Yeah good point @pubenvelope

Ive no direct experience in that industry, let alone like install an encoder! So there's a good deal of faith in accepting things at face value.

It sounds right, but that doesn't mean it is! Would also be curious if anyone in the know can corroborate.

@pubenvelope , I do not have direct broadcast encoder experience to be 100% certain. But here is how I am evaluating and thinking about Tony’s 2 comments:

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

I can relate to Tony’s comments based on my past life managing critical, single-point failure core network switches, in a decent-sized 24x7 ops organisation. That core network switch is the single hardware device through which ALL network traffic, into, within, and out of the company, runs through - really only ONE aspect of what the AIM encoders must do. If the switch fails, the company’s entire IT network, and operations given the high IT dependence, comes to a grinding halt until the issue is fixed. 

The core switches are costly and have a 5-7 year lifespan, but is replaced well before failure/EOL. The replacement is a massive 2-3 month planning exercise as the company goes without network access for up to 24 hours, longer if there are issues (a big deal for 24x7 operations). We thus have almost total inertia to switch to a competitor switch, even if it costs less - trying something new which could lead to network instability/availability, requires new technical skills, no immediate technical expertise/experience to address issues etc, being the key reasons.

Based on this experience, the analogy Tony gave that the encoders are like the “windows operating system”, it has the software smarts, the fact that ALL broadcasting traffic goes through that single encoder real time (no traffic = black screens), all AIM”s orchestration around captioning/translations going through that encoder, the encoder is significantly more critical to a broadcasters real time operations and significantly more complex, than the core network switch I was messing around with. It is not only a piece of replaceable hardware (or in the case of Alta and Falcon, software), it is the software “brains” in and out of iCap/LEXI, which in turn manages the orchestration complexity of live captioning/translations.

So, while I will never understand the actual technical position around encoders, I really have no reason to doubt Tony’s comments around encoder replacement complexity. I think a customer will only consider totally switching encoders to a competitor as part of an overall exercise to completely boot AIM out of its shop. 

On the second comment "..And the encoder industry is very niche, and far too small for a large challenger to enter and compete."

• I think Tony made that comment purely in the context of the EEG Physical Encoders - at $10.5k a pop, massive customer inertia to change - his comment makes sense. The business case won’t stack up for a competitor to try to barge in and compete on the EEG physical encoders at that selling price.
• Where it got confusing for me was when Tony talked about how easy it is to layer on the EEG proprietary software onto NEW encoders - he talked about Media Proxy, Grass Valley, Imagine, plus there is the Alta and Falcon encoders etc, and he also talked about future inorganic growth around other encoder networks
• This was where it crystallised for me that while the EEG encoders gave AIM the leg up to full automation of US broadcasters 3+ years back, the encoders themselves, today, is really not the critical piece of the puzzle nor is it the moat

It is not the encoders themselves, but rather the EEG proprietary software sitting inside an encoder (can be EEG Encoder Pro, Alta or Falcon) , which has the pipes into the iCap network, that is the technology stack moat. 

I thus do not think encoder competition, by itself, is a worry. There are and will be other encoders, doing a bunch of other things in customer networks.

What will absolutely be a worry and direct threat is if a competitor has (1) an encoder (2) software smarts that rival the EEG software on that encoder and (3) an iCap/LEXI-like setup OR an alternative approach to doing live voice captioning/translations minus iCap/Lexi. Even with this combined capability, they will still have to contend with deep customer inertia to switchover from a LEXI capability that is integral to their broadcast, which works fast, and is cheap. 

From my IT perspective, these are massive, massive technical hurdles to cross just to get to the starting line to compete with AIM on LIVE Video ...

Discl: Held IRL and in SM

Damn, we have a great brains trust! Good intel @jcmleng

@jcmleng you are like a dog with a bone this is great!

From your post:

• Where it got confusing for me was when Tony talked about how easy it is to layer on the EEG proprietary software onto NEW encoders - he talked about Media Proxy, Grass Valley, Imagine, plus there is the Alta and Falcon encoders etc, and he also talked about future inorganic growth around other encoder networks

This would only make sense if AIM were to capture these customers and move them over to the iCAP network (if not already on it) and onto LEXI SaaS. Any existing hardware encoders would need to be upgraded with the EEG proprietary software to allow this to happen. After all the hardware is just a mechanism that allows the software to run. As AIM moves further to software based encoders you would think it is becoming more hardware agnostic.

Mediaproxy looks like a good fit for AIM as real time broadcast stream analysis is what AIM are already doing. Not sure how this would meaningfully impact the business though.

