# Weekly SaaS / Software Podcast Recap - Week of June 21, 2026: SaaSpocalypse on Trial as SpaceX Buys Cursor for $60B

> SaaS and software podcast recap for the week of June 14–21, 2026. The week was dominated by the fight over whether AI is genuinely killing legacy SaaS, with SpaceX's $60B Cursor deal, Anthropic's Fable ban, and the seat-vs-consumption pricing reckoning running through nearly every operator interview.

## Weekly SaaS / Software Podcast Recap

### Week of June 21, 2026: SaaSpocalypse on Trial as SpaceX Buys Cursor for $60B

---

**Coverage window: June 14 – June 21, 2026**

**Top of mind:** The week was dominated by a single fight, whether AI is genuinely killing legacy SaaS ("SaaSpocalypse") or whether the narrative is running ahead of the data. The bears point to Wix at all-time lows and Adobe near multi-year lows; the bulls (notably Ramp's own spend data) say there is "no indication of even early signs of a slowdown" among incumbents. Layered on top: SpaceX's surprise $60B acquisition of Cursor reset the M&A playbook for devtools, the US government's de-facto ban on Anthropic's Fable model became the year's first capabilities-based AI regulation, and the seat-based-vs-consumption pricing reckoning showed up in nearly every operator interview (Salesforce's $3.6B Fin deal as the template).

## 1. Dominant Themes

**1. "SaaSpocalypse," real signal or narrative war?** The most-discussed topic, and genuinely contested. On **20VC (Jun 18)**, Benchmark's Rory O'Driscoll and "Ev" laid out a bear T-chart on Wix: all-time lows, 2026 guidance cut by $50M, 20% workforce cut (~1,000 people), trading at ~1x revenue, with coding agents (Lovable, Replit) replicating the product. Ev: "Any liquidity for pre-AI SaaS companies is top decile performance, any liquidity at all." **Chit Chat Stocks (Jun 19)** tagged Adobe and Wix as "AI losers" (Wix at a 25% FCF yield, pricing distress). On **Topline (Jun 14)**, Clari+SalesLoft CEO Steve Cox cited an MIT study that "95% of AI trials have ended in not being able to show proven ROI," and described failed AI-native startups landing on his desk after racing to $2–4M ARR then struggling with retention. The bull rebuttal came from Ramp's lead economist Ara Kharazian on **Big Technology (Jun 17)** using actual spend data: "80% of the market share for CRMs is just directly going to Salesforce," Figma "continues being one of the fastest growing vendors on our platform" despite Claude Design, and "there is no indication, at least in our data, that there are even early signs of a slowdown."

**2. The SpaceX/Cursor deal and a new software M&A playbook.** SpaceX's $60B acquisition of Cursor (built on Anthropic's Claude, ~$4B revenue run-rate at ~15x, reportedly -23% gross margin) hit at least six episodes. **This Week in Startups (Jun 18)**, Jason Calacanis called Anthropic's move "a platform stealing the application layer" (Microsoft/Lotus 1-2-3 parallel) and warned founders: "Do not take that deal. Do not trust OpenAI… he will pick the top five… and incorporate it as free product into their platform." **All-In (Jun 19)**, Chamath: "He essentially got Cursor for 15 billion… His business intellect is off the charts." **RiskReversal (Jun 19)**, Imran Khan framed it as SpaceX buying its way into enterprise coding against Codex and Claude Code. The subtext: coding-assistant market share is brutally transient (GitHub Copilot to Cursor to "QuadCode" within ~18 months).

**3. Anthropic's Fable ban, a regulatory Rubicon.** Commerce Secretary Lutnick restricted Anthropic's newly-released Fable model to US citizens (Export Restricted Act); Anthropic shut access entirely rather than partially comply, after AWS flagged a jailbreak within 90 minutes of launch. On **20VC (Jun 18)**, O'Driscoll: "It's the first time that the U.S. has ostensibly regulated an AI model based on capabilities… This is a Rubicon moment," and said Anthropic's IPO probability went from "90%+ certainty" to "significantly wrinkled." **The Information's TITV (Jun 17)** flagged the enterprise worry of "a de facto licensing regime for foreign employees" (~40% of AI paper authors are foreign-born). On **Odd Lots (Jun 19)**, Anthropic co-founder Jack Clark said engineers now write "eight times the amount of code they did in 2021 through 2024" and models can "functionally automate all of coding," pegging China as "six to 12 months behind."

