The same name, a very different machine
Two laptops sit next to each other on a retail page. Both say “NVIDIA GeForce RTX 4060”. One costs ₹94,990, the other ₹1,34,990. The obvious conclusion is that you are paying for the screen, the build, and the badge.
Sometimes you are. But often you are paying for something the listing does not mention at all: how much electrical power the graphics chip is permitted to draw.
NVIDIA does not sell laptop GPUs as fixed products. It sells a chip and a permitted power range, and the laptop manufacturer chooses where in that range to operate. For the RTX 4060 Laptop GPU the published range is 35 W to 115 W of total graphics power, with Dynamic Boost adding up to a further 25 W when the CPU is idle enough to spare it.
That is a span from 35 W to roughly 140 W. Four times the power budget, one product name.
Two machines from our catalogue
| Laptop | Sustained board power | Price |
|---|---|---|
| Lenovo LOQ 15 (i7-13650HX, RTX 4060) | 115 W GPU board power | ₹94,990 |
| Acer Predator Helios Neo 16 (i7-14650HX, RTX 4060) | 140 W GPU board power | ₹1,34,990 |
The Lenovo is configured at the top of the standard range. The Acer is configured at the top of the range plus Dynamic Boost headroom, and has the chassis volume and heatsink mass to hold it there.
Both are honestly described as RTX 4060 laptops. Neither manufacturer is doing anything improper. But the two machines will not produce the same frame rate in the same game, and nothing on a typical retail page tells you that.
Why sustained matters more than peak
Almost every laptop will hit an impressive power figure for the first sixty to ninety seconds of a load. The heatsink starts cold and absorbs heat faster than the chip produces it. Benchmarks that run for one minute capture this state, which is why short benchmark runs flatter thin laptops.
Then the heatsink saturates. From that point the machine can only dissipate as much heat as its cooling system moves to the air, and clock speeds fall until power draw matches that limit. This is the sustained state, and it is where the machine spends the whole of a two-hour gaming session, a long video export or an overnight render.
A laptop that peaks at 140 W and sustains 95 W is a 95 W laptop for any purpose you care about. This is why our specification tables record sustained figures rather than the boost numbers manufacturers prefer to quote.
What the extra power buys
The relationship between power and performance is not linear — it flattens as power rises, because higher clocks need disproportionately higher voltage. Going from 35 W to 80 W transforms a machine. Going from 115 W to 140 W is a real but smaller gain, in the region of ten to twenty per cent depending on the workload.
The gap widens the longer the load runs, because the better-cooled machine also holds its clocks steadier once the room warms up and the fans have been running for an hour.
Why retail listings never show this
Three reasons, none of them conspiratorial.
It is not a marketing-friendly number. “115 W” means nothing to most buyers, and a lower number looks worse next to a competitor even when the laptop is better in every other respect.
Marketplace templates have no field for it. Listing schemas carry processor, RAM, storage and GPU name. There is no board-power field, so even a manufacturer that wants to publish it has nowhere to put it.
NVIDIA stopped requiring it. Earlier generations used Max-Q branding to distinguish lower-power configurations. That naming was retired, and the disclosure requirement went with it.
How to find the number before you buy
- Read the manufacturer’s own specification sheet, not the retail listing. Look for “Total Graphics Power”, “Maximum Graphics Power” or “TGP”. Gaming sub-brands publish it more often than mainstream lines.
- Check reviews that run sustained loops. A reviewer who runs a benchmark for thirty minutes and reports the stabilised figure is measuring the thing that matters.
- Use thickness and weight as a proxy. At a given GPU tier, a heavier machine almost always carries a higher power budget. It is a crude signal but rarely a wrong one.
- Open NVIDIA Control Panel on a display unit. System Information reports the configured maximum graphics power on the machine in front of you.
The same problem, other components
Graphics power is the clearest example, but the pattern repeats. Processor names cover wide power bands too: the ₹46,990 Dell Inspiron 15 and the ₹68,990 ThinkPad E14 in our catalogue are both configured at 15 W sustained, while the ₹54,990 ASUS Vivobook 16 runs 45 W. All three are marketed as mainstream laptops with current-generation processors.
Storage behaves the same way. Two drives both described as “PCIe Gen4 NVMe” can differ several-fold in sustained write speed once their cache is exhausted, which is why our storage entries record sustained rather than burst figures.
The processor is competing for the same budget
GPU board power is only part of the picture, because in most gaming laptops the processor and graphics chip share one cooling system and often one power ceiling.
Manufacturers implement this as a combined budget. When a game loads the graphics chip heavily and the processor lightly, the system shifts power towards the GPU — this is what Dynamic Boost does. When a workload loads both, as many simulation and strategy games do, the two compete and each gets less than its headline figure suggests.
This is why the HP Victus 15 in our catalogue is recorded as “80 W combined CPU and GPU” rather than as two separate numbers. Quoting them separately would imply the machine can deliver both at once, which it cannot. Where a manufacturer publishes a combined figure, that is the honest number, and it is the one we record.
The practical consequence: a laptop with a high GPU board power and a weak cooling system may still disappoint, because the processor is being starved to feed the graphics chip. Benchmarks that load only the GPU will not reveal this. A game that loads both will.
What throttling actually feels like
Thermal throttling is rarely experienced as a number. It is experienced as inconsistency.
Average frame rate can look respectable while the experience is poor, because the average hides the pattern: strong performance for the first few minutes, then a drop, then oscillation as the machine repeatedly touches its thermal limit, backs off, cools slightly and pushes again. That oscillation is felt as stutter even when the average frame rate would suggest smooth play.
A machine with genuine thermal headroom behaves differently. It settles at a slightly lower clock and holds it, producing a lower peak but a flat, consistent line. Most people would rather have the flat line, which is why sustained figures and frame-time consistency matter more than the peak numbers in a specification table.
Two symptoms are worth knowing. If performance is fine for two minutes and then drops sharply, that is a cooling limit. If performance is lower from the very first second, that is a configured power limit — and no amount of cleaning or repasting will change it.
If you already own a lower-power laptop
Board power is set in firmware and cannot be raised. But several things recover performance that thermal conditions have taken away.
- Use the performance profile. Most gaming laptops ship in a balanced mode that caps power well below the hardware limit. The manufacturer’s own control application usually has a performance or turbo setting, and switching to it is the single largest free gain available.
- Stay plugged in. On battery, nearly every laptop restricts the graphics chip severely — often to a fraction of its rated power. A machine that feels slow unplugged may be entirely healthy.
- Clear the intakes. Vents are usually on the underside. On a bed, sofa or lap the machine is breathing its own exhaust. A hard surface, or a stand giving a few centimetres of clearance, is worth more than most cooling pads.
- Clean the fans annually. Dust in the fin stack is the most common cause of a laptop that has become progressively slower over two or three years, and it is reversible.
- Consider undervolting. Reducing voltage at a given clock lowers heat output, letting the machine hold higher clocks within the same thermal budget. It is well supported on most platforms but requires care and testing for stability.
None of these raise the ceiling. They stop you living below it, which for a machine a few years old is usually where the lost performance has gone.
The short version
A GPU name is a chip, not a performance level. Before you pay a premium for what looks like the same graphics card, find the board power. If the manufacturer will not publish it and no reviewer has measured it, treat the silence as information: it is very rarely the high-power configuration that goes unmentioned.