• Latest
  • Trending
  • All
The HMND 01 Alpha humanoid robot standing in an industrial facility in a black vest branded HMND, SAP and Martur Fompak, with two arms ending in parallel grippers and grey logistics totes stacked around it.

Humanoid HMND 01: what the robot unicorn really ships

19 July 2026
Answer card stating that Qwen-Image-2.1, released on 20 September 2026, ships open weights with a 7 billion parameter diffusion transformer, a Qwen3-VL 8B text encoder and an RGBA VAE totalling about 33 gigabytes in BF16, under the Qwen Research License that limits use to research or evaluation and requires a separate commercial licence, unlike the Apache 2.0 licence of Qwen-Image 1.0.

Qwen-Image-2.1 brings the weights back, but not the Apache licence

21 September 2026
Answer card stating that Ternary Bonsai 2 27B, released by PrismML on 17 September 2026 under Apache 2.0, packs Qwen3.8 27B into 5.95 gigabytes at 1.72 bits per weight, keeps 98.2 percent of the 14-benchmark average, about 75 percent on SWE-bench Verified and Terminal-Bench 2.1, and needs PrismML's llama.cpp fork to run.

Does Bonsai 2 27B really keep 98% of Qwen3.8 in 5.95 GB?

20 September 2026
Answer card stating that Jev 1.13 from TypeSafe AI is a decision model in early access since 15 September 2026 that returns typed probabilities instead of text, priced at 42 dollars per billion input tokens with output tokens free, answering in 70 to 500 milliseconds, with a 64K token request budget, text input only, and a documented list of things it does badly, including counting and dates.

Jev 1.13 bills $42 a billion tokens, and it can’t count

19 September 2026
Answer card stating that Qwen3.8-Omni-Flash launched on 17 September 2026 as an API only model on Alibaba Cloud Model Studio, taking text, images, audio and video in a 1M token context and returning text only, priced at 0.15 dollars per million input tokens for every modality and 0.47 dollars per million output tokens in the international regions, with no open weights published and the Qwen-Live Harness GitHub repository returning 404.

Qwen3.8-Omni-Flash bills audio at $0.15 and ships no weights

18 September 2026
Answer card stating that on 15 September 2026 AWS said it is unable to restore access to resources and data hosted exclusively in the Middle East Bahrain region me-south-1 and in the mec1-az2 zone of the UAE region, because the damage spanned multiple Availability Zones and exceeded what multi-AZ services are designed to withstand.

AWS can’t restore me-south-1, six months after the drone strikes

17 September 2026
Answer card stating that Google released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on 15 September 2026 at 3 dollars per million audio input tokens and 12 dollars out, that the thinking model requires asynchronous tools, and that Artificial Analysis scores it 82.6 on its Speech to Speech Quality Index.

Gemini 3.8 Live Extended Thinking rejects any tool that blocks

16 September 2026
Answer card summarising the Atria Dawn Preview release: 744B GLM-5.2 base, MIT licence, 1.5 TB BF16 and 756 GB FP8 checkpoints, 256K context, top on five of sixteen benchmark rows and trailing on SWE-bench Pro.

Atria Dawn Preview is 744B under MIT, and the BF16 weighs 1.5 TB

15 September 2026
Answer card stating that OpenAI released the Agents API in public beta on 10 September 2026 with no separate fee, billed through model tokens, tool calls and hosted sandbox time, with a choice of OpenAI hosted, self hosted or partner sandboxes, US only data residency and no Zero Data Retention support.

OpenAI’s Agents API has no fee, no ZDR and a one hour sandbox clock

14 September 2026
Answer card: Sakana Fugu Max at $2 and $6 per million tokens, Fugu Ultra v2 unchanged at $5 and $30, and Sakana saying Ultra v2 scores without Fable 5 or GPT-6 Astra in its pool.

