AI & ENTERPRISE
RTX 4090
Twice the memory, all the power.
The first dedicated RTX 4090 VRAM upgrade service in the United Kingdom.
From £1,000.
That's the starting price, for a card with all 12 of its original memory chips healthy, and your quote sets the exact figure.
Blackstone Repair is the first dedicated provider of RTX 4090 24GB to 48GB VRAM upgrade services in the United Kingdom, to the best of our knowledge, last checked on 12 October 2026. This service is not endorsed by or affiliated with NVIDIA Corporation.
From £1,000 when the original 12 memory chips are healthy and can be transplanted. Final price depends on parts required and is always agreed before work begins.
Your original core, clocks, and power limit are unchanged, and memory capacity is the only specification that increases, though the card comes back on a new board with a blower cooler. Every card is tested before it ships and the results are recorded on the test report we send with it. Full detail in our upgrade terms.
AI models, 8K timelines, and production renders all run out of memory before they run out of compute.
The RTX 4090 has the speed for all of it, but NVIDIA gave it only 24GB.
OFF THE CLOUD
AI that runs in the room with you.
A 48GB card turns a desktop into a serious AI workstation, with the models running on the machine in front of you.
The bigger the model, the smarter it is, and the more memory it needs.
A private assistant that reads everything you give it.
Legal contracts and case files, patient records, unpublished university research, client code under NDA. It answers from those documents without a single word leaving the machine.
Image and video generation with no meter running.
Flux at full precision and local video generation both outgrow a stock 24GB card. On 48GB they run without offload, overnight if the job calls for it, and the only cost is the electricity.
A coding assistant with the whole codebase in view.
48GB holds a capable model and enough context to keep a whole repository in view at once, not just the file you have open.
70B-class models, the tier where answers get good.
At Q4 quantisation a 70B model needs around 43GB. That tier never fits in 24GB, which is the reason this upgrade exists.
Why local beats the cloud for this
Private by physics
Nothing is uploaded, logged, or retained by a provider. Data cannot leak from a server it never reached.
A single cost
No per-token billing, no monthly tier, no invoice spike after a heavy week. Buy the memory once, run it flat out.
No terms attached
No rate limits, no model deprecations, no usage policies shifting underneath the work. The models you choose, configured your way, available offline.
Built to multiply
These are 2-slot blower cards, made for dense systems, so four of them in one workstation give 192GB of VRAM.
If everything you run fits in 24GB, you do not need this. The fit checker below tells you honestly.
What happens to your card
- 1
Register your interest
Tell us about the card and we reply ourselves, and a build then starts with a written quote, split into parts and our work, with the dates and the full terms.
- 2
Accept the quote
Paying the deposit accepts the quote and pays for the parts, and we then tell you when to send the card, which travels tracked and signed-for by Royal Mail with the cover figure shown in pounds before you choose a service, or comes by drop-off in Milton Keynes by appointment.
- 3
Teardown and validation
We photograph everything on arrival and check the core first, and if it cannot take the upgrade the deposit comes straight back.
- 4
The transplant
Hot air on a controlled preheat curve lifts your core and memory from the old board, with thermocouples tracking every stage. Cleaned and reballed, they move to the new clamshell board alongside the additional modules that double the capacity. The card that comes back has a 2-slot blower cooler, so a different shape and a different sound, and it exhausts out the back of the case instead of into it. The rebuilt card then goes through the full test protocol below.
TEST PROTOCOL
Tested before it ships
Test protocol v1.0, June 2026
- 5.1
Driver recognition
Standard NVIDIA drivers, no special configuration. The system sees 48GB or the card does not ship.
nvidia-smi reports 48GB at full PCIe link width - 5.2
Memory validation
Chip-level fault isolation across every populated memory channel, run at stock and again above stock. Zero errors tolerated at either setting.
every memory channel validated, zero errors tolerated - 5.3
Sustained burn-in
Sustained AI generation workloads at full power draw, with rest and cool-down periods, across a 48-hour test window.
AI workloads across a 48-hour test window - 5.4
Thermal mapping
Front and rear module temperatures measured at equilibrium. The clamshell design puts 12 modules on the rear, so we check them all.
memory junction temperatures within spec under sustained load - 5.5
The report
Every card ships with its own test report. It doubles as the warranty baseline: if the card ever comes back, that report is what we measure against.
signed, dated, ships in the box
Balance and tracked return
The balance, our work, is due only when your card has passed everything. It returns tracked and signed-for, with the cover figure shown in pounds before dispatch, in plain unmarked packaging, and we share the test results we record at completion.
Every report carries the date, the test versions, and the signature of the engineer who ran it.
As builds complete, their test reports, thermal captures and bench data publish here, card by card.
Same core, same silicon, twice the memory.
