Exowatt Pays $215k Median Despite Being a 2023 Startup

The Hiring Surge: Roles and Timelines

Exowatt, a thermal storage startup founded in 2023, is staffing for the moment its hardware hits the field. The question isn't whether the company is hiring — it's whether the roles on the board today match the deployment deadlines set when it raised $90 million (Wikipedia reported).

As of mid-2026, Exowatt employs roughly 90 people across Miami and Austin. That headcount sat against a customer backlog the company pegged at more than 85 gigawatt-hours (Wikipedia's figures put) in late 2024, a Series A closed in April 2025 that explicitly funded "commercial deployments scheduled to begin in 2025," and a product, the P3 modular solar thermal system, that promised up to 24 hours of dispatchable power (Exowatt's site showed) without rare earth minerals or grid interconnection. The hiring wave visible now is the execution layer for those commitments.

Zero G Talent's live board shows five salaried roles posted recently, all at lead or senior level, split between the two hubs. In Austin: Lead Mechanical Engineer: Structures ($150–215k), Lead Mechanical Engineer: Heat Exchangers ($180–215k), Lead Controls & Firmware Engineer ($125–185k), and Senior HR Business Partner. In Miami: the same structures role ($150–215k) and Lead Mechanical Engineer: Heat Exchangers ($150–215k). The board's salary band runs $135k–$215k with a $215k median (Zero G Talent's board data shows). Every engineering role is a lead position — not "senior," not "staff," but lead — signaling that Exowatt needs people who can own subsystems through production release, not just contribute to prototypes.

The internship slate, announced on LinkedIn in July 2026, broadens the aperture: openings across engineering, supply chain, business, and marketing in both cities. Vincent Walker, Head of People and Talent, is the named contact. That breadth, supply chain and business alongside engineering, mirrors a company transitioning from R&D to repeatable manufacturing and customer delivery.

The timeline pressure is legible in the role titles. Heat exchanger leads and structures leads don't appear when you're still iterating in the lab; they appear when you're tooling for volume. Controls and firmware at lead level means the thermal-to-electric conversion loop and the dispatch logic that makes the P3 "dispatchable" are hardening toward field deployment. The HR business partner seat in Austin, the newer hub, suggests the people infrastructure is being built to absorb the next hiring tier once these leads are seated.

Exowatt's founders, Hannan Happi (CEO, ex-Siemens, GE, Tesla) and Jack Abraham (Chairman, Atomic venture studio), incubated the company at Atomic and emerged from stealth in April 2024 with a $20 million seed (according to Wikipedia) led by Andreessen Horowitz, Atomic, and Sam Altman. The Series A a year later, led by Felicis with $35 million in debt from HSBC Innovation Banking, was explicitly a deployment round. The hiring board today is the lagging indicator of that capital converting to headcount — and the leading indicator of whether the 2025 deployment target holds.

Energy Systems Meet AI Infrastructure: The Hybrid Skill Set

Exowatt's technology sits at a collision point: a modular solar-thermal platform that captures sunlight with fresnel lenses, stores it as heat in a proprietary thermal battery, and converts it to electricity through a heat engine, all factory-built, truck-transportable, and designed to sit beside GPU racks that each draw as much power as 100 homes. The company's open roles read like a map of that collision. Lead Mechanical Engineer: Structures. Lead Mechanical Engineer: Heat Exchangers. Lead Controls & Firmware Engineer. Each title carries a dual mandate: deep energy-systems expertise, plus the ability to operate inside the constraints of AI infrastructure.

The mechanical roles reveal the thermal-core priorities. Heat exchanger design at Exowatt isn't academic; it's the interface between concentrated solar flux and a thermal battery that must hold temperature for 24-hour dispatch with minimal losses. The company went through more than 50 design iterations (Wikipedia found) before settling on its thermal storage approach. That history means candidates need fluency in transient thermal analysis, materials compatibility at cycling temperatures, and manufacturability at a target rate of one module every 16 seconds. The structures role adds seismic rating, wind loading, and the peculiar constraint of desert deployment on "Frontier Land" — open acreage across West Texas, New Mexico, Arizona, and Nevada where the sun is strongest and zoning fights don't start. Factory-built, truck-transportable, 30-year lifespan (Exowatt's data shows): every requirement pushes toward standardized, repeatable designs that survive field assembly without specialized crews.

