Transient thermal model of a MacBook Air 13.6″ M3 on the corrugated cooling wedge. Every case runs twice — on the stand and flat on the desk — so the gap between them is the product.
Click a part to isolate it. Click again to show all.
§1 — The numbers
30 °C room · 45 min sustained · wooden desk · 16 W (temperature) and 24 W (work)
| Desk | Bare | Stand | Saved | + Fans | Saved |
|---|---|---|---|---|---|
| Wood | 53.2 | 44.5 | 8.7 K | 36.6 | 16.7 K |
| Laminate | 52.1 | 43.8 | 8.4 K | 36.4 | 15.7 K |
| Glass | 48.9 | 40.8 | 8.1 K | 35.7 | 13.2 K |
| Steel | 48.5 | 34.3 | 14.2 K | 33.3 | 15.3 K |
Passively the stand is a conduction bridge, so it works best on a desk that can absorb heat. On steel it nearly matches the fan module.
§2 — Components
Nine parts, stand only — the fan module is separate. Three of them are sheet metal brazed into one stack, the way a car radiator core is built.
| Part | Spec | Mass | Why it exists |
|---|---|---|---|
| 1.0 mm 6061, black anodised | 89 g | The only surface touching the laptop. Carries most of the heat by conduction, then spreads it sideways into the folds. Thinning this saves weight but costs performance faster than any other part. | |
| 0.4 mm 3003, 14 mm pitch, 80 channels | 215 g | One part doing two jobs: 4,627 cm² of fin area, and the walls of the air channels. The zigzag is the heat exchanger — this is what makes the fan module worth attaching. | |
| 1.0 mm 6061 | 89 g | Closes the channels — without it the folds are open grooves, not ducts. Also presses flat against the desk, which is the main passive escape route for heat. | |
| aluminium extrusion | 35 g | Carry the laptop's weight so the thin core cannot crush, and set the wedge angle. Deliberately do not close the sides — air enters there as well as at the front. | |
| 30 mm aluminium | 20 g | The flat tall face the module mates against. Aligns the fan outlets with the channel openings and gives the magnets something rigid to pull on. | |
| N42 neodymium, Ø14 mm | 12 g | Tool-free docking that self-aligns. Their pull force also sets how well the module seals to the channels — too weak and air leaks around the joint instead of going through it. | |
| silicone, ≥2 W/mK, 0.5–1 mm | 9 g | The highest-risk part in the whole design. Fills the microscopic gaps between two nominally flat metal surfaces. Without it, an unclamped laptop resting under its own weight loses most of the contact benefit — and contact is where the passive cooling comes from. | |
| silicone, 4 off | 4 g | Stops the stand walking across the desk while you type. A stand that slides is a stand that loses contact. | |
| soft TPU, front lip | 2 g | The front lip sits directly against the chassis. Bare anodised aluminium would mark it within a week. |
Top skin, core and bottom skin are brazed as a single stack — CAB brazing, 3003 core between clad 6061 skins, exactly the radiator-core process. Rails and dock plate bond or rivet on afterwards, then the whole assembly is black anodised.
Brazing matters thermally, not just structurally: a glued joint would add a thermal resistance right where the heat has to cross.
Bare aluminium has an emissivity around 0.05; black anodising takes it to 0.88. That is a seventeen-fold change in the radiation path — meaningful at these temperatures, and it costs almost nothing.
Total: ~475 g. Dropping to 0.6 mm skins and a 0.3 mm core would reach ~350 g, at the cost of spreading performance in the top skin.
§3 — Target
Every Apple Silicon MacBook Air is fanless. Its entire thermal budget is the metal shell. That is the only reason this product has a job — and it is why a MacBook Pro is a bad target: fans already move the heat, so improving the shell barely registers.
| Room | Bare | + Stand | + Fans | Passive gain | Active gain |
|---|---|---|---|---|---|
| 18 °C | 16.5 W · 69% | 19.3 W · 80% | 22.6 W · 94% | +17% | +37% |
| 22 °C | 15.5 W · 64% | 18.5 W · 77% | 21.5 W · 89% | +19% | +39% |
| 26 °C | 14.8 W · 61% | 17.4 W · 73% | 20.5 W · 86% | +18% | +39% |
| 30 °C | 14.1 W · 59% | 16.6 W · 69% | 19.3 W · 81% | +18% | +38% |
| 34 °C | 13.2 W · 55% | 15.6 W · 65% | 18.3 W · 76% | +18% | +38% |
| 38 °C | 12.4 W · 52% | 14.7 W · 61% | 17.2 W · 72% | +18% | +39% |
The strongest single result in the study: the gain is flat at +18% / +38% across every room temperature. The product does not have a narrow operating window — it works the same in a cold office and a hot one.
