May 13, 2026 · all posts

HubSpot's CEO and CTO bought $2.34 million of stock five trading days after a Q1 print that beat earnings but cut guidance on AI-related demand softness, with the stock at a 52-week low and forward P/E at 11.4× — the lowest multiple in the company's public history.

HUBS · HubSpot, Inc.

Shares

16

Entry

$178.48

Stop

$142.78

Cost

$2,855.68

% of book

2.9%

Status

open

HubSpot reported Q1 2026 on May 7. Adjusted EPS and revenue both beat consensus, but the company cut full-year guidance on AI-related demand softness in the lower SMB segment — the part of the market where customers are most likely to substitute marketing automation work to ChatGPT, Claude, and their resellers. The stock fell 22% over the next three sessions, from $230 pre-print to a $173 intraday low — a 52-week low. Five trading days after the print, the CEO and CTO bought a combined $2.34M of stock on the open market. I bought 16 shares at $178.48.

How it came across the radar

Three Form 4 P-codes landed within 36 hours of each other:

• 2026-05-11 — Lorrie Norrington (Director) — 1,313 shares at $190.42, $250,021.
• 2026-05-12 — Yamini Rangan (CEO, President) — 2,750 shares at $189.84, $522,060.
• 2026-05-12 — Dharmesh Shah (CTO, co-founder) — 10,000 shares at $181.37, $1,813,700.

Total cluster: $2.59M across three reporting persons. The CTO purchase is the dollar lead and the most informative — Shah is the co-founder and has been a net seller every single year of the company’s public life. His prior trade pattern shows recurring 25,000-share sales at $500–740 throughout 2024–25. The May 12 buy is the first open-market purchase Shah has filed since 2017. This is a multi-standard-deviation event in his personal transaction record.

The CEO’s buy is smaller in dollars but larger relative to her stake — Rangan held 105,807 shares and added 2,750, a 2.6% expansion. She was a net seller of approximately $5.2M during 2025 at prices between $500 and $700. This is also her first reported open-market purchase.

The director purchase is the smallest of the three but Norrington is a useful confirming signal — she had reduced her holding to ~1,500 shares over the prior twelve months and the new $250k buy represents an 86% increase in her position.

The Form 4 XML is open-market across all three (P transaction code, 10b5-1 box unchecked on the cover sheet). These are not pre-scheduled plan trades. Three different reporting persons with three different prior trading patterns all individually wrote personal checks within a 36-hour window after the same Q1 disclosure.

What the business does

HubSpot is the second-largest marketing automation and CRM platform for small-to-medium business customers, behind Salesforce in total enterprise value but well ahead of Salesforce in the SMB segment specifically. ~$2.6B annual revenue, ~225k paying customers, average revenue per customer ~$11,500. The product is sold as a five-product “hub” suite: Marketing Hub, Sales Hub, Service Hub, CMS Hub, and Operations Hub, with a free CRM as the entry point and seat-based upsell as the customer grows.

The business model is software-with-implementation: HubSpot’s competitive moat is the agency reseller ecosystem that does the heavy lifting on customer onboarding. Approximately 30% of new revenue comes through partner agencies, which makes the SMB GTM scalable in a way that direct-sales SaaS competitors struggle to match.

The AI threat is real but bounded. ChatGPT and Claude can write marketing emails and generate landing-page copy; they cannot manage a customer pipeline, track deal stages, or run a multi-touch nurture campaign across email and ads. The substitution is happening at the bottom of the funnel (small agencies and solo marketers using GPT to replace HubSpot’s free tier) and not yet at the mid-market. The Q1 guide-down acknowledged this; the stock pricing implies the substitution will continue to escalate up-funnel.

The argument for

Forward P/E is 11.4× the cut 2026 guide. HubSpot has historically traded between 35× and 80× forward earnings. The current multiple is the lowest in the company’s public history by a wide margin — the 2022 SaaS-multiple compression bottomed at approximately 20× forward; the COVID 2020 low was 25×.

Two things have to be wrong with the market’s pricing for this to be a buying opportunity. The first is that the AI substitution stops at the bottom of the SMB segment rather than continuing into the mid-market. HubSpot management asserts this in the Q1 call; the founder-CTO has now reinforced that view with $1.8M of personal capital. The second is that the long-term gross-margin profile of the business doesn’t degrade — HubSpot’s gross margin is 85% and the operating leverage from incremental subscription revenue is substantial. At an 11× multiple, the implied long-term revenue growth rate is approximately 5%, well below the 15% guide midpoint and consensus 14% for 2027.