Fugu Max costs $2 and $6 while Fugu Ultra v2 runs without Fable 5

13 September 2026
Answer card stating that DeepSeek released DeepSeek-V4.1-Flash on 10 September 2026 as a 552 billion parameter mixture of experts model with a new causal encoder decoder architecture that activates 8 billion parameters on input and 16 billion on output, with native vision, a one million token context and MIT licensed weights, that the API model name is now deepseek-flash at 0.15 dollars per million input tokens and 0.60 dollars per million output tokens off peak, and that DeepSeek announced V4 Pro would be routed to V4.1-Flash from 14 September and reversed that on 11 September.

DeepSeek V4.1-Flash arrived, and the V4 Pro retirement lasted a day

12 September 2026
Answer card stating that Cognition released SWE-2 on 10 September 2026, a coding model post-trained from Kimi K3, scoring 50.0 percent on FrontierCode 1.1 Main against 50.9 percent for Claude Fable 5.1 and 27.3 percent on Terminal-Bench 4 against 55.8 percent, available only inside Devin.

SWE-2 trails Fable 5.1 by one point, and by 28 on Terminal-Bench 4

11 September 2026
Answer card for Meta Muse, free to 100 million tokens a week then $20 a month, launched 8 September 2026 for United States adults only, running in a dedicated per user virtual machine.

Does Meta Muse do enough to earn your inbox and a card on file?

9 September 2026
  • About
  • Contact
  • Privacy
  • Legal
Monday, September 21, 2026
  • Login
Packet Nebula
  • Home
  • Articles
    • Security
    • Network
    • Dev
    • Sysadmin
    • SEO
    • Email & DNS
  • Tools
    • Network tools: free, fast, no signup
    • Security tools: free, fast, no signup
    • Developer tools: free, fast, no signup
    • Sysadmin tools: free, fast, no signup
    • SEO tools: free, fast, no signup
    • Email & DNS tools: free, fast, no signup
  • Download
  • About
No Result
View All Result
Packet Nebula
No Result
View All Result
Home Dev

Humanoid HMND 01: what the robot unicorn really ships

by stephane
19 July 2026
in Dev
0
The HMND 01 Alpha humanoid robot standing in an industrial facility in a black vest branded HMND, SAP and Martur Fompak, with two arms ending in parallel grippers and grey logistics totes stacked around it.
494
SHARES
1.4k
VIEWS
Share on FacebookShare on Twitter

So a two-metre robot in a black vest stands in an auto-parts plant, moving grey totes from a pallet to a trolley, and that quiet scene is the actual story here. On July 17 The Information reported that Humanoid, the London startup behind it, raised a 150 million dollar first tranche at a valuation of 1.2 billion dollars. On paper that makes it the UK's first humanoid-robot unicorn. We've watched a lot of these machines get valued long before they get useful, so here's the honest split. The robot, HMND 01 Alpha, is real: 2.2 metres, 29 degrees of freedom, a swappable five-fingered hand or a parallel gripper, and two live proofs of concept already behind it. What isn't real yet is a fleet in production. It's still an Alpha, the wheeled Beta isn't due until the third quarter of 2026, and the big valuation is one outlet's reporting, not a number Humanoid has posted itself.

The short answer

Humanoid, a London robotics startup, hit a 1.2 billion dollar valuation on a 150 million dollar first tranche. Its robot, HMND 01 Alpha, is a real 2.2m dual-arm industrial machine with two live pilots behind it. But it’s still an Alpha, the wheeled Beta lands in Q3 2026, and the valuation is one outlet’s reporting. Big number, early product. Worth watching, not yet worth planning a warehouse around.

$1.2Bvaluation, per The Information
$150Mfirst-tranche Series A raised
2 POCslive pilots, no volume yet
Answer card: London startup Humanoid raised a 150 million dollar first tranche valuing it at 1.2 billion dollars per The Information, its HMND 01 Alpha is a 2.2 metre dual-arm industrial robot with 29 degrees of freedom that has run two live proofs of concept, with no volume deployment yet.
The one-card version. A unicorn valuation sitting on top of an Alpha robot with two pilots to its name.

