Firstly, apologies for the effusive flattery and praise towards Tony on my part during the interview. Very poor form, and not the kind of neutral, objective style that best serves your interests. So please keep my bias in mind, and be sure to push back in any areas that are warranted.

I just give a lot of credit to a business leader than can deliberately disrupt their own business when the structural winds of change are blowing, especially when such a move will hit revenues and free cash flows in the interim, and also when the payoff for the pivot and associated investments are uncertain.

Also, and Tony himself made this point, it's noteworthy that he's been pretty consistent in what he has said previously, and what he has delivered. That's not too common on the ASX, so credit where it is due.

Anyway, I wont rehash some of the key observations with this business. They've been made before (see past interviews and the company page).

A few of the key points from this discussion:

• R&D and product development is largely complete. They are now firmly focused on the ongoing commercialization of the product.
• No acquisitions are on the radar, and in fact Tony was very forthright on this.
• Their partner sales model will allow them to grow much faster and much cheaper than any direct sales model, and speed is critical given the land grab that is underway
• Voice is a far larger market than text (transcription), about 30x larger, which is a big part of how they arrive at their ambitious FY29 targets.
• They have some impressive directors on the board, which should not only add great experience, but also open doors.
• Looks like a NASDAQ listing is on the radar, and they have been moving towards this for some time. Most of their business is there and they have been presenting their accounts in a US friendly manner for this exact reason.
• AI adoption in live translation has hit an inflection point -- “voice is where text was 23 months ago.”
• ASX liquidity has significantly improved. This is interesting only in the sense that it puts AIM in play for larger institutional investors.

It feels like there's a lot of upside if the company can indeed deliver on their FY29 targets. But let's say they only get halfway there ($30m in EBITDA, or lets call it $15m NPAT) at a PE of 20, that's a market cap of $300m, compared with $180m today. Not a terrible margin of safety.

[Held]

AIM transcript.pdf

Ai-Media has just announced a new Long-Term Incentive Plan under which CEO Tony Abrahams will receive 1,000,000 Restricted Share Units (RSUs) over five years — 200,000 per year.

To earn them, all he needs to do is not walk away.

That’s it. No performance hurdles, no revenue or profit targets, no share price milestones. Just stick around, and every six months a new batch of RSUs will vest. Once vested, they convert into shares, which are then subject to a 12-month escrow.

I’m all for aligned executives, and those who deliver deserve to be rewarded. But Tony already owns ~35.9 million shares (about 17.2% of the company!) That’s an enormous stake. If alignment is the goal, he’s already there.

He also receives close to $500k per year (including super and leave entitlements). There were no short-term bonuses or equity grants last year, so perhaps this LTIP is meant to fill that gap.

Still, handing out performance-free equity to a founder-CEO with a massive existing stake feels… unnecessary.

Retention matters, but so does accountability. If you’re going to dilute shareholders by another 10 million shares over five years, some clear performance conditions seem like a minimum requirement.

Here are some notes from today's meeting with Tony.

His enthusiasm is somewhat contagious, but hopefully we can distill some objective insights from what he said.

Encoders = Trojan Horse

AI-Media's encoders (acquired via EEG) are deeply embedded in customer workflows (e.g. Channel 7, Foxtel).

Replacing them isn’t simple; it would require reconfiguring entire tech stacks, including broadcast automation systems.

These encoders come with proprietary firmware and network effects. clients already have them, so switching is costly and unattractive. And the encoder industry is very niche, and far too small for a large challenger to enter and compete.

Encoder + SaaS Model = Printer + Ink

Initial encoder sale is like selling the printer (~$10K each), but the real value comes from the ongoing SaaS revenue (the ink).

That takes the lifetime value of each encoder from ~$7K to ~$50K as they now drive much higher SaaS usage.

ICAP Network as the Toll Road

ICAP is the underlying data network that connects encoders to AI/ML models in the cloud. Its critical infrastructure no one else has at scale. It acts like a "toll road," monetized regardless of which AI engine is used.

Tech-Agnostic Integration

They can easily plug in APIs from any leading LLM (OpenAI, Anthropic, etc.). They don’t build the engine; they fuel it. In Tony's words, AI-Media is the fuel injection system, not the engine.

Moat in Live Video

All the tech is optimised for live video, which is hard to do well. Real-time accuracy + speed + low cost = hard trifecta.

Most competitors are focused on recorded content, but AI-Media is entirely focused on live, where there's real defensibility.

What Most People Miss

Headline Numbers Hide the Real Story. On the surface, total revenue looks flat or down, but the mix is shifting from legacy services (people doing the work) to scalable tech (machines doing it). Margins and cash flows are actually improving as low-margin services drop off.

Strong Cash Flow Despite Transition

Still delivering positive cash flow and EBITDA, despite deliberately cannibalizing their legacy business and investing for growth. Oh, and most investment is not capitalised, but rather expensed all upfront. Tony said he prefers to have a "clean" set of financials.

Transition = Growth Hidden in Deferred Revenue

SaaS sales are growing fast, but revenue is deferred (12 months upfront) and not yet recognized. Real growth is understated in reported numbers.

Encoder Ecosystem is Portable

They've proven they can run their software stack on third-party encoders (e.g. Grass Valley, Imagine, Media Proxy).

Opens up huge optionality for future partnerships or acquisitions.

Future SaaS-Driven Growth

SaaS now includes Lexi (captioning), Lexi Voice, Lexi Brew (LLM-enabled organizational comms). Lexi Brew lets orgs use private LLMs securely—like internal ChatGPT but with data privacy.

