Can Software Survive AI?

Summary

---

Podcast Summary: Exchanges – Can Software Survive AI?

Host: Alison Nathan, Goldman Sachs
Guest: Gabriela Borges, Software Sector Analyst, Goldman Sachs Research
Date: March 10, 2026

---

Episode Overview

This episode of Goldman Sachs Exchanges explores the existential risks and opportunities that Artificial Intelligence (AI), particularly generative AI, poses to the software industry. Host Alison Nathan interviews Gabriela Borges, who provides insights from investor sentiment to sector dynamics, with a focus on what determines winners and losers in an AI-disrupted software landscape. The discussion draws from Borges's recent contribution to Goldman’s "Top of Mind" report.

---

Key Discussion Points & Insights

1. Why AI Is Raising Existential Questions for Software (00:05–02:30)

• Recent AI advancements have spurred investor anxiety: AI has transitioned from a source of hype to a source of deep concern for software investors, especially following rapid improvements in coding automations and broader enterprise adoption.
  • Notable Example: Anthropic’s CLAUDE code, which expanded from helping developers to aiding everyday knowledge workers.
• The competitive landscape is intensifying: Major announcements from rivals like Anthropic and OpenAI led investors to question the defensibility of incumbent software companies.

Quote:
“We’re finally getting enterprise adoption of AI... It generally takes a fair number of years before you go from the ChatGPT moment in 2022... to actually starting to see the benefits in the enterprise ecosystem.”
— Gabriela Borges (01:39)

---

2. The Nature and Durability of Software Company "Moats" (02:30–06:00)

• Domain experience and data as defensible moats:
  Companies like CrowdStrike have built up valuable, proprietary security datasets over a decade, using human-in-the-loop methods for threat intelligence.
• Incumbents must prove their edge:
  It’s not enough for legacy software companies to simply possess data or experience; they must demonstrate that their proprietary strengths lead to better AI outcomes for customers, especially compared to new AI-first entrants.
  • Case in point: Salesforce’s “agentforce” vs. third-party tools.
• How to evaluate competitive durability:
  Borges describes a three-part checklist for enduring moats:
  1. Replatforming technology stacks to reduce legacy tech debt.
  2. Building organic innovation capabilities and M&A agility to cover gaps and stay competitive.
  3. Proving monetization capacity for AI features by tracking adoption and pricing power.

Quote:
“These are the types of things that we’re looking for today’s software leaders to figure out if they will still be a leader three, five, seven years from now.”
— Gabriela Borges (07:10)

---

3. What Will Stabilize the Software Sector? (07:26–09:00)

• Numbers must trump narrative:
  Stability will come when companies show positive, measurable improvement in key business fundamentals (“numbers need to contradict the narrative”).
  • Investors track metrics like ARR (annual recurring revenue), LTV/CAC, and AI usage among other KPIs.
• Sustained fundamental performance is key:
  The sector’s key metrics have deteriorated since their post-pandemic peak, but there are signs of recovery, as markets recently rewarded positive earnings performance.

Quote:
“You can start to see stabilization in the metrics…that type of engagement where you see the market respond favorably to positive data points… those dynamics tend to be pretty good for the sector to recover over time.”
— Gabriela Borges (08:15)

---

4. Investor Sentiment and Shifting Ownership (09:00–11:50)

• Value-based investors are entering:
  Historically, high-growth software stocks were seen as too expensive. Improved GAAP earnings and shifts in how companies treat stock compensation are making the sector more attractive to new cohorts of investors—particularly value and GARP (“growth at a reasonable price”) investors.
• Slower sector growth, but more efficiency:
  The typical company’s growth rate has dropped from over 20% to nearer 10%, which is changing the investor base. Companies, in response, are becoming more disciplined: expanding margins and using AI to boost productivity internally.

Quote:
“We’re finally at the point where you can actually start to do some interesting analysis around GAAP earnings power. And we think that’s bringing a set of investors into the space that historically have essentially written it off as being too expensive.”
— Gabriela Borges (09:35)

• Internal AI adoption:
  Software companies themselves are leveraging AI for efficiency—"drinking their own champagne."

---

5. How Software Companies Are Adapting to AI (12:08–13:42)

• Acquisition, not just innovation:
  Large software incumbents increasingly let venture-backed startups take big innovation risks and then acquire the best ones, integrating them into their platforms (e.g., Salesforce’s 10+ acquisitions in a year).
• Applying lessons from cybersecurity:
  The cybersecurity sector’s rapid innovation cycle—largely because of active adversarial threats—provides a model for how application software leaders might combine organic innovation and M&A for sustained strength.

