Megaport 827m AI Raise|SpaceX IPO Tests the Rotation

AI Inference or Momentum Trap?

Megaport raised $518 million from institutions in 48 hours, with a take-up rate of 99%. That number is the one the headlines led with, and on its surface it reads as straightforward confirmation of demand. What it does not explain is why that demand arrived at a $14.30 offer price — a 13.9% discount to Monday's close — and still produced essentially zero shortfall allocation.

A 99% take-up at a meaningful discount means institutions were not absorbing the offer out of passive obligation. They were making an active decision to add at a discount rather than stand aside. The position-pressure question that creates is not whether Megaport has demand — it clearly does — but whether institutions priced the four new AI infrastructure contracts at face value or discounted the execution risk embedded in $369.5 million of NVIDIA GPU capex that has not yet been deployed.

The $458.9 million in total contract value across those four deals is now the analytical load-bearing number. At a 99% institutional take-up, sell-side revision is already underway. Citi lifted its target 41% to $22.10 and raised FY27 EBITDA forecasts by 73% to approximately $255 million. That is a significant magnitude revision, and it tells you the prior models were not pricing AI inference contracts at this volume or margin structure.

What makes the Megaport move structurally interesting rather than just momentum is the inference thesis itself. Training AI models is a hyperscaler game. Inference — the moment a user actually queries a deployed model — can be distributed across a global network of connected data centres. Megaport operates 1,100 such locations across 31 countries. The Latitude.sh acquisition gave it a compute platform. The new contracts give it the contracted ARR base to justify the hardware. Combined group pro-forma ARR now stands at $662.9 million, with the compute division alone at $385.2 million of annualised recurring revenue.

Foreign institutional capital is the flow that moved first. The retail component of the $827.3 million offer does not open until 11 June and is sized at $309 million — meaning retail entry is structurally delayed, with new institutional shares not trading until 15 June. The participant structure at this moment is institutional-only on both supply and price discovery. That asymmetry matters for what happens when the retail component prices.

SpaceX Liquidity Pressure

The $827 million Megaport raise does not exist in isolation. It lands in the same week that Wall Street is running the SpaceX IPO roadshow — a $1.77 trillion valuation event targeting a Nasdaq debut around 12 June under the ticker SPCX.

SpaceX has been blocked from early S&P 500 inclusion. The index committee confirmed it will not waive the existing rules for megacap IPOs, meaning passive S&P 500 index funds — which represent the largest single pool of systematic equity demand in global markets — will not be forced buyers at listing. That refusal is not procedural. It means the SpaceX IPO must clear entirely through active and wholesale demand, without the $50+ billion passive fund absorption that S&P inclusion would generate.

That creates a specific liquidity demand on exactly the investor class that absorbed Megaport's institutional offer. Sharesies, Fidelity, and JPMorgan's wealth management arm are all actively distributing SpaceX access to retail and wholesale investors across ANZ. The minimum retail entry has been cut to USD 2,000. That breadth of distribution signals the deal is drawing from the same AI growth capital pool that institutional investors used to absorb Megaport's offer — the question is whether that pool has enough capacity to fund both events without price concession.

Yardeni Research's public framing — that supersized AI IPOs will not suck oxygen from markets — is the optimistic read. The structural counter is that SpaceX at USD 135 per share represents a direct competing bid for growth capital in the week Megaport's retail offer opens. Capital that enters SpaceX on 12 June is not in the retail entitlement pool on 11 June.

The flow signal to watch is not whether SpaceX lists successfully. It is whether Megaport's retail component achieves comparable take-up to the institutional 99% — or whether the shortfall rate rises, leaving the shortfall book to be allocated at the company's discretion, which would signal that the retail tranche is pricing at the margin of the AI growth capital pool's current depth.

Mining Rotation Risk

The capital rotating into Megaport and toward SpaceX is not appearing from nowhere. The ASX 200 fell 0.70% on the week, and within that index the sector decomposition is stark. Iron ore miners are absorbing sustained selling pressure as the spot price approaches USD 100 per tonne — the psychological threshold that triggers royalty and margin modelling resets across BHP, Rio Tinto, and Fortescue.

The structural source of that pressure is China's Mineral Resources Group, which now coordinates purchases for approximately 80% of Chinese steel mills. Annual contract negotiations concluded with CMRG pressing hard on price. Australian miners account for 65% of China's iron ore imports, but Guinea's Simandou project — backed by Chinese capital, expected to export 120 million tonnes annually at full ramp — represents Beijing's explicit diversification play. The first shipment departed earlier this year.

BHP's position this week crystallised the compounding risk. The company's Australian president warned publicly that Labor's industrial relations framework has made Port Hedland — the world's largest iron ore export terminal — more vulnerable to union action. A shutdown would remove roughly one million tonnes of export capacity per day at a moment when the price is already stressed to USD 100. The combination of buyer monopsony and seller labour disruption is a margin compression thesis that institutional models are beginning to price.

The capital flow that connects the Megaport raise to the mining selloff is not coincidental. Institutional rebalancing away from commodity exporters and toward AI infrastructure is the underlying rotation. The weight of evidence — 99% take-up on a discounted AI raise, iron ore near USD 100 with CMRG hardball, Port Hedland union risk unresolved — sits on the side of that rotation continuing rather than reversing.

The verification point is the Megaport retail take-up figure when disclosed after 11 June. If retail absorption matches the institutional 99%, the AI infrastructure thesis holds and the rotation deepens. If the shortfall rises materially — particularly in the same week the SpaceX IPO draws from the same capital pool — it is the first observable signal that the AI growth capital base in Australian markets has reached a near-term depth constraint, not a directional reversal of the rotation itself.