**4. Seat-based vs. consumption/outcome pricing, the model reckoning.** Salesforce's $3.6B acquisition of Fin (ex-Intercom) was the case study, O'Driscoll (**20VC, Jun 18**) called it "the golden path": Fin pivoted from seats to outcome pricing (99¢ per resolved intervention), and revenue went from $300M ARR at 7% growth to $400M at 25%. Snowflake CEO Sridhar Ramaswamy (**In Good Company, Jun 17**): "Most software companies charge per seat, but you charge for what people actually consume… In the world of AI, that's a big deal." Counterpoint from Ramp data (**Big Technology**): Adobe's metered usage is "still only like half a percent of their revenue," the transition is happening far more slowly than the narrative implies.

**5. Coding agents as an existential threat to the application layer.** Ramaswamy (**In Good Company, Jun 17**) was blunt: "I consider the model companies, things like the coding agents… to be our biggest competition… they represent the front door to computing," calling them "the biggest threat to all software," more than AWS or Microsoft. Perplexity's Aravind Srinivas (**20VC, Jun 15**) on why OpenAI prioritizes Codex: "Because that's where the money is… ChatGPT is a dominant consumer product but there is no money there… it's been commoditized."

**6. Hyperscaler capex and the ROI reckoning.** On **Monetary Matters (Jun 20)**, Jim Chanos modeled neoclouds like CoreWeave at "four or five, 6% returns on capital in the out years" even on heroic assumptions, calling it a finance business dressed as tech. Oracle (**Network Break, Jun 15**) posted $67B in AI infra contracts and RPOs of $638B (+363% YoY) but guided to $70B FY2027 capex and the stock fell ~10%. David Woo (**Monetary Matters, Jun 15**) argued hyperscalers are running an "earnings illusion," revaluing Anthropic stakes to offset capex, passing memory cost increases to customers. Aswath Damodaran (**Excess Returns, Jun 19**): "infrastructure investments… take 10 years to depreciate but could become obsolete in five."

**7. Enterprise SaaS brain drain.** Per **The Information's TITV (Jun 17)**, 85 Salesforce employees left for OpenAI and Anthropic in 2026, including ex-Slack CEO Denise Dresser (now OpenAI CRO) and the former Salesforce App Store CEO (now Anthropic VP Global Partnerships).

**8. Vibe coding and the "CEO hobby" problem.** Figma CEO Dylan Field (**Hard Fork, Jun 17**): "every startup founder, CEO in Silicon Valley is obsessed with Vibe Coding… It's driving their employees crazy." Kevin Roose: "taste is just the word we give to the stuff that the models aren't very good at yet." Figma's data rebuttal: Q1 2026 revenue $333M, +46% YoY, 139% NDR.

**9. Private SaaS stuck in the "messy middle."** The Clari+SalesLoft merger and OWN's $2.1B sale to Salesforce (**Run the Numbers, Jun 18**), framed as one of the rare clean COVID-era unicorn exits, not the norm.

**10. IPO frenzy.** **The Journal (Jun 15)** captured retail FOMO around SpaceX/OpenAI/Anthropic/Figma; WSJ's Spencer Jakab: "You can't make 100 times your money in a company that's already worth $1.75 trillion." OpenAI reportedly burned $3.7B in Q1 2026 ("more than half of its revenue") against $73B cash.

## 2. Key Debates

**Debate 1, is the SaaSpocalypse legitimate disruption or a bank run?** *Disruption camp* (20VC's O'Driscoll/Ev, RiskReversal's Dan Nathan, Chit Chat's Ryan Henderson): seat-based revenue is structurally impaired; Wix/Adobe at lows; Lovable/Replit and Claude Design taking share. *Durability camp* (Ramp's Kharazian, Figma's Field, Topline's Cox): Salesforce still wins ~80% of CRM spend, Figma at 139% NDR, AI-native rivals still tiny. The wedge: is Wix the "canonical example" (Ev) or just an outlier because website builders are uniquely AI-replaceable?

**Debate 2, seat-based is dead vs. consumption pricing is hard.** *Dead* (20VC, Topline): Fin's $3.6B exit is the template; Wix is the cautionary tale. *Hard to implement* (Big Technology, Snowflake): Adobe's metered model is still 0.5% of revenue; consumption requires demonstrated ROI. Ramaswamy's tell: even Snowflake calls navigating this its "number one challenge."