Quote:
“It makes a lot more sense for the venture capital community to make those types of bets and then for the software leader today to sit back and say, okay, let me take stock of where the cookies are crumbling so, so that I can figure out which of these technologies actually has staying power.”
— Gabriela Borges (12:43)

---

Memorable Moments & Quotes

• On existential fears vs. actual opportunity:
  “Are their concerns actually merited? And how can investors sort out the winners from the losers?”
  — Alison Nathan (00:10)
• On the pressure for incumbent innovation:
  “You have to prove to us as an investor community and as a customer base that you can actually deliver better AI experiences because of that [moat].”
  — Gabriela Borges (03:45)
• On the new investor base:
  “You actually get to some really nice valuation support for these stocks. And it’s attracting a cohort of investors that historically would not have looked at software as a sector.”
  — Gabriela Borges (11:38)

---

Conclusion & Forward Outlook

• Not a static moment:
  Both established and emerging companies are highly dynamic in response to AI’s disruption. The competitive landscape will continue to evolve, with leaders determined by the ability to modernize, innovate (organically or via M&A), and monetize effectively.
• Near-term stability driven by fundamentals:
  A clear recovery in business metrics, and the arrival of new investor profiles, could underpin the next phase of sector stability.
• Long-term winners will actively adapt:
  Those who best combine existing strengths with agile adaptation—whether via engineering, acquisitions, or internal efficiency—stand to thrive.

---

Segment Timestamps

• 00:05–02:30 – The AI disruption thesis and investor concerns
• 02:30–06:00 – Defining and testing software company moats
• 06:00–07:26 – What makes a moat sustainable?
• 07:26–09:00 – Data/metrics needed for sector stabilization
• 09:00–11:50 – Evolving investor sentiment and sector ownership
• 12:08–13:42 – How software companies adapt: innovation, M&A, and competitive lessons

---

Transcript

[A]

Does AI pose an existential risk to the software industry? Investors have been grappling with that question, leading to profound sell offs in software stocks this year. But are their concerns actually merited? And how can investors sort out the winners from the losers? I'm Alison Nathan and this is Goldman Sachs Exchanges. Today I'm joined by Gabriela Borges who covers the software sector with Goldman Sachs Research. We recently featured Gabriela in our latest Top of Mind report which explored these very questions. Gabriela, welcome to Exchanges.

[B]

It's a pleasure. Thank you for having me.

[A]

I think this is the first time, so I'm excited to have this conversation and it's been a very tumultuous time for the software industry. So great to have you. So, Gabrielle, let's just set the stage. Generative AI has been around for several years now, but the market narrative around AI seems to have really shifted from one of enthusiasm to one of concern and concern in particular in recent months about AI disruption risk, with the focus really on your sector, on the software sector. So walk us through why that is and what made investors rethink the prospects for software companies.

[B]

Yeah, absolutely. It's been a really interesting three months and if I were to pick out a couple of the milestones that really had investors level up the amount of attention that they were paying to software in particular and enterprise software even within that, I would say it started with significant improvements in coding algorithms. So that's things like Anthropic's CLAUDE code, which is a tool that helps you code more effectively if you're a developer. In January, Anthropic then put a wrapper around CLAUDE code that made it accessible to the average everyday knowledge worker like you and me. And so now I don't need to be a developer to be able to benefit from Generative AI. Right around the same time there are a number of announcements from. You have competitors in the space that essentially made investors take a step back and essentially say two things are happening at once. We're finally getting enterprise adoption off AI. It generally takes a fair number of years before you go from the ChatGPT moment in 2022, which was more of a consumer facing moment, to actually starting to see the benefits in the enterprise ecosystem. And then you had a number of announcements from Anthropic, OpenAI, some of the other competitors that suggested that competition was going to go up. And, and all of those things essentially came to a head in the last three months.

[A]

Interesting. So there have been real technological developments really ushering in this concern. But then when we talk about where the software companies sit among on this. So as you just said, increased competition. But we're hearing a lot about these moats and whether they are defensible. Talk to us about what those are and what they mean for, for the space.

[B]

There's a couple of examples that we think have been resonating with investors. If you think about the cybersecurity space as an example, a company like CrowdStrike, they're actually collecting data on endpoints. What is an endpoint? It's a mobile phone, it's a laptop, it's a server. And when they collect that data, they actually have a human in the loop doing reinforcement, learning and infusing threat intelligence, threat data and their own tradescraft into being able to label that data for cyber attacks. And CrowdStrike's been doing that for 10 plus years now. So the data set they have and the domain experience and the context that they have actually becomes really important. And you hear us using these words, domain experience, context a lot because we do think it's one of the more interesting moats that incumbents have. Now the question becomes okay CrowdStrike or okay Company ABC. If you have all this domain experience, you have to prove to us as an investor community and as a customer base that you can actually deliver better AI experiences. Because of that means that putting Sierra, which is a next gen startup in the private ecosystem on top of Salesforce versus using Salesforce's own AI technology, agentforce, Salesforce has to demonstrate that they can get a better outcome using agentforce. Similarly, Microsoft with Copilot, they need to prove that Microsoft Copilot can give you more enhancements to your Excel usage with an AI agent than simply using a tool that plugs in over the top like Anthropic Cowork. And so this is one of the questions that we've been pushing management teams on. Prove to us that you can actually use your moot and use your incumbency and use your domain experience to deliver better outcomes for the customer.

