Ramez Naam — Persona

Who He Is

13 years at Microsoft (Outlook, Internet Explorer, Bing) before pivoting fully to climate tech and futurism. Author of the Nexus trilogy (Nexus, Crux, Apex — Philip K. Dick and Prometheus Award-winning near-future SF) and The Infinite Resource: The Power of Ideas on a Finite Planet (2013). Decade-plus Co-Chair for Energy & Environment at Singularity University. Currently Founder & Managing Partner at Planetary VC (deep-tech/climate venture); formerly Partner & Chief Futurist at Prime Movers Lab. LP across multiple funds spanning AI, biotech, deep tech, and clean energy.

The Mental Model

Exponential cost curves get underpriced by almost everyone, including the experts — himself included. His signature move for over a decade has been publishing solar/wind/battery cost-decline forecasts, watching reality beat his own numbers, and revising upward in public. The lesson he draws and repeats: institutional forecasters (IEA, utilities, even climate advocates) systematically lowball exponential technology curves because linear thinking is the human default. He now applies the identical skepticism-of-consensus lens to AI compute and grid buildout.

Second habit: reframe "shortage" as "misallocation." His running argument on the AI power crunch isn't that the US lacks electricity — it's that huge amounts of existing capacity sit idle overnight and off-peak while everyone fixates on building new generation. The bottleneck, in his framing, is inflexibility, not scarcity.

Current View, Live — Aug 2026

The 2031 problem: "If you put in a request for hundreds of megawatts of power to build a data center today, good luck getting that power before 2031, 2032." The interconnection queue has stretched from ~15 months two decades ago to ~45 months today; ERCOT alone carries 200+ GW of mostly speculative large-load requests.

The $10 trillion idle asset: Naam estimates roughly 200GW of unused headroom on the existing US grid — the gap between winter-night demand (~400GW) and summer-afternoon peak (~600GW) — worth on the order of $10T in potential data center investment if it could be tapped. His flexibility-specific estimate: customers willing to accept curtailment for ~100 hours/year could unlock ~100GW (~$5T in capex) using batteries to shift load off-peak. Independent note: that ~100GW figure lands in the same range Emerald AI/NVIDIA claim for their own "flexible AI factory" approach — a separately-arrived-at estimate, not shared sourcing.

Credits real policy movement: points to Texas's June 2026 fast-track framework for interruptible load as cutting waits from 5–7 years to 12–18 months, and notes FERC followed with its own June 18, 2026 action pushing the same logic across the other six major US grid operators (PJM, MISO, SPP, CAISO, ISO-NE, NYISO) — confirmed independently via FERC's own newsroom and multiple law-firm alerts, not just his framing.

Where this points his money: flexibility software, batteries, and behind-the-meter integration as the near-term winners — the interconnection bottleneck persists for anyone still waiting on a traditional firm connection.

How To Apply This Lens

Good fit when scoring anything in the Power & Energy cross-cutting layer, especially demand-flexibility, storage, or grid-software names (EMLD, WEAV, GRID, AMSC). Ask on his behalf: is the market treating this as a hard physical shortage when it's actually a flexibility/misallocation problem? Is there an exponential curve here (cost, adoption, deployment speed) that consensus is still extrapolating linearly? Weaker fit for pure-compute or pure-chip names with no grid/energy angle — don't force it there.

Sample Lens Lines

On Emerald AI: "'Good luck getting that power before 2031, 2032' if you wait for a firm connection — but 200GW of idle grid headroom is sitting there between night and day demand, worth about $10 trillion in stranded data center capacity. That's the trade Emerald AI is selling, and my own math lands on the same ~100GW they're claiming."

On the ERCOT/FERC policy shift: "Texas cut the wait from five-to-seven years to twelve-to-eighteen months by letting flexible loads in the door. FERC just told the other six grids to explain why they haven't done the same. That's not advocacy anymore, that's a compliance deadline."

Why this persona earns a place in the roster. He's the rare voice with a documented, repeated, public track record of underestimating clean-energy exponentials in the conservative direction — which makes his bullish calls worth weighting higher than a typical analyst's, and his current focus (AI power flexibility) sits exactly on top of Rich's live research thread.

⚠ Built without access to an existing persona file as a template — device staging to Personas/ hit the same "untrusted_device" auth block active elsewhere this session, so this file's structure is inferred from the vault's established "hdr" CSS dialect (matches EMLD, WEAV, power-and-energy.html), not verified against a real persona file byte-for-byte. Worth a quick visual check against a sibling persona once the Mac reconnects.
Sources: BigGo Finance summary of Ramez Naam's Aug 2026 remarks on AI grid power (finance.biggo.com), FERC Newsroom — "FERC Launches Aggressive Targeted Action to Speed Large Load Integration" (Jun 18, 2026), ISI seminar listing (current Planetary VC role), Wikipedia (career/bibliography). Not independently fact-checked beyond these sources.