More parallel subagents made my pipeline slower. Here's the data.

A developer found that adding more parallel subagents to an ad-creative analysis pipeline increased latency rather than decreasing it, because context assembly before the LLM call became the bottleneck. With 8 subagents, aggregation consumed 61% of wall-clock time, and serializing JSON blobs took over 4 seconds. The fix was to have subagents write summaries to R2, reducing aggregation context from ~6,400 to ~1,100 tokens and cutting monthly costs from $207 to $38.

Adding a 7th subagent pushed my orchestrator latency from 22s to 31s — the opposite of what I expected. I'd been running a fanout pattern in my ad-creative analysis SaaS: spawn N subagents in parallel, collect results, merge into one verdict. The parallel part worked fine. Individual subagents finished in 9–12 seconds regardless of how many I spawned. The problem was everything after that. With 8 subagents, each returning ~800 tokens of analysis, the orchestrator was assembling a 6,400-token context before it could even call the LLM once. On Cloudflare Workers, serializing 8 JSON blobs into a single prompt string was taking 4+ seconds of pure CPU time before the first API call fired. The log entry that made it obvious: worker:orchestrator WARN aggregate context size=52480 bytes serialize duration=4312ms reason="context assembly backpressure" Measured across 3 weeks of production data: | Subagents | Total latency | Aggregation share | |---|---|---| | 2 | 14.2s | 18% | | 4 | 16.8s | 31% | | 6 | 22.4s | 47% | | 8 | 31.1s | 61% | At 6+ subagents, aggregation consumed more than half the wall-clock time. The fanout was fast. The funnel was the bottleneck. The fix wasn't reducing parallelism — it was changing what the orchestrator actually reads. Instead of passing full results to the aggregation LLM call, each subagent now writes to R2 on completion. The orchestrator pulls only a three-field summary struct per agent verdict , confidence , top signal . Eight agents still produce eight files, but the aggregation context dropped from ~6,400 tokens to ~1,100. Monthly cost for that one pipeline step: $207 → $38. The counterintuitive part: the bottleneck wasn't the LLM. It was the context assembly happening before the LLM even got called. I wrote up the full breakdown — including the R2 chunking pattern, the D1 counter approach for tracking partial completions without polling, and the KV-based loop guard for failed aggregation retries — over on riversealab.com.