What actually happened

Let’s separate the two things that got tangled in the headlines, because they’re not the same event. The news peg is money: The Information reported on July 17 that Humanoid closed a 150 million dollar first tranche of a Series A, at a valuation of 1.2 billion dollars excluding the fresh cash. The robot itself, HMND 01 Alpha, isn’t new. It debuted publicly at CES in January and has been running pilots since. So this week is a funding milestone bolted onto a machine that already existed.

That distinction matters more than it sounds. A unicorn valuation is a bet on where a company will be in five years, not a report card on what its hardware does today. And Humanoid is being pretty upfront about the gap. This is an Alpha. The company says it’s on track for commercial deployment within the next twelve months, which is startup-speak for “not deployed.”

The HMND 01 Alpha humanoid robot in a black vest branded HMND, SAP and Martur Fompak, standing between grey logistics totes with both arms ending in black parallel grippers.
Image: Humanoid (HMND 01 Alpha during the SAP and Martur Fompak logistics pilot)

Here’s the money picture in one look. The valuation is the loud number. The cash actually in the door is a lot smaller, and Humanoid is reportedly out chasing another 80 to 100 million dollars by September to top it up.

Bar comparison of Humanoid's reported 1.2 billion dollar valuation against the 150 million dollar first tranche of cash actually raised, on a linear scale in millions of dollars.
Valuation is a bet on the future. The first tranche is the money that actually landed.

The robot, honestly

Strip out the finance and HMND 01 Alpha is a genuinely capable-looking industrial machine, not a demo-reel prop. It stands about 2.2 metres tall, walks at up to roughly 7.2 kph, and carries 29 active degrees of freedom before you count the hands. It sees with 360-degree RGB cameras and a pair of depth sensors, which is the sort of setup you need to not knock over a stack of totes in a cluttered aisle.

The clever bit is the swappable end-effector. You can bolt on a 12-DOF five-fingered hand when you need dexterity, or a plain 1-DOF parallel gripper when you just need to grab a box and not drop it. In the SAP pilot photo above, it’s running the grippers, which tells you what the early real work looks like. Less “robot butler,” more “tireless picker.”

And the pilots are the part worth taking seriously. Humanoid has done two. One was a logistics-readiness trial at a Siemens plant. The other, earlier this year, put HMND 01 into a live automotive-logistics workflow with SAP and Martur Fompak: the robot pulled task instructions from an SAP AI agent, navigated to pallets, grabbed KLT boxes and delivered them to trolleys. Artem Sokolov, Humanoid’s founder and CEO, framed the point as robots “operating inside real production environments, connected to enterprise systems.” That’s the right thing to care about. A humanoid that talks to your ERP is more interesting than one that does backflips.

There’s a spec footnote worth flagging, because it’s the kind of detail that gets rounded away. The bimanual payload gets quoted at 15 kg on the spec sheet, but the actual SAP proof of concept ran at an 8 kg dual-arm limit. Both numbers can be true, a rated maximum versus what they ran in a live cell, but if you’re sizing this for real totes, the pilot number is the honest one to plan against. Ask what it lifts on your floor, not on the datasheet.

What’s still just a claim

The headline everyone’s repeating is that HMND 01 is “the only industrial humanoid robot on track for commercial deployment within the next 12 months.” Read that carefully. It’s Humanoid’s own positioning, not an independently verified fact, and the humanoid field is crowded with companies making near-identical timelines. Take it as ambition.

Checklist splitting what HMND 01 actually ships today from what is still a claim: a real 2.2m robot with 29 degrees of freedom and two live proofs of concept are solid, but it is still Alpha with the Beta due Q3 2026, the payload dropped from 15kg to 8kg in the live pilot, and the on-track-in-12-months line is Humanoid's own claim.
The split between what you can verify today and what's a forward-looking promise.

