CVD vs HPHT Yield Cost Comparison - HOLYCOME
A lab-grown diamond's wholesale price is set long before a jeweler ever sees the stone. By the time a buyer compares a two-carat CVD diamond against a two-carat HPHT diamond on a price list, most of the cost gap has already been decided inside the growth chamber: how many usable crystals came out of the run, how much of that rough survived cutting, and how much annealing, irradiation or heavy-element removal each stone needed. This guide breaks down CVD vs HPHT yield and cost the way Shenzhen factories actually track it -- per growth run, per rough carat, and per polished carat -- so you can tell when a price difference is structural and when it is just a marketing label.
Two growth chambers, two cost models
CVD (chemical vapor deposition) grows diamond by cracking a carbon-containing gas, usually methane, in a microwave plasma ball suspended above a small diamond seed plate. The chamber runs at low pressure and roughly 800 to 1,000 degrees Celsius, so a growth run is essentially an electricity bill plus a gas bill plus machine time. HPHT (high pressure high temperature) recreates the earth's growing conditions inside a cubic or belt press: carbon dissolves in a molten metal flux at about 1,400 degrees and 5 to 6 gigapascals, then precipitates onto a seed. The press is a heavier capital asset, draws more peak power, and needs metal catalyst cartridges that become consumables.
The consequence is that CVD economics scale with plasma quality and uptime, while HPHT economics scale with press cycles and catalyst cost. A CVD chamber that stays lit for 25 days straight with stable plasma is a good run; an HPHT press that finishes its cycle without a blowout is a good run. Both methods have improved sharply since 2020, and the per-carat cost gap between them has narrowed, but the shape of the cost stack remains different enough that buyers who understand it can negotiate harder.
What factory "yield" actually means
When a factory talks about yield, it usually means one of three numbers, and mixing them up is how buyers get misled. The first is growth yield: how much weight of sellable rough comes out of a run compared with the theoretical carbon deposited. The second is cutting yield: how much polished weight survives after sawing, bruting, blocking and polishing a given rough crystal. The third, which is the one that actually matters for margin, is graded yield: the share of polished carats that come out at a salable color and clarity grade, rather than in a bucket that must be discounted or re-sold as industrial material.
A crystal can have a great growth yield and a terrible cutting yield if it is riddled with internal stress that cleaves during sawing. It can have a great cutting yield and a terrible graded yield if it lands in a brownish tint that forces a heavy annealing step. Every number matters, and the best factory buyers ask for all three instead of accepting a single rough carat price.
CVD rough: yield profile and run economics
CVD grows as a flat rectangular plate, typically two to six millimeters thick once the run is mature. That geometry is good news for round and fancy shapes because the plate is already roughly parallel to the crystallographic plane a cutter wants. A single chamber can grow one plate per run, and modern fabs stack multiple plates in one plasma volume, which is where the per-carat cost advantage of CVD comes from. Power consumption for a mid-size CVD run typically sits in the 15 to 35 kilowatt range, and a run that produces 8 to 20 carats of rough from a well-tended chamber is routine.
The downside of CVD yield is stress and color. Early CVD plates came out brownish or grayish because of vacancy defects left behind during growth. Modern plates still carry a brownish or orange tint that requires a high-pressure high-temperature anneal to remove, and some stones also need electron irradiation followed by annealing to shift off-color tints into marketable near-colorless or fancy territory. That post-processing is not optional for a large share of CVD output; it is a line item in the cost stack. Internal defects in CVD tend to be graining lines, dark pinpoints and metallic flux remnants, rather than the feathery cracks common in HPHT, which means a CVD plate often cuts cleanly but must be graded honestly for inclusions.
HPHT rough: yield profile and press-cycle economics
HPHT grows a roughly cubic or octahedral crystal that already looks like a rounded diamond shape. That geometry is a gift to the cutter: a well-formed HPHT rough can often be fashioned into a round brilliant with less sawing loss than a CVD plate of equivalent weight. Press cycles run 70 to 120 hours depending on target size, and each cycle consumes a metal catalyst cartridge (iron, nickel or cobalt based), a graphite heater, and a seed. The press itself is a chunky piece of equipment, and wear on the die assembly is a real maintenance cost that shows up as downtime between cycles.