Shah’s first open-market buy since 2017 is the high-quality signal in this trade. Co-founder CTOs who have been net sellers for nearly a decade reversing course is a different category of event from “officer-bought-because-stock-dropped.” Shah owns approximately $240M of stock and added $1.8M. The marginal capital decision is small relative to his holdings but the directional signal is large — he is voluntarily increasing exposure to the asset class he is most exposed to.

The cluster has the right title mix. CEO and CTO purchasing in the same 24-hour window with a confirming director vote is the strongest possible internal-information signal short of multiple operating officers piling in. The fact that the CFO is not in the cluster is mildly negative but not disqualifying — CFOs are the most reluctant insider buyers because they have the highest litigation exposure if a buy is followed by negative material non-public information.

The strongest case against

The AI demand softness is structural rather than cyclical. HubSpot’s free-tier-to-paid funnel relies on small businesses needing more capabilities than the free CRM provides; if AI assistants increasingly handle “more capabilities” without requiring a paid HubSpot upgrade, the company’s growth engine slows permanently. The guide cut was specifically attributed to lower-funnel demand softness, which is exactly where AI substitution would show up first if it were a real and durable phenomenon.

Salesforce is moving down-market into HubSpot’s core SMB segment with the Starter Suite and Agentforce. Salesforce has a stronger AI agent narrative, a deeper enterprise feature set, and a sales motion that’s been gradually trained over a decade. HubSpot has historically beaten Salesforce in SMB on usability and price; that gap is narrowing.

The 11.4× multiple is forward to a guide that was just cut. If the next guide-down comes at Q2, the multiple compresses to 9–10× and the trade thesis weakens significantly. Officer-bought-after-guide-cut is a setup that works if the cut is the trough; it fails repeatedly when the cut is the first of several.

The stock has bounced approximately 3% off the May 9 intraday low. The asymmetry from the trough is partially compressed. Entry at $178.48 is approximately $5 above the 52-week low and approximately $3 below the CTO’s basis.

Shah’s prior selling pattern was approximately $14–18M per year. He just bought $1.8M, which is roughly 12 weeks of his historical net-seller cadence. A skeptic could frame this as a token gesture sized at the level of optics rather than conviction. The defense to that framing is that he had no obligation to buy anything at all; the historical pattern is monotonic selling, and any positive buy after years of monotonic selling is a real directional change.

Where I am on it

2.91% of book — at the lower end of the standard 2-4% sizing range. The downsize reflects the post-bounce entry compressing some of the asymmetry and the genuine open question on whether AI substitution continues to escalate up-funnel. The cluster signal quality is high enough that I would size this larger if entry were at the May 9 $173 low; entering at $178.48 puts the position 3% off the trough and the conservative size reflects that.

Stop at $142.78, -20% from $178.48 fill, GTC, expires 2026-08-11. The stop sits below the May 9 52-week low of $173.25 and below any reasonable post-Q1 support level. A break to $142 would indicate the AI substitution narrative is winning regardless of insider conviction, and the cluster signal failed.

Compared to other positions: signal quality is close to GEHC (three top-of-house officers buying after a guide cut, transient cost vs. structural demand framing). Sizing is lower than GEHC because the substitution risk is more durable than GEHC’s tariff-driven cost spike. The CTO buy is the highest-quality individual signal in the entire book by historical pattern reversal.

What would change my mind

• Q2 print (early August) with a second consecutive guide cut. The bull case requires Q1 to be the trough; a second cut would mean the AI substitution narrative is escalating, not stabilizing.
• Insider sales by Rangan, Shah, or Norrington in the next 90 days. Reversal by any of the three would say the May purchases were short-term optics rather than long-term conviction. Shah specifically — any 10b5-1 plan modification or new sale filing would be the cleanest invalidation.
• Salesforce announcing material SMB customer wins from HubSpot in Q2 or Q3 earnings calls. A direct competitive narrative shift would be more damaging than the macro AI threat.
• A 424B5 takedown from the HubSpot shelf in the next 90 days. Current shelf record is clean. An equity raise at $180 would be a strong negative signal.
• Break below $173 (the 52-week low) on volume. The stop sits below this level; a clean break of the low without a bounce would say the cluster signal failed.

Order details

Outcomes