[A]

But even if they can prove that in the near to medium term, this is not a static situation, there's a lot of dynamism, as we know, in terms of the evolution of technology and of these enterprises. So what will it take for you to be convinced that those moats will remain defensible or that AI enabled native companies and tools will catch up?

[B]

There's a couple of things that we've started doing in our due diligence and admittedly this has been part of our process for some time. Now, where I feel like it's really hard to be a good public company investor if you don't know what's happening in the private ecosystem. So we run a private conference, we spend a lot of time in the private company ecosystem and we're asking these private company startups the same questions. What is your moat? How is it defensible? Tell us about your roadmap, tell us about the customer experience. The other half of the equation is spending time with the CTOs of these companies, spending time with the engineering people at these companies. To go a little bit deeper on technical moats like the one we were just talking about with CrowdStrike. You put all of that together as a fundamental analyst, we care about the fundamentals. What that means is what are the leading indicators of your business? Things like bookings and billings, things like AI usage, the percentage of seats that you're monetizing. AI functionality in those types of statistics and Those types of KPIs go a long way to showing us that yes, company ABC is being successful with driving AI adoption and that AI adoption is differentiated enough that they can actually price and charge for it and monetize it properly.

[A]

But again, over time, will these moats that seem very sustainable at the moment endure?

[B]

We think selectively, yes. And so what are the things you need to do as a software leader today to make sure that your moat endures? We have a three step checklist, for lack of a better word. Number one is you have to replatform your current tech stack. Make sure that it's modern, make sure that the data flows nicely within it, and that you've minimized your legacy tech debt, which tends to compound if you've built your architecture over 10, 15, 20 years. The second thing you've got to do is actually figure out your organic roadmap. Make sure that you have the right engineering talent in place, the right engineering coding tools in place to be able to drive productivity up. And if there are gaps in your roadmap, that's okay. You don't need to be a leader in the space. You can be a fast follower. You need a really talented M and a team to be looking at the private company ecosystem and saying company ABC would be a really good product market fit for our existing roadmap. And then the third thing is this monetization point that we were just touching on. Do you have a way to track and measure usage in your install base such that you can actually turn on the monetization over time and actually price for that and that pricing Power is going to be a key litmus test as to whether your product is differentiated enough that you can actually charge for it. So the direct answer to your question is yes, selectively. The moats will endure. And these are the types of things that we're looking for today's software leaders to figure out if they will still be a leader three, five, seven years from now.

[A]

It sounds like it is going to take some time to make that determination and for these companies to figure it out. So in the meantime, you have investors who are owning these companies and these stocks. What will it take to for the sector to actually stabilize?

[B]

Yeah. So I'll come back to this idea of the numbers need to contradict the narrative. We do an analysis called Benchmarking Software where we look at ARR growth, that's annual recurring revenue and metrics like LTV to cac, lifetime value to customer acquisition cost. These are some of the key metrics that software investors use to assess the health of the business. And those metrics have actually been deteriorating for four years now. We had this really exciting period of time in 2020 and 2021 where a lot of digital transformation accelerated. And since that time there's been more of a digestion. A lot of these key metrics have been slowing down or getting worse. And so I think we're at a point now where you can start to see stabilization in the metrics. And one of the more encouraging things during earnings season is we started to see companies that actually put up decent results trade well after the print. And as someone who cares about fundamentals, as a fundamental investor, that type of engagement by investors caring about single stock ideas, that type of engagement where you see the market respond favorably to positive data points in an earnings report, those dynamics tend to be pretty good for the sector to recover over time. And within that, again, this is a focus on companies within the sector because we do think there will be separation between companies that get it and are transforming and companies that maybe are a little bit behind.

[A]

And as the year started out with a lot of investor concerns, we've seen a lot of volatility in the markets for other reasons. The what are you detecting about investor sentiment now? Is it evolving at all or are we still in a wait and see mode?

[B]

Some of my favorite conversations over the last couple of weeks actually have been value oriented PMs who say for the longest time I have been underweight or short these high priced data assets or data centric intangible assets. And a lot of metrics that people look to in the software world historically have been EV to free cash flow, which is, which actually doesn't treat stock based comp as a cash expense because it's not a cash expense or non GAAP earnings, which again you have this delta between GAAP earnings and non GAAP earnings tied to stock based compensation. We're finally at the point where you can actually start to do some interesting analysis around GAAP earnings power. And we think that's bringing a set of investors into the space that historically have essentially written it off as being too expensive. That's the beginning of hopefully what will be a more meaningful sentiment shift. I think phases two and phases three have to come over time. As we've just been talking about. Are the metrics getting better? Is the competition? Are there signs that the competitive intensity in this space is actually starting to moderate or increase or increase? Absolutely.