HPHT yield problems are different. The metal flux that carries the carbon can trap tiny metal particles inside the growing crystal, which is why HPHT rough often contains dark metallic inclusions that are opaque under the loupe and can only be avoided by careful orientation at sawing. Nitrogen uptake during growth produces yellow tint, and a share of HPHT crystals come out too brown or too gray to sell as near-colorless without treatment. On the plus side, a clean HPHT crystal can reach a vivid yellow or fancy green naturally, and those fancy colors command premiums that CVD usually cannot match without irradiation. For near-colorless white goods, HPHT has historically been slightly more expensive per rough carat than CVD, but the gap depends heavily on press utilization and local power pricing.
Side-by-side cost stack
The table below shows a typical Shenzhen factory cost stack for producing one polished carat of near-colorless (G to J) white goods, expressed as relative indices where CVD near-colorless white goods = 100. These are factory-typical ranges, not quotes, and real numbers move with power tariffs, chamber utilization and scrap rates.
| Cost component | CVD line (index) | HPHT line (index) | Notes |
|---|---|---|---|
| Growth power per run | 80-110 | 120-170 | HPHT press peak draw is higher |
| Consumables (gas vs catalyst) | 70-100 | 130-180 | Metal cartridges are a recurring cost |
| Machine downtime / maintenance | 90-120 | 120-160 | Press die wear is heavier |
| Post-treatment (anneal / irradiate) | 110-160 | 80-120 | CVD plates more often need color correction |
| Cutting yield loss | 100-130 | 80-110 | HPHT cube geometry cuts out more weight |
| Reject / downgrade rate | 100-140 | 90-130 | Varies heavily by run quality control |
| Composite polished cost | 100 | 105-130 | Wide run-to-run spread |
The takeaway is that CVD is usually cheaper to grow but more expensive to finish, while HPHT is more expensive to grow but often cuts out better. At the polished level the two methods land closer than most retail marketing suggests. The same dynamics are visible when you compare the broader lab grown diamond cost per carat economics that factories publish each year.
Post-treatment cost: the line item buyers forget
A large share of both CVD and HPHT rough leaves the growth chamber in a state that is not yet salable as near-colorless. CVD plates commonly need a high-pressure high-temperature anneal to remove brown vacancy color; some also need irradiation (electron beam or neutron) followed by a second anneal to push the tint to a marketable near-colorless or fancy yellow-green. HPHT crystals that come out brownish often respond to a lower-temperature anneal, and stones that carry heavy metallic inclusions may need nothing at all but will be graded lower for clarity.
Each treatment step costs money: chamber time, labor, and the risk that a stone cracks or fractures under the thermal cycle. A factory that can anneal in-house captures that margin; a factory that sends stones out pays a third party per stone. When you see two stones of the same size and grade priced differently, ask whether one was treated in-house and one was not, and ask whether the treatment is disclosed on the certificate. Reliable buyers always route treated and untreated stones to different SKUs because downstream appraisers will notice.
Size bands: which method wins at one, two and three carats
The cost comparison changes with target size, and this is where most buyers make the wrong default assumption. Below one carat, especially in 0.30 to 0.90 carat melee, CVD plates cut efficiently and the per-carat cost favors CVD; HPHT melee carries a higher press-cycle overhead that is hard to amortize over small stones. Between one and two carats, the two methods converge, and price usually comes down to clarity and color rather than growth method. Above two carats, and especially above three carats, the larger HPHT cube begins to cut out with less weight loss, while a CVD plate of comparable weight requires a very thick run that pushes growth yield down and defect density up.
