Category: Intel

Auto Added by WPeMatico

Posted on by

Intel explains what went wrong with its new Arrow Lake desktop CPUs

Intel’s new Core Ultra 200-series desktop processors, codenamed Arrow Lake, launched to seriously underwhelming reviews back in October. Now the American firm has collated five issues that it says are to blame for the difference between its own performance expectations and what reviewers experienced, with four of the five problems described as being already resolved by BIOS, Windows or application updates, and further performance enhancements to arrive in January 2025.

For context, our Core i9 285K and Core i5 245K review saw performance wins for the 285K against the outgoing 14900K in three titles – Crysis 3 Remastered, Dragon’s Dogma 2 and Forza Horizon 5 – while the 14900K retained its crown in the other eight games we tested, often by extreme margins.

Here are the five issues, their root cause and current status, quoted from Intel’s release:

Performance Topic Root Cause Status
Unusual scheduling, high run-to-run variation, low single-threaded scores, intermittent ~1.5x increase in DRAM latency, performance lower on Windows 11 24H2 vs 23H2 Intel mistimed deployment of OS power plan settings (“PPM”), which customises DVFS, core parking and C-states. This caused a 6-30% performance loss Solved in Windows 11 26100.2161 or newer
Intel Application Performance Optimiser (APO) not demonstrating expected performance results Missing PPM places CPU into state where APO profiles cannot apply, select reviewer BIOSes additionally set APO to disabled by default. This caused a 2-14% performance loss on APO-profiled titles Solved in Windows 11 26100.2161 or newer
BSOD when launching Easy Anti-Cheat titles on Windows 11 24H2 Known issue with Easy Anti-Cheat KMD and Windows 11 24H2, issue exacerbated by disabling Virtualisation-Based Security (VBS) Solved with new Easy Anti-Cheat driver distributed by Epic
Select performance settings misconfigured in some pre-release BIOSes Consistency of VIP performance settings not sufficiently checked by Intel, including ReBAR, Intel APO, compute ring frequency, IMC gear, sustained/transient power limits. This caused a 2-14% performance loss Solved in customer BIOSes now available
New BIOS performance enhancements Fresh optimisations developed for upcoming BIOS updates, currently in validation by Intel and its partners, single digit performance enhancement estimated among a 35-game geomean Motherboard BIOSes planned for January 2025

The table makes for interesting reading, and Intel’s community blog on the subject goes into further detail on each identified issue too. In it, Intel promises “a complete performance digest, inclusive of the January BIOSes” at CES early next year, so we should get a better idea of exactly the performance differentials we should expect with all five fixes applied.

We also saw extremely poor performance in Cyberpunk 2077 in particular, and that seems to have been solved in game update 2.2 if the patch notes are to be believed.

If you are one of the vanishingly small number of people to be running a Core i5 245K, Core i7 265K or Core i9 285K system, then you’ve probably already updated to the latest Windows and BIOS versions, but it’s worth doing so now if you haven’t – and again in mid January when further BIOS updates are made available – to ensure that your system is delivering maximum performance. Intel says that the January BIOSes can be “identified with Intel microcode version 0x114 and Intel CSME Firmware Kit 19.0.0.1854v2.2 (or newer).”

We’ll look to retest the Core i9 285K and Core i5 245K once these January BIOS updates are available, as we did see much worse performance than we expected and it would be nice to see what these proccessors are truly capable of. We’re expecting to hear more from Intel at that time – as well as rivals AMD and Nvidia – so stay tuned for our reports from CES in the new year.

Posted on by

Intel announces Arc B580 and B570 “Battlemage” graphics cards with XeSS 2 frame generation

Intel has announced its first “Battlemage” discrete graphics cards, the Arc B580 and B570. As well as being the fastest Intel GPUs ever made, the $249 B580 and $219 B570 are also equipped with Intel’s second-generation XeSS upscaling with frame generation and latency reduction. That’s a promising recipe, especially if Intel is able to continue shoring up its driver support for newer games as we approach the release dates of December 13th (B580) and January 16th (B570).

Before we get into the specs and features, let’s briefly recap the Arc story thus far. Intel released its first Arc discrete graphics cards in 2022, beginning with the Arc A380 and following up with the more powerful A750 and A770. The A380 was more of a proving ground, a novelty that offered excellent media capabilities but little raw grunt, while the A750 and A770 continue to offer reasonable value at roughly RTX 3060 levels of performance.

