Summary

US electricity demand is inflecting from a 2% to a ~10% CAGR — “broadly consensus,” Zach says, and “could be much higher” — driven by EVs, AI compute infrastructure, and industrial electrification. His dinner-party test: people guess 20-40% of new US cars are electric; the real answer is 7% (Europe ~25-30%), and he sees that going to 30-50% within five to ten years. Supply can’t respond: the interconnection queue holds twice today’s grid capacity, takes 5-10 years to clear, and only ~20-25% of queued projects get through.
The utility incentive structure is structurally broken: regulated utilities earn a return on equity on CapEx approved into rate base, so they have an “incentive to build, not to innovate” — ~40% of grid infrastructure predates the 1970s, and electricity prices have risen for two decades. The stakes per Zach: “there is no such thing as an energy-rich poor country,” and if the US doesn’t “work maniacally” to drive power costs down, “we’re gonna lose the race in AI, but we’re gonna lose the race in quantum and in biology.”
Base’s core insight is a utility-scale battery farm “chopped up into thousands of pieces” behind home meters — dodging the two constraints that cap utility-scale storage (the interconnection queue and transmission congestion) by deploying where the grid already exists, co-located with load. The customer pays $500 upfront plus $16/month, saves 10-20% on power, and gets whole-home backup versus $15-30k for a battery or generator — “cheaper than Netflix… It’s like a Costco membership.”
Unit economics: ~$6,500 net cost per installed battery (after a 30% ITC and the customer’s $500) generating ~$1,600/year in margin — $1,200 of it merchant energy trading at a modeled $40/kWh-year on a 30 kWh pack — for a ~4-year unlevered payback on a 10-15-year asset, targeted to 2-2.5 years by gen-3 in-house manufacturing. At 50% loan-to-value, the 20-30% unlevered IRRs “go up really significantly.”
The scary variable is merchant volatility: batteries earned ~$100/kWh in 2023 and ~$20/kWh in 2024, with a classic commodity boom-bust CapEx cycle behind it. Zach’s counter: “poles and wires move energy through space, batteries move energy through time” — batteries are fundamentally grid infrastructure, and utility partnerships (first: Bandera Electric, paying for fleet dispatchability and CapEx deferral) add contract-like utility payments alongside merchant exposure, with securitization — the Sunrun playbook — as the long-term financing path.
China leads LFP production: CATL, BYD, and Gotion are far ahead in LFP — “if you want to buy an LFP battery today you basically can’t do it outside of China” — and US attempts are a graveyard. But Zach calls the minerals scarcity fear “just fundamentally untrue” (citing Tesla’s Master Plan 3), and auto-OEM JV CapEx driving cell costs down is a trend Base benefits from strongly.
The company thesis is cultural as much as financial: “What SpaceX did to aerospace, what Anduril did to defense, Base wants to do to energy” — an engineering-led, vertically integrated entrant against incumbents where under 1% of employees work in R&D versus 20-40% at tech companies, with a longer-run “Palantir for utilities” play upstream.

Deep dive

1. The grid is “the most complicated machine ever built by humans” — and it’s aging out

Zach’s primer: the US actually runs three grids — the Eastern and Western Interconnects split by the Rockies, plus ERCOT in Texas — each with three components: generation, transmission (hundreds of thousands of miles of high-voltage lines), and distribution (millions of miles at neighborhood level). “I’m definitely gonna get this fact wrong, but” strung together the wires would reach the moon and back several times.
It’s a real-time supply-and-demand machine with effectively no storage — always-on, safety-critical, and “out of sight and out of mind for most people until it breaks.” In ~70% of the country, one investor-owned utility owns generation, transmission, and distribution in its territory.
The infrastructure is old: something like 40% was built before the ’70s, causing reliability problems and rising costs — every aged-out repair gets added back into the rate base. The historical arc: sub-5% of homes had power in the early 1900s; the New Deal (TVA, Hoover Dam) and post-war decades built the backbone of the American economy.

