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

Episode Summary

Executive Summary: The conversation centers on how going public has affected Cerebras and Planet Labs, and what their trajectories say about AI silicon and space as emerging infrastructure layers. Brad Gerstner frames both companies as proof that public markets can now reward long-horizon builders. Andrew Feldman argues Cerebras’ IPO mainly validated the team and sharpened execution, while Will Marshall says Planet’s public status improves credibility for governments and enterprise buyers. They also explore space-based compute, AI on Earth-observation data, and the idea that public markets may be reopening earlier for breakthrough companies.

Main Topics: The experience of going public (Priority: 5/5): Andrew Feldman and Will Marshall describe IPOs as validation events that bring capital, liquidity, scrutiny, and morale boosts, but do not change the core work of building the business. Public markets and long-term value creation (Priority: 5/5): Brad Gerstner argues that some of the biggest gains in venture-backed companies can occur after IPO, challenging the 'stay private forever' mentality. Planet Labs and the commercialization of space data (Priority: 5/5): Marshall explains Planet’s Earth-imaging network, the expansion of use cases beyond defense, and how public-company status reassures customers dependent on its data. Cerebras and the future of AI silicon (Priority: 5/5): Feldman outlines Cerebras’ thesis that AI requires a fundamentally different architecture than GPUs, with large chips and local memory to reduce data-movement bottlenecks. AI + space convergence (Priority: 4/5): The panel explores how large language models become more powerful when trained on real-world Earth data, and how space-based compute could become viable as launch costs fall. IPO timing and market pendulum (Priority: 4/5): The group discusses whether companies should list earlier, with the view that public markets may be shifting back toward earlier IPOs instead of extremely late-stage private funding.

Key Arguments: Going public is mostly a validation and liquidity event; it does not fix operational fundamentals or replace execution. Public status can improve credibility with customers, especially governments and large enterprises that need long-term continuity. A significant portion of a company’s total value can be created after IPO, so distributing shares too early may forfeit upside. AI expands compute into domains like images and language that computers historically handled poorly, increasing demand for specialized silicon. Cerebras’ big-chip architecture is designed to solve the core AI bottleneck: moving data between memory and compute. Planet’s Earth-imaging data becomes more valuable when combined with AI, enabling broader applications in agriculture, energy, civil government, and defense. Space-based data centers could eventually be cheaper than terrestrial ones if launch costs continue to fall and solar power in orbit remains highly efficient. The convergence of AI and space could create 'planetary intelligence' by feeding models real-world sensing data beyond text from the internet.

Data Points: Cerebras IPO pricing: $185 per share - Brad Gerstner says the IPO was priced at the top of the range, which was raised twice. Cerebras opening price: $320 per share - Gerstner notes the stock opened sharply above the IPO price. Cerebras market cap (mentioned): $50-$60 billion - Gerstner cites the company’s market value after the IPO move. Planet Labs stock move: from $5 to $50 per share - Gerstner describes roughly a 10x increase in Planet’s share price. Planet Labs public listing year: 2021 - Marshall confirms Planet went public in 2021 via SPAC. Planet Labs IPO valuation: $2 billion - Marshall says the SPAC valued the company around $2 billion. Planet satellite fleet: about 200 satellites - Marshall says Planet operates the largest Earth-imaging fleet. Earth imaging cadence: entire Earth every day - Marshall explains Planet images the full planet daily. Planet revenue from security: about 60% - Marshall says security-related customers make up roughly 60% of revenue. Earth observation market size: $75-$100 billion - Marshall estimates the market for Earth observation data plus AI applications. Launch cost today: just over $1,000 per kilogram - Marshall gives the current approximate cost to launch payloads. Launch cost threshold for space data centers: $200-$300 per kilogram - Marshall says this level could make space data centers cheaper than ground-based ones. Energy efficiency in sun-synchronous orbit: 5x more energy per solar panel - Marshall says orbital solar panels can collect about five times more energy than ground panels. Timeframe for lower launch costs: 2-3 years - Marshall expects launch costs to reach the space-compute threshold within a few years, especially with Starship. Potential compute horizon in space: within 10 years - Marshall predicts most compute may be placed in space within a decade. Cerebras performance claim: 15-18x faster than a GPU - Feldman says OpenAI uses Cerebras and it delivers much faster responses than GPU-based systems.

Pivotal Quotes: "I think the first thing is a lot of people asked us about how we got the timing right. Right. And I think the answer is by getting it wrong for a decade." — Andrew Feldman: Feldman describes Cerebras’ long path to IPO and the eventual emotional payoff for employees and families. "I think every single study shows that there is more money to be made, both in percentage and in what we care about, which is absolute." — Andrew Feldman: He argues that post-IPO value creation often exceeds pre-IPO gains, supporting holding shares longer. "Space and AI are really a match made in heaven. They're getting married, in fact." — Will Marshall: Marshall frames the strategic convergence between Earth observation, AI, and future space compute.

Implications: The discussion suggests public markets may again reward earlier-stage breakthrough companies, while AI and space are converging into new infrastructure categories. If launch costs keep falling, Earth data and orbital compute could unlock major new markets and reshape defense, science, and enterprise AI.