newsroom
$FIG · Google Stitch Just Destroyed Figma's Stock in One Update (Here's Why)
now playing · $FIG
$FIG···bearish· highunknown
Google Stitch decimates Figma valuation: down 80% since IPO
Google's new AI design tool Stitch generates production websites from napkin sketches in seconds, directly replacing Figma's core workflow for non-designers and causing a 9% singl…
$GOOGL···bullish· mediumunknown
Google's AI products dramatically underhyped despite superior execution
Gemini models and new releases (Stitch, AI Studio with Antigravity coding agent) are technically excellent and shipping rapidly, but Google suffers from a marketing awareness gap…
$RIVN···bearish· highunknown
Rivian loses $86K per vehicle; Uber partnership a decade too late
Rivian burned $3.6B on 42K deliveries ($86K loss/vehicle), has no autonomy division until 2024, no factory for R2, no software, and the Uber deal promises 50K robotaxis by 2031 —…
$UBER···neutral· mediumunknown
Uber hedging autonomy bets via Rivian and Nvidia partnerships
Uber is a great company with a bright future, but the Rivian deal looks strategic (securing hardware optionality) rather than competitive; the Nvidia/BYD stack creates a fragmente…
$TSLA···bullish· highunknown
Tesla's vertical integration and scale make it the only credible FSD winner
Tesla produces 4,700 vehicles/day (42K in 10 days vs Rivian's 50K in 5 years), owns full manufacturing and FSD stack, and has the data fleet to solve autonomy; partnership models…
$OPENAIbullish· highjosh kale
OpenAI pivots to coding/enterprise with Astral and PromptFlow acquisitions
OpenAI acquired Astral (Python tooling for 8-12M developers) and PromptFlow (security/debugging) to close the end-to-end software engineering gap vs Anthropic; winning the best co…
$ANTHROPICneutral· mediumunknown
Anthropic leads enterprise adoption (73% share) but Polymarket favors OpenAI for best coding model
Claude Code captures 73% of first-time enterprise AI purchases, yet prediction markets give OpenAI 94% probability of having the best coding model by March 31 — a narrative vs. re…
$CURSORbearish· highunknown
Cursor's Composer 2 claims beaten by chart crime: own benchmark, inverted axis
Cursor's self-reported benchmark (Cursor Bench) shows Composer 2 beating Opus 4.6, but the X-axis is inverted to fake cost efficiency, and Cursor lacks the capital to train a fron…
$MINIMAXbullish· mediumunknown
MiniMax M2.7 self-improves via 500 autonomous experiments, matches Opus 4.6 coding
Chinese lab MiniMax's M2.7 model allegedly matches Opus 4.6 and GPT-5.4 on coding after 100 rounds of autonomous self-improvement (30% gain), now handling 30-50% of the lab's rese…