The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →The GPU industry is moving in two directions at once. Jon Peddie Research (JPR) reported that total PC GPU shipments fell 7.5% sequentially in Q1 2026, while AI infrastructure, GPU servers and accelerator revenue continued to expand. The apparent contradiction is explained by market definition: a low-cost integrated PC processor, a desktop graphics board, an AI accelerator module and a rented cloud GPU are counted in different ways.
There is therefore no single, universally accepted “GPU market size.” The most defensible analysis keeps unit shipments, board share, semiconductor revenue, server spending and company segment revenue separate.
GPU market at a glance
| Metric | Latest figure | What it measures |
|---|---|---|
| PC GPU shipments, Q1 2026 | Down 7.5% sequentially | JPR; integrated and discrete PC graphics |
| PC GPU unit outlook, 2025–2029 | Approximately −3% CAGR | JPR forecast; units, not revenue |
| PC GPU installed base by 2029 | Approximately 3 billion units | JPR forecast |
| Desktop add-in-board shipments, Q1 2026 | Approximately 12 million | Discrete desktop boards; JPR |
| NVIDIA desktop AIB share, Q1 2026 | Approximately 90% | JPR estimate; not all GPUs |
| Worldwide AI spending, 2026 | $2.59 trillion | Gartner; all AI-related spending |
| GPU plus AI-accelerator revenue, 2028 | More than $150 billion | Gartner combined category |
| Worldwide server spending growth, Q1 2026 | 30.7% year over year | IDC; GPU servers were a major driver |
Sources: JPR PC-GPU update, JPR AIB report, Gartner AI forecast, Gartner semiconductor forecast and IDC server data.
What counts as the GPU market?
Before using a statistic, identify the product boundary and the unit of measurement.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware match#1 Best Overall
- Powered by Radeon RX 9070 XT
- WINDFORCE Cooling System
- Hawk Fan
- Server-grade Thermal Conductive Gel
- RGB Lighting
| Market layer | Typical metric | Examples of suppliers | Best use |
|---|---|---|---|
| Total PC graphics | Units shipped, installed base, attach rate | Intel, AMD, NVIDIA | PC adoption and platform trends |
| Desktop discrete add-in boards | Board shipments and unit share | NVIDIA, AMD, Intel | Gaming-card competition |
| Data-center GPUs and accelerators | Revenue, shipments, compute capacity | NVIDIA, AMD, cloud providers and custom-chip vendors | AI infrastructure |
| Professional visualization | Workstation units and revenue | NVIDIA, AMD, Intel | CAD, media, engineering and science |
| Embedded and automotive | Design wins, shipments and revenue | NVIDIA, AMD, Intel, Qualcomm and Arm ecosystem vendors | Edge systems and vehicles |
| Console and handheld graphics | SoC shipments and semi-custom revenue | AMD, NVIDIA and Arm-based vendors | Gaming ecosystems |
A company may also report a broader business than the chip itself. An AI server can include accelerators, CPUs, high-bandwidth memory (HBM), networking, storage, racks, cooling, software and services. Cloud GPU revenue is consumption revenue, not a count of chips shipped.
Units, revenue and average selling price
Unit data answers how many devices entered a market. Revenue data captures price, mix and systems. Average selling price (ASP) is the bridge: a market can ship fewer units yet grow revenue if buyers shift toward expensive AI accelerators or premium graphics cards. System revenue should never be substituted for GPU-chip revenue without a stated methodology.
PC GPU shipments and installed base
JPR’s Q1 2026 report covers PC graphics processors, including integrated and discrete parts. Shipments declined 7.5% from the prior quarter, a seasonal result that remained below the historical average. JPR’s current forecast calls for approximately 3% annual unit contraction from 2025 through 2029 and an installed base of about 3 billion units by 2029. These are unit forecasts, not a prediction that GPU-industry revenue will fall at the same rate.
Why PC units are mature
- Notebook and desktop replacement cycles are longer than during the pandemic upgrade surge.
- Integrated graphics in CPUs and system-on-chips satisfy many mainstream workloads without a discrete board.
