Yeah @lankypom, I agree with your view. I'm not sure if Google (and others) are a threat or an opportunity. I'm not sure anyone can know at this point.

I've taken the stance that I'll let the market tell me. I'll wait to see if Tony can keep the technology revenue growing at or near his goals. I'm open to both possible futures. If Tony can get to even 75% of his goals (for rev and margin) then Ai-Media will be mega. But I see an equally likely universe where the company kicks along with small growth primarily from legacy customers complying with regulations - in this case the market cap could halve or more.

I've taken a small position, about half my usual max position. I'll monitor the next couple of half year results, ready to bail if I sense strong headwinds.

Great summary @jcmleng, and love your coverage and insights for Ai-Media.

One correction, or clarification . . . I noted the professional services at US$200/hr (not US$20/hr). So, in blocks of US$20k. I thought this was noteworthy given they are pitching their new services as high end expert implementation support.

Even at this high hourly rate, I'm concerned about how this gets delivered efficiently. With Tony's goal of $150m for FY29, with 20% professional services, that's roughly 100,000 hours of professional services being delivered. At an ambitious 70% utlisation rate, that's an implementation workforce of about 70 people, plus people to manage those 70 people. Operating margin won't get close to the 80% he claims for tech.

@DrPete @jcmleng while on professional services revenue impact on margins, I think when Tony first announced these (earlier this year) he said he expected this revenue stream to generate a gross margin of 50% (I've pulled this from my model, which I updated when that revenue stream was announced as part of FY29, so I think that is correct.)

So you are right, margins will always be below the tech revenue margins.

For my valuation, I have Tech Revenue growing from %GM of 80% in FY26 to 83% in FY29, and for Services Revenue, I have this flat at 50%.

So, in one scenario, I have FY29 Tech Revenue of $100m and Services Revenue of $20m, which results in a Total Review of $120m, with a %GM of 77.5%.

At the "target" 4:1 split of $120m/$30m, at my FY29 margins, you'd get a %GM of 76.4%, and if there is no operating leverage in Tech Margins, so that it stays at 80%, then at the target, you'd get an overall %GM of 74%,

So, I think long term, $AIM will achieve %GM probably in the range 73% to 78%, depending on actual mix and margins achieved.

My point is, when tracking $AIM's progress it is quite simple and also helpful to model the two revenue streams.

Importantly, the valuation is much more sensitive to the Tech Revenue outcome (obviously).

Now, whether they can achieve that %Gross Margin for Services at that scale is another question. If they can, if would put them at the higher end of what is achieved in IT professional services consulting more generally (35% to 50%).

@DrPete, my calculator skills are not quite they used to be is my excuse!

$200/hr or $1,600/day is a nice blended rate for consulting - this would be an experienced/semi-experienced consultant level. From experience, good IT consulting/implementation/project margin is 40-45% margin, skewing towards 45%. Good BAU support margin is ~30-35% (lower skilled, more routine support/issues), skewed towards 30%. But resource cost will also vary accordingly - lower for BAU folk, higher for implementation folk. 

This is where the AIM flat rate block hours model is a really good one. Regardless of the nature of service, the revenue is a flat $200/hr. So if you get 40-45% margin on implementation, your margin on BAU support is potentially higher than that because of the lower cost resource.

This is very much in @mikebrisy's 35-50% range. Tony’s 50% is probably not a bad thumb suck number for the reasons above.

In terms of scaling, I suspect they will build a decent pool of contractor labour that will give them flex capacity, and they will be managed by project managers, who can either be AIM or contractors. I am thinking that the skillset will likely be skewed towards a more generic network, hardware infrastructure, technical support, that sort of space - not too hard to build a pool of these sorts of resources. I also suspect AIM will offshore this ...

Harder and costlier will be technical architects/deep skilled configurators who do the orchestration, configuration (all the AIM-specific bits), cyber security specialists - they must have a core group of these folk and spread them across the customers. 

If they get a good pool of resources with good flex capacity, Services will add to AIM’s moat. But it is critical that they get the resourcing model right - bad things will likely happen if they stuff that up with bad/cheap resources.

Will be keeping a close watch to see how they evolve this space. Tony has experience managing a pool of human translators but Jason will be key in building an IT services organisation. He sounds like he has done this before, so my confidence at this stage is high, but it will all come down to execution.

