Megaport (ASX: MP1) surges 28% on $254m AI compute contracts: What comes next

Megaport, the Brisbane-based network-as-a-service provider, jumped more than a quarter on Thursday after its US compute subsidiary Latitude.sh locked in three major contracts worth a combined US$182.9 million ($254 million). The deals are with two unnamed US technology customers running artificial intelligence and inference workloads. For retail investors watching MP1 on the ASX top gainers board, the question is no longer whether the company has an AI strategy. It does. The question is whether the strategy can actually pay back the capital it consumes, and what that means for the next 12 to 18 months.

What exactly did Megaport announce and why has the MP1 share price moved so hard today?

The MP1 share price opened sharply higher on Thursday, hit an intraday high of $13.51 (an early gain of around 37 per cent), and settled near $12.58, up 27.72 per cent on $79.2 million in turnover. The catalyst is a single ASX release. Latitude.sh, the compute platform Megaport acquired in late 2025, has signed three contracts spanning GPU, CPU, network and storage services. Two of the contracts run for 36 months and account for roughly 90 per cent of the headline value. The third runs for 24 months. Combined, they generate about US$65.2 million ($90.6 million) in annualised recurring revenue once fully deployed.

The market reaction reflects two things. First, the contracts are committed revenue regardless of customer usage. That is the structure long-only investors actually pay up for in a compute business, because it removes the on-demand pricing volatility that has weighed on hyperscaler-adjacent stocks. Second, one of the customers is already a Megaport network client, which validates the cross-sell thesis underpinning the entire Latitude.sh acquisition rationale. After a punishing year in which the stock had fallen from above $17 to a 52-week low of $7.36, Thursday’s move erases the year-to-date loss and resets the conversation around MP1.

How does the Latitude.sh acquisition explain Megaport’s pivot from network to AI compute infrastructure?

To understand the contracts, you have to understand why Megaport bought Latitude.sh in the first place. The acquisition was announced in November 2025 and closed before the end of last year. Megaport paid an upfront price of US$150 million, representing roughly 3.5 times the target’s annual recurring revenue at signing, with up to three contingent earnouts of US$50 million each tied to revenue performance. The total deal value could reach US$300 million ($459 million) over time. Megaport funded the upfront component through a $200 million institutional placement at $14.30 per share, a price that retail investors now look back on as a high-water mark.

The strategic logic was straightforward. Megaport already operated one of the largest software-defined network platforms in the world, connecting more than 1,100 data centres across 26 countries. Latitude.sh added on-demand CPU and GPU compute on top of that network fabric. The combined business pitches itself to enterprises and AI developers as a single platform where you can spin up dedicated compute and the private connectivity to use it. Chief Executive Michael Reid has framed this Thursday’s contracts as a “defining moment” for the merged entity, arguing that the shift in AI workloads from foundation model training toward inference and edge deployment plays directly to a globally distributed compute and network platform.

What does the $140 million capex commitment mean for Megaport shareholders and balance sheet risk?

The contracts require approximately US$101 million ($140.3 million) of incremental capital expenditure, primarily for NVIDIA GPUs, supporting compute, networking and storage hardware. Payback on the hardware is approximately two years, after which the assets remain inside the Latitude.sh pool and can either be renewed by the same customers or redeployed across the on-demand platform. Megaport has reaffirmed its FY26 revenue guidance of $302 million to $317 million and its EBITDA margin guidance of 21 to 24 per cent of revenue. Group capital expenditure guidance of $90 million to $100 million remains unchanged for the existing footprint, but the new contracts sit on top of that and lift the FY26 capex envelope by the full $140.3 million if hardware arrives before 30 June 2026.

Funding for the new spend comes from a combination of existing cash reserves and an upsized A$150 million debt facility. The market response on the day suggests investors are comfortable with debt-funded growth at this stage of the cycle, particularly because the contracted ARR is locked and the payback profile is short. The trade-off is leverage. Megaport is moving from an asset-light network platform with predictable free cash flow toward a more capital-intensive compute model. Retail investors should track net debt and interest cover at the August 2026 full-year results closely. If the FY26 result demonstrates clean delivery on the new contracts and disciplined gearing, the equity story holds together. If hardware delivery slips into FY27, revenue recognition moves with it, and depreciation and interest accrue regardless.

Why are analysts split between a bull case above $20 and a bear scenario near $5?

The sell-side has a wide range of views on MP1, and that dispersion itself is informative. Across the 15 analysts publishing 12-month targets ahead of Thursday’s announcement, the average price target sat at $15.72, with a high of $23.98 and a low of $9.00. Twelve of the 15 carry a Buy recommendation, three are at Hold, and there are no Sells. Citi has been the most bullish, with a target of $16.05. JPMorgan moved to Neutral earlier in 2026, flagging elevated execution risk around the simultaneous push into discretionary research and development, sales build-out, and infrastructure reinvestment. The H1 FY26 result, posted in February, included a $19 million net loss, which fed the bear case that Megaport has stretched itself thin.

The bull case is that the company sits at a structural inflection point. Group ARR reached $338 million at the half year, up 49 per cent year-on-year. Network ARR including India hit $272 million at 31 March 2026, up 23 per cent in constant currency. Latitude.sh on-demand ARR climbed 31 per cent in four months to US$58.7 million by 25 April, before counting either the April $35.4 million contract or Thursday’s announcement. If those growth rates carry through to FY27 and the new compute contracts hit their payback timeline, the company should reach a level of profitability that justifies the higher end of analyst targets. The bear case is execution. AI infrastructure demand can compress on pricing well before a 36-month contract reaches renewal, and a capital-intensive growth model funded by debt offers less margin for error than the original network-only Megaport.

