By Andrew Macken and Jessica Dharmasiri
Over the past several months, a powerful narrative has taken hold across global equity markets: that AI, and particularly the rise of AI coding agents is about to make traditional enterprise software obsolete.
The argument is straightforward: if AI can write code faster, cheaper, and perhaps even more capably than humans; why should customers continue to pay for incumbent software at all?
The market’s verdict has been swift and severe. Years of assumptions underpinning the technology sector – particularly around durability, long-term growth and pricing power – have been reassessed, resulting in an aggressive selloff of software stocks. Investors have dubbed this phenomenon the ‘SaaS-pocalypse’.
At Montaka, we have been studying this dynamic carefully. We believe the market’s conclusion that AI agents will rapidly displace established software platforms is substantially overstated. In fact, we believe that the indiscriminate nature of the sell-off may even create meaningful opportunities for investors who are willing to separate the noise from reality.
What the market gets right
Before explaining where we diverge from the prevailing market narrative, it is important to acknowledge what it gets right – because the market is not wrong about everything, in our view.
AI agents are real, and they are genuinely transformative. Tools like Claude Code have fundamentally altered the economics and speed of software development. Tasks that once required teams of engineers working over months can now be prototyped in weeks, or even days. The barrier to building functional software has dropped dramatically, and will likely continue to fall.
This matters – AI agents are absorbing a large and growing share of work that software engineers have traditionally performed. The ability to build has effectively been democratised.
Where we disagree with the market is the conclusion that is drawn from these facts. The market has taken the observation that code is becoming cheaper and easier to produce, and extrapolated it into a narrative that incumbent software platforms will be displaced. This logic is fundamentally flawed, in our view.
Where the market gets it wrong
The market’s misjudgement seems to stem from both a misunderstanding about where durable competitive advantages really lie and the practical constraints on deploying AI agents at scale.
-
Distribution is the competitive advantage, not code
Market participants that have sold-off substantially all software – and software adjacent – companies appear to be misdiagnosing a key source of competitive advantage that certain leading software platforms exhibit: scaled distribution. That is, a very large customer base into which existing software and associated datasets are embedded.
Montaka has never owned software companies for software’s sake. It is not about the code itself, and it never has been. We focus on strong and strengthening competitive advantages. Scaled distribution reflects a long history of everything from customer sales, implementation, integration, data governance, security, compliance, and employee training – none of which can be replicated by better code alone.
We agree that AI coding agents will deflate the value of raw code itself and increase its copyability. But this new world strengthens the value of distribution. If anyone can build a product, the question becomes: who can get it in front of customers, integrate it into their workflows, and earn the trust required for enterprise adoption. The answer, overwhelmingly, is the incumbents who already have it.
-
Security risks will limit the deployment of unconstrained AI agents
The market’s narrative has been captured by the idea of swarms of unconstrained AI agents being deployed internally by companies – largely in place of human employees. Montaka believes that given security concerns, enterprises are highly unlikely to allow AI agents to have unconstrained read and write permissions on their internal datasets and code base.
In practice, as we’ve heard from numerous enterprise customers and security experts, the reality of deploying agents into production is much more incremental. In order to deploy these securely and effectively, careful access controls to specific datasets and tools need to be granted, deliberate instructions on narrowly defined tasks need to be defined, and evaluation metrics and systems need to be implemented. Indeed, this is why it is taking Salesforce, ServiceNow and Microsoft much longer than anticipated to see large-scale deployments of their agentic offerings, though they are coming.
To be clear, enterprise customers are deploying agents into production. But they are reliant on existing traditional software tools; they are highly constrained for security and compliance purposes; and they take real time to implement, orchestrate, evaluate, and refine.
-
There won’t be enough electric power for a long time
It is important to keep in mind that the ‘compute intensity’ of using agents can be 100x, 200x, or even 500x higher than simply engaging with your standard ChatGPT or Gemini chatbot.
For the market’s implied vision of swarms of agents running 24/7/365, displacing human employees and traditional software, to take effect, the western world would surely need an unprecedented step up in electric power generation and grid capacity.
Montaka has conducted detailed investigation into the timelines associated with bringing new electric power generation and grid capacity online. There are substantial bottlenecks that, based on our analysis, will not be meaningfully alleviated for many years to come. The western world’s ability to substantially accelerate the supply of electric power capacity is extremely limited for much of the next decade.
