Key takeaways

  • FortifAI Limited is the renamed Mighty Kingdom Limited, a video-game developer that consolidated its shares one-for-five in July 2025, changed its ASX code from MKL to FTI, and completed the acquisition of FastAI and its exclusive Nol8 technology licence on 2 February 2026.
  • Nol8 is an FPGA-accelerated data-in-motion engine licensed from Israel's Technion. Version 1.0 launched on 29 June 2026 and has joined Megaport's AI Exchange, but only for proof-of-concept deployments, with general availability targeted for late 2026 and no Nol8 revenue reported to date.
  • All current revenue still comes from the legacy games studio. The company funded itself through A$5 million and A$15 million equity placements in 2026, and the shares have since traded below the A$0.715 April placement price.

FortifAI Limited is among the ASX small-cap board's more complete re-inventions. Until mid-2025 it was Mighty Kingdom Limited, an Adelaide-founded game developer that listed in April 2021, worked on licensed Power Rangers and Star Trek titles, and spent most of its listed life short of cash.

The change was mechanical as well as cosmetic. Shareholders approved a one-for-five consolidation in July 2025, the name became FortifAI and the ticker moved from ASX:MKL to ASX:FTI. On 15 December 2025 the company agreed to buy Perth-based FastAI Pty Ltd and an exclusive licence to a technology called Nol8, completing on 2 February 2026.

Latest announcement or development

On 29 June FortifAI launched version 1.0 of the Nol8 AI Data Plane. A day later it said Nol8 had joined Megaport's AI Exchange, a software-defined network spanning more than 1,100 data centres. The company stressed this covers proof-of-concept deployments, not commercial release, with general availability targeted for later in 2026.

On 8 July it lodged a notice of general meeting for 7 August in Perth. The resolutions are administrative: ratification under ASX Listing Rule 7.1 of shares already issued in the February and May placements, plus lead manager shares and Petra Capital options. Ratification restores placement capacity; it raises no money.

Recent financial and operational performance

The audited accounts still describe the old business. For the year to 30 June 2025, revenue was A$4.33 million, down about 15 per cent, with a net loss of A$1.99 million against A$6.99 million previously, an improvement driven by restructuring, not growth.

On a trailing basis to 31 December 2025, third-party compilations show revenue near A$3.01 million, a net loss of roughly A$0.81 million and operating cash outflow of about A$0.65 million. All revenue comes from the Mighty Kingdom studio. Nol8 contributed nothing, having nothing to sell.

Progress has been measured in people rather than customers. Yosef Keret, former Weebit Nano chief executive, joined the board on completion, and Kelly Herrell, previously head of data platform Hazelcast, took the top job on 1 June 2026.

Business model and principal assets

Nol8 is a data-in-motion processing engine developed at the Technion, the Israel Institute of Technology, and licensed through its research foundation. Rather than running pattern matching and rule evaluation in software on general-purpose processors, it compiles that work onto field-programmable gate arrays, so classification, governance and routing occur at wire speed as rules scale.

The intended model is consumption-based, priced per gigabit-second and sold through NeoCloud infrastructure providers, with management targeting gross margins above 95 per cent. Beyond the licence and the Celerriem subsidiary holding it, the only other asset of substance is the games studio.

Growth opportunities

The market case rests on the arithmetic of agentic AI: idle graphics processors, compounding unstructured data volumes, and governance rules multiplying as enterprises impose controls. From that, management derives a data-movement layer worth roughly US$120 billion by 2030.

Nearer-term opportunities are narrower and more testable. Converting Megaport proof-of-concept deployments into paying workloads would be the first real validation; reaching general availability by end-2026, as guided, the second. Signing NeoCloud partners to resell the product would avoid building a direct sales force.

Balance sheet, funding and cash flow

Funding has been entirely equity: A$5 million in February 2026 through 16,666,667 shares at A$0.30, then A$15 million in late April at A$0.715. Reported cash of A$2.89 million at 31 December 2025 predates both raisings, and the June-quarter Appendix 4C was not lodged at the time of writing.

Consideration for FastAI was also scrip: 155 million shares at a deemed A$0.11 plus 150 million performance rights vesting against technical and commercialisation milestones, with the Technion foundation taking a royalty on net sales. Shares on issue now sit near 345 million, and dilution has been heavy.

Principal risks

The central risk is plain. This is a pre-revenue technology asset attached to a small services business. Benchmarks are not customers and a proof of concept is not a contract. The end-2026 platform date is a target, not a commitment, and accelerated data processing has a long record of laboratory numbers that resisted commercialisation.

Competition spans hyperscalers, networking incumbents and well-funded private firms that often already own the customer. The technology is licensed rather than owned, adding counterparty and royalty risk, and core engineering sits in Israel. With no Nol8 revenue, further raisings look probable, and the earlier suspension under the old identity remains on the record.