First Watch Q2 2026: Solid Top-Line Growth, Razor-Thin Profits Keep Thesis Cautious

What happened

First Watch posted 3.4% same-restaurant sales growth and a 15.2% total revenue increase in Q2 2026, while opening 18 new system-wide restaurants across 15 states. Net income came in at just $2.3 million, underscoring how even healthy sales expansion yields little bottom-line profit. Adjusted EBITDA of $34.5 million reflects strong unit economics but heavy corporate-level costs and interest expense that consume most operating gains. The quarter’s 3.4% comps and aggressive unit additions align with management’s growth blueprint, yet the 7.9x leverage and sub-3% net margin remain critical friction points. In essence, the results confirm the business is growing, but they don’t yet dent the core bearish concern that the current valuation already prices in a much more profitable future.

Implication

Q2 2026 reinforces First Watch’s ability to grow unit count and comps, yet the negligible net income highlights the gap between restaurant-level success and corporate profitability. Investors should focus on whether EBITDA margins can climb into the low teens over the next few quarters, as the current ~20x EV/EBITDA multiple demands clear progress. Without a visible path to margin expansion and deleveraging, the stock is likely to remain range-bound, with downside risks if traffic turns or costs spike. The $13 attractive entry from our master report still looks more prudent than chasing at current levels. Any upside requires sustained 12%+ EBITDA margins, which have not yet materialized.

Thesis delta

The Q2 2026 print provides modest top-line encouragement, but the core thesis is unchanged: valuation remains stretched on real earnings power. Positive comps and unit growth keep the story intact, yet they fail to address the critical margin and leverage overhang. We maintain our cautious view, with no reason to raise the attractive entry level below $13 until EBITDA margins prove they can sustainably exceed the low double digits.