Australia | 10:00 AM
Megaport's strategic transition to an infrastructure-as-a-service provider has been reinforced and de-risked by the latest contract wins, underpinning further growth potential.
- Megaport's AI infrastructure contracts accelerate transformation to a global inference cloud platform
- GPU pool strategy converts on-demand AI compute to longer-term contracts
- Megaport offers differentiated exposure to neoclouds and data centres
By Danielle Ecuyer
Robust AI inference demand underpins growth
At the end of September, Megaport ((MP1)) announced three new AI infrastructure contracts with a total combined contract value of around $978.6m, including prepayments of around $322.6m.
Out of the three agreements, two are with new customers, with the total contract value encompassing GPU and CPU compute, network and storage for AI application and inference workloads.
Taking a step back from the latest announcements, Bell Potter, which initiated coverage of the stock, sought to remind investors Megaport is an infrastructure-as-a-service provider that operates a global platform for “deploying, connecting and scaling network, compute and storage on demand”.
Management has transitioned the company from a network-as-a-service provider, which facilitated and managed private high-speed network connections to data centres and cloud providers on demand.
The acquisition of Latitude.sh in late 2025 has brought forth exposure to compute-as-a-service, while partnerships with Wasabi Technologies and VAST have enhanced the offering to storage-as-a-service.
Megaport offers the neocloud stack of network, compute and storage for enterprises. Bell Potter considers Megaport as offering differentiated exposure on the ASX from other neoclouds like Firmus and/or Sharon.Ai --the former soon to list and the second is in preparation too-- as it is creating a globally distributed AI inference cloud which is less capital-intensive than building the physical AI factories or data centres.
Megaport leases space, i.e. racks and trays, in around 1,200 data centres around the world, including across 185 cities and 31 countries.
Against this backdrop, management created its GPU pool to attract new customers wanting immediate access to infrastructure, which can convert short-term usage to longer-term contracted deployments.
Megaport also indicated it will allocate GPUs already ordered for the GPU pool and has stated it will procure equipment to replenish the pool, which will add a further -US$252m (circa -$360m) in capex over FY27.
Both power and space have been secured for the new equipment needed to replenish the GPU pool.
What the new contract wins bring forth
Macquarie highlighted the announcement reinforced the GPU pool as a strategic advantage and, as identified by management’s announcement, capacity has the scope to initially underpin on-demand inference workloads, with scope to be channelled to longer-term contracts as opportunities come forth.
The analyst points out the pooled capacity paybacks are 16-22 months, supported by exposure to robust GPU pricing.
Significantly, most of the market volumes are characterised as long-term wholesale, i.e. over 50MW contracts.
A combination of strong relationships for smaller workloads enables Megaport to secure better deals, margins and paybacks than other neoclouds, with the broker pointing to a regular misconception in the market that all neoclouds sell at a spot price.
Management’s decision to allocate pooled GPUs to a new customer contract at a lower annual recurring revenue (ARR) per capex dollar is indicative of the focus on return. The customer’s prepayment of $282m to guarantee capacity has lowered the capex payback to around eight months, the broker explains.
Ord Minnett articulates the three contracts are expected to generate around US$163m ($232m) in ARR at full run rates by 4Q27, with all contracts secured via the subsidiary Latitude.sh.
The company pointed out contracts have a weighted average term of just over four years, with committed revenue over the fixed terms irrespective of customer usage.
Two contracts are with US-based technology companies, one existing, and the third is a new customer and publicly listed company deploying its own AI workloads.
Management highlights the new customers further diversify the customer base and reduce overall customer concentration risk. More than 85% of group pro-forma ARR is from North America.
Notably, the underlying trading momentum is also highlighted as remaining strong, with ARR billing above $500m currently. Network ARR achieved around $303m in August, with compute ARR at $201m as of September 22.
Macquarie emphasises network recurring revenue (NRR) remains robust and has gained momentum, accelerating to 116%, with constant currency growth of 29% y/y.
The growth is expected to be supported by ongoing “structural tailwinds” that are anticipated to rise, as existing cloud markets are fragmented and there is a growing focus on managing cloud costs with AI.
Management estimates it will have ARR of around $1.1bn and an annualised earnings (EBITDA) run-rate of over $650m with the full deployment of contracts that have been announced since April.
Network business brings growth and opportunities
Citi also points to the growth in the Network business as another positive.
Positively, Megaport was named as one of Nvidia’s Australian cloud partners to expand AI infrastructure in land, power and shell capacity.
The collaboration should bring forth better access to GPUs and an enhanced ability to sell to AI-native companies, the analyst points out.
Digging into the strategic importance of customer payments, Citi sees the prepayments as lower capital requirements.
Macquarie believes the prepayments are “meaningful” from a funding perspective. When aligned with the upsized debt facility of $845m from $825m, this analyst estimates management has total liquidity of around $2.2bn.
This compares to capex of around -$1.83bn, which will be used for delivering on strategic contracts, replenishing the GPU pool and possible US M&A, which was indicated at the FY26 result.
Backing out customer prepayments, Megaport’s funding is circa -$1.23bn for FY27 capex, which can be self-funded, leaving around $362m in available liquidity.
Bell Potter reminds investors Megaport raised $827m in equity in June/July and established a new $825m debt facility in August, which was upsized to $845m in September.
Management states Megaport is fully funded for all capex included in the updated FY27 guidance, including the strategic contracts and replenishment of the GPU pool, with around $362.2m in pro-forma liquidity.
Access to the full debt facility increases as revenue milestones are achieved and remains subject to agreed terms and conditions, including covenant limits.
Macquarie does flag upside risks to capex over the next 6-12 months from further wholesale contract wins. Bell Potter concurs sufficient liquidity is in place for current growth, but additional funding would be needed if further major contracts are secured.
Management also upgraded FY27 guidance, lifting revenues to $720m-$810m from $620m-$730m previously, and EBITDA margins to 42%-44% from 38%-40%.
Network revenue guidance was also lifted to $325m-$335m from $315m-$325m, while compute revenue guidance increased to $395m-$475m from $305m-$405m. Capex guidance increased to $1.78bn-$1.88bn from $1.28bn-$1.38bn, with the $500m increase reflecting the equipment costs for the new contracts.
The upgrades in margins were further evidence for Citi around robust operating leverage and incremental margin expansion. The contract wins were not considered a surprise given how strong demand is.
Morgans acknowledges the 25% FY27 earnings (EBITDA) upgrade, reinforcing strength across compute and network, both of which have continued to deliver above expectations.
The full story is for FNArena subscribers only. To read the full story plus enjoy a free two-week trial to our service SIGN UP HERE
If you already had your free trial, why not join as a paying subscriber? CLICK HERE