GPU Prices Are Up 50%+: What to Buy for Local AI Right Now

Why are GPU prices so high? Because of a memory shortage, not scalpers. AI datacenters are buying DRAM and GDDR faster than it can be made (IDC forecasts they may consume up to ~70% of world memory output in 2026), and every graphics card is mostly a memory purchase. On November 2, 2026 the cheapest in-stock RTX 50 cards we found sat 58-86% above launch price, the RTX 5090 Founders Edition was out of stock at $1,999 with partner cards from $3,340, and NVIDIA had raised the DGX Spark from $3,999 to $6,950 in two steps. The RTX 50 Super is still unannounced.

Updated November 2, 2026: every price below was re-read with its date. The July figures that this page first published are kept where they show the trend. The unified-memory boxes that were the value play in July ($3,099.99 ASUS GX10) have caught up with the GPU inflation, so the buying advice changed too.

The rest of this page is the detail: what things cost now, which numbers are confirmed versus leaked, and exactly what to buy depending on the models you want to run. To compare cards by price per gigabyte of VRAM, new and used, see cheapest GPUs per GB of VRAM.

---

Why Prices Are Up: The Memory Supercycle

Every GPU is mostly a memory purchase now. The industry calls it the memory supercycle: AI datacenter buildouts are absorbing DRAM and NAND at a scale consumer hardware has never competed with. IDC's forecast is the headline number — AI datacenters may consume up to ~70% of world memory output in 2026.

The knock-on effects are not subtle:

• GPU street prices are well above launch MSRP across tiers, not just at the flagship. Cheapest in-stock listings on November 2, 2026, against launch price:

• NVIDIA raised the DGX Spark twice. The Founders Edition went from $3,999 to $4,699 (+18%) on February 25, 2026, explicitly citing industry-wide memory supply constraints, and NVIDIA's marketplace listed it at $6,950 on November 2 (HotHardware reported the second increase on October 2, alongside new 64GB GB10 systems from October 23). Same hardware, higher price. When a vendor raises a price mid-cycle and says the quiet part out loud, believe the shortage is real.
• The RTX 50 Super refresh is still unannounced (more on that below).

Nothing between July and November suggested it clearing: prices rose further over those four months, not less. Memory contracts are the bottleneck, and the AI buildout writing those contracts shows no reported sign of slowing.

---

RTX 5090: The Street-Price Trap

Here is what the flagship costs (street prices; they move weekly):

Our advice is blunt: do not pay far over MSRP for a 5090 unless you are VRAM-bound today — meaning a model you need right now will not fit in the card you have. The 5090 is still the fastest thing you can put in a consumer PC, and for models that fit in 32GB it beats every unified-memory box on speed. But at $3,000+, you are paying datacenter-shortage prices for consumer silicon, with a leaked Super refresh that could reset the mid-range. For a single card head-to-head, see RTX 5090 vs 5080 for local LLMs.

If you need CUDA VRAM on a budget instead, the used market still matters, though it inflated too: a used RTX 3090 averaged $1,568 on resaleprices.com (fair asking range $1,499-1,606, November 1, 2026), against roughly $700-1,000 in mid-2026. See our RTX 4090 vs 3090 comparison and the RTX 3090 local AI guide — a used 24GB card plus our 24GB VRAM model picks covers a lot of real workloads without flagship pricing. AMD's 24GB RX 7900 XTX was cheaper used ($1,175) on the same readings.

---

The RTX 50 Super Delay

The refresh that was supposed to fix the mid-range VRAM problem is stuck — and it is stuck on the same memory shortage.

Status on November 2, 2026: still no NVIDIA announcement. What the supply-chain reporting says (this is reporting, not an official NVIDIA statement): cards reached at least one board partner, but the launch was put on hold over 3GB GDDR7 memory pricing, delayed rather than cancelled. The most recent report, from Benchlife via Notebookcheck (June 8, 2026, updated October 5), points to a CES 2027 (January) reveal, while noting earlier reports that still pointed to late 2026. With November here and nothing announced, January is the realistic earliest window.