The new B580 and B570 are aiming higher, with a mandate to perform well at a 1440p resolution and ultra settings. This makes sense, given that high refresh rate 1440p displays have now hit mainstream prices – eg you can pick up a 27-inch 1440p 165Hz monitors for less than £150, making it the new baseline.

intel arc b580 box
Image credit: Intel/Digital Foundry

To see this content please enable targeting cookies.

According to Intel’s numbers, the B580 Limited Edition (equivalent to an AMD reference or Nvidia Founders Edition card) is on average 24 percent faster than the old A750 at 1440p and ultra settings. That includes sub-60fps averages in games like Dragon’s Dogma 2, The Last of Us Part 1, A Plague Tale: Requiem and Total War Three Kingdoms; circa 100fps averages in games like Returnal, Cyberpunk 2077, The Witcher 3 – all with XeSS; and circa 165fps averages in Fortnite, F1 24, Diablo 4 and Doom Eternal – all with XeSS once again.

Intel also promises “best-in-class performance per dollar” versus Nvidia’s GeForce RTX 4060, with 1.32x average raster performance across 47 games, and 1.25x average RT performance across nine games. AMD’s entry-level Radeon RX 7600, meanwhile, is shown as achieving 1.06x and 0.88x performance versus the RTX 4060 across the same suite of games.

Internally, the BMG-G21 GPU and its Xe2 architecture delivers 70 percent extra performance per Xe-core and 50 percent extra performance per watt, mirroring improvements that Intel has made with its APUs as of late. These architectural improvements are quite wide-ranging, with 1.2x to 12.5x speed-ups to operations including compute dispatch, draw, pixel blend rate, mesh shader dispatch, vertex index cut, vertex processing, tessellation, ray triangle, ray trace and sampler feedback operations.

Arc B580 Arc B570
Xe Cores 20 18
Render Slices 5 5
RT Units 20 18
XMX AI Engines 160 144
Graphics Clock 2670MHz 2500MHz
Memory 12GB 10GB
Memory Interface 192 bit 160 bit
Memory Bandwidth 456GB/s 380GB/s
Peak TOPs 233 203
Total Board Power 190W 150W

XeSS is key to the equation here, with Intel stating that 150+ games now support the technology, and average performance improvements range from 22 percent to 80 percent at 1440p and ultra settings – with a 47 percent improvement being the average. XeSS is also now compatible with the DX11 and Vulkan APIs, rather than just DX12.

As well as the existing super sampling tech, the new XeSS2 includes frame generation (XeSS-FG) and a latency mitigation solution (XeSS-LL) – a necessary component for frame-gen, but something that’s also nice to have for more competitive games like first-person shooters. Demonstrating the features working together in F1 24, the B580 shoots from a 48fps baseline read-out to 93fps with XeSS balanced super sampling then 152fps when XeSS 2 frame generation is added to the mix.

The latency figures are also interesting, with 57ms of latency using native rendering, 32ms with XeLL, 19ms with XeSS + XeLL, and 28ms with XeSS, XeLL and XeFG.

intel arc b580 gpu
Image credit: Intel/Digital Foundry

Both XeSS frame generation and XeSS latency reduction require developer involvement, so adoption will take some time – but at least we won’t see a similar situation to Valve banning Anti-Lag+ users in Counter-Strike 2, as AMD’s universal latency reduction tech accessed game data without permission.

Ultimately, it’s a promising setup, but with Nvidia’s 50-series cards in the offing, Intel will need to continue to make big strides with their Arc graphics tech to stay competitive.

The company does seem to have kept its board partners on side for now, with Intel B-series GPUs on the way from Intel itself and six named partners: Acer (Nitro), ASRock (Steel Legend, Challenge), Gunnir (Photon, Index), Maxsun (I-Craft, Milestone), Onix (Lumi, Odyssey) and Sparkle (Titan, Guardian).

It’ll be fascinating to see how these cards perform once they’re available, and we’ll be following up in due course once we have had a chance to test them out.

Posted on by

Build a budget gaming PC for under £750 thanks to these Black Friday deals

We’ve already covered a RTX 4070 Super mini gaming PC for £1300 based around Black Friday discounts, but what if you want something even more affordable – something that doesn’t require paying a premium for an extra-small case, power supply and motherboard? Here, we’re going to achieve a powerful enough machine for 1080p gaming at high settings, yet spend less than £750.