For very large polished stones -- five carats and above -- HPHT has historically held the cost advantage, although thick-pad CVD technology has narrowed that gap. If your program is mostly solitaire centers at 1.5 to 2.5 carats, method matters less than grade; if your program includes a meaningful share of three-carat-plus stones, it is worth asking your supplier what share of those large centers came from each line. Fancy shapes shift the math again: ovals, pears and cushions cut more efficiently from plates with large, flat faces, while rounds and princess cuts benefit from the cube geometry of HPHT rough. A buyer running a mixed shape catalog should expect to see a mixed-method supplier rather than a single-method one.
Defect rates, scrap and the hidden cost of rejects
The most under-discussed cost in yield economics is the reject bucket. A growth run that looks successful on paper can still produce 15 to 30 percent of rough that ends up as industrial grit rather than gem material: plates that delaminated, crystals that cleaved during extraction, stones that landed at a clarity no retailer will stock, or stones whose color is too brown to sell without treatment that the factory decided not to run. Those lost carats have to be amortized across the salable carats, which is why two factories with identical listed rough prices can have very different polished prices.
A factory with good in-line QC -- microwave plasma monitoring for CVD, press pressure logging for HPHT, and rough mapping before cutting -- catches bad crystal early, before cutting labor has been spent on it. A factory without that infrastructure pays twice: once to cut a doomed stone and once to replace it. When you negotiate, ask for reject rate history by lot. A supplier that will show you that number is usually a supplier you can build a private program with.
Reject rates also move with lot age. A growth run pulled from a chamber that has been operating for hundreds of hours tends to be cleaner than a run pulled right after chamber maintenance, because the chamber wall coating and gas lines are still stabilizing. Experienced factories hold the first few runs after maintenance for internal grading rather than selling them as premium lots. If you are paying premium prices, ask whether the lot came from a stable chamber window or a fresh-maintenance window -- the answer explains a lot about why two supposedly identical stones grade differently.
Worked example: turning rough price into polished price
Take a rough crystal priced at $200 per rough carat. If cutting yield is 45 percent, one polished carat costs about $444 in rough alone. Add cutting and polishing labor, grading, treatment, packaging and certification, and a factory might land at a polished cost of $650 to $750 per carat before margin. If the same rough had a 55 percent cutting yield, the rough component drops to about $364 per polished carat, and the factory can either price more competitively or hold a wider margin. This is why a five-point difference in cutting yield is worth more than a ten-dollar difference in rough price.
Now run the same math for a crystal that needs treatment. If annealing and irradiation add $40 per polished carat, and the stone comes out one color grade lower than expected, the effective cost per polished salable carat jumps again. Buyers who only compare listed rough prices are usually comparing the wrong number; the correct comparison is the all-in cost per graded salable carat. For more on how factories think about this, see the breakdown of inclusion quality control and rough mapping.
Power geography and the hidden cost subsidy
Because both methods run on electricity, the location of the growth plant quietly sets a floor on per-carat cost. A CVD chamber drawing 25 kilowatts for a 120-hour run burns through 3,000 kilowatt-hours of power per cycle, and an HPHT press drawing 50 kilowatts through a 90-hour cycle burns through 4,500. At industrial tariff rates that difference is small; at subsidized tariffs it is large. Factories that sit in regions with lower industrial power rates can grow the same stone for materially less money, and that advantage flows straight into polished pricing. When you compare two suppliers with similar equipment but very different prices, the difference is often geography rather than technology.
Shenzhen-based production has other offsets: proximity to cutting, polishing and certification houses means rough does not travel thousands of kilometers between stages, which reduces handling damage, storage and lead time. A stone that moves from growth to graded polished in three days avoids the cracking, chipping and lost-lot problems that come with long logistics chains. For a buyer, the cost of a two-week delay is not just cash flow; it is the risk that a matched set goes out of match while stones sit in a warehouse.
How to audit a supplier's yield claims
Most suppliers will tell you their rough is high yield. Very few will show you the data. When you are sourcing seriously, ask for three specific items before committing to a large lot:
- A lot-level breakdown of rough input weight versus polished output weight, not just a final polished carat count.