Targeting $60M EBITDA by FY29

Organic plan in place (encoder expansion, Lexi suite, geographic rollout). But also exploring inorganic growth via M&A (especially targeting companies with large encoder bases but no SaaS layer.)

Land Grab Mindset

Acknowledges that first mover advantage matters in embedding into workflows.

Prefers low upfront M&A cost with earn-outs based on encoder conversion success.

Revenue breakdown

Two Main Buckets:

Services Revenue: Anything with a *human in the loop* (e.g. stenographers, respeakers).  

   - Charged hourly, recognized immediately.

   - This used to be 100% of the business at IPO.

   - Now shrinking rapidly, aim is for it to be just 20% of revenue, acting as a wrapper around tech.

Tech Revenue: Everything else (fully automated, AI-driven).  

   - Includes both **hardware sales (encoders)** and SaaS/software.

Within Tech, they break it down into:

  -Hardware (e.g. encoder boxes) – paid upfront, lumpy, not recurring.

  - SaaS – subscription or usage-based, recurring, higher margin.

But even SaaS Has Layers  

  SaaS includes:

  -Encoder-based SaaS (e.g. Alta, Falcon).

  -ICAP SaaS – toll-road style fees for using their data network.

  -Lexi SaaS – captioning and voice AI services, true growth engine.

  To isolate the “real” recurring growth engine, they exclude:

  - Hardware revenue.

  - Encoder-based SaaS.

  - ICAP SaaS (supporting legacy workflows).

  - Support services tied to hardware.

What’s left is the core Lexi SaaS, which is the cleanest indicator of long-term tech-driven success.

Re Deferred Revenue & Contract Liabilities: 

  - SaaS deals are increasingly paid 12 months upfront.

  - Revenue recognition is deferred, so it doesn’t show up immediately in the P&L.

  - This distorts comparisons and masks underlying growth, but cash is banked.

Accounting rules treat this as a liability, but Tony argues it’s a bit of a fiction: “It’s going to cost us $200K to deliver, but it shows up as a $2.7 million liability.”

That means the cost to fulfill is only ~10% of the deferred revenue.

He called it “completely misleading”, because it makes the company look like it owes more than it really does.

Anyway, i probably missed some stuff. But hopefully the AI helped me extract the key messages.

There's a lot about AI-Media that is easy to miss, so I wanted to try and elucidate the key aspects of the business, its offering and the competitive advantages it has -- as informed by today's conversation with co-founder and CEO Tony Abrahams.

But, to be honest, i'm not confident i've got things right so please correct me if you think I'm off base.

First, as @mushroompanda has already said, they aren't building AI models themselves. AI-Media’s technology relies on APIs (Application Programming Interfaces) to connect their software with these external AI engines.

AIM's products provide context to AI engines, such as metadata from broadcast streams (e.g., identifying speakers, locations, or specific program segments). This customization improves the accuracy of AI-generated captions by using additional data.

Their competitive edge comes from effectively embedding these AI models into a unique, customer-focused delivery system which integrates into existing workflows. It's more about ingesting audio/video feeds, extracting the relevant information, and sending it back in real-time so it cam be inserted/overlaid into the broadcast.

The acquisition of EEG (a provider of encoding hardware) was really a pivotal moment for AIM, giving it control over the critical hardware needed to feed audio data into AI models, allowing seamless integration with their cloud-based captioning services.

The total addressable market for AI language services is vast, estimated at around $70 billion annually. Tony pointed out that AI-Media’s current focus—live speech-to-text and live captioning—represents just a small fraction of this market, approximately 1%.

At the moment, around 90% of AI-Media’s operations are centered on live captioning and transcription, primarily driven by their LEXI solutions. However, Tony stressed that this is just the tip of the iceberg, indicating that the company is still in the early stages of tapping into the broader market potential.

A significant portion of the market opportunity lies in recorded media, which accounts for about 25% of the total market. This includes transcription and captioning for pre-recorded content like TV shows, movies, and online videos.

There are also considerable opportunities in broader AI language services such as voice dubbing, audio description for visually impaired audiences, and other multilingual services. This segment, which Tony identified as the larger $69 billion part of the market, involves using AI to handle voice processing, translations, and enhancing accessibility through audio descriptions.

The market for AI language services extends beyond broadcasting and includes government, enterprise, education, and entertainment sectors. Each of these industries has unique needs for language services, from live captioning and translations to complex workflows

Tony highlighted that one of the key drivers of market growth is the reduction in costs associated with AI-powered services compared to traditional human-based models. For example, automated audio description, which traditionally required 25 hours of human labor per hour of content, can now be fully automated, significantly reducing costs and making these services accessible to a broader range of content creators.

He also said that regulatory requirements for accessibility, such as mandatory captioning and audio descriptions, are increasingly pushing broadcasters and content providers toward adopting AI solutions. This trend further expands the market opportunity for AI-Media’s products.

As mentioned at the results briefing (or potentially earlier?) Tony outlined ambitious financial targets, including reaching $150 million in revenue and $60 million in EBITDA within five years. 

And he said they wouldn't need to raise capital to pursue this -- all the pieces are in place and growth can be driven by organic cash flow.

He certainly has his money where his mouth is, recently acquiring 5m more shares to lift his stake to almost 17% of the business (he bought at 31c -- not a bad trade so far!).

Anyway, it seems that AI-Media is a genuine market leader in a fast growing market that offers increased service and lower cost for its customers. It's well capitalised, cash flow positive, founder-led and expecting to 15x EBITDA in the next 5 years.

Shares are on ~27x EV/EBITDA

I don't presently hold in real life, but will be adding a watching position here on SM today.