# AIP and Brookfield are both bidding for Stack Infrastructure's Asia operations and the $30 billion price tag explains exactly what AI has done to data center valuations > Source: > Published: 2026-06-24 07:44:23+00:00 *Blue Owl's planned sale of Stack Infrastructure's Asia-Pacific data centers is current, but the reported buyer field points to GIP, Macquarie and IFM, not AIP and Brookfield. The $30 billion price tag still tells you the important thing: AI has made powered data center land in Sydney and Melbourne look like infrastructure royalty.* The number is big enough to make people stare at it and stop thinking. Don't. The Australian reported on June 17, 2026, that Blue Owl's Stack Infrastructure is preparing a sale process for its Asia-Pacific business at about $30 billion, with Morgan Stanley and Deutsche Bank advising. The names reported around the process are just as useful as the valuation: BlackRock's Global Infrastructure Partners is understood to be moving early, Macquarie is expected to work with GIP, and IFM Investors is seen as a logical contender. That is a different story from a loose market rumour about every AI infrastructure fund on earth lining up at the same door. The formal process has not started, according to The Australian's report, and Blue Owl has not announced a completed sale. So the honest version is narrower and stronger: the asset is being prepared for market, the expected price is roughly $30 billion, and the buyer pool is likely to be made of infrastructure investors large enough to swallow a regional portfolio whole. Stack's Australian assets explain why the sale is being talked about in that register. The Sydney campus spans more than 7.7 hectares and has 360 megawatts of capacity. Its Melbourne operations cover two campuses across 9.3 hectares with a combined 432 megawatts. That is almost 800 megawatts in Australia alone, before you even get into the wider Asia-Pacific footprint. You can argue over the final multiple. You can't pretend that kind of powered capacity is easy to recreate. Here's the thing about data centers in 2026: the market is not paying only for buildings filled with servers. It is paying for land, grid access, cooling, planning approvals and time. The last one matters most. If you need compute capacity for AI workloads, an asset that already has a path to hundreds of megawatts is not just more convenient than a greenfield project. It changes the calendar. That is why the Australian comparison matters. AirTrunk was sold to Blackstone in 2024 in a deal valued at about $24 billion. Now Stack's Asia-Pacific business is being talked about at about $30 billion. Those two figures sit close together for a reason. AI demand has changed what large infrastructure funds think existing hyperscale capacity is worth, especially in markets where power and permissions are the bottleneck. GIP's interest would fit its recent pattern. The Wall Street Journal reported last year that GIP and Spanish construction group ACS agreed to form a 50-50 venture to develop data centers, with an initial portfolio valued at about 2 billion euros. ACS also pointed to a pipeline of more than 11 gigawatts across North America, Europe and Asia-Pacific. That is not a side hobby. It is a direct bet that the picks and shovels of AI are going to include substations, long-lead electrical gear and campuses that local councils will actually allow to be built. Blue Owl has its own reason to test the market now. The Australian noted that Blue Owl Capital Corp, the firm's listed private credit vehicle, has been trading at a discount to reported net asset value, while Blue Owl said that vehicle is separate from Stack Asia-Pacific and has no ownership interest or economic exposure to those assets. Keep that sentence in the story because it is exactly the kind of dull fact that prevents bad analysis. A discounted credit vehicle is not the same thing as a forced sale of data centers. ## The Scarcity Is The Story You don't need to make this more mysterious than it is. Sydney and Melbourne are large cloud markets. Singapore remains constrained. Japan has land and power limits in the places where hyperscale buyers want to be. Johor Bahru has become more important partly because it sits near Singapore and can absorb some of the demand Singapore cannot. When governments want more sovereign AI capacity and cloud providers want more regional capacity, the existing operators with land and power move up the food chain. That is the point for builders and investors. The AI trade is not sitting only in Nvidia chips or frontier models. It is also sitting in the physical estate that lets those chips run. A model announcement can move quickly. A 360-megawatt campus in Sydney cannot be wished into existence next quarter. Read the $30 billion figure as a price on scarcity, not just enthusiasm. If the sale launches as reported, the real test will be whether the final buyer takes the whole Asia-Pacific portfolio and how much debt the market will tolerate against it. Until then, the clean conclusion is simple: the asset is current, the process is not closed, and the reported bidder field should be kept to the names actually reported. **Also read:** [ByteDance is seeking $20 billion offshore to fund an AI buildout that goes far beyond TikTok](https://startupfortune.com/bytedance-is-seeking-20-billion-offshore-to-fund-an-ai-buildout-that-goes-far-beyond-tiktok/) • [Ripple beats the MiCA clock and Luxembourg becomes crypto's new EU address](https://startupfortune.com/ripple-beats-the-mica-clock-and-luxembourg-becomes-cryptos-new-eu-address/) • [Nvidia's banned chips are selling for twice their price in China and Washington still thinks the export controls are working](https://startupfortune.com/nvidias-banned-chips-are-selling-for-twice-their-price-in-china-and-washington-still-thinks-the-export-controls-are-working/)