After Meta poached MongoDB's CEO, MongoDB's stock crashed 25%. If Big Tech sells enterprise AI directly, are the software stocks I own going to be OK?
2026-09-29
What Meta is targeting is software sold "per head"
On 9/28 Meta unveiled an enterprise platform made up of four parts: agents, business agents, developer APIs and coding tools, and Zuckerberg called it "the beginning of our next major business pillar" (Quartz, 2026-09-28). It brought in MongoDB's CEO to lead the effort, and MongoDB plunged about 25% in a single day (CNBC, 2026-09-28).
MongoDB's plunge was the result of two factors combined. One is a leadership vacuum: a CEO just 11 months into the job left a day before investor day. The other is a shift in the competitive landscape. However, since the company kept its fiscal 2027 guidance unchanged (FourWeekMBA, 2026-09-28), its business has not collapsed for now. Analysts also said the stock's reaction was overdone (Business Standard, 2026-09-28).
More broadly, the episode adds a new player to the "SaaS is dead" narrative that has run throughout 2026. The market's fear is that AI agents will break the pricing model of charging a license per employee. If one agent does the work of several people, companies cut seats. Indeed, Atlassian plunged 35% after reporting that enterprise seat counts fell for the first time.
How cheap has software already become?
| Period | Software Forward P/E | Relative to S&P 500 |
|---|---|---|
| 2020.5-2022.5 | 84.1x | About 4x |
| 2022.6-2024.6 | 43.2x | About 2x |
| 2025.7-2025.12 | 31.2x | Premium narrowed |
| 2026.1-2026.3 | 22.7x | Similar to or below the market |
(SaaStr, 2026-03-24)
The software ETF (IGV) fell about 30% from its September 2025 peak, wiping out nearly $2 trillion in market cap (SaaStr, 2026-03-24). It is the first time in history that software has traded cheaper than the market.
If rates pile on, who hurts more?
Software is the sector most sensitive to rates because its earnings are concentrated in the distant future. That's why software collapsed the most during the 2022 hiking cycle. Now it is being pressured by both AI substitution risk and a 10-year in the 5% range. So even within software, fortunes diverge.
- Most vulnerable: High-P/E names with seat-based pricing and still-weak profitability. Richly valued stocks like CrowdStrike (forward P/E about 180x), cited on Reddit as a put-hedge candidate, can be shaken hard by a single rate move
- Middle: Data infrastructure like MongoDB. The more agents there are, the more demand there may actually be for storing and retrieving data. Meta's agents will ultimately use a database somewhere too
- Relatively favored: Large platforms with usage-based pricing or already solid cash flow. Those selling agents directly, like Meta and Microsoft
"AI-disrupted software companies face very warranted decreases in valuations." — Orlando Bravo, Thoma Bravo (SaaStr, 2026-03-24)
What should investors do?
- There's no need to dump software ETFs all at once. Valuations have already come down to market levels. But if this week's PCE and jobs data cement an October rate hike, they could wobble again, so it's better to hold off additional purchases until after the 10/2 jobs report
- For individual stocks, first check the share of seat-based revenue and profitability. Loss-making stocks with a high share of seat-based pricing are exposed to both rate and AI headwinds
- For stocks like MongoDB that plunged on a leadership event, it's not too late to buy after confirming next quarter that the business is holding up. Until a new CEO is named, a strategy-vacuum discount will apply
- Meta itself has not yet disclosed pricing, launch dates or revenue targets for its enterprise business (FourWeekMBA, 2026-09-28), so it won't contribute to earnings any time soon. After a roughly 30% rally in September, it's too early to use this announcement alone as a reason to buy