Megaport sells software-defined network connectivity between data centres, letting enterprises plug into any cloud provider in under 60 seconds. At A$9.65 versus our fair value of A$7.05, the stock is overvalued by 27%. The core network business is genuinely excellent — but the market is pricing full success for a $410 million acquisition we have only five weeks of financial data to evaluate.

Best fit: Growth or Thematic investors with a 2–4 year horizon and high tolerance for binary outcomes. The network flywheel is real and accelerating — NRR at 111%, Americas ARR up 24% in constant currency, customer lifetime extending. For investors who believe AI infrastructure demand is structural and that Latitude.sh will deliver positive margins, the growth and thematic cases are compelling. The catch is that the current price leaves no room for disappointment.

Megaport operates the world's largest software-defined network-as-a-service platform, connecting 8,500 enterprises across 1,034 data centres in 30 countries. Customers pay monthly to provision high-speed connections between cloud providers — AWS, Azure, Google Cloud — without touching physical hardware. The model is capital-light at the margin, with 72% gross margins and customers who stay an average of 13 years.

The first half of FY26 demonstrated that the organic network business is firing on all cylinders. Net Revenue Retention inflected to 111%, Americas ARR grew 24% in constant currency, and EBITDA margins reached 26.2% — comfortably above full-year guidance of 21–24%. The story changed materially in December 2025 when Megaport completed the $410 million acquisition of Latitude.sh, a GPU-as-a-service compute platform, adding a second growth engine aimed at AI infrastructure demand.

The investment tension is straightforward. The network business justifies a premium. The compute acquisition does not yet — five weeks of consolidated data, no disclosed margins, and $345 million in goodwill make it impossible to assess with confidence. The market appears to be pricing full success for both. At A$9.65 versus our fair value of A$7.05, the stock is overvalued by 27%.

What happened? The H1 FY26 result was strong across every network metric that matters. NRR hit 111% — the highest in the company's history — meaning existing customers are expanding spend faster than any churn. EBITDA of $35.3 million at a 26.2% margin exceeded guidance. The Americas, Megaport's most strategically important region, grew ARR 24% in constant currency. Customer lifetime value rose 57% in constant currency terms.

What's next? FY26 full-year results in August 2026 are the most important near-term event. This will be the first disclosure of compute segment revenue at scale — Latitude.sh contributed only $5.8 million in the five weeks of H1 — and any margin data will be closely watched. Management has guided for $302–317 million in full-year revenue and 21–24% EBITDA margins, both of which our forecasts sit at the top end of.

The binary question beyond August is whether Latitude.sh hits its CY2026 revenue milestone of USD $52 million. If it does, and margins are disclosed as positive, the compute thesis gets its first concrete validation. If it misses, a goodwill impairment on the $345 million balance becomes a real risk. Meanwhile, the HQ cost line — growing at 28% annually against 20.5% of revenue — is the margin expansion bottleneck that management must address for the FY28 and FY29 EBITDA targets to hold.

The fair value of A$7.05 blends a probability-weighted DCF (61% weight, A$6.54) with trading multiples (39% weight, A$7.81). The gap to the A$9.65 market price is explained by three debatable assumptions: we deduct the $207 million Latitude.sh earnout liability as a debt-equivalent claim against equity — worth $1.10 per share — while many market participants appear to exclude it. We also apply a 12.2% discount rate reflecting acquisition uncertainty, versus the ~10.5% implied by the current price. And we use a 14x terminal EBITDA multiple, below the 16–18x the market appears to price in.

The single biggest risk is Latitude.sh. We have five weeks of financial data on a $410 million acquisition. The stub period showed a small loss on $5.8 million of compute revenue. If margins remain negative at scale, the $345 million in goodwill becomes impairment risk — a non-cash hit but a real signal. The bear case, assigned 25% probability, values the stock at $4.85 on exactly this scenario. Entry below $7.00 would provide adequate margin of safety for the network business alone, with compute optionality acquired for free.