- AI-capable PCs increasingly divide workloads among CPU cores, integrated GPUs and neural-processing units (NPUs).
- Memory prices, tariffs, product transitions and supply constraints can shift shipments between quarters.
- Gaming buyers may delay upgrades when high-end prices rise or performance gains are incremental.
JPR’s forecasts have changed materially as conditions changed: −6.1% to 2028 in Q1 2025, −2.9% in Q2, +1.5% to 2029 in Q3, +2.4% in Q4, and approximately −3% in Q1 2026. The sequence demonstrates forecast uncertainty, not a clean reversal of the installed-base trend. See the Q1 2025, Q2 2025, Q3 2025 and Q4 2025 updates for the publication vintages.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Attach rate is easy to misread
An attach-rate measure can exceed 100% because it counts integrated and discrete graphics across systems; it does not mean every computer physically contains more than one graphics card. Integrated graphics also dominate unit counts, so Intel’s broad graphics presence is not equivalent to leadership in high-end discrete boards.
Desktop discrete GPU market share
JPR’s Q1 2026 estimate covers roughly 12 million desktop add-in boards, with NVIDIA at approximately 90% and AMD and Intel making up most of the balance. The associated AIB outlook is approximately −3.3% CAGR to 2029. This denominator excludes integrated notebook graphics, consoles, most embedded products and data-center accelerators. It is therefore incorrect to describe the figure as “NVIDIA’s share of the entire GPU market.”
For comparison, a JPR-based summary reported approximately 44.28 million desktop discrete boards shipped during 2025 and NVIDIA at about 94% of the AIB market in Q4 2025. That is also a desktop-board statistic, not a measure of all gaming devices or GPU revenue. Tom’s Hardware’s summary provides that context.
Data-center GPUs and AI accelerators
Training large language models, generative-AI inference, recommendation systems, scientific computing, sovereign-AI programs and private enterprise clusters are expanding demand for accelerated servers. Cloud providers are adding capacity, while GPU virtualization allows several customers to share a physical accelerator.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
IDC reported 30.7% year-over-year growth in worldwide server spending in Q1 2026, with mass deployment of GPU servers a principal driver; IDC also cited a five-year server-market CAGR of 25.1%. Those figures include servers and their components, not standalone GPU revenue. Gartner forecast total worldwide AI spending of $2.59 trillion in 2026, up 47% year over year. That total includes systems, software and services, so it is not a GPU-sales estimate.
Gartner’s semiconductor forecast places combined GPU and AI-accelerator revenue at more than $150 billion in 2028, versus approximately $80 billion in 2024. Because the category combines GPUs and accelerators, it cannot be compared directly with JPR’s PC-unit CAGR.
What limits growth?
- HBM availability and advanced packaging capacity, including CoWoS-style capacity.
- Interconnect and networking throughput between accelerators.
- Electricity, cooling and rack-density limits in data centers.
- Export controls, tariffs and regional restrictions.
- Cloud-provider custom silicon and open software alternatives.
Supplier landscape
NVIDIA
NVIDIA combines discrete PC graphics with a data-center platform spanning accelerators, networking, systems and software. CUDA and its surrounding libraries can reduce migration costs for customers, while hyperscaler concentration and export controls remain commercial risks. NVIDIA’s reported Data Center business is broader than accelerator-chip revenue; consult its annual-report archive for each fiscal year’s segment definitions.
AMD
AMD reported 2025 revenue of $34.639 billion, including $16.635 billion in Data Center and $3.910 billion in Gaming. Data Center includes EPYC CPUs, Instinct GPUs and other products, so it is not pure GPU revenue. In Q1 2026, Data Center revenue was $5.8 billion, up 57% year over year, while combined Client and Gaming revenue was $3.6 billion, up 23%. AMD changed its reportable-segment structure in fiscal 2025; comparisons should use its retrospectively adjusted disclosures. Its 2025 filing also describes approximately $800 million in inventory and related charges tied to export restrictions affecting Instinct MI308 products. Sources: AMD 2025 Form 10-K/A and AMD Q1 2026 results.