The starting point for this build is the extremely affordable Ryzen 5 5600X, which has been reduced down to £99 – a great value for a six core, 12 thread Zen 3 processor with PCIe 4.0 support. It’s got ample power for entry-level gaming and general productivity, and the inclusion of the Wraith Stealth cooler in the box means we don’t have to spend any time, thought or budget on a third-party CPU cooler.

Category Product Deal Was
CPU AMD Ryzen 5 5600X (with Wraith cooler) £98.63 £113.80
GPU PowerColor Fighter RX 6750 XT £282.52 £300
Motherboard Gigabyte B550 AORUS Elite AX V2 ATX £100.98 £144.70
RAM Corsair Vengeance 16GB DDR4-3200 CL16 £28.98 £42.28
SSD WD Black SN770 2TB £92.99 £102.99
PSU Corsair RM850e (Used – Like New, Amazon Resale) £69.05 £86.32
Case NZXT H5 Flow (2024) £69.98 £89.99
Total £743.13 £880.08

Opting for a Ryzen 5600X means building on the venerable AM4 platform, with B550 motherboards offering the best balance of features and affordability. Luckily, there are options aplenty with on-board Wi-Fi and solid I/O for around £100-£120. One great option is the Gigabyte B550 Aorus Elite V2 which costs £101 from Ebuyer and comes with Wi-Fi 6, support for PCIe 4.0 NVMe SSDs and speedy USB-C. That means we need some DDR4 RAM, and this 16GB kit of Corsair Vengeance DDR4-3200 CL16 is just £26 from Amazon. That hits the sweet spot for price to performance, and gives plenty of headroom for gaming and most content creation workloads too. The WD SN770 2TB NVMe SSD for £93 from Scan provides solid capacity and excellent performance with up to 5150MB/s reads and 4900MB/s writes, respectively.

As for the case, I’m leaning towards the NZXT H5 Flow, which provides solid room to build in, a clean aesthetic and handy extras such as decent front I/O with wider cable channels to make it easier for ham-fisted fellows like me to route cables to a reasonable standard. It also comes with two 120mm fans pre-installed for convenience, too. From Amazon, it’s £70 for Black Friday. The Corsair RM850e PSU might be overkill for what’s inside, but provides a reliable, fully modular 80+ Gold unit with lots of headroom for upgrading, and for £69 from Amazon with the Amazon Resale 20 percent reduction, it’s hard to turn down.

The piece of the puzzle that’s quite difficult to decide on is with my GPU choice. I’m aiming for something in the £250-£300 range to keep this build at around £750 excluding Windows, and I’ve narrowed it down to either the £283 AMD RX 6750 XT, or £240 Intel Arc A770 16GB model. Both cards have their merits, with the Arc A770 packing in an extra 4GB of VRAM, better RT performance and support for AV1 encoding, while the RX 6750 XT performs better and benefits from better driver support. With these reductions in mind, the £336 (w/ code CHILLY5) RX 6800 also becomes a viable choice in and around that budget, which betters both of these cards.

Building this PC is probably what’s going to occupy my time over the festive period, and I’m not totally scared at all! I shall return later on with a build update to see how it all runs.

Posted on by

Where to buy AMD Ryzen 7 9800X3D: UK/US links and prices

AMD’s Ryzen 7 9800X3D is easily the best gaming CPU money can buy today, with hugely impressive performance gains against both the reigning champion 7800X3D and all other Ryzen processors on the market – while Intel’s recent releases are left further down the performance ladder. There are also major improvements in content creation performance and thermals, making the 9800X3D a stronger all-around choice than its predecessors.

With our 9800X3D review complete, it’s now time to round up all of the places you can find this hotly anticipated processor in the US and the UK.

Here’s our most recent data, including prices and availability for the biggest retailers on both sides of the Atlantic. Expect stock levels to be low for some time, based on the extremely high demand for this powerful processor!

AMD Ryzen 7 9800X3D – best UK prices and retailers

Retailer Price Availability
CCL £519 Pre-order – ETA 14 November 2024
Novatech £469 Pre-order – ETA 15 November 2024
Ebuyer £469 Pre-order – ETA 23 November 2024
Scan £469 Pre-order – ETA 29 November 2024
Overclockers £449 Pre-order – ETA December 2024
AWD-IT £529 Out of stock
Amazon UK TBA Not yet live
Currys TBA Not yet live

AMD Ryzen 7 9800X3D – best US prices and retailers

Retailer Price Availability
Amazon US TBA Not yet live
B&H Photo Video $479 Sold out, no restock ETA
Best Buy $479 Sold out, no restock ETA
Newegg $479 Sold out, no restock ETA
Newegg (bundles) $639-$999 Sold out, no restock ETA
2nd-gen 3D V-Cache from AMD
Image credit: AMD

Frequently asked questions

What motherboards support Ryzen 9000?