[A]

And so based on those conversations, do you think that the investor base in these companies in this sector is shifting?

[B]

The short answer is yes. And part of this is also tied to scarcity of growth. Where back in 2020 and 2021, the median company in this space was growing north of 20%. That number is much closer to 10% today. And so you're see seeing a set of investors come into these names that are more value oriented and also care about growth at a reasonable price. The classic GARP cohort. And that's a new set of conversations where these companies, to their credit, between 2021 and 2023, investors very loudly said we need to see more margin expansion. You have assets like Okta that went from 0% free cash flow margin to closer to 20% free cash flow margin in 18 months. We think these companies will actually respond in a similar way this time around. Investors are saying we really care about GAAP earnings. We, we don't like the stock based compensation dilution that's impacting your quality of earnings. The beauty of this is that software companies too are able to get more productivity out of their existing installed base of employees using some of these AI tools. They're drinking their own champagne. And so what that means is that OPEX growth going forward may actually be more tied to the cost of compute, which we know is coming down, rather than the cost of hiring and retaining talent. Of course that will still be a big part of any company value generation, but maybe to a lesser extent than before. For you put all of this together, you actually get to some really nice valuation support for these talks. And it's attracting a cohort of investors that historically would not have looked at software as a sector.

[A]

And that's just a really important point that you made, Gabriela, and you made it in the report as well that these companies are not standing still. They are responding, they are adapting these tools just like the native AI companies are utilizing these tools. A lot of dynamism within the companies themselves.

[B]

I do think this point is a really important one because it doesn't make sense if you are a software leader today to also be using your organic R and D to try to invest in all of these new and really exciting technologies coming down the pipe. It makes a lot more sense for the venture capital community to make those types of bets and then for the software leader today to sit back and say, okay, let me take stock of where the cookies are crumbling so, so that I can figure out which of these technologies actually has staying power and then you can use your incredibly healthy balance sheet to go and acquire the top two or three companies in this space. Salesforce has actually made north of 10 acquisitions in the last 12 months and fit them into your roadmap. Now once you fit them into your roadmap, the onus is on the company to actually make sure that they're really well, technically integrated, that you've got your salespeople trained and enabled to go cross sell into your installed base. But that is actually a playbook that's proven to be very successful over time. And maybe I'll use one more example here in the cybersecurity space where cybersecurity, the pace of innovation has been elevated for 10 plus years now because you're actually up against an active adversary, you're up against the hacking community. And so that means you've got to be able to see around corners. As companies like CrowdStrike and Palo Alto have leaned into M and A, that platforms have actually gotten stronger, their cost of customer acquisition has gone down and, and we think that's a really interesting parallel for what could now happen in the application software space. For names like Microsoft, Salesforce, Workday, Intuit, these are some of the best application software companies from the last decade. I'm really excited to see what they do from an M and A and organic innovation standpoint as well.

[A]

Lots to watch. Thanks so much for sharing your insights Gabriela.

[B]

It's my pleasure. Thank you for having me.

[A]

And thank you for listening to this episode of Exchanges which was recorded on March 10, 2026. Thanks for listening.

[C]

The opinions and views expressed herein are as of the date of publication, subject to change without notice and may not necessarily reflect the institutional views of Goldman Sachs or its affiliates. The material provided is intended for informational purposes only and does not constitute investment advice, a recommendation from any Goldman Sachs entity to take any particular action or an offer or solicitation to purchase or sell any securities or financial products. This material may contain forward looking statements. Past performance is not indicative of future results. Neither Goldman Sachs nor any of its affiliates make any representations or warranties, expressed or implied, as to the accuracy or completeness of the statements or information contained herein and disclaim any liability whatsoever for reliance on such information for any purpose. Each name of a third party organization mentioned is the property of the company to which it relates, is used here strictly for informational and identification purposes only, and is not used to imply any ownership or license rights between any such company and Goldman Sachs. A transcript is provided for convenience and may differ from the original video or audio content. Goldman Sachs is not responsible for any errors in the transcript. This material should not be copied, distributed, published or reproduced in whole or in part or disclosed by any recipient to any other person without the express written consent of Goldman Sachs. Disclosures applicable to research with respect to issuers, if any mentioned herein are available through your Goldman Sachs representative [e-mail] research hedge.HTML. goldman Sachs does not endorse any candidate or any political party. Copyright 2026 Goldman Sachs. All rights reserved.