Last week I told you about a company I found using the Macro Selloffs framework I'd just formalized — and I asked you to guess what it was.

A few of you got it right.

It's Salesforce. Ticker: CRM. I bought one share on June 10, 2026, at $173.08.

Here's the full story.

Let me run this through the framework I built last week.

Inciting Incident — Over the past year, AI capability has accelerated faster than almost anyone expected. Agents that could barely complete simple, scripted tasks a year ago are now handling multi-step workflows that used to require a human and a half-dozen software tools. Adoption has moved just as fast — enterprises running pilot programs twelve months ago are now deploying agents into production.

Market Reaction — That pace of progress spooked investors. The market concluded that if AI agents can now do what enterprise software used to require a human to operate, the software itself becomes redundant — and punished SaaS as a sector accordingly. By early June, that fear had a name: the "SaaSpocalypse."

CRM had been sliding since January 2026 — a slow, sustained decline that predates any single headline — and that decline sharpened considerably in the first two weeks of June. Two things compounded it specifically for Salesforce: soft guidance delivered on June 3rd gave bears something concrete to point to, and a Broadcom-driven selloff swept through the broader tech sector, dragging Salesforce down with everything else.

The result: CRM dropped roughly 13% in five trading sessions, touched a fresh 52-week low, and is now down more than 40% year-to-date — making it one of the worst performers in the entire software sector.

Overgeneralization — The assumption underneath the selloff is that SaaS is a single category — that every company built on a subscription software model is equally exposed to AI disruption, with the same thin moat and the same vulnerability to being replaced.

That's the generalization: it treats Salesforce the same as it would treat a five-year-old SaaS startup with no embedded workflows and no switching costs, just because both get labeled "SaaS."

The Gap Check — At this point, my BS-detector was going off. I've spent time inside companies that run on Salesforce — entire teams that eat, sleep, and breathe it. Sales operations, marketing analytics, customer success. Salesforce isn't a tool they use; it's how they work. The idea that this gets ripped out overnight because an AI agent can write a follow-up email felt wrong.

So I asked the simple version of the question: would Salesforce actually be replaced?

That's what I went to find out.

Before getting into the numbers, it's worth zooming out, because Salesforce has been here before. Twice.

The stock hit its first all-time high in late 2021, around $307, riding the pandemic-era SaaS boom. Then came 2022 — a brutal, broad unwind of high-multiple growth stocks, compounded by a slowdown following the Slack acquisition and a wave of executive departures. Salesforce fell nearly 48% that year.

By January 2023, activist investors — Elliott, Starboard, Third Point, ValueAct — all showed up within months of each other, pushing for profitability over growth. Salesforce responded: a 10% workforce cut, aggressive buybacks, and a "Year of Efficiency" that reshaped the margin story. The stock nearly doubled in 2023 alone.

The recovery kept going — but it wasn't a clean story. Even as the stock climbed, Salesforce's market share slipped, from 21.7% to 20.7%, as the overall CRM market grew faster (12.8%) than Salesforce's own revenue (9.5%). Still winning in absolute terms — more CRM revenue than Microsoft, Oracle, Adobe, and SAP combined — but growing slower than the market it dominated. By December 2024, fueled by continued margin expansion and the launch of Agentforce, Salesforce hit a second all-time high: $363, with investors betting the AI story would reaccelerate growth.

Worth sitting with that for a second: the same AI narrative that helped justify the second all-time high is now the reason the market is selling it off. Same story, opposite direction.

Since that December 2024 high, CRM has erased the entire AI-fueled rally and is now trading at levels last seen in early 2023. The question is whether June 2026 looks more like January 2023 — an overcorrection waiting to be unwound — or whether this time the concern is actually real.

Here's the gap, in full.

The SaaSpocalypse narrative assumes that AI agents will replace the software businesses currently running enterprise operations. Clean story. But it doesn't hold for Salesforce, for a few specific reasons.

The switching costs are enormous. The majority of Salesforce's revenue comes from large enterprise customers — Fortune 500 companies with $150,000 to $500,000+ embedded in Salesforce workflows, integrations, and custom builds. These aren't monthly subscribers who can churn with a click. Ripping out Salesforce means ripping out the operating system of an entire sales or service organization, retraining thousands of employees, and rebuilding integrations with SAP, ERP systems, and dozens of other enterprise tools. That is a multi-year, nine-figure undertaking. It doesn't happen because a competitor launched an AI agent.

I spoke briefly with an enterprise Salesforce practitioner, and they put it plainly: "Every Salesforce instance is custom to each account's process and tightly intertwined with SAP and ERP systems. Would be a huge cost and time commitment to rip out. That is a very tough sell at an enterprise company.”