**Debate 3, in the AI agent economy, does the model layer or application layer win?** *Model layer* (Srinivas, Calacanis): frontier models displace wrappers the way Anthropic displaced Cursor. *Application layer* (NEA's Tiffany Luck on Equity, Jun 17; Kharazian): "value really being created at all layers"; Figma's stickiness proves it. *Data platform* (Ramaswamy): consumption + data gravity is a different moat. The Ramp pattern underneath it all: model market share is transient, "that's the risk of every model company, that they go to zero in the next 18 months."

**Debate 4, can Salesforce/ServiceNow monetize AI agents?** *Bull* (TITV's Kevin McLaughlin, O'Driscoll, Chit Chat's Brett Schaefer): "Salesforce… has been very good at selling the next big thing… right now with AgentForce"; Schaefer thinks Salesforce "can probably double Finn's revenue" via existing-customer upsell. *Bear* (Srinivas, Henderson): institutions hold "$200 billion of Microsoft and Salesforce" and may rotate to Anthropic; Benioff "spending $300 million on Anthropic" signals dependence; "whenever Salesforce makes an acquisition… they spend too much." Overhang: 85 staff lost to the labs in 2026.

**Debate 5, Accenture: AI winner or loser?** A clean same-episode disagreement on **Chit Chat Stocks (Jun 19)**. Henderson: "You cannot own this at any price… There's no price I would own this at" (AI eliminates consultants; growth cut to 3-4%). Schaefer: "Consulting is a pretty durable industry, more durable than people think… I'm on the side that this works over the next 5-10 years" (16% FCF yield, ~5% dividend). Bloomberg Intelligence (Jun 18) noted the guidance cut sent the stock down a record ~20%.

**Debate 6, is the AI infrastructure build sustainable or a bubble?** *Bull* (Imran Khan, RiskReversal, Jun 19): "NVIDIA's earnings multiple is not like Cisco… trading at below market multiple"; ~80% growth, CUDA moat; "the Nvidia math says this isn't a bubble." *Bear* (Chanos, Monetary Matters, Jun 20): net-long AI but short neoclouds; CoreWeave returns ~5-6% on heroic assumptions. *Skeptic* (Damodaran): Mag-7 are now unfamiliar capex-heavy capital allocators. *Most specific* (David Woo): hyperscaler accounting is "an optical illusion about earnings acceleration… an income transfer from Microsoft to basically Micron."

**Debate 7, is Anthropic's ban a one-off or the new normal?** *Fixable* (O'Driscoll, 20VC): due-process case exists if rival labs aren't equally restricted. *New normal* (TITV's Leo Schwartz): "a de facto licensing regime for foreign employees." *Self-inflicted* (David Sacks, All-In): "Dario came to Washington… said that he had created a cyber weapon called Mythos… Mission accomplished"; Friedberg diagnosed "epistemic exceptionalism."

**Debate 8, Perplexity: AI search winner or "Google's first roadkill"?** *Bull* (Srinivas, 20VC, Jun 15): $20B, 45M users, >$500M ARR, revenue "more than tripled since the beginning of the year." *Bear* (Yaniv Bernstein, The Startup Podcast, Jun 16): "Google's first roadkill… I can't imagine Perplexity's traffic numbers are looking very good." Srinivas himself concedes "Google will be the token king… but they underestimated the importance of coding models."

## 3. Specific Names, Bull/Bear by Ticker

### Public

**SALESFORCE (CRM), Mixed/Bull.** O'Driscoll: Fin deal is "smart deal… the golden path" (Fin: $300M ARR/7% to $400M ARR/25% post outcome-pricing pivot). McLaughlin: selling AgentForce well. Schaefer: can "double Finn's revenue" via upsell; notes Benioff stopped selling stock in 2026. Bear notes: Henderson ("spend too much" on M&A), Srinivas (rotation risk, $300M Anthropic spend signals dependence). Ramp data: still ~80% of CRM spend, no AI churn. Repurchased $27B recently; 80,000 headcount; lost 85 staff to labs.

**ADOBE (ADBE), Bear.** Tagged with Wix as an "AI loser" (20VC, Chit Chat); near multi-year lows. Metered pricing still only ~0.5% of revenue (Big Technology). Modern Value Investing (Jun 20) floated it as a potential value play post-decline but with no firm bull thesis.