But as you point out, the question is whether a scale up zero to $30m of services revenue by 2029 is actually operationally feasible, and a good assumption to make ... with no actual cost numbers to work with, and looking at purely the number of labour hours required, it does feel rather ambitious. 

@mikebrisy, fully agree we need to watching these 2 streams separately/differently. Your $20m revenue feels more comfortable. I would be inclined to use a more prudent $15m revenue as a modelling start point, then adjusting upwards once we get some trend numbers.

I think Tony was happy with the recent interview we did here in SM so wanted me to do the moderating.

All it took was a little bit of praise and I was ready to agree to anything. Don't judge me! Haha

For full disclosure, there's no commercial arrangement (not that I'm above that kind of thing, but it wasn't on offer...sadly).

And yeah, I'm always a little wary of overly promotional investor relations too @twee.

Ha ha ha anything? A bit of praise and you fold like a lawn chair hey @Strawman? You sl… ;-)

I know, I know.. I'm not proud of myself :)

Tbh, I've always been a bit of a believer in saying "yes" often, especially when there's little cost if it does prove a waste of time. And most of the time it does. I can't tell you how many stupid investor related things I have had to endure over the last 25-odd years or so..

But occasionally you make a connection with someone, or pick up a few useful nuggets. It's that whole idea of giving luck a chance to strike, if you get me.

Nothing ventured, nothing gained!

Absolutely. Make the most of it @Strawman - hopefully we’ll get to see! Have a good weekend.

Is this a new side hustle (in the least) for our Grand Poobah @Strawman ? ;)

Great webinar. What a salesman.

And sometimes, the product being sold is also really good. It reminds me of the launch of NutriBullet. A blender? But then you watch the pitch a few times and you think, damn it I want one. 10 years later I use my NutriBullet every day.

It's the encoder. No one else can do what they do. 8 seconds verses 28 seconds. Trust. Accreditation. Workflow. Moat.

'The big tech companies are both our suppliers and our customers'. It doesn't make sense, but he keeps saying it, and each time with a few extra little details, and it seems possible. There are little fish in the ocean, that clean the teeth of the big fish.

There's no guarantee that their expansion and sales strategy works exactly as he laid out. And until it does there will be that doubt - how is it even possible to stay ahead of the curve? But someone has to be there. Cleaning the teeth.

@lastever - best analogy I've heard in a long time: little fish cleaning the big fishes' teeth.

That's one I'll remember.

With greater than 50% of the US broadcast market AIM are in a strong position to leverage their market access via their encoder base with an expanding product offering. LEXI Text is clearly going well for them and this is providing them with significant goegraphical expansion opportunities. LEXI Voice is a big a big part of their FY29 EBITDA $60m target but its success is not a lay down misere. This is an innovation race that wont stay still. As @twee pointed out in this post recently, Google, with deep pockets funding its research are innovating hard in this area and its novel solution to end to end speech to speech translation could be a technology paradigm shift that renders AIM's technology defunct. Maybe AIM hit lag limits with their technology solution. What I particularly liked about Google's solution is they have a "forced alignment algorithm" that gives the translated speech the same timing and rhythm as the original speech. This and the fact that the translated voice sounds the same as the original voice give a very natural sounding translated experience. Google is also doing this with a 2 second delay. I know that Google are playing in a difference space with its Google Meet and its end user to end user translation on its pixel devices but, as I said before, this is an innovation race and if AIM have an inferior product then their first mover advantage and encoder moat will be eroded pretty quickly as competitors move in. Potentially big ones like Google. I'm watching this one closely because it's a big holding for me.

That's impressive from Google@Shapeshifter. Hopefully another strawperson understands this better than me, because honestly, I still don't understand how AIM addresses this kind of advancement, but Tony must aware of it. He mentioned voice-to-voice today. Is the idea that they offer that google product on the encoder for the live-streaming niche? Or he's confident they can do the same thing. Anyone clear on that?

@Shapeshifter, @twee . Totally get the concern and and am also highly vigilant of this space. The question for me is: Are these Google improvements a thing to fear for AIM, or is it actually a positive thing for AIM?

There is no doubt that the hyperscalers already have better translation technology today than AIM, but that “better” is in the context of its own ecosystem. Meaning, if you have end-to-end Google in your setup, then you will benefit from any improvements in translation technology, today. And this will absolutely continue as Google continues to innovate. 