Will it fit and is it worth it
| Model | Quant | Needs | 16GB | 24GB | 32GB | 48GB |
|---|---|---|---|---|---|---|
| 13B classA capable assistant at full precision · 32GB and up | FP16 | ~26GB | does not fit | does not fit | fits | fits |
| 30B classStrong reasoning and document analysis · tight at long context on 24GB | Q4 | ~18GB | does not fit | fits | fits | fits |
| 30B classhigher precision, needs 48GB | Q8 | ~32GB | does not fit | does not fit | does not fit | fits |
| 70B classNear-frontier quality, the best on one card | Q4 | ~43GB | does not fit | does not fit | does not fit | fits |
| 70B classneeds multi-card | Q8 | ~75GB | does not fit | does not fit | does not fit | does not fit |
13B class
A capable assistant at full precision · 32GB and up
Quant
FP16
Needs
~26GB
30B class
Strong reasoning and document analysis · tight at long context on 24GB
Quant
Q4
Needs
~18GB
30B class
higher precision, needs 48GB
Quant
Q8
Needs
~32GB
70B class
Near-frontier quality, the best on one card
Quant
Q4
Needs
~43GB
70B class
needs multi-card
Quant
Q8
Needs
~75GB
70B-class at Q4_K_M quantisation, 8K context, needs about 43GB: fits with headroom on 48GB, does not fit on 24GB. Figures cross-checked against three published 2026 VRAM guides, as of June 2026.
Cloud, per year
£590
Upgrade, from
£1,000
Payback
20.3 months
If you only need occasional bursts, rent. If the card works most days, owning wins.
Cloud rate as of October 2026.
This assumes the card is already owned.
All roads lead to 48GB
The 48GB upgrade
Stock RTX 4090
RTX 6000 Ada 48GB
Cloud rental, 48GB class
DIY kit route
VRAM
48GB
VRAM
24GB
VRAM
48GB
VRAM
48GB, rented
VRAM
48GB, if the build works
Bandwidth
~900 GB/s
Bandwidth
~900 GB/s
Bandwidth
960 GB/s
Bandwidth
960 GB/s class4
Bandwidth
~900 GB/s, same design
Cost
From £1,0001
Cost
~£2,250 used market3
Cost
about £7,500 new6
Cost
~£590 a year at 4h/day4
Cost
board kit and modules, plus £2,000+ rework gear5
Warranty
90 days, modules included
Warranty
out of warranty, returns vary by seller
Warranty
manufacturer terms
Warranty
not applicable, rented
Warranty
none5
Note
send a card in1
Note
70B-class will not load2
Note
ECC, pro drivers
Note
data leaves the building
Note
slow, demanding work, one slip can end the core
The 48GB upgrade
VRAM
48GB
Bandwidth
~900 GB/s
Cost
From £1,0001
Warranty
90 days, modules included
Note
send a card in1
Stock RTX 4090
VRAM
24GB
Bandwidth
~900 GB/s
Cost
~£2,250 used market3
Warranty
out of warranty, returns vary by seller
Note
70B-class will not load2
RTX 6000 Ada 48GB
VRAM
48GB
Bandwidth
960 GB/s
Cost
about £7,500 new6
Warranty
manufacturer terms
Note
ECC, pro drivers
Cloud rental, 48GB class
VRAM
48GB, rented
Bandwidth
960 GB/s class4
Cost
~£590 a year at 4h/day4
Warranty
not applicable, rented
Note
data leaves the building
DIY kit route
VRAM
48GB, if the build works
Bandwidth
~900 GB/s, same design
Cost
board kit and modules, plus £2,000+ rework gear5
Warranty
none5
Note
slow, demanding work, one slip can end the core
- From £1,000 on a send-in RTX 4090 with all 12 original memory chips healthy, and the quote is always agreed before work begins.
- See the fit checker above for what fits where.
- eBay UK sold listings of working cards from 20 August to 4 October 2026, median of 63 sales.
- RunPod RTX 6000 Ada hourly rate, about £0.59 an hour, October 2026.
- A DIY build needs the board, twelve matched modules, BGA rework equipment and the skills to use them, and nothing covers it if it goes wrong.
- NVIDIA RTX 6000 Ada Generation 48GB: £7,499.99 new (PriceSpy UK best price), October 2026. The RTX 6000 Ada is a professional workstation card with ECC memory and certified drivers, differing from this upgrade in other respects.
Published third-party figures, as of October 2026. Our own measured numbers replace these as our builds complete.
Not for gaming. Games do not use 48GB.
Not for workloads that already fit in 24GB. The extra memory would sit idle.
Built in the UK
No borders, no customs
A card posted from the UK never leaves the country. No import VAT on the service, no duty, no seizure risk.
UK consumer law applies
The Consumer Rights Act 2015 covers this work. Your statutory rights survive everything in our terms.