- The share of polished carats that landed in your target color and clarity band versus the share that was downgraded or re-sold.
- Treatment history for the lot: annealed, irradiated, or untreated, and whether that treatment is disclosed on the certificate.
A supplier that produces these numbers quickly is running a modern factory. A supplier that hedges with generalities is probably amortizing a high reject rate across its best lots, and you will eventually feel that in your own margin. This is also why building a program with one core factory -- rather than hopping between brokers -- pays off: you get to see yield history over time, and the factory has an incentive to keep its reject rate low for you.
Pricing traps buyers fall into
The first trap is equating growth method with quality. A well-run CVD line outperforms a sloppy HPHT line on every metric, and vice versa. The label on the certificate tells you method; it does not tell you how well the press was operated. The second trap is comparing rough prices without adjusting for cutting yield. A cheaper rough that cuts out at 38 percent is more expensive per polished carat than a slightly pricier rough that cuts out at 52 percent. The third trap is ignoring treatment cost. A stone that needs irradiation is not free to finish, and if the treatment is not disclosed, your downstream appraiser will eventually ask why.
The fourth trap is over-focusing on color grade across a single method. A G-colored CVD stone and a G-colored HPHT stone look the same in a well-lit showcase, and the retail customer will not identify method. What they will identify is cut quality, face-up size and price. So margin competition is usually won on cutting yield and treatment cost, not on arguing that one method is more "authentic" than the other. For context on how these cost lines stack up across the year, the factory price per carat page tracks the same numbers from the sell side.
How HOLYCOME allocates rough between CVD and HPHT lines
At our Shenzhen facility we run both methods instead of forcing every program onto one line, because different SKUs win on different lines. Melee and sub-carat centers are mostly cut from CVD plates where plate thickness and growth uptime drive cost. Two-carat solitaire centers are split based on the specific lot: clean plates go to the CVD line, thicker high-clarity cubes go to the HPHT line. Three-carat-plus fancy shapes and fancy colors lean HPHT where the cube geometry and natural tint produce better value. Every stone is mapped before cutting, graded in-house, and then sent to an independent lab for the final certificate before it ships.
We also keep treated and untreated lots in separate SKUs, with the certificate disclosure matching the SKU. If you are building a wholesale program and want to see how a specific size band performs on each line, our team at service@holycome.com can send a lot-level breakdown including cutting yield, reject rate and treatment history for the past quarter. More background on how the two methods compare for end buyers is in our guide to CVD vs HPHT lab grown diamond quality, and the growth process itself is explained in how lab grown diamonds are made.
Buyers sometimes ask whether we sell only one method because that makes marketing simpler. The honest answer is that forcing every SKU onto one line costs you money: the wrong method for a given size band costs 5 to 15 percent more per polished carat than the right method, and that gap compounds across a year of restocking. Our recommendation is to let the lot data decide method allocation, then present the stone honestly to your customer rather than leading with growth-method branding.
The bottom line for buyers
CVD vs HPHT is not a quality contest; both methods produce excellent near-colorless white goods when operated well. It is a cost contest with three moving parts: growth run efficiency, cutting yield, and post-treatment spend. CVD tends to win on growth cost and melee, HPHT tends to win on large-stone cutting yield and fancy colors, and the all-in polished cost gap between them is usually narrower than retail marketing implies. When you compare suppliers, ask for yield numbers, reject rates and treatment disclosures rather than accepting a simple method label -- those numbers are where real margin hides.
For ongoing wholesale programs, the practical move is to stop buying "CVD" or "HPHT" as abstract categories and start buying specific lots with documented yield history. A lot that returns a consistent 48 to 52 percent cutting yield with a 10 percent downgrade rate is worth a higher per-carat price than a lottery rough price list, because your own downstream margin becomes predictable. That predictability is what lets you stock solitaire centers, match pairs for studs, and price rings with confidence instead of guessing at replacement cost each quarter.