Rank #4
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
Intel
Intel supplies integrated graphics through client processors, Arc discrete products and data-center or AI solutions. Intel’s broad Client Computing or Data Center revenue cannot be treated as standalone GPU revenue unless the company separately identifies it. Intel’s graphics share must therefore be labeled by denominator—total graphics units, discrete boards or revenue.
Gaming GPU trends
- Longer replacement cycles and high-end prices encourage buyers to retain prior-generation cards or use the used market.
- Ray tracing, upscaling and frame generation increasingly differentiate products beyond raw raster performance.
- VRAM capacity and memory bandwidth matter for high-resolution textures and AI-assisted features.
- 1440p and 4K adoption raises demand for premium cards, but consoles and handhelds compete for the same gaming time.
- Notebook GPUs and console or handheld SoCs are normally excluded from desktop AIB statistics.
AI-PC, edge and embedded graphics
AI PCs combine CPU cores, integrated GPU compute, an NPU and shared memory; premium models may add a discrete GPU. An AI-PC shipment is therefore not automatically incremental discrete-GPU demand. Automotive, industrial, professional-visualization and embedded products are often excluded from PC-GPU datasets even though they may be strategically significant and carry different prices, lifecycles and design-win economics.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Forecasts through 2029: how to compare them
| Forecast | Period | Metric | Inclusions | Interpretation |
|---|---|---|---|---|
| JPR PC GPU | 2025–2029 | Approximately −3% CAGR | PC GPU units | Mature or declining unit market |
| JPR desktop AIB | To 2029 | Approximately −3.3% CAGR | Discrete desktop boards | Separate from total PC graphics |
| Gartner semiconductor | 2024–2028 | About $80 billion to more than $150 billion | GPUs plus AI accelerators | Revenue category, not PC units |
| IDC servers | Q1 2026 and five-year view | 30.7% quarterly-year growth; 25.1% five-year CAGR | Server spending | Includes substantially more than GPUs |
For any new estimate, record the base year, end year, geography, nominal or real currency, units or dollars, chip/board/system/service boundary, and whether integrated graphics or AI accelerators are included. A standard calculation is CAGR = (ending value / beginning value)^(1 / years) − 1.
How to read GPU market statistics
- Identify the denominator: total PC graphics, desktop AIBs, data-center accelerators, consoles or another segment.
- Check whether the number is units, revenue, ASP, installed base or spending.
- Determine whether it covers a chip, board, server, rack, cloud service or company segment.
- Check the quarter or fiscal year, geography and forecast publication date.
- Look for integrated graphics, notebook parts, consoles and AI accelerators in the inclusion rules.
- Keep reported results, estimates and forecasts separate.
- Do not use user surveys such as Steam hardware prevalence as shipment market share.
Frequently Asked Questions
How large is the GPU market in 2026?
There is no single defensible total. JPR measures PC units, Gartner measures a combined GPU-and-AI-accelerator revenue category, and IDC measures server spending; each answers a different question.
Best Value
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Is the GPU market growing or shrinking?
PC GPU units are forecast by JPR to decline approximately 3% annually from 2025 to 2029, while data-center infrastructure and combined GPU-and-AI-accelerator revenue are expanding.
Does NVIDIA have 90% of the entire GPU market?
No. Approximately 90% is JPR’s Q1 2026 estimate for desktop add-in boards, not integrated graphics, notebooks, consoles, embedded products or data-center accelerators.
Why can PC GPU shipments fall while industry revenue rises?
Integrated PC demand can weaken while expensive AI accelerators, GPU servers and premium boards raise average selling prices and revenue.
Are AI accelerators included in GPU statistics?
Sometimes. Gartner’s forecast explicitly combines GPUs and AI accelerators; JPR’s PC shipment data and desktop AIB share are different categories.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →The Bottom Line
The reliable conclusion is a segmented one: mature or declining PC GPU units coexist with rapidly expanding AI and data-center accelerator spending. Always state the market boundary, metric, period and source before comparing a GPU statistic.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