Covered both ways
Shipping cover is shown in pounds before you choose a service, with the cover figure for the return leg shown the same way before dispatch.
You talk to the engineer
The person who quotes your card is the person who builds it and signs its report.
Sending a card abroad means customs both ways, with possible import VAT and duty to sort out, while a card posted from the UK stays inside the UK on a tracked Royal Mail parcel.
Sending a card abroad means customs both ways, with possible import VAT and duty to sort out, while a card posted from the UK stays inside the UK on a tracked Royal Mail parcel.
PRICING
What the price includes
From £1,000.
That's the starting price, for a card with all 12 of its original memory chips healthy, and your quote sets the exact figure.
About £600 of the price is a deposit that pays for the parts and getting them to us, at October 2026 prices, and the other £400 or so is for the work itself, paid once the card has passed testing.
The deposit comes back in full if the card can't take the upgrade, if we can't get the parts, or if you pull out before we start the transplant, and it doesn't once the transplant has started, because the parts are fitted to the card by then.
- Teardown, core and memory validation, photographed on arrival
- New 48GB clamshell board and 2-slot blower cooler
- 12 additional Micron D8BZC modules fitted alongside the original 12
- Firmware configured for the doubled memory
- The full test protocol and a signed report in the box
- 90-day warranty covering our work and the modules we fit
We keep the original board and cooler, which is part of how the price works.
Postage both ways is set out on the quote, with the cover figure in pounds.
A failed original chip is replaced at £40 and added to the quote with your agreement, and any other small part the card needs is agreed with you before we fit it.
Memory prices keep changing, so these figures are as of October 2026, while the quote carries the exact ones and stays fixed once it's accepted, apart from a failed chip or small part that we agree with you first.
Because we are new, reputation matters more to us right now than margin, so this price is lower than it will be once our work speaks for itself.
Card faulty? A fault on the board or in the memory is often fine, since the board is replaced and we can fit new chips, but a fault in the core means the upgrade can't go ahead, and we only know once we've checked the card.
Warranty at a glance
Returned tracked and signed-for, in plain unmarked packaging.
Register your interest
We reply to each one ourselves, and nothing is booked or charged by registering.
Upcoming: the RTX 4080 SUPER 16GB to 32GB upgrade.
See the 4080 SUPER upgradeQuestions, answered properly
No. The card runs on standard NVIDIA drivers and reports 48GB out of the box.
Not on compute or bandwidth, which are unchanged, though the blower cooler on the new board gets loud at full load, so it's worth knowing going in, and the report that goes back with the card shows its own measured numbers, since real-world results vary by workload.
The 48-hour burn-in exists to answer this. For business-critical deployments we recommend a spare card, as we would with any hardware.
We do not have one yet, honestly, because we have just started. Nor will we invent a number before the builds add up. What stands behind every card instead: every memory channel validated, the 48-hour test window before shipping, and the 90-day warranty. Real figures publish here as they accumulate.
Often, because a fault on the board doesn't matter when the board is replaced and a dead memory chip is replaced at £40, but a fault in the core means the upgrade can't go ahead, and then the deposit comes back in full.
Yes, any board-level shop with the right tooling can. We are the best placed though, since we build the card and keep its baseline test report.
You can. The comparison above prices that route too. It needs the board, twelve matched modules, BGA rework equipment, and the skills to use them, with no warranty if it goes wrong. If you have all four, you do not need us.
No. Games rarely use anywhere near this much VRAM. The upgrade is for memory-bound AI and compute work.
Yes, it is your card. Our terms ask one thing: describe it accurately as a modified card when you sell it. The test results we record help the sale, since the next owner gets the same record.
Cover figures are shown in pounds next to every shipping option before you choose a service. For a high-value return we agree the service and its cover figure with you before anything ships. Full details are in the shipping section of our terms.
For several cards, tell us how many on the form above and we quote each one separately, each with its own deposit, then build them one at a time, so the dates stretch with the number of cards.
If our checks show the card cannot take the upgrade, the deposit comes back in full and you choose between having the card back, with the postage at your cost, or donating it to us. If a build fails on our bench despite reasonable care on our part, we give back everything you have paid and return the card at our cost, but not its value, as clause 6 of the upgrade terms sets out.
Twice: a deposit when you accept the quote, which pays for the parts, then the balance for our work once the card has passed testing. The deposit comes back in full if the card can't take the upgrade, if we can't get the parts, or if you pull out before we start the transplant.
Heat, mainly: the transplant takes controlled heat cycles, with a small but never-zero risk to the core and the memory. We publish the honest picture, risk by risk, on its own page:
What can go wrong and what we do about itSame approach, 16GB to 32GB, from £850, and you can register your interest on its page, linked at the bottom.