ICON PLC (ICLR) — Technical state

As of 2026-09-17 — re-computed nightly across the coverage universe. Closelook diary readings, not advice.

Uptrend stack intact

Trades +1.4% vs the 50-day and +15.2% vs the 200-day average, 16.8% below the 52-week high and +154.9% off the 52-week low; a golden cross printed 39 sessions ago. Computed from daily closes.

Ranked #49 on today’s Directional Flow — Rising fastest list (scan of 2026-09-17).

The moving-average stack

vs 20-day+1.3%
vs 50-day+1.4%
vs 200-day+15.2%
52-week band16.8% off the high · +154.9% off the low
Streak1 consecutive down session
CrossoverGolden cross 39 sessions ago
50-day crossprice crossed up through the 50-day 4 sessions ago
Stochastic%K 62.4 · %D 79.6 · bear cross

Continue: the ICLR chart & overview · the signal board

FAQ · from the current data · as of 2026-09-17

Quick answers

What is the current Closelook signal on ICLR?

Uptrend stack intact — Trades +1.4% vs the 50-day and +15.2% vs the 200-day average, 16.8% below the 52-week high and +154.9% off the 52-week low; a golden cross printed 39 sessions ago. Computed from daily closes. As of 2026-09-17; the signal board re-computes across the coverage universe nightly. A research diary reading, not advice.

Is ICLR above its 50-day and 200-day moving averages?

As of 2026-09-17, ICON PLC trades +1.4% above its 50-day and +15.2% above its 200-day average, with the 20-day +1.3% away.

How far is ICLR from its 52-week high?

16.8% below the 52-week high and 154.9% above the 52-week low, as of 2026-09-17.

Mentioned on Closelook

Where ICLR appears in the diary — 2 pieces, newest first. Every link is our own writing.

Long reads

Google and RSI — The Loop Nobody Has ClosedReport · 13 Sept 2026The International AI Safety Report 2026 devotes a section to it, and ICLR 2026 ran a workshop with recursive self-improvement in the title.

Dailies

The Opex Handover — Why the AI Trade Is Rotating, Not Rolling OverDaily Pulse · 30 Jun 2026Google's KV-cache compression (ICLR 2026) cuts inference memory roughly 6× and speeds attention up to 8× on H100-class hardware — software-only, training-free, with no measurable accuracy loss.