Today’s AI Top Pick: UPST

Upstart is the cleanest setup on the board today. All three available timeframes point up with meaningful magnitude: 4h fc_short +42.11%, 1d fc_short +4.90% / fc_mid +46.75% / fc_long +71.90%, and 1wk fc_short +17.16% / fc_mid +31.89% / fc_long +76.64%. That's rare multi-timeframe agreement combined with a kronos bullish_prob of 1.0 and a near_term_bullish of 0.8 — the model isn't just calling a long-dated bounce, it's calling immediate strength too. Critically, you're NOT chasing. UPST sits at position_in_21bar_range_pct of just 5.94 (4h), 15.82 (1d), and 4.13 (1wk) with a weekly drawdown of -25.33% from the high. Entering near a range low with a bullish forecast is exactly the asymmetry we want. Contrast that with MNDY (73-81 in range) or SMCI (92 on 4h with a fully bearish forecast) — those are extended and rolling over. Fundamentals clear the screen: fwdPe 14.1, peg 0.65, salesYoY 46.3%, epsNextY 171.16%, profitMargin 4.84%, targetUpsidePct 50.3%, shortFloat 31.09% (biggest squeeze fuel in the pool). The near-term catalyst that seals it: on 2026-09-11 CEO Paul Gu bought 50,000 shares for $1.3M — insider conviction at the lows, right when the kronos setup is firing. Waiting for a lower price risks missing the squeeze; the range-low entry IS the pullback. ADTN and PAR score close behind on pure forecast magnitude, but ADTN carries a -14.7% post-earnings hangover headline and negative margins (-1.99% profit, -0.18% oper), and PAR has weekly range positioning at 73.4 (a bit extended) with a -14.52% profit margin. UPST offers the best combined package of tape, valuation, catalyst, and asymmetry today.