TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
←
▶ 23:00 · $NVDA · Feldman: Building GPU-like architecture has near-zero odds of beating Nvidia
episode briefing
All-In Podcast

The IPO Comeback: Why Tech Giants Are Finally Going Public | All-In Liquidity IPO Panel

2026-06-06 · 6 company · 4 thematic
sentiment
4 bull1 bear1 neu
speakers
will marshall

Co-founder and CEO of Planet, the satellite-imaging company that operates a large fleet of Earth-observation satellites.

brad gerstner

Founder and CEO of Altimeter Capital, a technology investment firm, and host of the BG2 podcast. Known for concentrated growth bets and widely-read public memos on big tech.

andrew feldman

Leads Cerebras, a public company (CBRS) building wafer-scale AI inference chips; previously co-founded SeaMicro (acquired by AMD); describes $25B backlog and global data center buildout at unprecedented scale.

quote
“if you want to be 20 times better than somebody, right? Your architecture can't look like them, right? It it it can't. They have they have enjoyed and and eaten all the lowhanging fruit. So, if you build a GPU, the odds that you're better…”
—andrew feldman
now playing · $NVDA
$PL···bullish· high· poswill marshall
Marshall: AI + daily earth imaging unlocks $75-100B market for planetary intelligence
Planet's 200-satellite fleet images entire Earth daily creating a time-series dataset; applying LLMs to this imagery unlocks agriculture, energy, civil government, and security ap…
$PL···bullish· high· poswill marshall
Marshall: Space-based data centers become economic when launch costs hit $200-300/kg
Launch costs have dropped 10x in a decade to ~$1,000/kg; Starship trajectory suggests $200-300/kg in 2-3 years, making orbital data centers cheaper than terrestrial due to 5x sola…
$NVDA···bearish· highandrew feldman
Feldman: Building GPU-like architecture has near-zero odds of beating Nvidia
Nvidia has captured all low-hanging fruit in GPU architecture; any competitor building a similar architecture cannot achieve 20x differentiation. Cerebras chose wafer-scale + near…
$CBRS···bullish· high· posandrew feldman
Feldman: Cerebras wafer-scale architecture delivers 15-18x inference speedup vs GPUs
Cerebras solved the memory-compute bottleneck by building a dinner-plate-sized chip with memory adjacent to compute, enabling real-time AI inference that GPUs cannot match; OpenAI…
$GOOGL···bullish· mediumbrad gerstner
Gerstner: Google hasn't sold a single Planet Labs share since IPO
Google remains Planet Labs' largest single investor and has not distributed any shares post-IPO, signaling long-term conviction in the space data and planetary intelligence thesis.
$SPCX···neutral· mediumbrad gerstner
Gerstner: SpaceX IPO would need quadrillion valuation to match historical public-market upside
Historical mega-caps (Google, Meta, etc.) went public at few billion, not trillion; SpaceX at current ~$200B+ valuation has already captured most upside privately. For public inve…