Controls and firmware sit on the other side of the interface. The P3's heat engine converts stored heat to electricity on demand, which means the control system must respond to data-center load profiles that spike and plateau with training runs and inference bursts. Exowatt's own communications describe digital-twin technology that predicts maintenance weeks ahead, thermal management that adapts to weather patterns automatically, and control software that "learns and optimizes based on actual load patterns." That language points to model-predictive control, real-time optimization, and firmware that runs on embedded hardware with deterministic latency — skills more common in aerospace or automotive power electronics than in traditional solar EPC firms.

AI infrastructure fluency is the third leg, and it's not optional. Happi has said the company was created in response to "growing energy demands from artificial intelligence and data centers," identifying a gap for renewable solutions that provide reliable baseload power. The numbers he cites are specific: a typical 100-megawatt data center equals 100,000 homes; U.S. data-center demand could exceed 600 terawatt-hours by 2030, over one-eighth of national generation. Candidates who pass the screen understand what those figures mean for power density, ramp rates, and the economics of behind-the-meter versus grid-interconnected deployment. They know why a GPU cluster can't tolerate the voltage sag that a commercial building shrugs off. They've read the interconnection queue studies and can explain why "dispatchable renewables" is a category distinct from solar-plus-storage.

The hybrid shows up in the manufacturing plan, too. Exowatt targets domestic supply chains: raw materials sourced and product manufactured in the U.S., with distributed fabrication near sunbelt data-center corridors. That means mechanical engineers who design for domestic tooling, controls engineers who specify components with secure lead times, and both fluently discussing how module cost curves behave when volume hits 10 million units over five years. The company's backlog, tripled from 30 to over 90 gigawatt-hours between September 2024 and October 2025, translates to 7 million modules in the near term and 72 million over five years. Every role touches that scale.

No single discipline owns the full stack. A thermal engineer who treats the data center as a black box will miss the ramp-rate requirements that size the heat engine. A firmware engineer who optimizes for steady-state efficiency will miss the cycling duty that kills battery longevity. A power-systems specialist who ignores the modular manufacturing constraint will design a system that can't hit the 16-second takt time. The screen filters for people who have already worked at the intersection, or who can demonstrate, with specific project evidence, that they think across the boundary.

Inside the Screen: Technical Tests and Cultural Fit

Vincent Walker, Exowatt's Head of People and Talent, serves as the named contact for the company's current hiring push, including the internship cohort opened across Miami and Austin in those roles. That single data point is the most concrete anchor the public record offers for how Exowatt structures its candidate evaluation. Beyond Walker's title, the company has not published a detailed breakdown of its interview stages, scoring rubrics, or pass-through rates. What can be reconstructed comes from the roles themselves, the hybrid domain they sit in, and the values leadership repeats in public channels.

The first-party board data shows five salaried openings with bands ranging from $135k to $215k. Each role demands a different technical depth. A mechanical lead for heat exchangers will face thermodynamics and fluid-dynamics scrutiny; a controls and firmware lead will be tested on real-time embedded systems, communication protocols, and safety-critical code. The HR partner role signals that people-operations rigor is being built in parallel. There is no single "coding test" that covers this spread; the technical assessment is role-specific by necessity.

What unifies the technical screen is the product: the P3, a modular thermal-energy-storage system that captures solar heat, stores it at temperature, and dispatches electricity on demand for up to 24 hours. Candidates for engineering roles must demonstrate they can work across the boundaries that P3 embodies: thermal-mechanical design, power electronics, controls firmware, and grid-interconnection standards. A system-design conversation for a controls engineer will likely probe how they handle state estimation when the storage medium is molten salt or concrete rather than a battery; a mechanical lead may be asked to trace thermal-stress cycles over a 30-year design life. The research does not confirm the exact format (take-home, whiteboard, live debugging), but the product architecture makes cross-disciplinary fluency a non-negotiable filter.