Grass Valley is a broad solution for live media production across corporate, government and broadcast. Grass Valley was puchased by Beldin (NYSE: BDC) for US$220m in 2014. From Wiki:

• Grass Valley manufactures television production and broadcasting equipment. Its product line includes cameras, live production switchers, replay and highlights products, media storage, editing, asset management, modular infrastructure, routing, conversion equipment, control, monitoring, and automated playout, sold as hardware, software, and software as a service (SaaS).

Again not sure how this would fit into AIM's goals. Perhaps this was just a technical proof-of-concept but it was interesting to hear Tony spitball on this.

@Shapeshifter , hahaha, I badly needed a distraction from the news, no better way than to focus on your company and its moat!

Thanks for the links to Mediaproxy and Grass Valley.

Grass Valley looks equipment centric (inputs into the broadcast) and "recording"-centric (vs live video) and thus feels quite far away from what AIM does in terms of orchestration of the live video broadcast feeds. So lumping it in the "encoder" bucket feels totally strange. But the editing/replay/highlights may make sense in terms of use of the iCap network.

Mediaproxy is much closer, but it is stream analysis/compliance reporting centric. As it would have to have a broadcast feed go through the solution, it does make sense to put EEG software on it but instead of "encoding", it also feels more like an exercise to monetise the iCap network, rather than compete head-to-head with AIM's live video orchestration.

So I think you raise an important point ... the comment around "layering on EEG software on these ...", (I struggle to see them as "encoders") is perhaps really about increasing iCap/LEXI usage rather than adding more "encoders" into the mix. And yes, the hardware becomes increasingly moot over time ...

It is perhaps easier to think about the "EEG software" as the REAL technical enabler for iCap/LEXI monetising, not the encoder:

• an encoder minus the EEG software is useless to AIM as there is no pipe to the iCap network, thus no revenue.
• a non-encoder solution (GrassValley, Mediaproxy) with EEG software should theoretically enable access to iCap/Lexi = revenue

A note to myself is to be vigilant in how Tony uses the words "encoder" and "encoder network" and to put them in proper context!

The good thing is that we probably need not worry about these technical semantics for now as Tony has parked "other encoder networks" for inorganic growth later, to chase Enterprise and Government. But being more precise on this topic now will make it easier to understand the evolution of encoders later!

I had the same thought @jcmleng but my assumption is TOTAL encoders = hardware encoders + software encoders.

Surely it's in the name?

And despite the revenue recognition being different (upfront vs SAAS) both act as the same driving force for LEXI uptake which is the end game.

@Shapeshifter and @jcmleng I thought the same, but I am not 100% confident.

Here's my reasoning.

Yesterday when Tony gave his Russian Doll analogy, he at one point said Encoders was one layer, and within that are Software Encoders + Hardware Encoders. This was as a step along the way to peeling back to the kernel of SaaS Lexi Revenue.

So, whenever I see # Encoders, I am seeing = [ # Encoder Pro (HW) + # Falcon (SW) + # Alta (SW) ]

In the AGM speech, Tony refers to

"The installed base of 5,000 EEG Encoders in US/Canada was critical to the successful rollout of LEXI to capture 50% of total US/Canada professional live Broadcast captioning market at 90% gross margins, while successfully instigating charging on third party use of the iCap Network."

If you go right back to the presentation in 2021 announcing the acquisition of EEG, the slide below makes clear that the "Encoder" can be "cloud or on-prem hosted", i.e., hardware (on prem) or software (cloud).

I think the diagram below is unambiguous. "Video Content" has to flow through the "Encoder" which is represented as a device, but we know there are software and hardware pathways for this before it ends up in the iCap Cloud.

It makes sense to use this definition, because then # Encoders is a complete driver of Tech Revenue (i.e., everything not involving a human).

(As a side note, I observe that we are seeing a similar thing in $AIM as we are in $AD8. A transition from providing hardware with the software embedded and bundled with it, to unbundling hardware and software, providing software that can run on customer/OEM hardware. This creates an initially lower revenue, higher margin business, but expands the addressible market to being hardware technology agnostic. In both cases the value driver is the number of customers who are using the technology, and then converting this into a subscription service.)

So, with this logic, the relevant metric to report to shareholders is TOTAL ENCODERS.

@Shapeshifter , you would think so and agree! Not a biggie in the overall scheme of things, but wanting that clarity as that subtlety might help in understand encoder trends down the road and my making sense of potential confusion around revenue. Some raw thoughts around this:

• As more broadcasters (US or Global) move to the 2110 standard, Alta encoder volumes should rise and EEG encoders correspondingly fall which should show up in falling hardware revenue but rising SAAS revenue
• Will be interesting to see if the EEG encoder improvements evolve towards the 2110 standard or will it stay as it
• If EEG encoders do not evolve to 2110, does this mean that the physical EEG encoder manufacturing eventually get sunsetted over the coming years, and AIM moves all customers to an Alta