Megaports 827M AI Bet|Customers or Competitors?

The Day the ASX Stood Still for One Stock

On a day when the Australian sharemarket recorded its worst weekly performance in nearly a month, with the S&P/ASX 200 finishing down seven-tenths of a per cent, one company did not get the memo. Megaport, listed on the ASX under the code MP1, surged more than thirteen per cent to finish at $18.77 — the standout gainer across the entire index. The reason was a capital raise so large, and a contract announcement so sudden, that analysts spent much of the session simply trying to work out what had just happened.

Four new artificial intelligence infrastructure contracts appeared within forty-eight hours, with a combined value of approximately $458.9 million Australian. To fund the capital expenditure required to deliver those projects — and to establish what the company is calling a GPU pool — Megaport launched an $827.3 million entitlement offer. The institutional component, priced at $14.00 per share, was completed with a take-up rate of approximately ninety-nine per cent. That is as close to unanimous as institutional markets ever get. Chief Executive Michael Reid framed the moment plainly: "We are at the forefront of an accelerating inflection point across the industry. As use cases shift from AI foundation models to inference and the edge, Megaport is becoming an essential platform for powering the applications of tomorrow."

That confidence landed well. The question is whether the price it is landing at makes sense.

The Number That Does Not Fit

Here is the tension that the session's price action does not resolve. Institutions subscribed to Megaport's raise at $14.00 per share. The market then valued the stock at $18.77 — a thirty-four per cent premium to the subscription price on the same day. But the discounted cash flow valuation published by analysts puts fair value at $10.25. That is not a rounding error. It means the stock is trading at eighty-three per cent above what conventional cash flow modelling says it is worth, and thirty-seven per cent above even the discounted entry point institutions just paid. The company reported a net loss of $20.25 million in its most recent period.

The unstated premise that the institutions and the market are both taking as given is that Megaport's AI infrastructure contracts will compound into durable recurring revenue before hyperscaler competition internalises the same work. That premise is doing enormous lifting in the valuation, and it is not evidenced by historical cash flows — it is entirely a forward bet on the durability of third-party AI infrastructure demand. The contracts are real, but they did not exist a week ago. The speed of their appearance is itself part of the argument: AI build-out is happening so fast that Megaport's contracted revenue pipeline materialised before analysts had time to price it in. That velocity is the bull case. It is also what makes the valuation so hard to anchor.

The retail entitlement tranche is still open. Existing shareholders who do not subscribe face dilution from the approximately thirty-six million new shares the institutional component will issue, with the retail tranche adding further. The decision retail investors now face is whether to commit fresh capital at $14.00 — below where the stock trades but above what the models say it is worth — into contracts that carry no operating history.

The Hyperscaler Risk That Does Not Go Away

There is a structural question buried inside Megaport's business model that the contract announcements sharpen rather than resolve. The customers signing these AI infrastructure deals — large United States technology providers — are the same hyperscale operators that have spent the better part of a decade building their own private networking and interconnection capacity. Amazon Web Services, Google, and Microsoft have each invested heavily in proprietary infrastructure that, in theory, reduces their long-run dependence on third parties.

The bull case says AI inference and edge deployment is so fragmented, so geographically distributed, and so urgent that no hyperscaler can build fast enough on their own. Megaport's CEO is articulating that case explicitly, and at ninety-nine per cent institutional take-up, the argument is being accepted. The bear case says that once the immediate capacity crunch eases, the same companies who signed these contracts will quietly internalise the work — the same pattern that has played out in cloud storage, content delivery, and compute over the past decade. The one-month share price return of eighty-four per cent and the three-year total shareholder return of one hundred and twenty-nine per cent suggest the market has firmly chosen the bull case. But a company can be directionally correct and still be overvalued at a specific price point.

To understand why institutions subscribed at ninety-nine per cent despite that risk, it helps to look at the broader AI environment this week. Anthropic — the San Francisco AI company whose Claude model now authors more than eighty per cent of the code merged into Anthropic's own codebase — filed for a United States initial public offering with a valuation of $965 billion. Its run-rate revenue crossed $47 billion in May, up from $14 billion in February. In a striking piece of timing, the same company simultaneously called for a global pause on building the most powerful AI systems, warning that the latest models show signs they could escape human control. The company compared the challenge to nuclear arms control, but harder to verify — AI training is easier to hide than a missile silo.

The paradox is that the most credible voice in AI safety is also the fastest-growing AI company in the world, and it is rushing to list publicly before any pause could be implemented. That contradiction — caution and acceleration existing simultaneously at the top of the industry — is precisely the environment in which four large AI infrastructure contracts can appear from nowhere and push a mid-cap ASX stock up thirteen per cent in a single session. It explains why ninety-nine per cent of institutions subscribed. It does not explain what happens if the AI build-out pace slows.

What the Next Thirty Days Will Tell Us

The verification point for Megaport's positioning is not a date — it is a price. If the stock holds above $14.00 through the close of the retail entitlement offer and into the commencement of trading for new shares on approximately 15 June, the market is signalling that the gap between subscription price and current price is sustainable, and that institutions are not planning to sell into the retail offer window. If the stock trades back toward the $14.00 level — or below it — it signals that the gap between institutional subscription and open-market sentiment is narrowing in a direction that disadvantages retail participants who subscribed late.

The broader condition for continuation is that AI infrastructure demand from hyperscalers remains externally sourced rather than internalised. Every new contract Megaport signs with a large United States technology provider is evidence for the bull case. Every announcement from AWS, Google, or Microsoft of expanded proprietary capacity is evidence against it. Those signals are continuous, not binary. The market's eighty-four per cent one-month re-rating assumes the former is the dominant signal for the foreseeable future.

The condition under which this logic breaks is not a single event. It is the participant structure. If the ninety-nine per cent institutional take-up was driven by short-term positioning into the share price pop — rather than by conviction in the long-run revenue durability — then the real test begins when the entitlement offer closes and the lockup on new shares lifts. The stock's behaviour in the fortnight after 15 June will say more about the quality of that institutional endorsement than any analyst note published today. A stock that holds above $16.00 through mid-June suggests the positioning is durable. A stock that retraces toward $14.00 suggests the subscription rate reflected opportunism rather than conviction.

The rest of today's session offered its own set of unresolved tensions. Northern Star Resources fell more than seven per cent despite gold trading near historic highs at $4,378 per ounce, continuing a pattern that has frustrated gold-sector investors for months. ANZ and Macquarie cut fixed mortgage rates in a week when the Reserve Bank was being asked to explain whether government spending is making its rate decisions harder to execute. Resolute Mining fell sharply on Mali production disruptions. Against that backdrop, Megaport's thirteen per cent gain felt less like a market event and more like a referendum on whether Australian institutional capital has decided that AI infrastructure is the trade of 2026. The vote was ninety-nine to one. The question is whether the retail entitlement — still open — carries the same conviction, or just the same price.