The other soft spot is the money itself. A first tranche means the round isn’t closed. The 1.2 billion dollar valuation comes from a single exclusive report, and Humanoid hasn’t posted the raise on its own channels as of this writing. None of that means it’s wrong. It just means the number you’re seeing quoted everywhere is reporting, and the round is still in motion. Worth remembering when the figure gets treated as gospel by the next twenty aggregators.

What it changes for you

If you run warehouse or manufacturing logistics, this is a “start a conversation” signal, not a “issue a PO” one. The interesting thread isn’t the valuation, it’s that a humanoid ran a real pick-and-place loop wired into SAP. That integration-first approach is the thing to track across the whole sector, and it pairs with the hardware side we’ve been following: robot brains are getting cheaper too, like NVIDIA’s mainstream Jetson Thor modules heading downmarket. Cheaper compute plus enterprise-connected pilots is the combination that eventually makes these things pencil out.

If you don’t touch physical logistics, this is upstream weather. It’s the same capital wave pushing money into custom AI silicon and data centers: real, large, and slow to reach your desk. A UK humanoid unicorn is a data point that the money believes physical AI is next. Whether the robot earns it is a 2027 question.

The honest read: HMND 01 is a real robot doing real pilots, and the valuation is a real bet that it scales. Both can be true. Just don’t confuse the second for the first. Check back when the Beta ships and someone other than Humanoid counts the deployments.

Sources: the funding and valuation come from The Information’s exclusive, “Humanoid, UK-Based Robot Maker, Becomes a Unicorn” (July 17, 2026). Robot specs and the pilot details come from Humanoid’s own writeup of the SAP and Martur Fompak proof of concept, with additional spec detail from MassRobotics. The 15 kg spec-sheet payload versus the 8 kg pilot limit is noted where the sources differ; the “only humanoid on track in 12 months” line is Humanoid’s own claim and is not independently verified.

Frequently asked questions

What is Humanoid's HMND 01?

HMND 01 is an industrial humanoid robot built by Humanoid, a London startup. The current model, HMND 01 Alpha, is a dual-arm mobile manipulator aimed at warehouse and factory work like picking, kitting, machine feeding and moving totes. It stands about 2.2 metres tall, has 29 active degrees of freedom excluding its hands, and can be fitted with either a five-fingered hand or a simple parallel gripper. Humanoid calls it the UK's first industrial humanoid robot.

How much did Humanoid raise, and what is it worth?

The Information reported on July 17, 2026 that Humanoid raised a 150 million dollar first tranche of a Series A round, valuing the company at 1.2 billion dollars before the new money. That crosses the unicorn line. The report also says Humanoid is looking to raise another 80 to 100 million dollars by September. Humanoid has not published the round on its own site, so treat the exact figures as reporting rather than a confirmed filing.

Is HMND 01 actually deployed anywhere?

Not in volume. Humanoid has run two proofs of concept. One was a logistics-readiness trial at a Siemens plant. The other, in early 2026, was a live automotive-logistics run with SAP and Martur Fompak, where the robot took instructions from an SAP AI agent and moved KLT boxes to trolleys. Those are real pilots in real facilities, but a pilot is not a production fleet. Humanoid says it is on track for commercial deployment within 12 months, which is a target, not a shipment.

What are HMND 01's specs?

HMND 01 Alpha stands roughly 2.2 metres tall and moves at up to about 7.2 kph. It has 29 active degrees of freedom excluding end-effectors, 360-degree RGB cameras and two depth sensors. End-effectors swap between a 12-DOF five-fingered hand and a 1-DOF parallel gripper. Bimanual payload is quoted at 15 kg on the spec sheet, though the live SAP proof of concept ran at an 8 kg dual-arm limit. Humanoid says the whole Alpha was built in about seven months.

Can I buy an HMND 01?

Not yet, and no public price has been posted. HMND 01 Alpha is a development and pilot unit. A wheeled Beta is scheduled for the third quarter of 2026, and Humanoid frames general commercial deployment as a next-12-months goal rather than an order you can place today. If you run warehouse or manufacturing logistics and want in early, the realistic path is a pilot conversation, not a purchase order.