2. The rate-base machine pays utilities to build, not to innovate

Your power bill is roughly half the cost to generate, half the cost to move the power. Utilities propose CapEx to their public utility commission, it gets added to rate base, and they earn a regulated return on equity on it: “the more CapEx you deploy, the more return you generate for the shareholders.” Show up with unproven technology and the PUC calls it too risky — so rate bases and electricity prices balloon while innovation stalls.
Generation is a pure cost game: “Electron’s electron. There’s no special electrons” — levelized cost of energy (LCOE) is the core metric, and the lowest-cost generator wins. Pure-play retailers are “really just energy brokers” — 10-20% gross margin, sales-and-marketing-led, “notoriously not very good businesses” with a checkered history of blowing up on risk.
Gentailers — Vistra, NRG, Constellation, Calpine — pair a long book (generation) with a short book (retail): selling fixed-price retail power is a short position, owning generation the hedge. “That’s been a strong business model, and you can see that in the stock prices.”
The Texas exception: California deregulated in the late ’90s, Enron’s price manipulation “put a real quick stop to deregulation across the country,” and Texas carried the torch — though “deregulation is kind of a misnomer”; it’s a market structure creating competition. Today 80% of Texas is competitive (TDSPs like Oncor and CenterPoint barred from owning generation or retail) and 20% remains regulated munis and co-ops — making Texas “this laboratory for energy innovation” and the national leader in wind and solar.

3. No energy-rich poor countries — cheap power decides the AI race

The chart that hooked Zach as a student: energy consumption per capita against GDP per capita, “one of the strongest correlations in economics.” His categorical version: “There is no such thing as an energy-rich poor country.”
Cost thresholds define industrial viability — desalination, heavy machining, manufacturing only pencil once power gets cheap enough. Net of subsidies, solar is now the lowest-cost marginal source of power — though “energy is a geographically defined problem,” and he expects solar to win in more of the planet over the coming decades based on cost curves.
The geopolitical stake: the US is in “a regime of increasing electricity prices,” while China builds infrastructure to push power — and therefore compute — costs down. “If we don’t work maniacally to build out the infrastructure in this country… we’re gonna lose the race in AI, but we’re gonna lose the race in quantum and in biology and the next couple areas of innovation that are inevitably energy consumptive.”

4. Demand goes from 2% growth to 10% — and the queue can’t clear

Zach’s dinner-party test: ask ten people in a major US city what share of new cars sold this year are electric and “most people will guess twenty percent, thirty percent, forty percent. The real answer is seven percent” (Europe ~25-30%, China above that). His call: 7% goes to 30% then 50% over five to ten years, with long-haul trucking and the AI compute build-out — visible in interconnection queues, Microsoft/Three Mile Island, Oklo, the CoreWeave IPO — stacking on top.
Electricity demand has compounded at roughly 2% for fifty years; “it’s broadly consensus that that CAGR is gonna go from two percent to ten percent. Could be much higher.” Patrick’s gloss — worth keeping: 10% isn’t a high rate, but off this base “the absolute amount of marginal new demand… is crazy.”
Supply can’t respond: the interconnection queue holds roughly twice the capacity already on the grid, takes 5-10 years depending on state, and is polluted by “financial shenanigans” — developers filing multiple applications and funding whichever clears first. Only on the order of 20-25% of queued projects actually get through. Causes: transformer shortages, regulatory politics, and plain “trucks and crews and poles and wires execution.”

