LED vs. HPS Grow Lighting: Energy Cost Comparison

The number that matters: photons per joule

Grow lights are not compared in lumens or watts. The metric is photosynthetic photon efficacy, PPE, in micromoles of photons per joule of electricity, written µmol/J. Higher is better, and it is the only lighting figure that converts directly to a utility bill.

Cited efficacy figures:

Fixture type	PPE (µmol/J)	Source
1,000 W double-ended HPS with electronic ballast	1.66 to 1.70	Nelson and Bugbee, PLOS ONE 2014 [nelson-bugbee-2014]
Mogul-base (single-ended) HPS	about 1.02	Nelson and Bugbee 2014 [nelson-bugbee-2014]
Ceramic metal halide	about 1.46	Nelson and Bugbee 2014 [nelson-bugbee-2014]
Fluorescent	about 0.95	Nelson and Bugbee 2014 [nelson-bugbee-2014]
Best LED fixtures tested in 2014	1.66 to 1.70	Nelson and Bugbee 2014 [nelson-bugbee-2014]
DLC horticultural QPL minimum, V3.0 (effective March 31, 2023)	2.30	DesignLights Consortium [dlc-hort-v3]
DLC horticultural QPL minimum, V4.0 (effective April 18, 2025)	2.5	DesignLights Consortium [dlc-hort-v4]

Two points from that table. First, in 2014 the best LEDs merely matched double-ended HPS, which is why early adopters were disappointed and why the "LEDs don't work" reputation lingers in the industry. Second, the floor has moved. The DLC now describes its V4.0 threshold as more than 45 percent above the most efficacious non-LED option, the 1,000 W double-ended HPS luminaire, and expected the higher bar to delist about 11 percent of the lower-performing products that qualified under V3.0, with final delisting on January 5, 2026 [dlc-hort-v4]. Fixtures on today's list start at 2.5 and many are listed above 3.0.

The Northwest producer survey found that most respondents still ran high-intensity discharge lamps in flower rooms, typically around 1,000 W each [nwpcc-cannabis], so the efficacy gap in the table is the gap most operating facilities are still sitting on.

Efficacy to electricity: the arithmetic

The chain from fixture to bill is short.

1. Decide the light intensity you want on the canopy, in µmol per square meter per second (PPFD), and the canopy area.
2. Multiply to get total photon flux needed, in µmol per second.
3. Divide by the fraction of fixture output that actually lands on the canopy.
4. Divide by fixture efficacy to get watts.
5. Multiply by hours per year to get kWh.

Because efficacy is in the denominator, going from 1.7 to 2.5 µmol/J cuts lighting kilowatts by 1 minus 1.7/2.5, which is 32 percent. Going to 3.0 cuts them by 43 percent. That holds at any intensity, as long as you deliver the same photons in both cases.

Heat: the second bill

Nearly every watt that enters a fixture leaves the room as heat, whether it first passed through the canopy as light or came straight off the ballast and arc tube. So a room that draws 32 percent fewer lighting watts puts roughly 32 percent less heat into the space, and the HVAC system runs less to remove it. Denver's guide documents a vegetative room that replaced 72 single-ended HPS fixtures totaling 72,000 W with 72 LED fixtures totaling 35,640 W, an immediate 36,360 W reduction, with a corresponding cut in air conditioning tonnage [denver-bmp-energy-2019].

The other half of the HVAC story does not follow the lights. Plants transpire based on light, temperature, and vapor pressure deficit, and at the same PPFD they release about the same water under LED as under HPS. Dehumidification load therefore falls far less than sensible cooling load, and a room that used to get free reheat from HPS may need supplemental heat in the lights-off period to keep dehumidifiers efficient. Our pages on dehumidification load and HVAC sizing for grow rooms cover that trade-off; it is a design question, not a reason to stay on HPS.

A worked comparison: 1,000 square feet of flowering canopy

Every figure below is an input we chose so you can trace the math. Replace them with your own spec sheets and your own tariff.

