“We are finished”
www.bbc.com
www.bbc.com
themachinemaker.com
Why and when EMASS is a competitor to BrainChip
Why they are not always competitors — and where BrainChip still has advantages
What to watch — when EMASS becomes a real competitive threat
My view: EMASS is a conditional competitor — not a full replacement threat, but one that could seriously limit BrainChip’s reach if execution succeeds
Let’s hope this start up don’t end up going how this start up did
A few days ago our partner, MosChip, announced its collaboration with EMASS to implement an edge AI SoC, designed for ultra-low-power “always-on” edge devices.
EMASS is targeting the same markets BrainChip has pitched for years; drones, cameras, embedded sensors, small robots, anomaly detection, industrial edge devices, etc.
Because the SoC is a drop-in chip it means MosChip's customers can bring products to market faster and they won't have to spend any money to licence IP.
I asked ChatGPT how it sees the competition between BrainChip and EMASS. See below for response.
View attachment 93651
MosChip Partners with EMASS for Silicon Implementation of Next-Generation Edge AI SoC | Machine Maker - Latest Manufacturing News | Indian Manufacturing News - Latest Manufacturing News | Indian Manufacturing News - Machine Maker
MosChip Technologies, a key provider of silicon and product engineering services, has announced its collaboration with EMASS on the development of the ECS-DoT Edge AI System-on-Chip (SoC)—an ultra-low-power processor built for always-on intelligence in wearables, drones, industrial IoT...
ChatGPT 5
EMASS becomes a competitor to BrainChip when:
In such scenarios:
- The use-case is edge AI / low-power inference / always-on sensor fusion / wearables / IoT / drones / embedded devices, where BrainChip also targets with Akida.
- The buyer prefers a full SoC solution that “just works” — i.e. a ready-made chip (like ECS-DoT) rather than licensing IP and integrating it in a custom SoC design.
- The performance/efficiency trade-offs EMASS claims (low power, acceptable inference performance) are “good enough” for the customer’s needs — then the neuromorphic advantages BrainChip offers (e.g. spiking, temporal/event-based processing, maybe lower latency or lower long-term power) may be less relevant.
So in these domains, yes — EMASS is a commercial threat and competitor.
- EMASS reduces barriers to entry for OEMs — no IP licensing, no custom integration, less engineering complexity.
- The “good enough, cheap, fast” proposition may out-compete BrainChip’s “specialised but more complex / niche” value — especially for low-cost IoT and commodity devices.
- EMASS could capture volume markets where BrainChip might struggle, shrinking the addressable market for Akida.
There are conditions under which EMASS and BrainChip are not direct competitors, or where BrainChip likely retains an edge:
• Different technical approaches and potential performance envelope
Thus, BrainChip retains advantages in high-demand or niche use-cases (e.g. event vision, robotics, autonomous sensors, temporal sensor fusion, low-latency control, on-device learning, neuromorphic workloads).
- BrainChip’s value is often in neuromorphic processing — spiking/neural-network architectures optimized for temporal data, event-based sensors, sparse compute, on-device learning, streaming data, sensor fusion.
- EMASS’ ECS-DoT is an edge-AI SoC — presumably conventional (or at least not explicitly “spiking neuromorphic”) — more like a standard inference accelerator for low-power devices. That means:
- For static sensor inference, image classification, simple ML tasks, EMASS might be “good enough.”
- For temporal, event-driven, always-on sensor fusion, hard real-time, or low-latency streaming workloads, Akida (BrainChip) might perform better — or may be more efficient in long-term deployment.
• Business model difference — IP licensing vs full SoC
This difference means that for high-volume, low-cost commodity devices, EMASS might win.
- BrainChip sells IP + design integration + partners, aiming for custom SoCs and OEM partners.
- EMASS sells ready-made chips / SoCs, which is simpler for many OEMs — but less customizable.
But for specialised, performance-sensitive, or custom products, BrainChip’s IP-driven model still makes sense.
EMASS becomes a major threat to BrainChip if:
- ECS-DoT (or future EMASS SoCs) prove robust, low-cost, and widely adoptable — especially in IoT, wearables, sensors, drones.
- Their “always-on, ultra-low power” claims hold up in real-world deployments.
- They expand their SoC lineup across nodes (not just one chip), capturing device makers looking for drop-in AI hardware.
- They begin targeting markets BrainChip aims for — consumer devices, mass-market IoT, industrial sensors, etc.
- BrainChip doesn’t broaden its partner network or fails to secure enough SoC-integration deals.
If all that happens, EMASS could significantly shrink the market for Akida-based products — especially in low-margin, high-volume segments.
If I were advising BrainChip, I'd say: assume EMASS will eat some of the low-end / mass-market pie — but fight hard for the high-value, neuromorphic-native niches that EMASS can’t easily serve.
- For edge-AI commodity devices, EMASS is likely to be more attractive than BrainChip (cheaper, easier, faster to adopt).
- For specialised, neuromorphic-heavy, temporal or sensor-fusion tasks, BrainChip still holds strong advantages.
- BrainChip should treat EMASS as a firm competitor for the “low-cost edge AI” market but maintain that their differentiation lies in neuromorphic strengths, custom SoC/IP flexibility, and temporal/spiking performance.
I hope you didn’t turn into a shorter after seeing the share price keep falling…(or let’s say, I hope you’re not running for a different team now)
The whole “market share in AI” argument is tricky anyway. If someone says “this will reduce BrainChip’s share of the market,” well, we’ve known for a long time that BrainChip won’t be the only player. And we don’t even know how big the total AI market will ultimately become.
