NVIDIA's 64GB DGX Spark goes on sale from six PC makers on October 23 at $4,999, about $2,000 to $4,000 below in-stock 128GB units today. Two clustered units cost more than one 128GB machine, so the purchase makes most sense for agents on models that fit in 64GB.
Perspective Coverage
13 publishers
- Builder
- Builder 50%
- Operator
- Operator 26%
- Investor
- Investor 24%
Reality
- Evidence58
- Adoption
- Insufficient
- Hype gap+30
- Incentives60
- Confidence62
Microsoft will pitch AI that runs on the PC itself at its October 7 event while the RTX Spark Surface Laptop Ultra still lacks a price and ship date. Windows teams due for a refresh have good reason to hold orders until both are public.
Reality
- Evidence45
- Adoption
- Insufficient
- Hype gap+30
- Incentives60
- Confidence40
Portable Computer's harness is real engineering. The bill of materials starts at a $4,800 DGX Spark or a 24GB RTX card, which turns local agents into a hardware order.
Perspective Coverage
3 publishers
- Builder
- Builder 33%
- Operator
- Operator 40%
- Investor
- Investor 27%
Reality
- Evidence55
- Adoption
- Insufficient
- Hype gap+20
- Incentives70
- Confidence60
PAIR proxies Ollama and LM Studio, so the agent keeps seeing one connection and no harness code changes. The adoption cost moves to disk, because a node is only eligible if it already has the exact model downloaded.
Perspective Coverage
7 publishers
- Builder
- Builder 45%
- Operator
- Operator 38%
- Investor
- Investor 17%
Reality
- Evidence62
- Adoption
- Insufficient
- Hype gap+30
- Incentives72
- Confidence60
The largest token allowances in China go to developers, not shoppers. The real pressure on Western AI pricing sits in API list prices that run 60% to 90% below OpenAI and Anthropic.
Reality
- Evidence55
- Adoption25
- Hype gap+30
- Incentives60
- Confidence55
Tom's Hardware puts the M5 Ultra at up to 1.2 TB/s of memory bandwidth over as much as 256GB of soldered unified memory, so the capacity that decides which weights stay resident is settled at checkout.
Perspective Coverage
3 publishers
- Builder
- Builder 42%
- Operator
- Operator 35%
- Investor
- Investor 23%
Reality
- Evidence55
- Adoption
- Insufficient
- Hype gap+30
- Incentives55
- Confidence55
Alibaba's 27B scores 52 on the Artificial Analysis index from a 17GB quantized file. Filling its 262,144-token window needs roughly 16 GiB of KV cache on top of that, so the file size is the smaller half of the sizing question.
Reality
- Evidence30
- Adoption15
- Hype gap+45
- Incentives50
- Confidence32
A principal engineer built the same EV-charging invoice service twice and timed a coding agent through nine cumulative features on each. The two setups differed in more than layering, and he says so.
Reality
- Evidence45
- Adoption
- Insufficient
- Hype gap+12
- Incentives25
- Confidence50
Alibaba's 27B model fits a 32GB card with 15GB to spare. Tom's Hardware still had to pick between llama.cpp's full 262K window at half-hour prefill and a supported vLLM deployment capped at 32K.
Reality
- Evidence60
- Adoption38
- Hype gap+15
- Incentives55
- Confidence58
NVIDIA Research rebuilt the serving stack for someone else's open-weight video model and got a five-second clip out in 1.653 seconds, but the default profile is lossy and the dense comparison run is the honest baseline.
Reality
- Evidence62
- Adoption45
- Hype gap+30
- Incentives72
- Confidence58
PAIR cut a five-subagent inbox task from 18 minutes to 8 minutes 48 seconds across three machines, which is just over two times the speed for three times the hardware, and every box in NVIDIA's demo cluster was one NVIDIA sells.
Reality
- Evidence32
- Adoption15
- Hype gap+25
- Incentives72
- Confidence45
Chinese banks and telcos are packaging inference the way airlines package miles, though the analyst closest to the trend calls the consumer bundles a supply-led experiment whose users never see a balance.
Reality
- Evidence54
- Adoption61
- Hype gap+14
- Incentives71
- Confidence57
Portable Computer runs Perplexity's agent on 27B models on your own hardware. Paid tiers only, Linux first, Windows in September, and a 24GB VRAM floor most desktops miss.
Reality
- Evidence32
- Adoption12
- Hype gap+42
- Incentives66
- Confidence40
The bandwidth gap is 4.4x and the capacity gap is 4x, which is why these two boxes are not really competing. One decides whether a model fits; the other decides whether it is usable.
Reality
- Evidence38
- Adoption32
- Hype gap+25
- Incentives42
- Confidence35
Portable Computer keeps agent work on hardware you own and asks permission before any single step escalates. The entry fee is 24GB of VRAM and a paid subscription.
Reality
- Evidence36
- Adoption13
- Hype gap+31
- Incentives74
- Confidence57
Portable Computer runs the agent's control plane on device and charges only when a task goes to the cloud. The part that decides when that happens is also the part nobody has priced or locked down.
Reality
- Evidence42
- Adoption12
- Hype gap+34
- Incentives68
- Confidence45
JetBrains has taken the assembly work out of running a coding agent offline. What it could not take out is the hardware, and that is now the thing deciding who adopts.
Reality
- Evidence56
- Adoption18
- Hype gap+16
- Incentives72
- Confidence63
Alibaba's new 27B model defaults reasoning_effort to xhigh. Simon Willison measured 22,276 reasoning tokens and 21 minutes for one SVG that took two minutes with reasoning off.
Reality
- Evidence68
- Adoption28
- Hype gap+8
- Incentives42
- Confidence62
A dev.to guide argues local LLM capacity planning collapses into one napkin equation. Run it first and the hardware shortlist writes itself, tier names and all.
Reality
- Evidence42
- Adoption20
- Hype gap+24
- Incentives55
- Confidence38
A single-box test in Japan put 76 tokens/s next to 4.4 tokens/s, then found the deciding variable elsewhere: which models held the output format and which invented reassurance.
Reality
- Evidence52
- Adoption22
- Hype gap−8
- Incentives45
- Confidence55