The leaked specs, labeled as leaks:

Why this matters for buyers: a 24GB RTX 5080 Super would gut the case for paying $3,340+ for a 5090 just to get VRAM. That leaked lineup is the one genuine reason current prices could normalize. But the timing is unknowable, it hinges on the exact memory pricing that caused the delay, and waiting indefinitely for a leak is not a plan. Decide based on what you need to run this quarter.

---

Unified-Memory Boxes: The Math That Changed

In July a new category was winning the price-per-gigabyte fight: small boxes with 128GB+ of unified LPDDR5X that CPU and GPU share. By November they had caught the same inflation as the GPUs. They still hold models no consumer card can, but they are no longer cheap.

ASUS Ascent GX10 — $5,999 (was $3,099.99 in July)

The GX10 carries the same NVIDIA GB10 Grace Blackwell chip and 128GB LPDDR5X as the DGX Spark. In July it was in stock at $3,099.99 (1TB); on November 2, 2026 NVIDIA's marketplace listed it at $5,999 (1TB), with the 4TB model at $7,999 and out of stock. That is still ~$950 less than the DGX Spark ($6,950 on the same day) for the same silicon, so if you want the NVIDIA stack in this category, this remains the box.

The mid-2026 review consensus on the Spark/GB10 platform is worth internalizing: its value is memory capacity, not speed. A single RTX 5090 is faster for anything that fits in 32GB. The GB10 niche is 70B-200B-class models at usable speeds — the models a consumer card simply cannot hold. The software story is maturing fast: vLLM published an official DGX Spark guide on June 1, 2026. NVIDIA's CES claim of "up to 2.5x inference gains" from software is a vendor claim — treat it as such.

AMD Ryzen AI Halo Developer Platform — $3,999.99 (unchanged)

Launched July 10, 2026, this is AMD's own-brand Strix Halo mini-PC: Ryzen AI Max+ 395 (16 Zen 5 cores, Radeon 8060S with 40 CUs), 128GB LPDDR5X-8000, 2TB SSD, 10GbE, Wi-Fi 7. It is still listed at $3,999.99 at Micro Center (exclusive, in-store only), with ROCm, LM Studio, and ComfyUI preconfigured — Windows 11 Pro or Linux, same price. AMD positions it directly against the DGX Spark for up-to-200B-parameter local models. In July it cost $900 more than a GX10; in November it is about $2,000 less, with 2TB storage and 10GbE in the box. You trade CUDA for ROCm.

A 192GB version with the new Ryzen AI Max+ PRO 495 (Gorgon Halo, up to 160GB allocatable as GPU memory) became orderable at Micro Center on October 12, 2026 for $7,299.99, per Born City (October 14). For the chip itself, our Strix Halo / AI Max+ 395 guide has the full picture.

Strix Halo minis — cheapest 128GB, if you shop carefully

The mini-PC market caught the same inflation as GPUs. Prices read from each vendor's store on November 2, 2026 unless noted:

At $3,919 on sale, the MS-S1 Max is the cheapest 128GB box we found in November, just under AMD's own $3,999.99 Halo — about the price of one street-price 5090, with four times the memory. More options in our best mini PC for Ollama roundup and the best local AI server guide.

Apple — the laptop path, and now the desktop one

The MacBook Pro M5 Pro/M5 Max line (announced March 3, 2026, available March 11) reaches 128GB unified memory at 614GB/s on the M5 Max (M5 Pro: up to 64GB at 307GB/s), with a Neural Accelerator per GPU core, per Apple. It is the only way to carry 128GB of model memory in a backpack. On the desktop, Apple announced the Mac Studio M5 Max (from $2,499 at 36GB, up to 128GB at 460 or 614GB/s depending on GPU) and M5 Ultra (from $5,499 at 96GB, up to 512GB at 1.2TB/s) on August 25, 2026, shipping from September 22. Apple's prices are list prices that did not move with the street market. Details and configs in our Apple Silicon AI buying guide.

And the ceiling, for context

The NVIDIA DGX Station (GB300, 784GB-class coherent memory) is orderable via OEMs — ASUS, Dell, GIGABYTE, MSI, Supermicro — at ~$97K-$125K with 4-12 week lead times as of July 2026 (an MSI XpertStation WS300 was listed at $96,995.99 on CDW then; not re-read in November). Not a consumer product; it is here so you know where the category tops out.

