Sequoia's David Cahn: AI IPOs will reset all private valuations — and energy needs deep IP, not just capital

Sequoia's David Cahn: AI IPOs will reset private valuations — and energy needs deep IP, not just capital

Cahn's opening frame is simple: venture capital has changed shape. Five or six years ago, a hundred interesting companies competed for attention. Today, most of the conversation concentrates on the top ten. That concentration of capital into OpenAI, Anthropic, xAI, and a handful of others has effectively restructured how the market works — and the first wave of AI IPOs is about to force a reckoning.

The IPO reset

The core problem with late-stage AI valuations right now is the absence of liquid comps. In 2021, investors could look at public CrowdStrike and Datadog to underwrite private Databricks. No equivalent framework exists for Anthropic or OpenAI today. Cahn argues that changes the moment the first major AI company goes public. If Anthropic is first out, its trading multiples will immediately provide a valuation anchor for every private AI company behind it — which is either clarifying or painful depending on where you sit.

SpaceX is the closest existing bellwether. Cahn notes it has held up well, and the liquidity unlocks that spooked analysts ahead of time turned out to be non-events; the stock actually traded up roughly 7% on what was supposed to be the big unlock day.

The transition won't be a one-day wave. Lockups, LP decisions to hold, and RIA allocations will spread the impact. But the optionality of liquidity changes behavior well before actual cash flows back into seed rounds.

“The most devastating thing to the frothy hype and the traction that we're seeing with AI and the funding of it from VCs would be, if Anthropic and or OpenAI have a blip in their public exits, their IPOs... That would be catastrophic, and it would immediately reset virtually all valuations in the AI world. And probably deprive half the companies that are raising money from access to capital.”

Data center buildout: servers, steel, and power

Cahn frames the full infrastructure stack around three pillars he outlined in a blog post roughly 18 months ago: servers (chips), steel (industrial components), and power. The market is well-educated on the compute layer now. The steel and industrial side is getting attention — Elon Musk making his own gas turbines being the visible example. Power remains the hardest constraint, with effectively no spare capacity available in the next 24 months and CapEx planning cycles that stretch two to four years.

Why energy looks more like SpaceX than Neo Cloud

The Neo Cloud model was primarily a financing innovation: buy Nvidia chips, finance them through Blackstone or a similar vehicle, and operate at scale. There's real execution skill involved, but the IP is thin. You could, Cahn says, start a Neo Cloud tomorrow if you understood the deal mechanics.

Energy is different. The companies worth owning have spent years building IP that is genuinely hard to replicate — and that's what separates them from capital-deployment plays.

Form Energy

Sequoia's most recent energy investment is Form Energy, a grid-scale iron-air battery company. The founder, Mateo Chang (identified in locked facts as a former Tesla energy business leader and described by Cahn as one of the top three battery experts alive), has spent nine years developing the technology. Iron and air are both cheap inputs, which is the core logic: abundant, low-cost materials for grid-scale storage. Form recently signed a project with Google.

Cahn's thesis is structural: more solar and wind on the grid means more batteries. A company with nine years of accumulated IP is well-positioned to scale into that demand, particularly as AI data centers create urgent new power needs.

Sequoia's energy portfolio also includes Valor, a nuclear investment led by partner Shaun Maguire, who wrote a hardware manifesto years before hardware investing became fashionable. His framework — hardware companies spend years building potential energy through hard R&D before it converts to kinetic energy and stock prices move — is the lens Cahn applies across the energy stack.

The capital-raising test for hardware founders

One constraint Cahn flags for seed-stage hardware founders is whether they can plausibly raise a billion dollars. Even exceptional IP doesn't survive if you can't get there. He tells founders directly: if he can't see the path to a billion in capital, he won't invest, because the company will get consolidated before it reaches scale.

The tell he looks for is whether he'd go work for the person. That framing captures founder quality more efficiently than any checklist: an employee betting on a company is putting 100% of their equity into one bet, so the judgment is high-stakes and instinctive. Beyond that, clarity of thought matters. Form Energy's pitch is a good example: iron-air batteries use cheap inputs; storage is needed because the sun doesn't shine at night. Simple logic, coherently delivered, is what allows a founder to recruit talent and raise capital across many rounds.

Relationship time also matters. Cahn had known Form Energy's founder for six years before Sequoia invested. Maguire knew his nuclear founder for years before committing. These aren't decisions made in a single pitch.

Timeline synchronization

On physical-world AI acceleration, Cahn's most practical point is about timing. Starting an AI power company today is probably too late — the need is immediate, and the companies already positioned have a multi-year head start. Nuclear is the opposite: the inflection is likely in the 2040s, so an early-stage company has time to mature. Matching the company's development timeline to the market opportunity's inflection point is, in his view, one of the most important calls an investor makes.

The broader shift toward hardware investing among VCs reflects a structural concern about the digital layer: timelines move so fast in software that moats are hard to build. Hardware companies that survive the multi-year R&D gauntlet can end up with durable competitive positions. The terminal version of that logic is the machine that makes the machine — re-industrialization at scale, where Cahn sees the most significant long-term investment opportunity.