News / AI PCs
AI PCs & local-AI supercomputers: shipping now vs coming (2026)
The unified-memory wave that lets one quiet desktop run a 70B model is here — GB10 boxes are shipping, Ryzen AI Max is spreading, and a consumer RTX Spark plus AMD's Medusa Halo are on the horizon. Here's what's real today and what's still a leak.
By Evan Cole · Last updated July 17, 2026
A year ago, running a 70-billion-parameter model at home meant a tower of GPUs and a small space heater's worth of power draw. In 2026 it means a box the size of a paperback stack. The reason is unified memory: a single large pool the CPU and GPU both address, so a ~65–200 W desktop can hold a large model entirely in memory. That shift is what created the “AI PC” — and the pace of new hardware is fast enough to be worth tracking. Here's what's confirmed and shipping, and what's still a rumor.
Shipping now
- NVIDIA GB10 “DGX Spark”-class boxes are shipping. NVIDIA's DGX Spark went on sale in late 2025, and partner machines are now live from ASUS (the Ascent GX10), MSI (EdgeXpert), Dell, HP, Lenovo, Acer and Gigabyte. All pair a GB10 Grace-Blackwell chip with 128 GB of unified memory and NVIDIA's DGX OS / CUDA stack. Supply is constrained and pricing moves; the ASUS Ascent GX10 is the most affordable GB10 machine on Amazon.
- AMD Ryzen AI Max+ 395 “Strix Halo” is spreading fast. The 128 GB-capable APU now ships in the Framework Desktop, Beelink GTR9 Pro, Minisforum MS-S1 Max, GMKtec EVO-X2, HP Z2 Mini G1a, Corsair's AI Workstation and more — the value story for big local models. AMD has even published guidance on clustering multiple Ryzen AI Max nodes for trillion-parameter mixture-of-experts inference.
- Apple M5 shipped — in laptops. The M5 MacBook Air and M5 Pro/Max MacBook Pro arrived in early 2026. M5 desktops (Mac mini, Mac Studio) are not out yet, so the current desktop local-LLM Macs remain the M4-family Mac mini and the M4 Max / M3 Ultra Mac Studio.
- Intel Panther Lake (Core Ultra Series 3) launched at CES 2026 — Intel's first 18A-process mobile chip, with an upgraded NPU — pushing the thin-and-light “AI PC” laptop tier forward (though, as ever, the NPU matters far less than memory bandwidth for large local models).
On the horizon (rumor / roadmap)
- NVIDIA RTX Spark (rumored, Computex 2026). A consumer Windows-on-Arm PC platform co-designed with Microsoft — reportedly up to 20 cores, an RTX 5070-class GPU and up to 128 GB — with laptops said to arrive around Fall 2026 from Dell, ASUS, HP, Lenovo, Microsoft and MSI, starting near $3,000. It would be NVIDIA's push into mainstream Windows PCs. Specs and pricing are unconfirmed.
- AMD “Medusa Halo” (leak, ~2027). The rumored Strix Halo successor — Zen 6 plus RDNA 5, with a leaked 384-bit LPDDR6 bus said to raise memory bandwidth by roughly 80%. More bandwidth is exactly what these machines need most. Leak-level only; treat the numbers as unconfirmed.
- NVIDIA “Vera Rubin Spark” (roadmap, 2027–2028). NVIDIA's own published roadmap points to a GB10 successor on LPDDR6, with lower-cost variants planned — but that's availability years out, not a 2026 purchase.
- Apple M5 Ultra / M5 Mac desktops (expected, unannounced). An M5 Mac mini and Mac Studio are widely expected in late 2026, possibly slipping into 2027. Unannounced as of July 2026.
At a glance: now vs next
| Platform | Status | What it means for local AI |
|---|---|---|
| NVIDIA GB10 (DGX Spark / GX10) | Shipping | 128 GB unified + CUDA stack, buyable now |
| AMD Ryzen AI Max+ 395 | Shipping | Cheapest 128 GB unified for big models |
| Apple M4 Studio / mini | Shipping | Fastest turnkey (up to 546 GB/s), zero setup |
| NVIDIA RTX Spark | Rumored ~Fall 2026 | Consumer Windows-on-Arm AI PC |
| AMD Medusa Halo | Leak ~2027 | LPDDR6, ~+80% bandwidth (unconfirmed) |
| Apple M5 desktops | Expected late 2026 | Next-gen turnkey Mac, unannounced |
Don't want to wait? What to buy today
If you want to run local models now rather than wait on a rumor, one machine per silicon family covers the range — all verified on Amazon at research time (re-check the seller and price, which move on these niche systems):
- ASUS Ascent GX10 — NVIDIA GB10 · CUDA stack — Workstation-class. Check price on Amazon →
- Beelink GTR9 Pro — Ryzen AI Max+ 395 · 128 GB value — Upper mid. Check price on Amazon →
- Apple Mac mini (M4 Pro) — Apple M4 Pro · turnkey — Mid. Check price on Amazon →
Full picks, verdicts, pros/cons and a speed/memory comparison are in our best AI PC for local LLMs guide and the AI PC benchmark.
Frequently asked questions
Can I buy an NVIDIA GB10 AI supercomputer right now?
Yes. NVIDIA's DGX Spark and partner boxes — including the ASUS Ascent GX10, MSI EdgeXpert, and others — began shipping in late 2025 and are on sale in 2026, though supply is thin and prices move. The ASUS Ascent GX10 is currently the most affordable GB10 machine on amazon.com. Dell, Acer, Lenovo and HP GB10 units exist but sell mainly through their own channels rather than Amazon.
