Odd Lots (Bloomberg)
Episode Summary: Jared Sleeper on Which Software Companies Will Survive the "SaaSpocalypse"
Release Date: February 19, 2026
Hosts: Joe Weisenthal & Tracy Alloway
Guest: Jared Sleeper (Partner, Avenir)
Episode Overview
This episode ("Which Software Companies Will Survive the 'SaaSpocalypse'") dives into the current turmoil facing listed software/SaaS companies. Amid steep share price declines and a climate of AI-driven disruption, Joe and Tracy invite Jared Sleeper, an experienced software investor, to break down what’s actually at risk in the sector, which structural factors matter, and what the future might hold for different players in the software industry.
Key Discussion Points and Insights
1. Why Are Software Stocks Crashing?
- Backdrop: The software ETF IGV is down sharply, Salesforce and Atlassian have lost significant value. Talk of a “SaaSpocalypse” is rampant.
- Joe’s Framing: The arrival of easy-to-use AI tools—like Claude code—suggests “if any old fool can write software, maybe these companies... don’t have much value” (01:19).
- Jared’s Context: Most investors don’t understand the nuts and bolts of enterprise software, which leads to panic selling and little market support during downturns (05:16).
“It’s one of those Rorschach test kind of sectors where almost no one's logged into Salesforce and clicked around... So when there's panic, there's not a lot of support for the stocks.”
— Jared Sleeper [05:16]
2. How Did the SaaS Business Model Become So Entrenched?
- Tracy’s Question: Why didn’t companies just build these tools in house?
- Jared’s History Lesson: Third-party software is cheaper to buy than to build/maintain in-house, especially given integrations and the ecosystem of support. SaaS became about scale — building once, selling thousands of times, at a fraction of the cost of a single engineer (06:52).
3. Integration and Complexity: Are AIs a Threat?
- Integrations: Historically, third-party consultants helped businesses stitch together complex systems (like SAP, Oracle, Salesforce). Network effects and “herd familiarity” (standardization of user experience) are major lock-in forces (07:56, 13:12).
- AI and Claude Code: Now, generative AI can handle some integration coding, but many system transitions are human problems—people need to understand and manage the context behind their existing data (10:17).
"The challenge of writing code for integrations is going away... but that's not the bulk of the challenge. ...Within most organizations that's a human problem."
— Jared Sleeper [10:17]
4. What Are (and Aren’t) Customers Actually Buying When It Comes to SaaS?
- Code is not the only product. Customers pay for:
- Support & Reliability;
- Brand and Trust (“herd familiarity”): Standardization, comfort, and onboarding ease (e.g., people choose Zoom not because Teams is unavailable, but because everyone knows Zoom) (12:25).
- Regulatory Knowledge and Ecosystem: For example, DocuSign's value accrues from its legal/regulatory reliability and broad integrations, not just e-signature code (23:13).
5. AI's Real Threat: Terminal Value Anxiety
- Short Term: Financials for many SaaS companies are “pretty good”—growth has slowed but not cratered.
- Long Term: Market anxiety is about “terminal value”—the risk that, within a few years, code generation and AI agents will fundamentally obsolete current workflows, reducing current SaaS companies to zero-value (14:35).
“There are two arguments against software... One is the world's going to stay the same, but software just gets a lot cheaper. ...The second is the world's about to get really weird and the way that knowledge work happens is going to change.”
— Jared Sleeper [13:54]
6. Who Is Most At Risk?
- Enterprises vs. SMBs: Enterprises with high customization and dedicated IT are most likely to rip out existing systems in favor of AI-native (and thus threaten SaaS incumbents); small businesses are much less likely due to inertia and switching pain (21:32, 22:18).
- Jared’s anecdote: His family’s grocery store changed POS systems for the first time in decades and won’t switch again soon.
- Specific Software Examples: Seemingly “simple” tools like DocuSign may be protected by scale, brand, and regulatory knowledge, even as free alternatives exist (23:13).
7. The Bull Case: Results-Based Pricing & AI-Native SaaS
- The ultimate upside could be results-based pricing: Instead of charging per user/per month, software may deliver business outcomes directly and price accordingly.
- Example: Intercom’s “Fin” AI support agent grew ARR dramatically (25:04).
- Software companies could capture much greater value by automating tasks previously performed by expensive humans (26:51), especially if they move quickly.
