Why HubSpot ($HUBS) is a Buy | The ‘AI Victim’ Trade is Mispriced

The market thinks AI will hollow out SaaS incumbents. HubSpot’s unit economics and a massive insider buying cluster tell a different story.

Photo by Conny Schneider on Unsplash

DEEP DIVE | Powered by Leeway AI

The Thesis: HubSpot at $198 is priced as collateral damage in the AI wars. The fundamentals say otherwise: +23% revenue, NRR re-expanding, $750M FCF guide, and an insider buy cluster at $185. Q2'26 in August is the next read.

The Analysis

The market has filed HUBS under ‘AI victim’ — a SaaS incumbent about to be hollowed out by Salesforce Agentforce and Microsoft Copilot. The multiple (~15x) reflects that verdict. The numbers do not.

Q1'26 printed +23% revenue, NRR re-expanding to 103.5%, FY26 EPS guide $13.04–13.12, ~$750M FCF guide, $1.8B cash, no debt, and a 25% FCF margin. The 500+ seat cohort is up 5x YoY. Breeze credits +67% QoQ. This is not a company unraveling — this is a high-quality platform caught in a cohort-wide de-rating.

The behavioral tell sharpens it. On May 11–12, Rangan, Shah, and Norrington bought $2.6M of stock at ~$185 after 18 sequential prior sales. Founders rarely flip the tape unless they see a floor. That is a Bayesian prior, not a guarantee — but it is the cleanest one available.

The Structural Bet Underneath

HubSpot repriced its Customer Agent from $1.00/conversation to $0.50/resolved-outcome — deliberately absorbing model-cost risk in exchange for adoption velocity and data-graph entrenchment before Salesforce and Microsoft establish mid-market workflow gravity.

With frontier inference costs deflating ~5–10x/year cumulatively, that $0.50 fixed price gets more accretive on a 12–18 month lag. It is a calculated short of inference-cost volatility.

⚠️ Three Risks That Deserve Honest Weight

1. Unmonetized Compute Drag: The ~35% of unresolved tickets generate unmonetized compute drag that token deflation has to outrun.
2. Usage-Based Revenue Discounts: Usage-based revenue historically trades at a discount to ARR (e.g., Twilio, Snowflake post-2022) — meaning eventual Breeze disclosure could compress the multiple rather than expand it.
3. Tier-Mix Masking: Core Seat +90% YoY without tier-mix disclosure could be Starter/freemium-skewed, masking Pro/Enterprise stagnation.

The genuinely under-priced threat is not the visible enterprise giants. It is GoHighLevel-style white-label agency platforms quietly eroding the bottom-of-funnel where HubSpot’s customer-count narrative lives. This is invisible in current disclosures, but worth watching closely.

The Timing & Catalysts

Salesforce only broke out standalone Agentforce ARR ($800M) at Q4 FY26, which means competitive disclosure pressure on HubSpot builds through late 2026 into 2027, not immediately.

📅 Q2'26 (August) and INSPIRE (September) are the next reads, but neither carries a formal commitment to break out Breeze. The catalyst is probabilistic, not scheduled.

The honest synthesis: the multiple does not need a disclosure miracle. It needs HubSpot to remain HubSpot for 24 months while two exogenous things resolve — continued inference deflation and the SaaS cohort exiting de-rating. Both are probable. Neither is certain.

The floor is structurally identified at $173 (5Y low, insider-buy zone, GAAP-profitable, FCF-supported). The near-term ceiling is capped by cohort beta.

Leeway Scenarios (12 Months)

🐂 BULL (24% Probability) | Target: $305 Q2'26 NRR ≥104%, INSPIRE delivers Breeze KPI granularity, inference costs keep deflating, cohort exits de-rating on Fed pivot or High-Yield spread relief. Multiple re-rates from 15x to 20–22x on FY27 EPS $15.50−16.50. The ‘AI victim’ framing collapses.

⚖️ BASE (45% Probability) | Target: $240 Steady beat-and-raise. NRR creeps to 104%, FY26 EPS $13.20−13.50, insider-buy zone holds as structural bid. Cohort de-rating exhausts gradually. Multiple expands modestly from 15x to 17–18x. Recovers roughly half the post-Q1'26 selloff over 12 months.

🐻 BEAR (22% Probability) | Technical Level: $172 Inference costs deflate too slowly, unresolved-ticket compute drag erodes Customer Agent economics, non-GAAP gross margin slips to 80−81%. Breeze disclosure triggers usage-revenue discount (Twilio analogue). GoHighLevel visibly erodes bottom-of-funnel. NRR stalls at 102−103%. Retests $173.

🐻 TAIL RISK BEAR (9% Probability) | Technical Level: $140 Credit-led stagflation fires: HY OAS >425bps, SMB bankruptcies spike, NRR breaks below 100%, passive deleveraging cascades through software. Multiple compresses to 11–12x on cut FY27 EPS. Breaks $173 floor on flow, not fundamentals.

My Verdict

• ⏱️ 6 Months: Neutral (Cohort Beta Dominates, No Scheduled Catalyst)
• 📅 1 Year: Overweight (Perception-Shift Trade, Asymmetric Risk/Reward)
• 🔭 5 Years: Overweight (Consumption-Execution Re-Rating)

The Takeaway: The 6-month outlook is honestly neutral — Q2'26 prints into a fragile credit backdrop with no formal Breeze disclosure scheduled, and cohort beta is the binding constraint.

The 1Y and 5Y are where the asymmetry lives. Each quarter Breeze grows +67% QoQ and Customer Agent resolution holds at 65%, the ‘AI victim’ framing mechanically loses probability. The structural floor at $173 is doing real work, validated by the May insider cluster.

I don’t need a disclosure miracle here — I need HubSpot to keep being HubSpot while inference deflation and cohort de-rating exit do their work. Both probable. Neither guaranteed. That’s the trade.

📉 Full Data & Interactive Charts: leeway.tech/en/aktienanalyse/HubSpot

Disclaimer: This is not financial advice. Based on Leeway AI model outputs.