---

What to Buy, By Model Size

Match the box to the models you actually run. That is the whole game in 2026:

Two honest caveats. First, unified-memory boxes are slower per token than a 5090 for anything that fits in 32GB — buy them for capacity, not speed. Second, what does 128GB actually unlock? Frontier-class open-weight models like GLM-5.2 are exactly the class of model this hardware exists for. Browse the full picture on our hardware hub and the best GPUs for AI ranking.

---

What Is Coming Next

Gorgon Halo (Ryzen AI Max PRO 400 series) — announced ~May 20, 2026, and now on sale in AMD's own 192GB Ryzen AI Halo box ($7,299.99 at Micro Center since October 12, above): Max+ PRO 495 (16-core/40CU with the new Radeon 8065S), PRO 490 (12-core/32CU), PRO 485 (8-core/32CU), with up to 192GB LPDDR5X-8533 and up to 160GB allocatable as GPU memory (~273GB/s, +7% bandwidth). AMD claims one SoC can load a 300B-parameter FP4 model — "a first for any single SoC that is not a Mac Studio" — which is a vendor claim until third parties test it. AMD said OEM systems from ASUS, Lenovo, and HP would follow in Q3 2026; we did not re-check their availability for this update. If you want 192GB, this is the platform to look at.

Medusa Halo (Ryzen AI MAX 500) — Zen 6 + RDNA 5, LPDDR6, ~460GB/s. Rumor-stage only, realistically 2027-2028. Do not plan purchases around it.

RTX 50 Super — see above: unannounced as of November 2, CES 2027 the most-cited window, all per supply-chain reports.

---

Verdict

The late-2026 market rewards buyers who know exactly which models they want to run — and punishes everyone paying street price out of habit.

1. Do not pay far over MSRP for an RTX 5090. The cheapest in-stock partner card was $3,340 (+67%) on November 2; the $9,000+ listings are indefensible unless a model you need today will not fit in your current VRAM.
2. The RTX 50 Super is still unannounced. Leaked 24GB mid-tier cards are the one credible force that could normalize prices — CES 2027 most-cited. Waiting on a leak is not a strategy.
3. For big models, unified memory still wins on capacity, but no longer on price alone. Ryzen AI Halo at $3,999.99 and the MS-S1 Max at $3,919 are now the value 128GB boxes; the GX10 ($5,999) and DGX Spark ($6,950) roughly doubled since early 2026.
4. For models that fit in 32GB, a fast GPU still wins on speed. A used 24GB card remains the best value in raw tokens per second per dollar.

Buy capacity for the models you cannot fit. Buy speed for the ones you can. And in a memory supercycle, never buy either at panic prices.

---

Sources

• NVIDIA marketplace, graphics cards and personal AI supercomputers — Founders Edition stock, RTX 5070 Ti and 5060 Ti partner prices, DGX Spark $6,950 and ASUS GX10 $5,999 (read November 2, 2026)
• NVIDIA developer forum: DGX Spark price change — $3,999 to $4,699, February 25, 2026
• HotHardware: DGX Spark 64GB — second Spark price increase and 64GB systems from October 23 (October 2, 2026)
• Walmart RTX 5090 and RTX 5080 listings — $3,340 and $1,579 (read November 2, 2026)
• videocardprices.com RX 7900 XTX — $1,820.99 new (October 17), $1,175 used; resaleprices.com used RTX 3090 tracker (November 1)
• Micro Center: AMD Ryzen AI Halo — 128GB $3,999.99; 192GB PRO 495 $7,299.99
• Minisforum MS-S1 Max, Beelink GTR9 Pro, GMKtec EVO-X2 store pages (read November 2, 2026)
• Apple Newsroom: Mac Studio with M5 Max and M5 Ultra (August 25, 2026)
• Notebookcheck on the RTX 50 Super timing (June 8, 2026, updated October 5)
• Tom's Hardware — GPU street-price tracking and memory-shortage reporting
• TechPowerUp — GPU specifications database and market coverage
• VideoCardz, ServeTheHome, Wccftech — RTX 50 Super supply-chain reporting and Strix Halo mini-PC coverage (leak-grade where noted)
• NVIDIA developer forum, Micro Center, ASUS eShop — first-party pricing for DGX Spark, Ryzen AI Halo, and Ascent GX10

---

FAQ