Inside the half-trillion-dollar firm backing AI’s backbone

EQT isn’t a household name in Australia but it probably should be. Headquartered in Stockholm, Sweden, and operating for over 30 years, EQT is the second-largest equity firm globally by capital raised according to PEI  (behind only KKR and ahead of Blackstone) and the largest private equity firm outside of the United States.

Managing just under half a trillion Australian dollars in assets under management (AUM), the firm is deeply tied to the Wallenberg family, a legendary sixth generation industrial dynasty that still controls 30% to 40% of the Swedish stock market capitalisation.

Melbourne native Ken Wong, Partner and Head of Asia Pacific Infrastructure at EQT Partners, joined the firm in 2018, establishing EQT’s Sydney headquarters in 2020. Today, Wong is based in Hong Kong in in pursuit of a structural growth opportunity that he believes is just getting started: the colossal infrastructure build-out required to power the booming demand for AI.

“We are just at the start of this journey," Wong says. "The use cases for AI and therefore the look-through demand on what's required is enormous."

Wong’s insights are driven by proprietary visibility across EQT’s global portfolio of roughly 300 companies, providing a front-row seat to how businesses are deploying AI to transform their operations.

“When you look at the exponential curve of the use cases that we see in our portfolio companies, whether it be healthcare or technology or even within our infrastructure portfolio, it may not just be the cost savings but the efficiency is absolutely phenomenal.”

In this episode of The Rules of Investing, Wong explains EQT's deep ties to Australia, his perspective on the scale of the AI infrastructure build-out, and how EQT’s investment process allows them to capitalise on opportunities across multiple jurisdictions.

Tune into the full interview via the players or read a summary below 

EQT's physical "picks and shovels" strategy

Instead of trying to guess which software platform or large language model will ultimately dominate the tech sector, EQT focuses entirely on the underlying physical infrastructure required to keep them running. 

They are investing heavily in core physical assets, specifically data centres and renewable energy platforms, that provide the essential foundation for artificial intelligence to function globally. By building the underlying physical environments, they generate stable infrastructure returns without exposing themselves to direct technology platform risks.

“Our strategy is a picks-and-shovels strategy effectively," Wong says. "We are just providing the infrastructure for those players to power their LLMs. We don't actually need to make a bet on who's going to win those platform wars.”

Real estate vs. infrastructure: data centre returns

While some market participants view data centres as passive real estate investments, EQT treats them as active infrastructure due to the operational ecosystem value-add required, such as securing complex power and fibre solutions.

The underlying business model relies on highly secure, 20+ year "take or pay" leases signed with hyperscale tenants who hold credit ratings stronger than the US government. 

EQT provides the physical building shell, passes through all direct energy costs to the tenant, and secures built-in inflation protection, creating a reliable and defensive cash flow profile.

“They effectively contract with us and usually we'll have an anchor tenant before we've even broken ground on building a data centre for 20 plus years on a take or pay in terms of capacity," Wong notes. "We have got CPI protection and energy cost pass through.”

The multi-year bottleneck

The primary speed limit on the global AI build-out is not chip availability or land supply, but the massive strain on local electrical grids. In regions like the United States, data centre projects can face agonising wait times of seven to eight years just to get a stable connection to public utilities.

EQT circumvents this severe bottleneck by matching its digital footprint with its proprietary renewable energy and storage assets, packaging self-sustaining, off-grid "gigawatt style" campuses that can be fully constructed and online within a tight 12-month window.

“Grid connectivity, grid connectivity and grid connectivity. That is the biggest bottleneck that we see to why data centres are not actually being built out as fast," Wong explains. "The capacity would be, it's probably double, triple, if not more, if we were able to connect data centres to the grid more quickly.”

The energy transition and electrification

A powerful combination of macroeconomic tailwinds, geopolitical friction, and shifting consumer behaviour is accelerating the global transition toward renewable energy. Driven by the desire for domestic energy independence following global supply shocks, nations are moving rapidly away from fossil fuels.

Critically, this shift is structural rather than purely ideological, as plain vanilla solar combined with battery storage is now fundamentally cheaper to build and run per kilowatt than traditional coal-fired power stations. 

This green energy boom directly intersects with the "electrification of everything," from domestic heating to electric vehicles.

“A renewable plus battery solution is on a gigawatt or kilowatt basis cheaper to produce today than a coal powered fire station," Wong says. "Things being cheaper is a very strong motivator for people that the look-through demand of consumers to cause significant investment.”

The most-misunderstood opportunity: AI infrastructure

The broader market frequently commits the error of grouping all AI investments into a single speculative technology basket, obsessing over individual software failures or structural collapses at prominent tech start-ups.

Wong argues that the true underappreciated opportunity resides entirely within the physical layer of the ecosystem. 

While predicting specific software winners remains highly volatile, the physical infrastructure that powers any and all artificial intelligence platforms remains an essential utility with severe structural under-supply, making it an exceptionally mispriced growth vector.

“Most misunderstood is AI infrastructure, not AI," Wong says. "It's the picks and shovels to what powers AI.”

“I think most of our competitors probably understand that on the infrastructure space, but I think the average person down the road puts all of AI in that same bucket of ‘is OpenAI going to collapse or not and what's the implications’.”