Should I wait for the rumored RTX Spark or AMD Medusa Halo?
Only if you don't need a machine now. The RTX Spark is a rumored consumer Windows-on-Arm PC platform expected in laptops around Fall 2026, and AMD's Medusa Halo (a Strix Halo successor with a wider LPDDR6 bus, leaked at up to ~80% more bandwidth) is roadmap-level for roughly 2027. Both are unconfirmed on final specs and pricing, and history says first-gen availability is constrained. If you want to run local models today, current GB10, Ryzen AI Max and Apple M4 machines already do the job; if you can wait a year for more bandwidth, watch this space.
Are Apple M5 desktops out for local AI?
Not yet. Apple's M5 shipped in laptops (the M5 MacBook Air and M5 Pro/Max MacBook Pro) in early 2026, but M5 Mac mini and Mac Studio desktops have not been announced as of July 2026 — they're expected later in 2026, possibly slipping into 2027. For a desktop local-LLM Mac today, the M4-family Mac mini and the M4 Max / M3 Ultra Mac Studio are the current options.
Sources
Confirmed specifications and shipping status are drawn from manufacturer pages: NVIDIA's DGX Spark (GB10, 128 GB unified, 273 GB/s) and Apple's Mac Studio and Mac mini spec pages for Apple memory bandwidth. AMD Ryzen AI Max (“Strix Halo”) specifications and the mini-PC landscape are reported by ServeTheHome, Tom’s Hardware and NotebookCheck; independent local-LLM throughput figures come from ServeTheHome, Level1Techs and LMSYS. The rumored items (RTX Spark, AMD Medusa Halo, Vera Rubin Spark, M5 desktops) are leak- or roadmap-level, attributed to the hardware press and labelled as unconfirmed above. This page was produced with AI assistance; every confirmed spec was cross-checked against a primary source before publication.
More hardware news
All news →GPU rumors
AMD RDNA 5 (UDNA) & Radeon RX 10900 XT: rumors and release dateAMD's next-gen UDNA / RDNA 5 is rumored to return to the high end — a 36 GB GDDR7 flagship aimed at the RTX 6090, targeting mid-2027. Two leaks split on the specs (96 vs 154 CU); we show both. Sourced, labelled as rumor.
Updated July 17, 2026
CPU rumors
AMD Zen 6 vs Intel Nova Lake: next-gen CPU rumorsThe next desktop CPU race — rumored Ryzen 10000 (Zen 6) vs Core Ultra 400 (Nova Lake): 24 vs 52 cores, 6GHz vs 288MB bLLC cache, both eyeing CES 2027. Sourced, confirmed-vs-rumor kept honest.
Updated July 15, 2026
GPU rumors
RTX 6090 specs, rumors, and release dateWhat's actually leaked about NVIDIA's next flagship (Rubin / GR202) — plus the DRAM crisis that may push it to 2028, and where the price rumors now sit. Every figure sourced.
Updated July 15, 2026
GPU rumors
RTX 50 Super refresh: rumors, delays, and release dateThe rumored RTX 5070 / 5070 Ti / 5080 Super and why the high-VRAM refresh may be delayed indefinitely — or cancelled. Sourced, honest, NVIDIA-confirms-nothing.
Updated August 8, 2026
Related buyer's guides
All reviews →Buyer's guide / GPUs
RTX 5090 vs RTX 5080: which should you buy?
The 5090 doubles the 5080's VRAM, compute and price. For gaming that's rarely worth it; for VRAM- or CUDA-bound creator and AI work it often is. A sourced head-to-head plus a clear buying call.
Read the comparison →
Buyer's guide / GPUs
RTX 5070 Ti vs RX 9070 XT: which should you buy in 2026?
Roughly tied at 1440p, the 5070 Ti edges 4K and wins ray tracing, creation and local AI; the 9070 XT wins raster-per-dollar — a gap the 2026 memory shortage widened. Sourced head-to-head plus a verdict by use case.
Read the comparison →
Buyer's guide / GPUs
Is 8GB of VRAM enough in 2026? (RTX 5060 Ti 8GB vs 16GB)
Same chip, twice the memory: the RTX 5060 Ti 8GB vs 16GB isolates VRAM as the only variable. It holds up at 1080p esports and falls off a cliff almost everywhere else. Sourced benchmarks, plus ray tracing, frame gen and local-AI VRAM costs, and a clear buy rule.
Read the guide →
Buyer's guide / GPUs
DLSS 4 vs FSR 4 in 2026: which upscaler is actually better?
DLSS still leads on image quality and game support; AMD counters with universal AFMF 2 frame generation and lower prices. The accurate, sourced head-to-head (DLSS 4.5 vs FSR 4/Redstone) — and whether it should sway your GPU.
Read the comparison →
Buyer's guide / GPUs
1440p vs 4K gaming in 2026: which should you game at?
1440p is the 2026 sweet spot for most; 4K is for immersion and creators with the GPU budget. The honest breakdown — and, unlike the monitor sites, the exact GPU tier to pair with each resolution and refresh target.
Read the comparison →
Buyer's guide / Monitors
OLED vs IPS for gaming in 2026: which should you buy?
OLED wins the picture — perfect blacks, per-pixel HDR, fastest motion and now the highest refresh rates; IPS wins brightness, price and zero burn-in. A neutral, evidence-based breakdown with the honest 2026 burn-in verdict — and the GPU tie-in the monitor sites skip.
Read the comparison →