AI vendors are building on top of Salesforce, not instead of it. This is the part the narrative gets backwards. The companies building AI agents are plugging into Salesforce's data layer, not competing with it. Salesforce is where the customer data lives. AI needs that data to work. The more AI proliferates across enterprise, the more valuable the underlying data infrastructure becomes.

Agentforce — Salesforce's own AI agent platform — is already at $1.2 billion in ARR, up 205% year-over-year. That's not the trajectory of a company being disrupted. That's a company eating the disruption from the inside.

The bear case has a specific mechanism — and it's worth taking seriously. The real concern isn't that Salesforce disappears. It's that AI agents replace human workers, and since Salesforce charges per human seat, revenue per customer shrinks even as AI capability grows. That's a legitimate transition risk. But the direction of travel matters more than the near-term choppiness. Salesforce is already moving toward consumption-based agent pricing — charging per agent action rather than per human seat. Individual agent pricing will likely be lower than human seat pricing. What changes is volume. As AI gets cheaper, agent deployment accelerates. Fewer human seats, far more agent seats, with greater capability per dollar spent. The short-term billings numbers reflect that transition in real time — billings grew only 3.6% last quarter against a 20%+ historical norm. That deceleration is real. But it's more consistent with a pricing model in transition than with enterprise customers heading for the exits.

The contracted future revenue is the anchor. There is $67.9 billion in total contracted future revenue sitting on the books — revenue that's already been signed, committed, and is owed to Salesforce. Current RPO (cRPO, the next 12 months) is $33.6 billion, up 14% year-over-year. That's enterprise customers renewing and expanding, not quietly walking toward the exit. 90% of the Fortune 500 are customers. 39 analysts rate it a Buy, with an average price target of $254.99.

The valuation has dislocated severely from fundamentals. CRM is currently trading at roughly 8x forward EV/EBITDA. The software sector average is approximately 30x. Q1 FY2027 revenue hit $11.13 billion — up 13% year-over-year. Non-GAAP operating margin reached a record 34.8%. EPS of $3.88 beat estimates by 24%. The market is pricing in near-zero enterprise growth. That requires Fortune 500 companies to actively freeze or reduce Salesforce spend — inconsistent with every piece of switching cost data available.

The bottom line. Forget Agentforce, forget AI re-rating, forget the growth story entirely. Just take the enterprise anchor and model it as a slow-growing utility. Based on my own modeling, I get a share price of around $200. The stock is currently trading around $158. Everything else is upside.

The narrative says "AI replaces SaaS." The reality is "AI needs SaaS, and Salesforce owns the data layer it needs."

This is the section that would have saved me from LULU.

Last time, I had a thesis. I had an entry. I had no exit. So when the signals came, I had nowhere to put them, and I held a 62% gain all the way down to a 73% loss.

This time, the exit is written down before I need it.

Add to position — $145: If CRM hits $145 — just below the current 52-week low of $146.32 — I add meaningfully to the position, but only if I can confirm the continued selling is sympathy-driven and not a fundamental signal. The question I'll ask before pulling the trigger: is this still the SaaSpocalypse narrative, or has something real changed? Agentforce ARR stalling or confirmed enterprise churn would keep me from adding regardless of price.

First trim — $230: When CRM hits $230, I sell half. That's a 33% gain from entry, enough to lock in a meaningful return and take the emotional pressure off the remaining position. From there, the rest rides with house money.

Full exit — $255 to $280: This is the Wall Street consensus range — 39 analysts, average target $254.99. When the position approaches that range, I sell the remainder. The mispricing has closed. The thesis has played out. I'm done, regardless of what the next headline says.

Thesis-break exits — full exit, any price, immediately:

That last one is important. A price drop alone doesn't trigger the exit. The question is always whether the price reflects something real. If CRM dips on a market-wide selloff and the enterprise revenue is intact, I hold. If CRM breaks down because customers are actually leaving, I'm out the same day.

One honest flag: Insider activity is mixed. Parker Harris sold over 134,000 shares earlier this year. Starboard Value — one of the activists who helped force the 2023 efficiency push — has fully exited. I don't dismiss that. But insider sales are noise as often as signal, and Starboard's exit may say more about their fund mandate than about the company's trajectory. Counterpoint: Bill Nygren at Oakmark went public on June 25th saying Salesforce is not done growing and is actively redirecting cash flow to buy back shares. Reasonable people disagree here. I've checked it against my thesis-break criteria and none of them are triggered.

Timeline: 6 to 12 months.

Over the last three weeks, we've built something together.

And next week, I'm opening it up.

I'm launching the model portfolio — a live, trackable record of every position, entry price, thesis, and exit triggers, updated in real time. If you've been reading along and want to follow the actual decisions as they happen, not just the postmortems, that's what's coming next.