5. Base: a utility-scale battery farm “chopped up into thousands of pieces”

The product: Base becomes your power company in deregulated Texas, installs its battery on your home for $500 upfront and $16 a month, uses it as a grid resource when the grid is up, and backs up your home when it’s down — while saving customers 10-20% a month on power. The alternative is a $15-30k battery or a generator that’s loud, smelly, and needs maintenance. “We joke that there are no sexy electrons. People just want their bill to go down and their lights to stay on.”
The system-level view: 99% of grid storage is utility-scale — an asset class backed by Blackstone (IPA Power), BlackRock (Jupiter), Apollo (Broadreach) with tens of billions of CapEx — but capped by two constraints: 5-10-year interconnection and transmission congestion (you pick a node and “kind of inevitably get it wrong” by the time you clear the queue). Base deploys “where the grid already exists… co-located with the power load,” circumventing both.
Operationally: charge roughly midnight-4am, discharge into the 5-8pm summer peak or the winter morning ramp, 365 days a year, plus ERCOT ancillary-services qualification in progress. The claim: lower landed $/kWh than utility-scale, and higher revenue per kWh — because the same asset also earns retail power margin and the monthly resiliency fee.
Downstream adjacencies: with a battery and inverter on the home — “really like a computer on the circuits” — solar is the obvious next product, then smart EV charging, electric water heaters, and heat pumps, all optimized to the same North Star: “Lower the bill, keep the lights on.”

6. Unit economics: a four-year payback engineered toward two

Today it costs ~$10,000 to get a battery in the ground (hardware BOM, two electricians on site 4-6 hours, customer acquisition). A 30% ITC — 40% in the energy communities where most installs sit — plus the customer’s $500 nets to ~$6,500 unlevered. Income: ~$200/year from the $17 monthly fee, ~$200/year retail margin (10-15% on an ~$1,800 annual bill), and ~$1,200/year trading — a modeled $40/kWh-year over ten years on the 30 kWh gen-2 pack. That’s $1,600 on $6,500: “on the order of a four-ish year payback… people can check my math.”
The strategy is “developing a compounding cost advantage through vertical integration” — design, manufacture, install from own warehouses, own on balance sheet, sell and trade the power. Gen 2 takes cost to ~$8,000 (3-year payback); gen 3, possibly with their own factory, to ~$6,000 (2-2.5 years) — on 10-15-year useful-life assets. “In a commodity industry, your North Star has to be delivering the commodity to the customer at the lowest price possible.”
Leverage is the kicker: contracted ERCOT batteries get loan-to-values in the 70s-80s; merchant more like 30-50%. Base’s $1,200 trading stream is merchant today, but the $400 of customer payments and retail margin “look a lot more like contracts” — get LTV to 50% and the 20-30% unlevered IRRs “go up really significantly.”

7. The scary variable: merchant cash flows swing $100 to $20 per kilowatt-hour

Patrick zooms in on the outsider’s fear — what does a six-year drought of 2024s look like? Zach doesn’t dodge: commodities have boom and bust years, driven by “notoriously unpredictable” weather (home HVAC drives ERCOT price swings) and a reflexive CapEx cycle — 2023’s ~$100/kWh profits pulled in a wave of battery build-out, 2024’s weather didn’t show up and low volatility delivered ~$20/kWh.
The reframe worth keeping: “Poles and wires move energy through space, batteries move energy through time.” Batteries are more akin to T&D infrastructure than to wind and solar — fundamental value in voltage control, frequency response, and CapEx deferral, not just arbitrage, as VPPs (virtual power plants) go mainstream.
The second business is utility partnerships, first with Bandera Electric, which pays Base for dispatchability of a fleet deployed in its territory — lowering cost-to-serve and deferring transformer upgrades — with more Texas and national partnerships to be announced over the next year. Longer-run: utilities spend sub-1% of headcount on R&D versus 20-40% at tech companies, so “there’s a play here to kind of be the Palantir for utilities” — their outsourced R&D function.