Assumptions:

Input	Value	Note
Canopy	1,000 sq ft = 92.9 m²
Target PPFD	1,000 µmol/m²/s	A common flowering target; a design choice, not a fact about your crop
Fraction of fixture photons reaching the canopy	85 percent	Assumed; depends on fixture optics, mounting height, and wall reflectance
HPS efficacy	1.7 µmol/J	Double-ended, per Nelson and Bugbee [nelson-bugbee-2014]
LED efficacy, two cases	2.5 and 3.0 µmol/J	The DLC V4.0 floor [dlc-hort-v4], and a typical better-than-floor fixture
Photoperiod	12 hours per day, 365 days	4,380 hours per year
Electricity price	14.19 cents per kWh	U.S. average commercial, June 2026 [eia-epm-5-6-a]
Cooling energy	1 kWh of HVAC electricity per 3 kWh of heat removed	A round coefficient of performance for planning
Demand charge	12 dollars per kW-month	Assumed

Photon requirement: 1,000 µmol/m²/s × 92.9 m² = 92,900 µmol/s on the canopy. Divide by 0.85 for optics: 109,300 µmol/s at the fixtures.

Lighting kW:

• HPS: 109,300 ÷ 1.7 = 64,300 W = 64.3 kW
• LED at 2.5: 109,300 ÷ 2.5 = 43,700 W = 43.7 kW
• LED at 3.0: 109,300 ÷ 3.0 = 36,400 W = 36.4 kW

Annual lighting kWh (× 4,380 hours):

• HPS: 281,600 kWh
• LED at 2.5: 191,400 kWh (saves 90,200)
• LED at 3.0: 159,400 kWh (saves 122,200)

Cooling kWh avoided: the lighting kWh saved is heat not added. At 3 kWh of heat per kWh of HVAC electricity, LED at 2.5 avoids about 30,100 kWh of HVAC use; LED at 3.0 avoids about 40,700.

Dollars per year:

	HPS	LED 2.5	LED 3.0
Lighting energy	39,960	27,160	22,620
Cooling energy attributable to lights	13,320	9,050	7,540
Demand, lights plus one third for HVAC, × 12 months	12,340	8,390	6,990
Total	65,620	44,600	37,150
Saving vs. HPS		21,020	28,470

So on this set of assumptions, 1,000 square feet of canopy saves roughly 21,000 to 28,000 dollars a year moving from double-ended HPS to a DLC-listed LED. Scale linearly: 10,000 square feet is ten times that. The saving is smaller where power is cheap (Texas' commercial average was 8.66 cents in June 2026) and larger where it is not (California, 27.33 cents) [eia-epm-5-6-a]. The LED retrofit ROI calculator lets you change every input.

What this example leaves out, in both directions: HPS lamp replacement costs (a recurring HPS expense), LED fixture cost (a large up-front LED expense), any yield change, and the dehumidification and reheat effects described above. The RII benchmarking study of 84 indoor facilities found LED-flowered operations 34 percent better on energy per canopy square foot and 80 percent better on grams per unit of energy than double-ended HPS operations [rii-led-study-2020], which suggests real-world results include a yield effect on top of the efficacy math.

Capital cost and payback

The reason LED adoption lagged was never the operating cost. Nelson and Bugbee found LED fixtures cost five to ten times more per photon than HPS in 2014 [nelson-bugbee-2014]. Prices have fallen a long way since, but a full flower-room retrofit is still a six-figure decision for a mid-size facility.

To turn the example above into a payback, assume (our number, not a market fact) LED fixtures cost 1.10 dollars per watt installed. The 43.7 kW LED case costs about 48,000 dollars per 1,000 square feet. Against a 21,000 dollar annual saving, simple payback is a little over two years before rebates, with the demand-charge portion depending on whether a ratchet delays it (see demand charges explained).

Rebates and the DLC list

Utility efficiency programs in most states on this site pay for horticultural LEDs, and almost all of them key eligibility to the DLC horticultural QPL. ComEd's business program, for example, includes LED grow lights meeting DLC horticultural listing and efficacy requirements among its standard incentives [comed-ee-business]. Programs generally require pre-approval before purchase, cap incentives at a share of project cost, and pay per fixture or per kWh saved. The FAQ on LED retrofit and rebate questions covers the paperwork.

Regulation is starting to point the same way. Massachusetts caps horticultural lighting power density at 36 watts per square foot for most cultivator tiers, with 50 watts allowed for Tier 1 and Tier 2, and offers an alternative compliance path for facilities whose fixtures are all on the DLC QPL with efficacy at least 15 percent above the QPL minimum [ma-935-cmr-500-120]. At 1,000 µmol/m²/s and 85 percent optical efficiency, the worked example's HPS room draws about 64 W per square foot, nearly double the cap, while the 3.0 µmol/J LED room draws about 36. That is not a coincidence; the rule was written around LED efficacy.