Threshold guide · the $/W band

Solar cost per watt: cash vs loan price benchmarks.

Updated May 16, 2026. Sources verified the same day. The cash and loan benchmarks below use Berkeley Lab's latest distributed-solar data update.

Take the cash price on your solar proposal. Divide by the system's DC size, in watts. The number you get is the single cleanest cost comparator anyone can hand you. In Berkeley Lab's latest data update¹, 2024 residential PV-only systems landed at a median $3.50/W for cash purchases and $4.70/W for loan-financed systems. That gap is the story: the same roof can look fair on cash pricing and expensive once the loan structure is baked in.

So the first question is not “is this solar quote cheap?” It is: which price am I looking at? Cash $/W compares the system. Loan $/W compares the financing structure as well.

Why cash $/W, and not financed $/W

Cash $/W is the comparator that cuts cleanly across installers. A financed total isn't the same metric — it bundles APR-based financing cost with what may also be a dealer fee baked into the cash quote (use the solar loan calculator). Cash $/W is what you would actually pay an installer on the day of installation if you wrote a check; it's what LBNL's report uses to separate cash purchases from loans; it's the number you need before judging the financing.

Compute it on every quote you receive. The math is one division. The spread across three quotes is more revealing than any of the individual numbers in isolation.

The 2024 benchmark, in two sentences

$3.50/W cash median; $4.70/W loan median: the latest Berkeley Lab snapshot makes the cash-vs-loan split visible. A quote that looks close to market on cash price can move into expensive territory once the financed total is the number being compared.

The older 2024 Tracking the Sun summary still matters for range-setting: its 2023 state-median residential band ran from $3.20/W to $5.20/W on a cash basis, and its top-100 installer medians ranged from $2.60/W to $5.90/W². Translation: there is no single national “fair price.” There is a cash benchmark, a loan benchmark, and a state/installer spread wide enough to make the second quote worth getting.

What moves $/W inside the band

Four real factors push the same installer's price up or down by meaningful amounts. None of them are conspiracy; all of them are worth asking about.

1. System size. Smaller systems carry higher $/W because fixed costs (permits, truck rolls, design fees) get amortized over fewer watts. LBNL documents a $1.0/W difference in median price between the smallest and largest residential systems².
2. Equipment tier. Microinverters or DC optimizers (used on 93% of 2023 residential systems) cost more than string inverters. Premium-tier panels (high-efficiency monocrystalline) cost more than standard. Battery inclusion is a much bigger lever — Berkeley Lab isolates a meaningful per-watt premium for systems that include storage, separate from the PV-only headline number.
3. Site complexity. Steep-pitch roofs, multi-array layouts (e.g., split between south and west exposures), ground mounts, and main-panel upgrades all push $/W up. Berkeley Lab's regression treats ground-mount systems as a distinct cohort with a measurable premium over flush roof mounts. A 200-amp main-panel upgrade isn't in the LBNL headline numbers but commonly adds $1,500–$3,500 to the install — which is $0.18–$0.42/W on an 8.4 kW system.
4. State-fixed effects. LBNL's regression model isolates a $2/W range across state fixed-effect variables alone — meaning the same equipment in the same configuration on a similar roof can cost meaningfully different amounts depending purely on what state you're in. The drivers are cost of living, retail electricity rates, incentive program design, solar insolation, and permitting friction.

Where $/W can mislead

Three quiet ways the comparator misleads. Worth knowing before you rank three quotes side by side.

Low cash $/W. A quote well below the cash median is not automatically a steal. The price may not include permitting, inspection, interconnection, a main-panel upgrade, or sales tax — each of which can show up at contract signing or first invoice. Ask for the all-in cash price before comparing.

High loan $/W. Above the loan median, there may be a real reason — premium equipment, complex install, steep roof, small system in a high-cost market — but it is also where dealer-fee inflation tends to live. If the quote is financed-only and the cash price is not on the page, the elevated number may be telling you more about the loan markup than the equipment.

Identical $/W, different systems. Two proposals at $4.10/W can pencil out to very different 25-year lifetime costs. A microinverter system trades higher upfront cost for better partial-shade tolerance and panel-level monitoring. A standard string inverter system trades lower upfront cost for simpler maintenance but worse performance under shading. $/W is the cost comparator; performance comparison requires looking at production estimates against PVWatts for your specific site.

The 10-minute price test

Three steps. None require any tool the salesperson controls.

1. Compute cash $/W for every quote you have. Divide cash price by DC kW × 1,000.
2. Compare cash quotes against the latest Berkeley Lab cash median ($3.50/W), then compare any loan quote against the loan median ($4.70/W). A wide cash-to-loan jump is a financing question, not a panel question.
3. Run the quote calculator on the best one and on the median one. The verdict logic is the same band you just used, and it surfaces other thresholds (production, payback, financing gap, tax-credit timing) at the same time.

One more thing: your state still matters

The national cash and loan medians are useful, but state pricing still matters. The 2023 state-median band of $3.20–$5.20/W shows how wide the geography spread can get. LBNL publishes state-level breakdowns in its Tracking the Sun data tables at trackingthesun.lbl.gov). If your quote is $4.80/W in California, you may be high relative to local cash pricing. The same $4.80/W in a high-cost northeastern market may be closer to the local median. Both directions matter for negotiating leverage.

The framing question for any cost-per-watt conversation: relative to my state's 2023 median, is this number high, low, or normal — and if it's high, what specifically about my install justifies the spread? If the answer is a list of real items (steep roof, MPU, complex layout), that's a price you can negotiate against. If the answer is hand-waving, the spread is the markup.

Where $/W stops helping

The financed-side markup that often sits behind a low APR (see the loan calculator). The lifetime-cost picture, which depends on production estimates and utility-rate escalation that don't show up in $/W. Whether the system is sized correctly for your roof and usage (that's the system-size question, not the $/W question). For the audit that connects all of these, the quote calculator runs every threshold above and surfaces the financed-side gap and the production sanity check in one pass.

1. 1. Berkeley Lab, "U.S. Distributed Solar and Storage: 2025 Data Update" (October 2025; data through year-end 2024). Residential PV-only systems installed in 2024 had a median gross installed price of $3.5/W for cash purchases and $4.7/W for loans. The same update reports a 2024 median residential PV system size of 7.7 kW and a 14% battery attachment rate. Verified 2026-05-16. emp.lbl.gov · 2025 Data Update (PDF) ↩
2. 2. Berkeley Lab, "Tracking the Sun, 2024 Edition" (Executive Summary, August 2024; data through year-end 2023; sample ~3.7M U.S. distributed PV systems). State-level median residential installed price in 2023: $3.20–5.20/W. Loan dealer-fee gap, verbatim: "a large portion of residential systems are loan-financed, and installed prices reported for these systems likely include dealer fees, adding anywhere from 5-50% to the total up-front price paid by the customer." Verified 2026-05-16. emp.lbl.gov · Tracking the Sun 2024 (PDF) ↩

Next: Test the financed price — when the cash $/W and the loan total tell different stories.