**FIGMA, Bull (valuation caveat).** Chip Stock Investor (Jun 16): Q1 2026 revenue $333M (+46% YoY), 139% NDR, $89M FCF (27% margin), $1.6B cash/no debt, FY26 guide ~$1.42B (+35%), "one of the better software earnings reports," but "still trading for… about 75X expected next-12-month free cash flow." Kharazian: "an extremely durable software vendor" despite Claude Design. Field (Hard Fork): "very, very bullish on design"; flagged board member Mike Krieger resigning days before Claude Design launched.

**SNOWFLAKE (SNOW), Bull (CEO-led).** Ramaswamy (In Good Company, Jun 17): half the addressable Global 2000 are customers, 13,000+ customers, consumption model as AI-era advantage, but names coding agents "our biggest competition" and survival "my number one challenge." Products: Cortex Code, Snowflake Intelligence, MCP.

**DATADOG (DDOG), Bull.** Network Break (Jun 15): Dash announcements, bring-your-own-cloud, Federated Log Search (Databricks/ClickHouse/Snowflake), Bits AI auto-remediation, AI Guard runtime security. Current CTO Trevor Marshall (RiskReversal, Jun 17): "better than if we were to build it ourselves… massive run on a big quarter."

**ORACLE (ORCL), Cautious/Bear.** Q4 FY26: revenue $19.2B (+21%), cloud $9.9B (+47%), $67B AI infra contracts, RPOs $638B (+363%), $43B debt raised, $70B FY27 capex, FY27 guide $90B revenue/$8.05 EPS; stock −10%. Hollingsworth: "don't sell products you don't have yet." Nathan: "if you don't have the cash flow… and you have to raise debt like Oracle does… it's no bueno."

**PALANTIR (PLTR), Mixed/Bear.** Nathan (RiskReversal, Jun 15): "valuation got a little fat… now starting to trade with the SaaS apocalypse theme… down like 30 some percent," though acknowledges AIP as "special sauce."

**MICROSOFT (MSFT), Bearish/Cautious.** Marshall (RiskReversal): Azure growth "high teens… down dramatically," Copilot uptake "less than 16, 17%" of 400M base, "we weren't seeing incremental productivity." Woo: lost OpenAI exclusivity; Copilot adoption "worse than expected." Khan (Jun 19, more constructive): Nadella "no longer capacity constrained, I'm energy constrained."

**NVIDIA (NVDA), Strong Bull (consensus).** Khan: "~60.5x next year… below market multiple," ~80% growth this year, CUDA moat, "this isn't a bubble." Chanos: net long, "~15x on 2027 EPS… much cheaper than Intel." Woo more cautious: dominant in training but inference "much less complex," where AMD/Broadcom/Intel can compete.

**CLOUDFLARE (NET), Neutral/Mentioned.** Only referenced re: CEO Matthew Prince's public spat with Vinod Khosla; no investment thesis.

**ACCENTURE (ACN), Contested.** See Debate 5. Henderson hard bear, Schaefer contrarian bull (16% FCF yield, ~5% dividend); stock −20% on guidance cut.

**ROKU (ROKU), Acquisition target.** Schaefer cautiously positive on a reported Fox acquisition at $22B (~$160/share, ~30% premium); Henderson mixed on standalone operating inflection.

### Private

**ANTHROPIC, Complex/Contested.** *Bull:* Kharazian, "now the most popular AI model used by US businesses" (41% of US firms vs OpenAI 39.5%); Clark (Odd Lots) on automating coding. *Bear:* Srinivas values it at "$1 to $1.5 trillion" but warns "in 6 or 12 months… they won't even be around" if Claude Code is their win; All-In hosts cite "evasiveness and immaturity"; Damodaran flags broken unit economics ("$6,000 an hour to use Claude Fable"); Woo notes ARR "$9B in December to $42–44B" boosted by "token maxing" pre-ban. Compute reportedly from xAI at $1.25B/month (90-day out). IPO probability "significantly wrinkled" by the Fable ban.

**OPENAI, Cautious/Mixed.** *Bull:* NEA's Luck sees a path to trillion-dollar-plus; filed confidentially for IPO. *Bear:* Srinivas ("ChatGPT… there is no money there… commoditized," "isn't ready for an IPO"); Calacanis (tokens become "commoditized… like hard drives and bandwidth"); Bernstein ("they suck at product"). Q1 2026 burn $3.7B vs $73B cash; hired ex-Slack CEO as CRO.