It is our assessment that there won’t be enough electric power available for the kind of large-scale disruption the market is now implicitly forecasting – even if there were no security concerns.
In our view, this power-constrained world favours the hybrid approach – between using probabilistic agents (whose outputs can vary each time they run) for some tasks, and traditional deterministic software (which executes the same way every time) for other tasks. And this is precisely the approach being pursued by the enterprise customers of the leading software platforms, such as ServiceNow, Salesforce, and Microsoft.
A period of price-to-value divergence
Markets, in many ways, can be narrative machines. Right now the dominant narrative is that AI will spell the end of enterprise software as we know it. We believe this narrative is wrong – not in its premises, but in its conclusions.
Yes, AI agents will make it easier and cheaper to write software. But the competitive advantages of leading software platforms are not a function of how hard their code was to write. They are a function of decades of trust, integration and institutional dependence that no agent, however capable, can replicate overnight.
Montaka’s portfolio reflects our convictions. Our holdings in leading enterprise software platforms like Microsoft, Salesforce and ServiceNow were never predicated on the scarcity of code. They were, and remain predicated on the enduring value of distribution, integration and the mission-critical role these platforms play in many of the world’s largest organisations.
We believe, in time, that the market will come to agree with our view that this is a period of significant price-to-value divergence. And when it does – as fundamental earnings power historically dictates – we believe that patient investors will be rewarded.
Note: Montaka is invested in Microsoft, Salesforce and ServiceNow
Andrew Macken is the Chief Investment Officer and Jessica Dharmasiri is the Chief of Staff to the CIO at Montaka Global Investments. To learn more about Montaka, please call +612 7202 0100 or leave us a line at montaka.com/contact-us
Podcast: Join the Montaka Global Investments team on Spotify as they chat about the market dynamics that shape their investing decisions in Spotlight Series Podcast. Follow along as we share real-time examples and investing tips that govern our stock picks. Click below to listen. Alternatively, click on this link: https://podcasters.spotify.com/pod/show/montaka
The Other Side of The SaaS-pocalypse
By Andrew Macken and Jessica Dharmasiri
Over the past several months, a powerful narrative has taken hold across global equity markets: that AI, and particularly the rise of AI coding agents is about to make traditional enterprise software obsolete.
The argument is straightforward: if AI can write code faster, cheaper, and perhaps even more capably than humans; why should customers continue to pay for incumbent software at all?
The market’s verdict has been swift and severe. Years of assumptions underpinning the technology sector – particularly around durability, long-term growth and pricing power – have been reassessed, resulting in an aggressive selloff of software stocks. Investors have dubbed this phenomenon the ‘SaaS-pocalypse’.
At Montaka, we have been studying this dynamic carefully. We believe the market’s conclusion that AI agents will rapidly displace established software platforms is substantially overstated. In fact, we believe that the indiscriminate nature of the sell-off may even create meaningful opportunities for investors who are willing to separate the noise from reality.
What the market gets right
Before explaining where we diverge from the prevailing market narrative, it is important to acknowledge what it gets right – because the market is not wrong about everything, in our view.
AI agents are real, and they are genuinely transformative. Tools like Claude Code have fundamentally altered the economics and speed of software development. Tasks that once required teams of engineers working over months can now be prototyped in weeks, or even days. The barrier to building functional software has dropped dramatically, and will likely continue to fall.
This matters – AI agents are absorbing a large and growing share of work that software engineers have traditionally performed. The ability to build has effectively been democratised.
Where we disagree with the market is the conclusion that is drawn from these facts. The market has taken the observation that code is becoming cheaper and easier to produce, and extrapolated it into a narrative that incumbent software platforms will be displaced. This logic is fundamentally flawed, in our view.
Where the market gets it wrong
The market’s misjudgement seems to stem from both a misunderstanding about where durable competitive advantages really lie and the practical constraints on deploying AI agents at scale.
Distribution is the competitive advantage, not code
Market participants that have sold-off substantially all software – and software adjacent – companies appear to be misdiagnosing a key source of competitive advantage that certain leading software platforms exhibit: scaled distribution. That is, a very large customer base into which existing software and associated datasets are embedded.
Montaka has never owned software companies for software’s sake. It is not about the code itself, and it never has been. We focus on strong and strengthening competitive advantages. Scaled distribution reflects a long history of everything from customer sales, implementation, integration, data governance, security, compliance, and employee training – none of which can be replicated by better code alone.
