Intelligence at the Edge. Always-On. Private. Physical.
Laguna Hills, California
- Tweets 1,616
- Following 266
- Followers 9,568
- Likes 542
📦 AI hardware doesn’t have a math problem. It has a delivery problem.
Sequora starts from one question: what should you actually accelerate?
📖 Read the blog: bit.ly/4yyKvm1
3
17
294
🔋 Your biggest chip is doing the smallest job.
Watching a sensor all day to confirm nothing happened shouldn’t need the big processor awake. Give that job to a small sensing brain and let the big one sleep.
📖 Read the blog: bit.ly/3UwrcM7
#EdgeAI
19
78
2,015
🔋 Generative AI in a wearable is a memory problem, not a compute problem.
Attention cache grows with every token, and memory traffic sets the power ceiling on a battery.
BrainChip Sequora™ IP holds history in a fixed-size state.
Link below 👇
bit.ly/4B1M12l
3
21
81
1,488
BrainChip retweeted
Sponsored Podcast - Embedded Executive LIVE at @embedded_USA: @BrainChip_inc
Rich Nass and Steve Brightfield, Chief Product Officer at brainchip, talk tech at embedded world North America #AI #ewna
embeddedcomputing.com/techno…
3
15
659
📣 Today, we announced BrainChip Sequora™ Core IP, a generative AI edge processor licensed as verified RTL.
🔋 Wearables: generative AI inside a battery budget.
📡 Defense: signal processing and generative AI on one core, no cloud.
Full announcement: bit.ly/3VNHAYS
3
25
105
2,078
Thanks Raúl Parada and @CttcTech for putting Akida into a year-long wildfire detection simulation.
Modeled result: a solar drone with Akida logged 4,200 patrol hours and ran 87% on sunlight. Simulation, and the paper is upfront about it.
Link below.
doi.org/10.1016/j.iot.2025.1…
6
21
101
2,652
BrainChip is proud to sponsor Semiconductor Australia in Sydney on 22 October.
CEO Sean Hehir is presenting and joining a panel, and our booth shows IP, silicon, boards and reference platforms you can access today.
2
16
69
1,817
Hear the failure before it stops the line.
AkidaTag runs vibration anomaly detection AI on the machine itself. No cloud round trip. Nothing to stream.
Pre-orders open.
bit.ly/4dedJii
1
15
68
1,438
📄 A new FOI survey on neuromorphic radar and radio, by Hamrell, Sjögren and Ovrén, names Akida among commercial chips and cites our AFRL radar work. Worth a read 👇
doi.org/10.1088/2634-4386/ae…
2
22
93
2,002
📡 Ground control is not a real-time system. In deep space the round trip is minutes to hours.
So the spacecraft decides. Detect, extract, classify, discard. Terabytes in, a small payload out.
📖 Read Dr. Jonathan Tapson's blog
bit.ly/4AOTcLr
4
22
83
1,917
Always-on tire monitoring, processed at the wheel ⚡
Akida Pico IP classifies tire pressure and vehicle load in a single streaming pass, with inference in the microwatt range.
Architecture, integration options, and validation results in our new white paper
Link below👇
bit.ly/4i7qgqp
3
23
96
2,152
AkidaTag learns you. Not the other way around.
On-device learning in a tag about the size of a key fob. Prototype your own activity recognition and nothing leaves the device.
Pre-orders open.
bit.ly/4dedJii
24
86
1,561
Everyone's racing to add more compute to the edge. We think that's the wrong axis.
Our CEO, Sean Hehir, gets into why on this week's EE Journal Fish Fry, event-based processing means Akida only computes when something actually changes, not nonstop.
bit.ly/4ypoZkl
1
15
83
1,730
You can't always be there.
Prototype something that can.
AkidaTag runs motion AI on the device, worn as a pendant, on the wrist or arm.
No camera. No cloud.
Pre-orders open.
bit.ly/4dedJii
3
18
76
1,701
How much power does your chip burn just to tell you nothing happened?
BrainChip's Akida runs always-on sensing in the milliwatt range so the host can actually sleep, then wakes it only when something real occurs.
Link below.
bit.ly/4daHi4f
#EdgeAI #Akida
1
26
89
1,629
the industry keeps solving edge AI by adding more compute. we solved it by needing less
AkidaTag runs real-time listening and detection on a battery, for days, with zero cloud calls
it's up for pre-order. link below
bit.ly/4dedJii
1
23
92
1,704
A few days out and still thinking about the conversations from @embedded_world North America.
Three demos, three days, one question we kept hearing: why isn't more edge AI actually running at the edge?
5
12
86
2,204
🤯 We split one video frame across six neuromorphic chips and got a better result than any single chip could give.
Our CPO's new blog, "Clone It, Build It, Run It," breaks down exactly how, plus the number behind it.
📖 bit.ly/46LhGaA
1
22
97
3,170
Our CPO Steve Brightfield joined Rich Nass of Embedded Computing Design at embedded world North America.
Full conversation is up now: bit.ly/EmbeddedComputingDesi…
Still at booth 6623 through Sept 24, come see it live.
21
85
2,365