**CURSOR / ANYSPHERE, Bull (acquired).** Calacanis: ~$4B run-rate, 15x, -23% gross margin; Anthropic's internal "Skunk Works" undercut them. Chamath: "got Cursor for 15 billion… unbelievable." SpaceX provides Colossus compute. But Kharazian notes "QuadCode has majority of the market" already.

**PERPLEXITY, Contested.** See Debate 8. Srinivas strong bull ($20B, 45M users, >$500M ARR, IPO target 2028); Bernstein hard bear ("Google's first roadkill").

**DATABRICKS, Mentioned/Bull.** PowerLaw holds a stake (TITV); named with Snowflake as moving to specialization; Datadog deepening integration.

**CANVA, Mentioned/Bull.** PowerLaw holds a stake; no specific metrics.

**CLAY, Bull.** Kareem Amin (Invest Like the Best, Jun 16): valued >$4B, "one of the fastest-growing software companies of the last few years… went from one to 100 in two years," usage-based pricing, RevOps target market.

**MISTRAL, Sovereign-AI Bull (qualified).** O'Driscoll: "in a good place… even if it's not the best model," raising €3B at $20B, ARR ">half a billion." Ev skeptical on model quality but credits the inference-platform pivot.

**CLICKUP, Bull.** COO Gaurav Agarwal (Topline, Jun 21): ">$300M ARR," replaced 22% of workforce with AI, "3 to 1 ratio between AI agents… and employees," 600 marketing assets/week via agentic workflows.

**GATHER AI, Bull (physical-AI SaaS).** CEO Sankalp Arora (SaaS Interviews, Jun 17): ~$15M ARR, 170% annualized NDR, 2.5x YoY, Series B $40M led by Smith Point Capital (Keith Block).

**CLARI + SALESLOFT, Bull (merged).** CEO Steve Cox (Topline, Jun 14): "building the world's first predictive revenue system," MCP server live, Copilot into SalesLoft by July.

**RIPPLING, Bull (mentioned).** Ben Ling (Bling Capital, This Week in Startups): portfolio co., reinvested at Series A; "people are very thrilled to be investors."

**AIRTABLE, Mixed.** Calacanis: "5X or 10X" of a fund but "a hundred Airtable killers that are AI first now"; Ling bullish (portfolio co., launched Super Agent Jan 2026).

**DOCUSIGN, Qualified Bull.** Ex-CFO Mike Dinsdale (TITV): "the value… is absolutely not the e-signature component, which is commoditized. It's the operational component… embedded in enterprise workflow."

**DEEPSEEK, Skeptical / short-term bull.** Kharazian: "a little overrated… only about 0.4% of businesses are using it" (prior spike faded fast). Calacanis sees it for cost-driven routing ("your cost is going to essentially move to free").

**xAI, Mixed.** Sopp (RiskReversal): model "probably not as useful as some of the other models," losing "$1B/month" pre-leases. Bernstein: "a massive failure as a business." Khan (more bullish on cloud): "fourth largest cloud provider in no time," $2B from Google/Anthropic deals, but "what's so special about it?"

**CORWEAVE, Bear.** Chanos (Monetary Matters, Jun 20): "the old Magnetar guys… if Blackstone is in your business… you're in the finance business"; ~5-6% returns on heroic assumptions.

**GOOGLE / ALPHABET (GOOGL), Bull (contested).** Kharazian: "extremely underrated… might end up being one of the big winners no one's talking about," Cloud "~60% a quarter," can offer better routing. Srinivas: "the token king… full-stack TPUs," but "far behind the frontier" on coding models. Woo cautious: 60% of earnings from cyclical advertising, $80-85B equity raise for capex.

**META (META), Bearish/Cautious (consensus).** Sopp: "their models suck… they get punished until there's more revenue in the door." Srinivas: capex "doesn't make sense" vs. a 6-8% ad bump; Micron "might be more valuable than Meta in the next 6 to 12 months." Khan owns it (17.5x this year) but Zuckerberg "may be over-investing because he's over-ambitious."

**MICRON (MU), Contested.** Khan: HBM "sold out possibly to the end of 27." Srinivas (contrarian bull): could top Meta in value within 6-12 months. Woo (bear): HBM pricing power "very short-lived," new capacity not online until 2027-28.