8. Margins, Stock-Based Compensation, and Real-Profit Questions
- Stock-Based Compensation (SBC): U.S. SaaS firms often ignore the true cost of SBC, reporting only “non-GAAP” margins, masking their real profitability (34:34).
- The median public software company has only a 5% net profit margin on a true (GAAP) basis.
- Layoffs and Headcount: AI is making code-writing massively more productive—raising pressure for layoffs (especially of less adaptive employees). Management teams may use layoffs both to cut costs and to retain top AI talent (36:49).
9. Demand for Coders, Skillsets, and Human Advantage
- Coding jobs will evolve from writing code to managing and guiding AI code. Adaptability and sociability become key skills, as AI eats “isolated, commoditized” work (39:10, 41:11).
- Quote:
“If you like working on problems in isolation, not socially with other people... that's probably a pretty tough place to be. Yeah, it's going to be a more social world.”
— Jared Sleeper [41:11]
10. Why Are All Data and Index Companies Selling Off, Too?
- Companies like Moody’s or S&P sell data products that also look vulnerable but remain entrenched for now. The trend is for investors to “shoot first, ask questions later” about AI risk (43:17, 44:09).
11. Investor Behavior and Panic Patterns
- Professional hedge funds and quant shops can't afford the risk of appearing “caught holding the bag,” exacerbating mechanical selling and volatility (44:09, 45:42).
"The scary thing... is because it's not fundamental... They have no idea when it will stop. ...You're one OpenAI model release or Anthropic model release away from more fear."
— Jared Sleeper [44:09]
Notable Quotes & Memorable Moments
- "SaaSpocalypse... My Cassastrophe."
— Jared Sleeper joking about the sector’s rolling puns (05:05) - "I always think of Salesforce as... I’m not really sure what they do, but, yeah, it’s just ugly."
— Tracy Alloway (01:06) - "For the first few quarters, the management team of Chegg, you know, had their heads in the sand. But then it became clear that it really was existential."
— Jared Sleeper (17:20) - "Intercom... got very aggressive about building out an AI product... they’ve really re-accelerated their business."
— Jared Sleeper (25:04) - "Adaptability... constantly trying and testing these tools, making sure you’re staying at the cutting edge... being aware of what’s human."
— Jared Sleeper’s advice to software workers (39:10) - "If you like working on problems in isolation, not socially... that's probably also a pretty tough place to be."
— Jared Sleeper (41:11)
Timestamps for Important Segments
- 00:36 — Current state of SaaS stocks; why are they tanking?
- 04:23 — Jared Sleeper’s background and sector experience
- 06:52 — Why SaaS exists & how software became “productized”
- 08:29 — Integrations: the sticky pain of changing enterprise tools
- 10:17 — The real human challenge of integration in the AI era
- 13:12 — What makes software sticky: “herd familiarity” and support
- 13:54 — Two market fears: deflation or total disruption
- 15:12 — SaaS by the numbers: growth and net retention post-pandemic
- 17:20 — The “Chegg” case as leading indicator for AI disruption
- 18:18 — The role of proprietary data and context for AI agents
- 21:32 — Spectrum of SaaS: from broad platforms to vertical/minimal tools
- 23:13 — DocuSign: Why “simple” tools aren’t easily replaced
- 25:04 — Intercom’s AI pivot and results-based pricing in practice
- 34:34 — Real profitability: the Stock-Based Compensation debate
- 36:49 — The case for pending layoffs in SaaS
- 39:10 — Future of coding jobs and human value
- 44:09 — Trading behavior, market structure, and investor panic
Conclusion
The episode paints a nuanced landscape: Software companies face unprecedented uncertainty—not because their short-term financials have collapsed, but due to deep anxiety over the "terminal value" that AI and automation might soon eat their business models alive. Investors are fleeing at the first sign of trouble, and only those firms with strong network effects, trusted brands, and the ability to reinvent themselves with AI-native features/business models may survive. Ironically, the very “stickiness” and complexity of legacy enterprise software is itself both a moat—and a challenge in an age where code is cheap but context and trust are priceless. The "SaaSpocalypse", then, isn't sudden death—it's a strange new world in which adaptability, brand, and data matter more than ever, and even the most dominant players can't rest easy.
For those who haven't listened:
This episode is a timely, accessible masterclass for anyone interested in how enterprise software actually works, the dynamics behind today's SaaS market meltdown, and the early contours of an AI-driven business future.