Cultural fit, meanwhile, is explicitly tied to Exowatt's stated mission: domestic manufacturing, energy independence, and "American dynamism" framing that appears repeatedly in founder communications and LinkedIn posts from mid-2024 through 2026. Happi's background — Siemens, GE, Tesla — and Abraham's venture-studio model (Atomic) both signal a preference for operators who have shipped hardware in regulated, capital-intensive environments. The company's public messaging frames energy as a national-security issue and emphasizes "no foreign supply chains." Candidates who cannot articulate why domestic sourcing matters for lead-time control, ITAR-adjacent compliance, or resilience will miss the values alignment bar, regardless of technical score.

The gap between what the research confirms and what a candidate experiences is real. Exowatt has not released an interview guide, and third-party accounts (Glassdoor, Blind, candidate write-ups) are sparse for a 51–200-person company that emerged from stealth only in April 2024. For applicants, the actionable inference is clear: prepare to defend design decisions in the language of both thermal-mechanical engineering and power-systems integration, and be ready to explain why building in the U.S., with U.S. supply chains, changes the engineering trade-offs you make. The screen selects for people who have already made those trade-offs.

Why the Grid Ran Out of Headroom

The last time U.S. electricity demand grew this fast, the iPod hadn't launched. Between 2005 and 2025, GDP more than doubled while total annual electricity use crept up just half a percent per year on average. That flatline shattered in 2019. Since then, demand has climbed 6 percent — from 3,813 to 4,051 terawatt-hours in 2025 — and the curve is bending sharply upward. The Energy Information Administration now forecasts retail sales above 4,230 TWh in 2027, roughly 10 percent higher than 2023. A Columbia University study aggregating utility filings to FERC puts the five-year annual growth rate at 4.7 percent. The driver is singular: data centers.

Their share of U.S. electricity has more than doubled, from 1.9 percent in 2018 to 4.4 percent in 2023. By 2028, the Department of Energy projects 6.7 to 12 percent, 325 to 580 TWh, two to three times 2023 levels. The Electric Power Research Institute sees 9 percent by 2030. Globally, Gartner expects 26 percent consumption growth in 2026 alone, pushing data center electricity to 565 TWh and power demand to 132 gigawatts, a 27 percent jump from 2025. By 2030, worldwide data center power could hit 290 GW. AI-optimized servers will account for nearly one-third of that consumption in 2026 and surpass conventional servers by 2027. GPU-accelerated deployments for AI have already more than doubled data center energy demand between 2017 and 2023, even as computing efficiency improved.

The constraint is no longer chips. It's power. "Surging demand for compute-intensive AI workloads is driving unprecedented data center power growth, while AI capacity is now constrained by power availability, making data center power security the new battle ground for scaling and protecting margins in the global AI race," said Linglan Wang, director analyst at Gartner. The grid is signaling distress: PJM's summer 2025 capacity auction cleared at $333 per megawatt-day, up from $29 a year earlier. ERCOT could see 14 percent load growth in 2026. Investor-owned utilities requested a record $18.2 billion in rate-base increases in 2025, more than double the 2019 figure. Residential prices rose more than twice the inflation rate in 2025. The bottleneck is physical: generation, transmission, and firm, dispatchable power that can run 24/7 without carbon penalties.

That is Exowatt's addressable market. The company's open roles map directly to the physics of the problem. The board's salary band clusters at a $215k median, signaling that Exowatt prices this talent at a premium, correctly, because the skill set is hybrid by necessity.

Google's 2025 sustainability report illustrates the customer pull: 37 percent year-over-year electricity growth, 250 percent since 2019, with data centers consuming 42 million megawatt-hours, comparable to the total demand of New Zealand or Denmark. Google matched 100 percent of consumption with renewable purchases for the ninth straight year and signed 12 GW of net-new clean energy in 2025, its largest annual total. Yet it also pursues advanced nuclear, enhanced geothermal, long-duration storage, and natural gas with carbon capture. The hyperscalers need every firm, clean megawatt they can secure, and they need it yesterday.

Exowatt's hiring surge is not organic growth. It is a forced response to a market that has moved from "planning" to "desperate." These roles, mechanical structures, heat exchangers, controls firmware, are the hardware layer that makes dispatchable, clean power work at data center scale. Candidates who understand both the thermodynamic constraints of a 100 MW liquid-cooled cluster and the control-loop latency a grid operator demands are the ones who clear the screen. The market pressure isn't abstract. It's in the auction prices, the rate cases, the 90 GW of data center load driving 55 percent of 2024's growth forecast. Exowatt is hiring because the grid ran out of headroom, and the only way to add more is to build systems that didn't exist five years ago.