8. China leads LFP production — but minerals aren’t the constraint

Battery basics: cell → module → pack. EVs grew up on NMC (nickel manganese cobalt) — energy-dense, high C-rate, light. Storage has flipped to LFP (lithium iron phosphate) — heavier, slower to charge, “but it’s a lot safer… way less prone to a thermal runaway, which is a fancy way of saying fire.” Zach calls sodium-ion “super promising” and watches iron-air for long duration.
China bet on LFP early: CATL, BYD, and Gotion are far ahead, with Samsung, SK, and LG racing to catch up. “If you want to buy an LFP battery today, you basically can’t do it outside of China.” The US record is a graveyard — US companies including American Battery Factory and others ran out of money before production scale; Northvolt also unraveled, though it was not US-based — because LFP plants are “not quite semiconductor complexity, but they’re pretty darn close”: billions of dollars and years to production scale.
On raw materials, Zach is categorical: the minerals worry is “just fundamentally untrue” — he points to Tesla’s Master Plan 3 on crustal abundance of lithium, iron, manganese, cobalt. Extraction is genuinely hard, “but we will not live in a world where we’re unable to build incremental battery capacity because of some kind of mineral constraint.” Meanwhile auto-OEM JVs plowing CapEx into cell factories keep driving costs down — a trend Base benefits from strongly.

9. The capital stack is half the company: from venture debt to securitization

Zach’s abstraction: one way to see Base is batteries behind the meter; another is “a yield curve. It’s just a cash flow stream” — matched to whichever capital provider will underwrite it. Since Base isn’t “insanely profitable,” its tax credits get monetized via transfer or tax-equity partnership flips — a mature market proven out in residential solar.
Debt comes in two flavors: project finance (SPVs, lending against cash flows, priced on merchant-versus-contracted mix and geographic diversity) and asset-backed securitization — “typically the largest, most liquid, and lowest cost part of the capital markets.” Sunrun “nailed the securitization and tax equity playbook”; SunPower and Sunnova did it “less successfully, and that’s kind of all unraveling.”
Today’s debt is venture-style, collateralized by balance-sheet cash and the ability to raise equity; as the fleet grows that breaks, and the destination is “almost certainly the securitization market.” The operating proof accumulating: almost 1,500 batteries in the ground, deploying 20 a day, going to 50 a day over the next couple of months.

10. Founding story and the excellence Zach has seen firsthand

The path: college years building anaerobic digestion systems in India — “this isn’t a business. This is like a cool project” — then Blackstone PE, where a 2018 lithium-mine carve-out forced a decade-long view on lithium and surfaced the core insight: solar-plus-storage’s marginal cost would undercut coal and gas, incremental grid capacity would be solar and storage, and storage was badly underbuilt. After Zach left Blackstone, IPA Power was bought; he carried away the two-constraint thesis.
At Thrive — “high conviction, high concentration, high involvement,” and “diversification is the enemy of returns” — an Anduril diligence factory tour introduced him to co-founder Justin, ex-head of manufacturing at Anduril who led SpaceX’s Starbase build. Three months of nightly 9pm calls started with Zach’s line: “if you built a really high-quality battery pack assembly business in the US, it’d be really hard to lose over the next decade.”
The company thesis in one sentence: “What SpaceX did to aerospace, what Anduril did to defense, Base wants to do to energy” — the engineering-led, R&D-driven entrant in a large industry of entrenched, non-technical incumbents. His operating credo, borrowed and lived: “everyone’s definition of good is just the best they’ve personally seen” — Blackstone’s analytical rigor, Josh Kushner optimizing “for the next deal” rather than the last dollar, and hires from Starlink and Tesla (Dino, thirteen years, original Roadster) setting the bar.
The material about Zach’s dad and Dell is the episode’s best color: he invites a sell-side analyst into board meetings to pitch the bear case — “don’t hold any punches. Give us your best shot.” Parenting as “demanding and supportive… pleased but not satisfied”: the 99 on a math test met with “What happened to the last point? You ran for the touchdown and you fumbled the ball on the one-yard line,” and the red car at the top of the hill that drives away as you reach it — “that red car is still driving away from him,” forty years in.

Episode 146 - June 27, 2025 Market Overview July 22, 2025