So what is the bear case for PME?
Thinking on paper here…
Many on here will know PME better than me, what am I missing?
Essentially the bear case seems to be that with the recent progress of AI capabilities, PME’s formidable strength has become their fatal flaw.
CEO Sam Hupert alluded to this on the H1 call. They are a capital light software only installation. Further that they have only lost customers infrequently but when they did it was based on price, not functionality.
Monkeys and typewriters – The AI bull case
If it’s true that an infinite number of monkey pounding and slapping on the keys of an infinite number of typewriters would eventually produce the complete works of William Shakespeare, what does this imply for AI?
An army of AI agents / bots could spin up even more AI agents / bots and this swam can generate code very quickly at very low cost for as long as it takes to adapt, improvise and overcome until they have software as good as the solution that any SaaS business provides?
If this is true for any software business, the most lucrative targets likely will be hit first. Alternate AI projects will also be swarming so multiple competing solutions would likely be in a race to get there first. As more arrive at this destination, price would come down as competition swings from almost non-existent to intense.
How long all of this might take is anyone’s guess but the train has left the station and the direction of travel does not seem in question.
Your margin is my opportunity
However, someone / something will have to kickstart this AI swam in the direction of PME’s software and while this is relatively cheap compared to humans coding, it is not costless. The compute and opportunity cost will need to be factored in before pointing AI capabilities at a problem.
With PME’s margins, dominance, software only status, what better target to aim for?
Contracted revenue
Sam also mentioned on the H1 that they have over $1bn in contracted revenue.
It’s hard to believe hospitals will be trying to rip up contracts compelling them to pay this over time.
As hospitals are primarily concerned with healthcare, not AI prophecy, it’s also hard to believe that the current pipeline of contracts will not convert at close to historical rates.
Sam said the current pipeline remains "robust" and has grown in both volume and diversity. Historically their Win Rate has been 80% or higher in major US tenders.
So they have plenty of high margin revenue coming down the pipe in terms of contracted and potential future wins.
The current contracted revenue should translate to about $0.5bn in NPAT over time.
That’s a lot of potential funding for AI spend of their own.
Barriers
Switching costs – not likely all that high, Sam said when they install PME, radiologists are more productive from day 1, suggesting here is not a lengthy training period and an alternative solution would likely be just as easy to use. I’m likely oversimplifying this.
Economies of Scale –The scale that AI seems capable of at relatively low cost suggests AI has the potential to undermine pure software companies who have historically enjoyed Economies of Scale. Not likely quick or easy for challengers to build scale though.
Network effects – Likely strong but not insurmountable by AI solutions that are a lot cheaper. Training doesn’t seem intensive, so it’s not like there’s an army of radiologist drilled in the specifics of Visage to the detriment of other tech.
There are a few genuine barriers I think PME have.
IP – this is the big one and the target AI would be aiming for. Software will be the initial target, but there are other IP / intangible barriers in the form of reputation, relationships, etc that should work in PME’s favour, especially given the conservative nature of healthcare. A slow sales cycle and thus far sticky customers would be a big test for would be challengers.
PME seem to have embraced AI early and are looking to be the platform that peripheral solutions can sit on, including AI innovations from others.
Management say that the more PME improves, the further they move ahead of their competition - although I think part of this might be that they are the only native could solution. AI natives will be looking to use this playbook against them.
If their dominance can make them the de facto industry standard, they will be much harder to replace. However it doesn’t seem clear to me that this is a winner takes all market, and they only have 10% of the US market today.
Hospitals seem unlikely to spin up their own solution, so a credible AI threat would likely come from an AI specialist or an adjacency that gets repurposed to replace Visage.
Hardware providers embedding Agentic AI solutions in their equipment might be a trojan horse to reduce the power of Visage in the market?
Peripheral threats
Hackers - I’m reliably informed that hacking should become a lot easier with AI. Makes sense. While AI should also help defend against attacks, hackers only need to be successful once, defence needs to work 100% of the time. AI’s prophesised capabilities could also give cover to hackers who find their way into parts of PME’s software code.
Margin Pressure - Even if PME are able to keep their growth ticking along, they may be less inclined to increase prices as much as they have in the past.
Continued price increases may be justified and commercially acceptable given the productivity improvements PME’s software delivers but it would also make them a bigger target for AI solutions to aim for.
Increased internal spend on AI and other defence costs could also impact margins.
Outlook from here
Pipeline should also hold up in the near term and continue generating growth for PME.
Management seem alive to the threats and opportunities of AI and have a strategy to make the most of it.
If they weren’t already, they will now be aware of the market’s concerns, so messaging should improve from here.
It’s hard to think that such a dominant, well run, deeply entrenched industry leader for a mission critical service will be wiped out or even severely dented by AI any time soon.
I will be looking for evidence that management are taking steps to proactively defend themselves from the coming onslaught (assuming it does materialise over time). They will need to do this beyond just building AI solutions of their own. This will mean doing things that AI can’t do and entrenching their dominance beyond what a newer, better software solution could replace.
Otherwise their cloud native solution than now dominates the market could be eroded by an AI native solution(s) over time.
Disc: Held