We agree that AI coding agents will deflate the value of raw code itself and increase its copyability. But this new world strengthens the value of distribution. If anyone can build a product, the question becomes: who can get it in front of customers, integrate it into their workflows, and earn the trust required for enterprise adoption. The answer, overwhelmingly, is the incumbents who already have it.
Security risks will limit the deployment of unconstrained AI agents
The market’s narrative has been captured by the idea of swarms of unconstrained AI agents being deployed internally by companies – largely in place of human employees. Montaka believes that given security concerns, enterprises are highly unlikely to allow AI agents to have unconstrained read and write permissions on their internal datasets and code base.
In practice, as we’ve heard from numerous enterprise customers and security experts, the reality of deploying agents into production is much more incremental. In order to deploy these securely and effectively, careful access controls to specific datasets and tools need to be granted, deliberate instructions on narrowly defined tasks need to be defined, and evaluation metrics and systems need to be implemented. Indeed, this is why it is taking Salesforce, ServiceNow and Microsoft much longer than anticipated to see large-scale deployments of their agentic offerings, though they are coming.
To be clear, enterprise customers are deploying agents into production. But they are reliant on existing traditional software tools; they are highly constrained for security and compliance purposes; and they take real time to implement, orchestrate, evaluate, and refine.
There won’t be enough electric power for a long time
It is important to keep in mind that the ‘compute intensity’ of using agents can be 100x, 200x, or even 500x higher than simply engaging with your standard ChatGPT or Gemini chatbot.
For the market’s implied vision of swarms of agents running 24/7/365, displacing human employees and traditional software, to take effect, the western world would surely need an unprecedented step up in electric power generation and grid capacity.
Montaka has conducted detailed investigation into the timelines associated with bringing new electric power generation and grid capacity online. There are substantial bottlenecks that, based on our analysis, will not be meaningfully alleviated for many years to come. The western world’s ability to substantially accelerate the supply of electric power capacity is extremely limited for much of the next decade.
It is our assessment that there won’t be enough electric power available for the kind of large-scale disruption the market is now implicitly forecasting – even if there were no security concerns.
In our view, this power-constrained world favours the hybrid approach – between using probabilistic agents (whose outputs can vary each time they run) for some tasks, and traditional deterministic software (which executes the same way every time) for other tasks. And this is precisely the approach being pursued by the enterprise customers of the leading software platforms, such as ServiceNow, Salesforce, and Microsoft.
A period of price-to-value divergence
Markets, in many ways, can be narrative machines. Right now the dominant narrative is that AI will spell the end of enterprise software as we know it. We believe this narrative is wrong – not in its premises, but in its conclusions.
Yes, AI agents will make it easier and cheaper to write software. But the competitive advantages of leading software platforms are not a function of how hard their code was to write. They are a function of decades of trust, integration and institutional dependence that no agent, however capable, can replicate overnight.
Montaka’s portfolio reflects our convictions. Our holdings in leading enterprise software platforms like Microsoft, Salesforce and ServiceNow were never predicated on the scarcity of code. They were, and remain predicated on the enduring value of distribution, integration and the mission-critical role these platforms play in many of the world’s largest organisations.
We believe, in time, that the market will come to agree with our view that this is a period of significant price-to-value divergence. And when it does – as fundamental earnings power historically dictates – we believe that patient investors will be rewarded.
Note: Montaka is invested in Microsoft, Salesforce and ServiceNow
Podcast: Join the Montaka Global Investments team on Spotify as they chat about the market dynamics that shape their investing decisions in Spotlight Series Podcast. Follow along as we share real-time examples and investing tips that govern our stock picks. Click below to listen. Alternatively, click on this link: https://podcasters.spotify.com/pod/show/montaka
This content was prepared by Montaka Global Pty Ltd (ACN 604 878 533, AFSL: 516 942). The information provided is general in nature and does not take into account your investment objectives, financial situation or particular needs. You should read the offer document and consider your own investment objectives, financial situation and particular needs before acting upon this information. All investments contain risk and may lose value. Consider seeking advice from a licensed financial advisor. Past performance is not a reliable indicator of future performance.
Related Insight
Share
Get insights delivered to your inbox including articles, podcasts and videos from the global equities world.