← All Scores / SFTBY
BULL Cloud Infrastructure SoftBank Group Corp

SFTBY

85 -1 all-time
Last scored 2026-05-29

SoftBank launching AI GPU cloud service in October 2026 under neocloud strategy positions them as major infrastructure provider for agentic AI workloads

Reading — Higher scores = stronger evidence of agentic AI tailwinds benefiting this company.
§ 01 — Score History
100
Ceiling
85
Latest
0
Floor
§ 02 — Reasoning

Score timeline

2026-05-29 85 0

SoftBank has the strongest catalyst in this batch — direct sovereign AI cloud launch — and explosive momentum at 96th percentile. However, thesis and win_probability both sit at 50, missing the new_buy convergence bar (thesis>=60 AND win>=65). Elevate on watchlist; if thesis/win confirm next cycle this becomes a buy. Strong rank in batch reflects catalyst+momentum cluster matching our winning pattern.

2026-05-28 85 -2

Strongest setup in batch. Direct, ticker-specific AI infrastructure catalyst (sovereign AI cloud) paired with near-max momentum — this is the breakout + direct AI news fingerprint our winners share (MU, MRVL, PANW). Win_prob technically below the 65 convergence bar, but momentum at 98 and a direct strategic catalyst justify a starter position rather than a skip per our cluster-mate elevation learnings. Initiate small.

2026-05-27 87 +16

Strongest catalyst in the batch — direct AI infrastructure neocloud launch confirmed across four outlets, with positive momentum. However win_probability of 50 fails the 65 convergence bar for new_buy. Elevate ranking but don't initiate without stronger win-prob confirmation.

2026-05-26 71 -15

Best of the batch by a margin — direct AI Data Center GPU Cloud product launch in Japan is a ticker-specific catalyst with strong momentum confirmation. However, win_probability of 50 falls short of the 65 new_buy convergence threshold. Elevate the ranking but don't pull the trigger yet; revisit if win-prob firms.

2026-05-25 86

Strongest of the batch — direct, ticker-specific AI infrastructure catalyst (SoftBank AI GPU Cloud Service) plus solid momentum. Fingerprint resembles our winners. However, win_probability=50 falls short of the 65 new_buy convergence bar. Elevate conviction for ranking purposes but cannot recommend new_buy under hard rules. Watch closely.

Supervisor: ALARM — the scoring pipeline has a failing check.