Clearing the Bar: What Successful Applicants Show

Exowatt's open roles, two Lead Mechanical Engineer positions for structures, two for heat exchangers, a Lead Controls & Firmware Engineer, and a Senior HR Business Partner split across Austin and Miami, reveal a hiring bar built on a specific intersection: deep thermal-mechanical systems expertise paired with the control-loop fluency that dispatchable energy for AI workloads demands. Candidates who clear the screen do not treat "energy" and "AI infrastructure" as adjacent bullet points; they demonstrate how the two domains constrain each other in a 24-hour dispatch cycle.

Start with the mechanical roles. The structures and heat-exchanger leads both sit in a $150k–$215k band, and the job titles signal the physics Exowatt cares about: modular thermal storage that can charge from intermittent renewables and discharge steadily into a data-center load profile. Successful applicants show they have sized pressure vessels, managed thermal cycling fatigue, and designed for transportable modularity, not just static plant equipment. A resume that lists "heat exchanger design" without specifying working fluid, delta-T targets, or fouling mitigation strategies gets filtered. The same holds for structures: call out seismic or wind-load certification for relocatable skids, not just ASME code compliance for permanent installations.

The Lead Controls & Firmware Engineer role ($125k–$185k) is where the AI-infrastructure fluency becomes non-negotiable. Exowatt's product must respond to data-center power ramps that can swing tens of megawatts in seconds — a control problem that looks more like a GPU cluster's transient response than a utility peaker plant. Candidates who pass the technical screen walk through a specific example: a real-time controller they shipped that balanced state-of-charge across distributed storage modules while maintaining voltage stability under step-load changes. They name the RTOS, the sampling rate, the safety-integrity level, and the hardware-in-the-loop test setup. Vague "embedded systems experience" does not substitute.

Cross-domain translation is the differentiator. The strongest applications include a one-page addendum or cover-letter section that maps a past project's energy-system constraints, round-trip efficiency, degradation per cycle, start-up time, to the equivalent AI-infrastructure metrics: tokens per watt, checkpoint-recovery latency, or rack-level power capping. This is not marketing fluff; it proves the candidate can speak the language of the customer (the data-center operator) while engineering the supply side. Exowatt's own site frames the product as "powering AI with 24-hour dispatchable energy," and hiring managers screen for that framing instinct.

Location flexibility matters more than most applicants assume. Four of the five technical roles list Austin first; Miami appears as a secondary site for the mechanical positions. Candidates who explicitly note willingness to relocate to Austin, or who already live in the Texas energy-AI corridor, signal lower onboarding friction. The Senior HR Business Partner role, also Austin-based, suggests the company is building its people-operations muscle there, so engineering candidates who reference Austin's talent ecosystem (Tesla, Samsung, Oracle, the University of Texas) demonstrate local network awareness.

Finally, quantify the hybrid work. A mechanical engineer who writes "reduced thermal resistance 18% across 12 heat-exchanger iterations, cutting charge-time variance from 22% to 6% under variable solar input" gives the reviewer a dispatch-relevant metric. A firmware engineer who states "implemented model-predictive control on STM32H7, achieving 99.2% state-of-charge tracking accuracy at 1 kHz loop rate under 50% step-load transients" hands the hiring panel exactly the evidence they need. The salary bands (median $215k per board data) reflect the scarcity of people who can produce those numbers. Tailor every bullet to the dispatchable-power-for-AI constraint; everything else is noise.

Where Exowatt's Bar Stands Apart

Exowatt sits at a rare intersection: a hard-tech energy startup that speaks the language of AI infrastructure buyers, backed by the same capital table that funds frontier model companies. That positioning shapes a hiring bar distinct from both traditional renewables developers and pure-play AI labs. Three vectors define the difference — founder pedigree, deployment philosophy, and the thermal-versus-battery technical bet — each of which filters for a specific hybrid profile.