AMD’s X670E, X670, B650E and B650 motherboards all support Ryzen 9000 out of the box, so you won’t need to upgrade if you’ve got a Ryzen 7000 system already. X670E offers PCIe 5.0 throughout, X670 offers PCIe 5.0 for storage and graphics, and B650 offers PCIe 5.0 for storage only. In general, as you progress from B650 to X670E, expect more features, more PCIe lanes and higher prices. Newer generation X870E and X870 motherboards are also available, but these tend to offer relatively minor spec bumps with no major new features and the same processor compatibility, so we recommend sticking with the cheaper 600-series motherboards for 2024 at least.

What memory is best for Ryzen 9000?

As was the case with Ryzen 7000, Ryzen 9000 is a DDR5-only platform, so your older DDR4 RAM won’t work here. Instead, you’ll need to pick up a new kit of DDR5, which starts at DDR5-4800 and goes up to DDR5-6400 and beyond. We recommend DDR5-6000 with the lowest latency you can find – CL30 or lower is great, but higher latencies are still OK.

The usual RAM buying advice applies here. First, get two or four sticks to ensure you’re running in dual channel mode, as using a single stick suffers from a severe performance penalty. Secondly, make sure you enable XMP or EXPO in your motherboard’s BIOS to ensure that you’re getting the rated speeds – you can check your current RAM speed with a free tool like CPU-Z or Windows 11’s Task Manager.

What CPU coolers support Ryzen 9000 / AM5?

While most CPU coolers now support AM5 out of the box, some older models designed for AM4 that don’t use the default AMD backplate may not be compatible. Therefore, it’s best to look for AM5 support specifically, which is available from most popular brands including Corsair, Noctua, NZXT, Thermalright, Cooler Master, Be Quiet!, etc.

Well, there you have it – the best prices for AMD’s all-new Ryzen 7 9800X3D and a little advice too. Let us know if you spot one of these CPUs for a lower price, and stay tuned to @dealsfoundry on Twitter for more PC deals as we find them.

Posted on by

AMD Ryzen 7 9800X3D review: obliterating the competition

It’s been a rough old time when it comes to benchmarking CPUs, I can tell you that. The Ryzen 9000 release back in August was let down by relatively meagre performance improvements and an unready software ecosystem, while Intel’s 285K and 245K launch last month showed a significant performance regression versus 14th-gen – despite a genuinely interesting shift in architectures. Thankfully, AMD is here to save the day with the Ryzen 9 9800X3D.

As you might expect, the 9800X3D is the fastest gaming CPU we’ve ever tested, but what you might not expect is just how much faster it can be in some titles versus the 7800X3D, 9700X and 14900K – and how much of a step forward it is outside of games versus its X3D predecessor.

What I particularly like about the 9800X3D is that it does something a bit different to the 7800X3D and 5800X3D before it. Rather than adding on the extra L3 cache above the CCD, AMD has popped it underneath this time around, a move that sees temperatures go down and clock speeds go up. The result is a chip that is a much better all-rounder than its predecessors, while still delivering the kind of game-changing performance that makes X3D chips worth waiting for.

2nd-gen 3D V-Cache from AMD
Image credit: AMD

To properly get to grips with the level of performance on offer, we’ve run the 9800X3D through our full gamut of 11 game benchmarks at 1080p, 1440p and 4K, including some of the most demanding single-player (Dragon’s Dogma, Starfield, Cyberpunk) and multiplayer (CS2, F1 24) releases. Our power analysis and DDR5 testing also returns from our recent Intel Core Ultra 200S reviews.

Before we get into all that though, it’s worth taking a closer look at the specs here – and whether the changes we see versus the 7800X3D and 9700X are reflected in our synthetic and content creation benchmarks.

In short, the 9800X3D offers higher base clock speeds and much more L3 cache than the 9700X, but it’s also worth noting that the TDP has climbed to 120W – versus 65W and optionally 105W – for the 9700X. Boost speeds are still 300MHz slower on the X3D chip, but versus the 7800X3D it’s clear that the new on-bottom V-Cache location has delivered some nice gains – plus 500MHz (!) when it comes to base clock and plus 200MHz when it comes to boost. For context, we only saw a 100MHz leap in boost clock when going from the 7700X to the 9700X, so that new design does make itself known.