But this superior Google translation technology is not deployable into a Fox live broadcast feed, in its present form, as it was not designed for use in those environments. Which is where AIM comes into play through the encoder/iCap Cloud/Lexi/Lexi infrastructure and the orchestration around these components with the customer’s other media infrastructure and all piped through the broadcasters live network pipe etc. That is not a space that the hyperscalers play in at all, today.

If the hyperscalers intended these translation capabilities to be used in an enterprise scenario, and enabled it plug-and-play, AIM would already be completely stuffed and irrelevant today as the orchestration and infrastructure should be completely redundant, technically. The customer would have to decide AIM or Google/Microsoft etc. Right now, I suspect this is not even a conversation as there is no viable technical alternative. 

The question then is, what will change in the next 5 years in this enterprise scenario, assuming Google continues to aggressively innovate in speech-to-speech translation. I see 2 scenario’s:

1. If Google continues to focus on innovating for its own ecosystem, then the likely outcome is “no change for enterprise” - they will still need the AIM infrastructure/service/orchestration as there will be no practical way to use the new Google technology in an enterprise setting, just like today

2. The more likely development is that for AIM to stay relevant with its customers, it will be forced to innovate and follow the path that Google is now taking and/or adopt the Google technology and bake it into the AIM ecosystem - this would be a win-win for both customer (access to better technology and service) and AIM (moat tightens further).

What would be a very serious, thesis breaking development is if any of the hyperscalers suddenly move away from the current “my-ecosystem-only” focus and pivot its innovation to/into the enterprise/B2B space, and directly challenge AIM’s infrastructure/orchestration. The minute this smells like a possibility, I will be out of AIM in a jiff.

I am thus hyper vigilant not so much on the technology innovation itself (more of it is better for everyone), but on whether the hyperscalers pivot these innovations into the enterprise space. That for me, is the bigger thing to focus on, not the innovation itself.

Good points @Shapeshifter and good comments on the broadcast side @jcmleng, as I said in the other thread, my take is nuanced. I don't think the delay matters as much in broadcast (already a delay built in) but I think it matters more for the United Nations interpretation application and other enterprise applications.

I see the broadcast customers as dinosaurs competing with tech native companies, AIM role is to prolong the life of its customers in a world where video everything is growing online but legacy broadcast is dying and ripe for consolidation. So the good thing here is that you don't need the latest tech nor will your customers be able to change their workflow to use something else but the bad thing is your customers could go extinct.

@lastever I don't think google product will be an off-shelf product in the short-term since they are ahead of everyone else. Google/alphabet motivations for products like google meet and buds is probably just to retain users on android, a defense against apple/microsoft reducing its ad-reach. It's interesting how they integrate it into youtube as well. By that logic they are unlikely to publish the model or target enterprise.

On the live event side, there's few things to unpick and I'm still not totally there. AIM often boast of doing the AWS live events, this is interesting because Amazon sells transcription services and has its own streaming platform. It's not that they couldn't do it if they wanted but Amazon's focus was to improve their transcription service rather than build specific applications. This relies on having the encoder, which is probably only viable when you have a big budget and the full sound setup. I personally have been to small conferences where some of the presenters are remotely presenting from a zoom call, it's a bit janky but you could imagine a presenter being on a google meet call now and people at the conference who didn't speak the presenter's language joining the call to get their language of choice. No encoder required and you still have the translation happening just once for consistency and it's arguably secure too. Now, Tony mentioned in our last interview that's there's no unmixing of the audio being done, well, in my experience Microsoft Teams is pretty good at damping the dog barking/baby crying in my meetings so they are achieving a similar effect. AIM's moat really relies on their being an encoder and that's most useful for those with traditional broadcast/sound workflows, their current customers in the US (with more potential customers in Europe).

Google is not only evolving the software rapidly but also the hardware and this is starting to elimilate the need for cloud based translation processing.

For example with the Google Pixel 10 phone the processing now happens on device and looks like this:

Audio capture on your device -> AI processors run the speech to speech models locally ie the translation encoding happens directly on your phone -> Google cloud only for packet routing -> audio delivery and decoding on the target device

My point is you have the phone doing all the heavy lifting - a small device that fits in your pocket. For end to end user translation this is going to be hard to beat from Google especially with their newer AudioLM speech to speech translation model with low latency.

This technology is evolving rapidily. Can a small Australian company with a AUD$174m market cap match it with a US$3.8t behemoth? I feel like the tech is important because it solves the problem. LEXI Voice needs to be a good product otherwise it wont be successful. Is LEXI Voice good enough?