FortifAI’s NOL8 proves its AI mettle across Megaport’s global network

FortifAI's NOL8 platform passed live testing across Virginia, Tokyo, and Sydney via Megaport's global network, delivering byte-identical data governance results and processing 34 terabytes per day compared with 2.6 terabytes for its legacy software benchmark. FortifAI CEO Kelly Herrell cited a ten-fold cost advantage, and the company will now pursue global scaling with design partners across Megaport's network.

FortifAI’s NOL8 proves its AI mettle across Megaport’s global network NOL8 delivered byte-identical data governance results across live testing in Virginia, Tokyo and Sydney The platform processed governed data at 34 terabytes per day, compared with 2.6 terabytes for FortifAI’s legacy software benchmark FortifAI will now pursue global scaling with design partners across Megaport’s network Special Report: FortifAI’s NOL8 platform has passed its biggest test yet, showing that it can govern AI data consistently across three continents while delivering substantial speed and cost advantages over legacy software. The live testing connected environments in Virginia, Tokyo and Sydney through Megaport’s global network. The exercise covered pre-embedding, inference and agent-to-agent use cases. These are key steps in how AI systems prepare information, generate answers and communicate with other AI agents. Across every geography and use case tested, FortifAI’s ASX:FTI https://stockhead.com.au/company/fortifai-fti/ NOL8 platform returned the same governance output, byte for byte. That’s crucial for large organisations running private AI. This is when sensitive data and AI workloads are kept within controlled infrastructure, rather than handed entirely to public AI services. One policy, global network NOL8 is described as an AI Data Plane: the layer that controls, prepares and delivers data to AI systems while enforcing governance policies as the information moves. The testing showed that one policy could be applied consistently wherever an AI workload was running. This gives enterprises greater freedom to move data between cloud providers, while retaining governance control. FortifAI CEO Kelly Herrell said the results provided “applied proof” of the NOL8 V1.0 engine. “When we took the same workloads to three continents across the Megaport global network, NOL8 returned the same answer, byte for byte.” She cited a ten-fold cost advantage. NOL8’s in-region engine latency was measured at three milliseconds, in line with earlier testing. Across continents, distance was the only variable recorded. The test environments were hosted in public clouds, while Megaport provided the connectivity linking the locations. More governed data for the dollar The economic results were just as important as the technical ones. NOL8 governed data at a rate of 34 terabytes per day, compared with 2.6 terabytes per day for FortifAI’s benchmark legacy software implementation. NOL8’s throughput also declined by less than 1%, when the number of governance policies increased by 400%. FortifAI said the tests did not reach NOL8’s processing ceiling. This meant the reported performance figures represented measured floors at the tested workloads, rather than the platform’s maximum capacity. When throughput was measured against hardware cost, NOL8 delivered more than ten times the governed data per dollar of infrastructure, compared with the CPU central processing unit -based software benchmark. That could become increasingly important as enterprises grapple with the cost and power requirements of expanding AI infrastructure. Getting more work from expensive GPUs The same testing found NOL8 could remove up to 64% of data from the AI pipeline before it reached the graphics processing units, or GPUs. GPUs perform the heavy computational work. Filtering out unnecessary data earlier means those expensive chips can spend more time processing useful information. FortifAI said this could allow every megawatt of GPU capacity to deliver up to 2.8 times more useful AI work. At the same time, NOL8 itself consumed a fraction of the power required by a CPU-powered software pipeline. Megaport CEO Michael Reid said enterprises increasingly wanted the flexibility to run AI across clouds, GPU providers and their own data centres. “As AI infrastructure becomes more distributed, enterprises will need new ways to move data privately between these environments while maintaining consistent governance.” What comes next FortifAI will now move towards globally scaling NOL8 V1.0 with design partners across Megaport’s network. That network connects major global hyperscalers, neocloud GPU providers and private data centres. This gives NOL8 a potential route into enterprises running AI workloads across multiple infrastructure environments. The next commercial test will be converting the platform’s technical performance into expanded design-partner deployments and, ultimately, paying customers. FortifAI cautioned that the timing, expansion and conversion of those design-partner arrangements are not assured. Performance may also vary across production workloads, data types and governance policies. Still, the latest program gives FortifAI its strongest technical evidence yet that NOL8 can operate across global AI infrastructure while maintaining the speed, governance consistency and cost efficiency large enterprises will demand. This article was developed in collaboration with FortifAI, a Stockhead advertiser at the time of publishing. This article does not constitute financial product advice. You should consider obtaining independent advice before making any financial decisions. Related Topics UNLOCK INSIGHTS Discover the untold stories of emerging ASX stocks. Daily news and expert analysis, it's free to subscribe. By proceeding, you confirm you understand that we handle personal information in accordance with our Privacy Policy https://stockhead.com.au/privacy-policy/ .