental growth we’ve seen in top markets.”

Data centres sit at the intersection of power and communications infrastructure, and years of investment in established tier-1 hubs such as Ashburn, Virginia, have created powerful network effects. The flip side, Lawler warned, is that “all the top markets are effectively out of space and power,” forcing developers and investors toward tertiary or even frontier markets to secure capacity.

Tammi Fisher, managing director, real assets, at Australia’s Future Fund said refinancing risk and rental renewal risk hadn’t yet been tested through a full cycle.

“We’ll talk to our partners and some of them are definitely more cautious about that, like how do you price that risk that’s maybe 10-15 years out in data centre locations that are highly customised, super-spec’d for a specific tenant. How do you think about that end-of-life risk, if there is end of life? You’ve got others who are much more optimistic about that, and, you know, and there’s so much growth in it,”

The booming energy needs of data centres is also transforming once “sleepy” infrastructure holdings, which now require large capex programs and different operating skill sets from management.

AI builds on existing cloud demand

While AI is attracting significant attention, hyperscale cloud computing remains the underlying demand driver for data centres, Lawler said. For example, Amazon Web Services generated more than $100 billion in revenue last year with about 30 per cent operating margins.

“You have this existing wave of cloud computing demand, and then you have a second entirely independent wave of demand coming from AI. They’re independent, and very importantly, they’re not cannibalizing each other.”

However, one way to de-risk investments in data centres was to structure the deal so that returns were not dependent on the unpredictable future path of AI. Hartviksen said the loans they were making were typically five-to seven-year maturities, with a lot of customisation in the space allowing for enhanced terms.

“We don’t need to think about the technology risk or what these data centres are going to look like beyond the term of our loans, which is another reason we find it very attractive.”

Lawler said Blue Owl owns the building, including the mechanical, electrical and plumbing infrastructure that will support a data centre’s use cases, but it does not own servers, networking equipment, or GPU chips.

“I don’t think anyone here knows exactly what AI is going to look like in seven to 10 years, but what we want to be able to do is look at our investment and say over this 15-year lease term, even if the residual value is effectively zero – so an enormously draconian scenario just by virtue of the lease payments and all the different structure that we’ve built in – we can still generate a positive and attractive, risk-adjusted return for our partners.”

Be part of the conversation next year – register now for the Fiduciary Investors Symposium Stanford 2026. Asset owners only.

Our next event is at the University of Oxford, November 4-6 2025. Register today – asset owners only.