Start with the cap table. The Seed round brought those investors together; the Series A added Felicis Ventures leading $35 million equity (Wikipedia's data shows) alongside $35 million debt (according to Wikipedia) from HSBC Innovation Banking. That mix, venture studio, top-tier VC, AI-lab leadership, and institutional debt, signals a commercialization timeline measured in quarters, not decades. Traditional project-finance solar shops hire for PPA negotiation and tax-equity structuring. Exowatt's board data shows roles like Lead Mechanical Engineer – Heat Exchangers at $180k–$215k and Lead Controls & Firmware Engineer at $125k–$185k, bands that align more with SpaceX or Tesla energy than with NextEra or AES.

The deployment philosophy sharpens the filter further. Bill Flood, Head of Engineering with 35 years in aerospace, frames the company's bet explicitly against space-based and underwater data-center concepts: "Space is for what can't be done on Earth. Powering AI isn't one of those things." That stance — desert "Frontier Land" in West Texas, New Mexico, Arizona, Nevada — means Exowatt hires for rapid, modular, ground-based deployment on American supply chains. Candidates who have only modeled grid-interconnection studies in PJM or ERCOT often lack the civil-works, permitting, and logistics experience the company needs to ship containerized thermal batteries to remote sites and commission them in weeks. The LinkedIn posts emphasize "no foreign supply chains" and "permission" as the real holdup, a hiring signal for people who have navigated Bureau of Land Management leases, county zoning, and factory-to-site logistics, not just interconnection queues.

Then there is the thermal-storage bet itself. Latitude Media's September 2024 critique noted "we've seen it before": concentrating solar thermal with thermal energy storage has a history (Crescent Dunes, Ivanpah, various molten-salt pilots). Exowatt's differentiation is modularity, 24-hour dispatch, and a product (P3) designed for data-center load profiles rather than utility peaking. That technical choice filters for engineers who understand heat-exchanger fouling, thermal-cycle efficiency at scale, and the control systems that translate a data center's fluctuating demand into thermal-battery charge/dispatch cycles. Battery-storage veterans from Fluence or Tesla Megapack bring relevant power-electronics and BMS experience but often lack the high-temperature thermal-fluid and materials background. Conversely, concentrated-solar-thermal alumni from BrightSource or Abengoa know the thermodynamics but may not have architected the firmware-defined dispatch that an AI cluster expects. Exowatt's screen looks for the overlap.

The founder backgrounds reinforce the bar. Happi's path, Siemens, GE, Tesla, spans utility-scale generation, industrial power electronics, and automotive-grade manufacturing. Abraham's Atomic studio model means the company was incubated with a "ship product, not paper" operating rhythm. That combination produces a culture where the same structures role (posted at $150k–$215k in both Austin and Miami) is expected to own the design from CAD through first-article inspection in a contract fab shop, then integrate with the controls team's firmware release cycle. Pure mechanical engineers from traditional OEMs often struggle with the software-defined iteration speed; pure software engineers from AI labs rarely grasp the thermal-mechanical constraints that dominate the bill of materials and the reliability model.

Geography matters too. Dual headquarters in Miami and Austin, 51–200 employees split across both, forces a distributed-team discipline that many early-stage hard-tech companies avoid by co-locating. The internship program posted in July 2026 (those disciplines across both cities) and Walker's role as Head of People and Talent suggest a people-ops maturity unusual for a two-year-old startup. Candidates who have only worked in single-site skunkworks or fully remote SaaS teams may underestimate the coordination overhead of shipping physical hardware from two hubs to desert deployment sites.

In aggregate, Exowatt's hiring bar reads: thermal-systems depth at Tesla/SpaceX/automotive grade, firmware-level comfort with real-time dispatch, supply-chain and deployment experience in remote U.S. locations, and the communication cadence to keep Miami mechanical and Austin controls aligned. Peers in frontier energy-AI, whether battery integrators targeting data centers, next-gen geothermal developers, or small-modular-reactor startups, each emphasize a subset. Exowatt demands the intersection. That is why the board shows five salaried roles open simultaneously with narrow, high-band titles rather than a broad "energy engineer" requisition. The screen is not higher in absolute terms; it is narrower in dimensionality.

The Kicker

These roles, heat exchanger leads, controls leads, an HR partner in Austin, are the output of a talent-first approach. The next time the board updates, the names on it will be the people who made the 16-second takt time real.

Working in frontier tech? Zero G Talent tracks the openings: see every open Exowatt role, browse frontier tech jobs, the companies hiring, and the people building the field.