Tags: aiautomationhardwarehumanoidnewsrobotics
Share198Tweet124
stephane

stephane

  • Trending
  • Comments
  • Latest
Answer card: Proton Lumo 2.0 is private by policy, not by locality. Saved history is locked so even Proton cannot read it, but the prompt is decrypted on a Proton EU server to answer it, then forgotten.

Proton Lumo 2.0 review: how private is it, really?

3 September 2026
The Agentic Coding section of the official Hy4 preview benchmark appendix published by Tencent, a table comparing Hy3 and Hy4 preview against DeepSeek V4 Pro 0813, Qwen 3.8 Max, GLM 5.3, Kimi K3, GPT 5.6 Sol and Claude Opus 5 across SWE-bench Multilingual, SWE-bench Pro, DeepSWE, three SWE Atlas tasks, SWE-Marathon, Terminal-Bench 2.1, NL2Repo-Bench, CyberGym, ProgramBench, PostTrainBench and Harbor-Index.

Tencent’s 770B Hy4 tops one benchmark row in 46

3 September 2026
Answer card: Qwen 3.7 Max is API-only and cannot run locally yet; the open Qwen models (Qwen 3.6 27B, qwen3:8b to 32b) run offline via Ollama.

Qwen 3.7 local: what you can actually run offline

22 June 2026
Answer card: JWTs are not encrypted, anyone can read them; the signature proves who issued the token, not who may read it.

Are JWTs encrypted? No, and the difference will bite you

0
Answer card: a random 8 character password falls in under 2 hours offline, while 16 random characters hold for 1.4 trillion years at the same speed.

How long does it take to crack a password in 2026?

0
Answer card: three DNS records decide if your mail lands or bounces; SPF lists allowed senders, DKIM signs messages, DMARC sets the failure policy.

SPF, DKIM and DMARC explained: the records your email needs

0
Answer card stating that Qwen-Image-2.1, released on 20 September 2026, ships open weights with a 7 billion parameter diffusion transformer, a Qwen3-VL 8B text encoder and an RGBA VAE totalling about 33 gigabytes in BF16, under the Qwen Research License that limits use to research or evaluation and requires a separate commercial licence, unlike the Apache 2.0 licence of Qwen-Image 1.0.

Qwen-Image-2.1 brings the weights back, but not the Apache licence

21 September 2026
Answer card stating that Ternary Bonsai 2 27B, released by PrismML on 17 September 2026 under Apache 2.0, packs Qwen3.8 27B into 5.95 gigabytes at 1.72 bits per weight, keeps 98.2 percent of the 14-benchmark average, about 75 percent on SWE-bench Verified and Terminal-Bench 2.1, and needs PrismML's llama.cpp fork to run.

Does Bonsai 2 27B really keep 98% of Qwen3.8 in 5.95 GB?

20 September 2026
Answer card stating that Jev 1.13 from TypeSafe AI is a decision model in early access since 15 September 2026 that returns typed probabilities instead of text, priced at 42 dollars per billion input tokens with output tokens free, answering in 70 to 500 milliseconds, with a 64K token request budget, text input only, and a documented list of things it does badly, including counting and dates.

Jev 1.13 bills $42 a billion tokens, and it can’t count

19 September 2026
  • About
  • Contact
  • Privacy
  • Legal

Copyright © 2026 Stephane Cardon.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Home
  • Articles
    • Security
    • Network
    • Dev
    • Sysadmin
    • SEO
    • Email & DNS
  • Tools
    • Network tools: free, fast, no signup
    • Security tools: free, fast, no signup
    • Developer tools: free, fast, no signup
    • Sysadmin tools: free, fast, no signup
    • SEO tools: free, fast, no signup
    • Email & DNS tools: free, fast, no signup
  • Download
  • About

Copyright © 2026 Stephane Cardon.