Will Fastly's LTM net retention rate drop below 105% in any quarter reported by Q3 2026 earnings?

NRR at 110% is 500bps above the 105% threshold with a strong upward trajectory (102%->106%->110%). For NRR to breach 105%, gross churn would need to roughly double from ~1.3% to ~6%+ while expansion holds, or expansion would need to collapse from ~11.3% to below ~6.3%. Neither is base case over 3 quarters. The usage-based model (76% of revenue) creates vulnerability to macro shocks -- a customer traffic decline of ~5% across the base could compress expansion significantly -- but the declining gross retention (30bps/year) is not accelerating fast enough to threaten 105% absent a shock.

NRR is an LTM metric, so Q1 2025 data rolls off and gets replaced by Q1 2026. Even if Q1 2026 is weaker due to seasonal patterns or macro headwinds, the LTM math requires Q1 2026 to be dramatically weaker than Q1 2025 to pull the trailing twelve months below 105%. The 800bps improvement over 4 quarters provides an enormous buffer. The company discontinuing annual retention rate disclosure could suggest they anticipate gross retention deterioration, but they are keeping NRR disclosure, which signals confidence in the headline number staying strong.

The strongest bear case centers on top customer concentration (34% of revenue, top 10 growing +28%). Loss of a single top-3 customer representing ~5-7% of revenue could mechanically reduce NRR by 300-500bps. Combined with background gross retention decline, this could push NRR toward 105%. However, multi-product switching costs (50% on 2+ products) make sudden churn of a major customer unlikely in a 3-quarter window. The probability of a specific top customer event AND it being severe enough to breach 105% is low but represents the primary non-trivial risk path.

NRR at 110% with a 105% threshold means a 500bps buffer on a strong upward trend. Even if the trend flattens entirely, NRR would need to actively reverse course. The gross retention decline of 30bps per year would take years to erode the expansion buffer at current rates. Multi-product adoption (50% on 2+ products generating 75% of revenue) provides structural support for retention. Enterprise customer net adds (+32 YoY) indicate continued land-and-expand execution.

The trajectory speaks clearly: 102%->106%->110% over four quarters is acceleration, not deceleration. The momentum is upward, driven by multi-product expansion and enterprise additions (+32 net adds). For NRR to drop 500+ bps in 3 quarters would require a significant negative shock -- major customer loss, macro-driven traffic collapse, or competitive displacement. None of these are signaled in the analysis facts. The gross retention decline is a long-term monitoring concern but not an imminent threat to the 105% level.

The unresolved committee debate about passive vs. deliberate expansion is the key uncertainty. If NRR expansion is primarily traffic-driven rather than product-driven, it is more vulnerable to macro slowdowns. In a recession scenario, usage-based revenue (76%) could compress, expansion could stall, and NRR could drop. But even in that bear scenario, going from 110% to below 105% in 3 quarters is aggressive. A more realistic bear case might see NRR drop to 106-108%, still above the threshold. The probability of a macro shock severe enough to cause a 500+ bps NRR decline within 3 quarters is low.

Strong uptrend (102%->106%->110%) with 500bps buffer to threshold. Multi-product adoption at 50% provides switching cost protection. Enterprise net adds positive. Very unlikely to breach 105% in 3 quarters absent a major exogenous shock not currently visible in the data.

NRR accelerating with enterprise customers growing and multi-product adoption at 50%. Gross retention decline is minor at 30bps per year. The 105% threshold is far below current 110% level. Would need unprecedented deterioration to breach in 3 quarters.

Usage-based model (76% of revenue) creates some macro vulnerability, and top customer concentration (34% of revenue) means one large loss could hurt. But the 500bps buffer and strong trajectory provide substantial protection. Even with pessimistic assumptions, most scenarios keep NRR above 105% over this timeframe.