Wolfspeed Q4 Preview (19 August 2026): The $30m a Quarter of Depreciation That Vanished — and Why a -27% Gross Margin Is Not a Demand Signal

Wolfspeed reports fiscal fourth-quarter results after the close on Wednesday 19 August, guided to revenue of $140-160 million with gross margin expected to stay negative. The headline will almost certainly be another loss, and the headline will almost certainly be the least informative thing in the release. Two mechanisms are running underneath it in opposite directions: a bankruptcy that mechanically removed roughly $30 million a quarter of depreciation from the cost line, and a demand base that is being swapped, painfully and incompletely, from electric vehicles to data centre power. The print is a progress report on that swap.

What is being reported, and what it is not comparable to

Wolfspeed confirmed on 5 August that it will report fiscal fourth-quarter and full-year 2026 results on Wednesday 19 August after the close, with a call at 5:00 p.m. Eastern.

One structural warning first, because it will otherwise corrupt every comparison in the coverage. Wolfspeed emerged from a prepackaged Chapter 11 on 29 September 2025, eliminating roughly $4.6 billion of debt. Accounting rules then required fresh-start treatment: every asset and liability remeasured to fair value on that date. Fiscal 2026 is therefore reported as a Predecessor stub — 30 June to 29 September 2025 — plus a Successor period running from 30 September 2025. The two are, in the company's own words, not comparable with what came before.

The arithmetic that is safe to do: revenue for the first three quarters of fiscal 2026 was $196.8 million (Predecessor) plus $318.7 million (Successor), or $515.5 million, against $560.6 million for the same three quarters of fiscal 2025. The guide implies a full year of roughly $655-676 million.

The two product lines are living in different economies

The single most useful table in Wolfspeed's third-quarter release is the product line split, and it is barely covered.

Materials — bare silicon carbide substrates sold to other device makers — has fallen more than a third year on year and has now been flat at almost exactly $50 million for two consecutive quarters. That is the shape of a product that has found a price floor rather than a demand recovery. Substrates are where capacity expansion, much of it Chinese, has been heaviest and where a wafer is closest to a commodity.

Power Products, the finished devices, fell only 6.9% year on year. The sharper sequential drop from $118.3 million was explained by the company itself when it guided: customers had pulled orders forward from the Durham fab ahead of its planned closure, some had dual-sourced during the bankruptcy, and electric vehicle demand was weaker. Two of those three are one-off.

The margin: an accounting gift and a utilisation problem

GAAP gross margin was -27% in the March quarter, against -46% in December. A nineteen-point improvement in a single quarter looks dramatic. Most of it is not operational.

Two things sat inside the December figure that were never going to repeat in full. The company disclosed $23 million of inventory fair-value step-up — a fresh-start artefact that inflated cost of revenue once as revalued inventory sold through — and $48 million of underutilisation costs. Strip only the step-up out of December and the margin was closer to -32%, so the genuine sequential improvement is around five points, not nineteen.

The larger and more durable change is depreciation. Fresh-start accounting cut property and equipment from $3,916.5 million at 29 June 2025 to $717.1 million at 29 March 2026 — roughly $3.2 billion of carrying value removed. Wolfspeed told investors in February that this, net of additional intangibles, would reduce ongoing depreciation and amortisation "by approximately $30 million per quarter, which will be fully realized as inventory turns."

That has begun to show. Quarterly depreciation and amortisation was $30.9 million in the March quarter against $53.9 million a year earlier. On $150 million of revenue, $30 million of removed depreciation is worth about twenty percentage points of gross margin — a margin improvement that arrives without a single additional wafer being sold.

What remains is the honest problem. Underutilisation costs are charges for capacity that exists, is staffed and is depreciating but is not producing. At $48 million against a gross loss of $40.0 million in the following quarter, the implication — inferred, not disclosed for the March quarter — is that the entire gross loss is the cost of empty fabs, and that the products themselves contribute positively. If that is right, the path to breakeven is volume through lines that already exist, which is a very different problem from a product that cannot be made profitably.

Where the new volume is supposed to come from: 800 volts

The demand story Wolfspeed is now selling is not electric vehicles. It is the rewiring of the AI data centre, and the mechanism is specific enough to check.

Nvidia has been pushing the industry from traditional 54-volt in-rack distribution to 800-volt direct current, because the physics of the older approach stops working at megawatt scale. Nvidia's own engineering write-up puts it starkly: delivering one megawatt to a single rack at 54 volts would need up to 200 kilograms of copper busbar, and the rack busbars in a one-gigawatt site could require up to 200,000 kilograms of copper. Moving to 800 VDC is projected to cut copper use by about 45% and lift end-to-end efficiency by roughly 5%, with the architecture aimed at deployments from 2027 and more than eighty ecosystem companies building to the specification through the Open Compute Project.

Higher voltage is exactly where silicon carbide beats silicon. Switching losses in a converter rise with voltage, and wide-bandgap devices hold their efficiency where silicon MOSFETs and IGBTs start to give it away — which is why the 800 VDC transition is a content opportunity for a company that makes little else.

On 6 August, Wolfspeed and LITEON announced that LITEON had qualified Wolfspeed silicon carbide for its 800 VDC sidecar power and compute rack power supply platforms, drawing on Wolfspeed's 200mm manufacturing platform. LITEON's general manager for cloud infrastructure, John Chang, framed it as strengthening the company's "ability to meet the performance, efficiency, and reliability requirements of next-generation AI infrastructure."

Qualification is genuine. It is also not revenue, and Wolfspeed's own risk factors flag the distance between design-ins, design-wins and volume shipment as an open question. Which is why the number to watch is the sequence rather than the announcement: AI data centre revenue grew about 50% sequentially in the December quarter and about 30% in the March quarter, both from a base the company describes only as "modest".

The balance sheet buys the time

None of the above matters if the company runs out of runway before the swap completes, so the cash lines are the third thing to read.

Net capital spending of $5.0 million in a quarter is the number that captures the change of posture most completely. This is a company that spent years building Mohawk Valley and Siler City and has now stopped almost entirely — capital expenditure was already down more than 90% year on year in the December quarter, with spending limited to previously committed investments. Long-term debt stands at $922.2 million with $798.3 million of convertible notes, against $1.16 billion of cash, and in the March quarter the company refinanced about $476 million of first-lien debt, cutting total debt by $97 million and estimated annual interest expense by $62 million.

Free cash burn near $90 million a quarter against $1.16 billion of liquidity is roughly three years of runway at the current rate, before any improvement. That is the resource the transition is being funded from.

What to actually watch on Wednesday

Three things, in this order — and none of them is the loss per share.

A revenue print inside $140-160 million tells you almost nothing on its own; the range was set three months ago and is wide enough to absorb a lot. A gross margin materially better than -27% with utilisation costs unchanged would mean the depreciation relief has fully washed through inventory, which is a one-time step, not a trend. A gross margin better than -27% and higher revenue would mean utilisation itself is improving, which is the only version that compounds.

The broader read-across is to the same theme covered in the repricing of mature-node power chips — AI infrastructure generating demand at the unglamorous end of the semiconductor stack — and to Coherent's fourth quarter, where the same build-out showed up as a boom in one segment and a contraction in the other inside a single company. Wolfspeed is the same story caught earlier, with a weaker balance sheet and a harder legacy business to escape.

Wednesday is also the day the July FOMC minutes are published, which is where the macro attention will sit. For the dollar and the other seven majors, that is the event that matters; a silicon carbide maker's utilisation rate is not one of the five factors that the meter scores. Keeping those two categories apart is most of the discipline.

Educational macro context only — not investment advice.