The RA project is very relevant. Without digging in to the...

The RA project is very relevant. Without digging in to the reports, there are 24 telcos handed to RA for go-live and at approx 20 are live. Thats around 500 physical GPU's and 6,000 vGpu's at max capacity running Windows games at 720P and 30+ fps.

There havent been any unusual business impediments but it has been challenging to get the product versions rolled out to tackle the larger markets economically (India and Indonesia in particular). Version 2.8 added Android support in VM's but Version 3.0 thats being finished now adds Windows/Linux/Android support in containers and consolidates the streaming engine for video/sound/controller into a shared engine. It also employs the new AMD encoder cards for server I/O and that allows us to push a lot of the buffer moves to pointers in the card drivers rather than memory ops at the application layer. Long way of saying that version 3.0 is the one that really packs a performance punch across all game libraries.

Cloud gaming is like video was. We moved from video players to DVD and each time there was a tech shakeout. VHS/Beta, Bluray/[Toshiba] then the internet got fast enough and Netflix killed all that. In gaming its PS5, Xbox, PC, Nintendo and the industry punters say that by 2028 you wont care because youll play in the cloud. But that outcome is driven by performance and economics This next version of our softwre should deliver performance at a price point that opens up the volume markets to game publishers to deliver form the cloud.

The same tech can be used to launch AI applications but the value proposition for that is much different until AI driven interactive video is a common thing. Almost everyone on earth has played a video game but not many of us have personal AI trainer bots counting our pushups yet. When thats default, we'll be ready.

FY25 will see ARR reported across three sectors of Telco, SaaS and GPU for the full year. It will be the first year we could do that and show a clear trend.