In every industry, countless companies compete with each other. That doesn’t mean they can’t still be successful. Sure, it’s disappointing when another company seems to come out of nowhere and achieves more than BrainChip despite all the patents and partnerships.
But if BrainChip finally delivers solid customer revenue, that would send a clear signal to the market. Until that happens, we’ll have to deal with comparisons and discussions around other companies.
Hi Bravo.A few days ago our partner, MosChip, announced its collaboration with EMASS to implement an edge AI SoC, designed for ultra-low-power “always-on” edge devices.
EMASS is targeting the same markets BrainChip has pitched for years; drones, cameras, embedded sensors, small robots, anomaly detection, industrial edge devices, etc.
Because the SoC is a drop-in chip it means MosChip's customers can bring products to market faster and they won't have to spend any money to licence IP.
I asked ChatGPT how it sees the competition between BrainChip and EMASS. See below for response.
View attachment 93651
MosChip Partners with EMASS for Silicon Implementation of Next-Generation Edge AI SoC | Machine Maker - Latest Manufacturing News | Indian Manufacturing News - Latest Manufacturing News | Indian Manufacturing News - Machine Maker
MosChip Technologies, a key provider of silicon and product engineering services, has announced its collaboration with EMASS on the development of the ECS-DoT Edge AI System-on-Chip (SoC)—an ultra-low-power processor built for always-on intelligence in wearables, drones, industrial IoT...
ChatGPT 5
EMASS becomes a competitor to BrainChip when:
In such scenarios:
- The use-case is edge AI / low-power inference / always-on sensor fusion / wearables / IoT / drones / embedded devices, where BrainChip also targets with Akida.
- The buyer prefers a full SoC solution that “just works” — i.e. a ready-made chip (like ECS-DoT) rather than licensing IP and integrating it in a custom SoC design.
- The performance/efficiency trade-offs EMASS claims (low power, acceptable inference performance) are “good enough” for the customer’s needs — then the neuromorphic advantages BrainChip offers (e.g. spiking, temporal/event-based processing, maybe lower latency or lower long-term power) may be less relevant.
So in these domains, yes — EMASS is a commercial threat and competitor.
- EMASS reduces barriers to entry for OEMs — no IP licensing, no custom integration, less engineering complexity.
- The “good enough, cheap, fast” proposition may out-compete BrainChip’s “specialised but more complex / niche” value — especially for low-cost IoT and commodity devices.
- EMASS could capture volume markets where BrainChip might struggle, shrinking the addressable market for Akida.
There are conditions under which EMASS and BrainChip are not direct competitors, or where BrainChip likely retains an edge:
• Different technical approaches and potential performance envelope
Thus, BrainChip retains advantages in high-demand or niche use-cases (e.g. event vision, robotics, autonomous sensors, temporal sensor fusion, low-latency control, on-device learning, neuromorphic workloads).
- BrainChip’s value is often in neuromorphic processing — spiking/neural-network architectures optimized for temporal data, event-based sensors, sparse compute, on-device learning, streaming data, sensor fusion.
- EMASS’ ECS-DoT is an edge-AI SoC — presumably conventional (or at least not explicitly “spiking neuromorphic”) — more like a standard inference accelerator for low-power devices. That means:
- For static sensor inference, image classification, simple ML tasks, EMASS might be “good enough.”
- For temporal, event-driven, always-on sensor fusion, hard real-time, or low-latency streaming workloads, Akida (BrainChip) might perform better — or may be more efficient in long-term deployment.
• Business model difference — IP licensing vs full SoC
This difference means that for high-volume, low-cost commodity devices, EMASS might win.
- BrainChip sells IP + design integration + partners, aiming for custom SoCs and OEM partners.
- EMASS sells ready-made chips / SoCs, which is simpler for many OEMs — but less customizable.
But for specialised, performance-sensitive, or custom products, BrainChip’s IP-driven model still makes sense.
EMASS becomes a major threat to BrainChip if:
- ECS-DoT (or future EMASS SoCs) prove robust, low-cost, and widely adoptable — especially in IoT, wearables, sensors, drones.
- Their “always-on, ultra-low power” claims hold up in real-world deployments.
- They expand their SoC lineup across nodes (not just one chip), capturing device makers looking for drop-in AI hardware.
- They begin targeting markets BrainChip aims for — consumer devices, mass-market IoT, industrial sensors, etc.
- BrainChip doesn’t broaden its partner network or fails to secure enough SoC-integration deals.
If all that happens, EMASS could significantly shrink the market for Akida-based products — especially in low-margin, high-volume segments.
If I were advising BrainChip, I'd say: assume EMASS will eat some of the low-end / mass-market pie — but fight hard for the high-value, neuromorphic-native niches that EMASS can’t easily serve.
- For edge-AI commodity devices, EMASS is likely to be more attractive than BrainChip (cheaper, easier, faster to adopt).
- For specialised, neuromorphic-heavy, temporal or sensor-fusion tasks, BrainChip still holds strong advantages.
- BrainChip should treat EMASS as a firm competitor for the “low-cost edge AI” market but maintain that their differentiation lies in neuromorphic strengths, custom SoC/IP flexibility, and temporal/spiking performance.
Nothing is 100% facual until heard from brainchip.
Agreed.
Go Brainchip
Ah well we are all sorted then... 2047 here we come
If MosChip is a partner of BrainChip then it is possible that MosChip have produced drop in chips containing BrainChip IP to sell to EMASS. MosChip have previously stated that they want BrainChip IP. I don’t have the link as it was several years ago but somebody might remember what I’m referring to?
Filthy rich by then no doubtAh well we are all sorted then... 2047 here we come