This is also the most expensive Ryzen 7 X3D chip ever shipped, with a $479 RRP versus the $449 7800X3D and 5800X3D – though (potentially deliberate) supply issues have meant that the 7800X3D is far from the bargain it once was.

CPU design Boost Base L3 cache TDP Price
Ryzen 9 9950X Zen 5 16C/32T 5.7GHz 4.3GHz 64MB 170W £555/$584
Ryzen 9 9900X Zen 5 12C/24T 5.6GHz 4.4GHz 64MB 120W £383/$383
Ryzen 7 9800X3D Zen 5 8C/16T 5.2GHz 4.7GHz 96MB 120W $479
Ryzen 7 9700X Zen 5 8C/16T 5.5GHz 3.8GHz 32MB 65W £309/$324
Ryzen 5 9600X Zen 5 6C/12T 5.4GHz 3.9GHz 32MB 65W £229/$249
Ryzen 9 7950X3D Zen 4 16C/32T 5.7GHz 4.2GHz 128MB 120W £550/$595
Ryzen 9 7950X Zen 4 16C/32T 5.7GHz 4.5GHz 64MB 170W £429/$487
Ryzen 9 7900X3D Zen 4 12C/24T 5.6GHz 4.4GHz 128MB 120W £443/$569
Ryzen 9 7900X Zen 4 12C/24T 5.6GHz 4.7GHz 64MB 170W £323/$395
Ryzen 7 7800X3D Zen 4 8C/16T 5.0GHz 4.2GHz 96MB 120W £398/$476
Ryzen 7 7700X Zen 4 8C/16T 5.4GHz 4.5GHz 32MB 105W £264/$269
Ryzen 5 7600X Zen 4 6C/12T 5.3GHz 4.7GHz 32MB 105W £189/$207
Ryzen 5 7600 Zen 4 6C/12T 5.1GHz 3.8GHz 32MB 65W £167/$198
Ryzen 9000 benchmarking setup
A very clean and not at all dusty test rig, featuring a Ryzen 9000 processor, RTX 4090 Founders Edition, ASRock Taichi X670E and the Alphacool Aurora Eisbaer 240mm AiO. | Image credit: Digital Foundry

All testing was performed on fresh installs of Windows 24H2 with the latest chipset drivers, BIOS updates and Nvidia’s 565.90 graphic drivers installed (to allow us to reuse the mountain of data we collected for our Core 285K and 245K reviews). Resizable BAR and core isolation were enabled on each system.

Our CPU test platform is based around the Nvidia GeForce Nvidia GeForce RTX 4090 Founders Edition graphics card and Trident Z5 Neo DDR5-6000 CL30 RAM. Each CPU is accompanied by a 1000W+ power supply from Corsair or NZXT, a high-end 240mm or 360mm AiO and a suitable premium motherboard: the ASRock X670E Taichi for AMD Ryzen 7000 and 9000, the MSI Z890 Meg Ace for Intel Core Ultra 200S, the Gigabyte Aorus Z790 Master for 12th and 14th-gen Intel Core and the Asus ROG Crosshair 8 Hero for AMD Ryzen 5000 – with Trident Z Royal DDR4-3600 CL16 as DDR5 RAM is not supported on this platform.

Storage duties are handled by a range of PCIe 4.0 or PCIe 5.0 SSDs, including the 4TB Lexar NM790 for our AMD Ryzen 9000 system.

Now, let’s move onto the content creation benchmarks to get a better sense of how the 9800X3D advances on its predecessors. Even if you don’t plan to use the 9800X3D to crank out videos or do 3D modelling, this sort of testing also sets out the maximum uptick we’re likely to see in gaming workloads and gives you a more general sense of all-core performance.

Cinebench 2024 (1T) 2024 (MT) R20 (1T) R20 (MT)
Ryzen 5 3600X 77 578 485 3654
Ryzen 7 5800X3D 95 915 546 5746
Ryzen 9 5900X 98 1171 610 8393
Ryzen 5 7600X 114 845 744 5814
Ryzen 7 7700X 118 1127 758 7609
Ryzen 7 7800X3D 112 1074 688 6988
Ryzen 9 7900X 116 1605 776 11196
Ryzen 9 7950X 121 2004 784 14272
Ryzen 5 9600X 132 935 850 6358
Ryzen 7 9700X 130 1172 862 7851
Ryzen 9 9800X3D 134 1342 823 8938
Ryzen 9 9900X 135 1784 879 12617
Ryzen 9 9950X 138 2235 866 15850
Core i5 14600K 120 1400 777 9420
Core i7 14700K 127 1987 818 13614
Core i9 14900K 133 2107 875 15297
Ultra 5 245K 128 1435 841 9864
Ultra 9 285K 144 2386 895 16055