What is the milestone timeline between today’s announcement and Megaport’s next major catalyst?

The phasing matters because it determines when revenue and earnings actually show up in the accounts. Megaport has placed orders for the GPU, network and storage hardware now, with delivery expected in late FY26 or early FY27. Deployment will occur on a phased basis starting in the first half of FY27, which means roughly the September 2026 to December 2026 quarter for the first activations. Revenue recognition follows deployment, not contract signing, so the headline $90.6 million ARR number begins to flow into reported results from H1 FY27 onwards.

Between now and the next results event, the calendar is short. The full-year FY26 results land in August 2026 and will be the first hard test of the combined Megaport and Latitude.sh group. Investors will want three things on that release. First, evidence that the underlying network business has continued to compound the 23 per cent constant currency ARR growth seen in March. Second, an FY27 guidance range that incorporates the new compute contracts cleanly, with disclosed assumptions on deployment timing and gearing. Third, clarity on whether the contingent consideration milestones tied to the Latitude.sh acquisition are now in line of sight, because hitting those would imply the deal pays for itself faster than the original three-year structure assumed.

How are retail investors on HotCopper and Twitter/X talking about MP1 after the contract announcement?

Retail conversation around Megaport has shifted in tone over the last two weeks. The April $35.4 million Latitude.sh contract had already triggered a 9 per cent intraday rally, and at that point most retail commentary on HotCopper was characterising MP1 as a beaten-down rebound trade with the AI infrastructure thesis still unproven. Citi’s $16.05 target was the most frequently cited bull anchor in the threads. Thursday’s announcement is materially larger than the April deal and is the first piece of news that gives retail bulls a clean answer to the standard bear question, which is whether Megaport can actually win contracts at the scale needed to justify the Latitude.sh purchase price.

On Twitter/X, the cashtag $MP1 has seen elevated traffic through the Asian session, with the conversation centred on three points. The first is the comparison with global AI infrastructure peers and whether the ASX is properly pricing exposure to the same thematic. The second is the question of whether today’s gain is short-covering or the start of a genuine re-rating, given that MP1 had been one of the most heavily-shorted technology names on the ASX through early 2026. The third is the August results event and what guidance management can credibly issue without overpromising on FY27 deployment. The dominant retail framing is that the contracts validate the strategy. The unresolved retail question is whether the next leg requires fresh news or whether August alone is enough.

How does Megaport’s AI compute pivot compare with the broader ASX technology and global AI infrastructure trade?

Megaport sits at an unusual point on the ASX technology map. It is not a pure software-as-a-service growth story like Xero, which posted FY26 results on the same day and saw shares fall 9 per cent despite a 31 per cent revenue lift. It is not a hardware or semiconductor name. It is closest in profile to global AI infrastructure plays such as Equinix or Digital Realty in scale terms, and to a wider cohort of compute-as-a-service operators in business model. The ASX gives investors limited direct ways to express the AI infrastructure trade, which is one reason MP1 has been volatile. Capital flows in and out of it as the global narrative on AI capex shifts.

The macro backdrop favours the Megaport thesis on inference workloads, which require lower latency and geographic proximity to end users than training workloads. That is precisely where a globally distributed compute and network platform earns its premium versus a centralised hyperscaler. The countervailing risk is that the inference market itself becomes commoditised faster than the 24 to 36 month contract durations Megaport has signed. That is the scenario the bears at JPMorgan have been pricing, and it is the reason the stock can carry both a $23.98 high target and a $9.00 low target inside the same analyst consensus.

What are the key takeaways from the Megaport share price surge on the $254 million Latitude.sh AI compute contracts?

• Megaport (ASX: MP1) jumped 27.72 per cent to $12.58 on $79.2 million in turnover after Latitude.sh signed three contracts worth a combined US$182.9 million ($254 million) with two US AI customers.
• The contracts add roughly US$65.2 million ($90.6 million) in annualised recurring revenue, with around 90 per cent of total contract value sitting in 36-month commitments and the balance running for 24 months.
• Incremental capex of US$101 million ($140.3 million) is required, primarily for NVIDIA GPUs and supporting hardware, funded through existing cash and a new $150 million debt facility, with a stated payback of approximately two years.
• FY26 revenue guidance of $302 million to $317 million and EBITDA margin guidance of 21 to 24 per cent of revenue have been reaffirmed. Deployment of the new hardware begins in the first half of FY27, meaning the new revenue stream lands in reported results from late 2026 onwards.
• Analyst consensus across 15 brokers sits at $15.72 with a $23.98 high and a $9.00 low. JPMorgan moved to Neutral earlier in 2026 on execution risk, while Citi remains the most bullish at $16.05. Twelve of 15 analysts retain a Buy.
• The next hard test is the FY26 full-year result in August 2026, where investors will look for clean delivery on the contracts, a credible FY27 guidance range, and evidence the Latitude.sh acquisition is on track to hit its contingent consideration milestones.
• The retail thesis hinges on whether the August result confirms today’s announcement as the start of a genuine re-rating or whether the share price needs further contract wins to sustain its move back through the $13 level.