| MacBook Air 13″ M2 / M3 / M4 | best fit |
| MacBook Air 15″ | good — more shell, runs cooler |
| iPad Pro + Magic Keyboard | fanless, unproven fit |
| MacBook Pro (any) | poor — already has fans |
The 13″ is the hottest case and the highest volume. Start there; a 15″ variant is a footprint change, not a redesign.
| Light load, 8 W — bare | 42.2 °C |
| Light load, 8 W — stand | 37.6 °C |
| Light load, 8 W — fans | 33.3 °C |
Yes — 4.6 K even at light load, which is the difference between a warm palm rest and a comfortable one. But the sustained-work numbers above are the real sales argument.
§4 — Market
| Funnel step | Assumption | Result |
|---|---|---|
| Active Macs worldwide | est. | ~100 M |
| … that are notebooks | ~80 % | ~80 M |
| … that are MacBook Air | ~55 % | ~44 M |
| … Apple Silicon, therefore fanless | ~65 % | ~29 M |
| … doing sustained heavy work | ~12 % | ~3.4 M |
| … aware of throttling & would pay | ~8 % | ~275 K |
| Year-1 reachable via niche channel | 5–15 % | 14 K – 41 K units |
| Retail price assumed | $89 |
| Gross margin assumed | 45 % |
| Conservative — 14 K units | $1.2 M rev · $0.55 M GP |
| Optimistic — 41 K units | $3.7 M rev · $1.65 M GP |
| Pool growth per year | ~9 M new Airs |
This is a real niche hardware business, not a unicorn. Crowdfunded-accessory scale: single-digit millions in year one, a growing installed base, and a defensible technical story.
The addressable pool grows by roughly 9 M fanless Airs a year, and Apple shows no sign of putting a fan in the Air. The thermal problem is structural, not a bug that gets patched.
| Comparable | Typical price | What it does |
|---|---|---|
| Plastic laptop riser | $20 – 35 | lifts, no thermal claim |
| Aluminium stand | $40 – 70 | lifts, some spreading |
| USB fan cooling pad | $30 – 60 | loud, tethered, generic |
| This product | $79 – 99 | measured watts, cordless, modular |
The premium has to be earned by a published, repeatable benchmark — “holds 19 W instead of 14 W” — not by the word aluminium.
§5 — Value & risk
Passively, channel velocity is 2.3 mm/s — under a watt of airflow. Almost all of the 8.7 K comes from metal touching the laptop and conducting heat away, not from moving air.
The corrugation only earns its complexity under forced air. Lead with the active product; passive is the silent fallback.
Zero front lift — wrists stay put. Contact, not clearance — conducts into metal instead of lifting into air. One core, two modes.
§6 — Roadmap
Rear fan module bolted to a plain flat aluminium plate. No corrugation, no tooling. Isolates the only two things that carry heat: contact conduction and forced airflow.
Cost: a few hundred. Removes most of the risk.
Repeatable load, thermocouples at three points on the bottom case, chip temperature and package power logged together. Wood and steel. Tool the corrugated core only if active mode validates.
Watch watts, not just degrees — a cooler machine gets faster, which reads as a null result if you only watch temperature.
Battery runtime, acoustic target, magnet strength and safety, the contact pad, industrial design. Then the 15″ variant.
The hard engineering here is mechanical and acoustic. The thermal question is already answered.
§7 — Method & limits
Reading the chart. The shaded band between the curves is the product. A curve flattening at 100 °C on the chip view means that machine has started throttling — holding temperature by doing less work.
What is solved. Two transient cases step together. Solids are RC nodes — chip, chassis, 20 slices of the stand. Air is marched slice by slice. Flow comes from a pressure balance: buoyancy vs viscous loss when passive, fan curve vs system curve when active. The desk is a semi-infinite solid under transient flux.
Anchor. One constant is fitted so a bare machine on wood reaches 46 °C case / 95 °C chip after 45 min at 16 W and 22 °C — what these machines measure at.
Limits. Not CFD. Temperature along the channel, not across the machine — no hot spots, no recirculation. Conductances ±30–40%. The comparisons are trustworthy; the absolute numbers inherit the anchor.
Status. Nothing here has been built or measured. Every figure is a hypothesis with a number attached.