Starting with our Cinebench test results, including the popular R20 and more recent 2024 release, the 9800X3D shows a small uptick in single-core performance (130 vs 134) and a much larger ramp up in multi-core performance (1342 vs 1172) versus the 9700X, most likely attributable to its higher power limits or the extra L3 cache. Note that we tested the 9700X at its stock 65W setting, though a 105W mode was made available later and in our 9700X review we saw it deliver similar performance to the 9800X3D – 136 single, 1280 multi.

The bigger differential comes when examining the 9800X3D versus the 7800X3D. Here, scores are up around 20 percent for single and 25 percent multi-core in the two Cinebench versions, putting the 9800X3D half-way between the eight-core 7700X and 12-core 7900X – not bad.

While Cinebench is meant to reflect a 3D modelling and animation task in Cinema 4D, our Cinebench test reflects a real use case for us: transcoding a Patreon video file into H.264 and H.265 (HEVC). Here, we measured the 9800X3D sucking down a healthy 259W at full tilt, versus the 191W-max 9700X, but the new CPU also turns in an impressive 59.59fps H.264 and 27.44fps HEVC average transcode frame-rate. That’s 15 percent faster than the 9700X, admittedly while drawing 35 percent more power.

Handbrake H264 (fps) HEVC (fps)
Ryzen 5 3600X 26.66 10.80
Ryzen 7 5800X3D 42.00 18.71
Ryzen 9 5900X 57.59 23.83
Ryzen 5 7600X 41.29 18.31
Ryzen 7 7700X 53.27 23.65
Ryzen 7 7800X3D 49.63 21.54
Ryzen 9 7900X 78.35 32.59
Ryzen 9 7950X 98.58 41.68
Ryzen 5 9600X 42.51 19.77
Ryzen 7 9700X (191W max) 51.80 23.79
Ryzen 7 9800X3D (259W max) 59.59 27.44
Ryzen 9 9900X 82.96 35.33
Ryzen 9 9950X 103.25 44.97
Intel Core i5 14600K 59.42 25.39
Intel Core i7 14700K 80.26 31.07
Intel Core i9 14900K (476W max) 85.06 35.08
Intel Ultra 5 245K (286W max) 61.05 26.88
Intel Ultra 9 285K (362W max) 97.17 38.44

Comparisons versus the 7800X3D are also impressive, with a 27 percentage point advantage in the HEVC test. We’re still some way from the likes of the 12-core and 16-core AM5 designs, but it’s still an impressive showing for a gaming-focused part and one that proves the efficacy of AMD’s new bottom-cache design.

Now let’s get into the turkey and Yorkshire puddings of our testing, the gaming benchmarks. We have 11 games in all.

AMD Ryzen 7 9800X3D analysis

Posted on by

Intel reportedly lost PlayStation 6 chip deal to AMD

Intel
reportedly
lost
Sony’s
PlayStation
6
chip
contract
to
Advanced
Micro
Devices
in
2022,
missing
out
on
a
potential
$30
billion
in
revenue.

This
is
according
to

Reuters,
with
sources
claiming
that
the
deal
fell
through
after
Intel
and
Sony
failed
to
settle
on
“how
much
Intel
stood
to
take
from
each
chip
sold.”

The
sources
alleged
that
internal
projections
from
Intel
suggested
that
Sony’s
console
business
“could
have
pumped
roughly
$30
million”
into
the
firm
had
the
contract
gone
through.

Reuters
also
noted
it
could
have
financially
boosted
Intel’s
foundry
segment,
which
formally
launched
earlier
this
year.

In
response
to
the
claims,
an
Intel
spokesperson
said:
“We
strongly
disagree
with
the
characterisation
but
are
not
going
to
comment
about
any
current
or
potential
customer
conversations.
We
have
a
very
healthy
customer
pipeline
across
both
our
product
and
foundry
business,
and
we
are
squarely
focused
on
innovating
to
meet
their
needs.”


AMD
has
had
a
design
contract
with
Sony
since
2014,
having
made
custom
chips
for
both
PlayStation
4
and
PlayStation
5.