And additional I got this answer
Alright, let’s imagine I’m Qualcomm, having just acquired Edge Impulse to boost my edge AI game. The question is whether I’d let BrainChip, with its Akida neuromorphic tech, still play a role in my plans. Here’s how I’d think it through, keeping it strategic and straightforward.
First off, as Qualcomm, my goal is to dominate edge AI—think IoT, phones, industrial gadgets, you name it. Buying Edge Impulse gives me a slick software platform to help developers build and deploy AI models fast, especially on my Snapdragon and Dragonwing processors. It’s a big win for making my hardware more appealing and accessible. But BrainChip’s Akida is a different beast—it’s hardware, specifically neuromorphic IP, designed to mimic the brain, sip power, and handle AI tasks right at the sensor. So, do I still need them?
Here’s the deal: Snapdragon’s my bread and butter—fast, versatile, and already powering millions of devices. It’s a general-purpose champ with beefy AI capabilities via its Hexagon DSP and Adreno GPU. Akida, though, is a specialist—ultra-low power, event-based, and great for niche edge cases like always-on sensors or tiny battery-powered gadgets. If I’m Qualcomm, I’ve got to ask: does Akida fill a gap Snapdragon can’t, or is it just overlap?
Let’s break it down. Snapdragon’s awesome for high-performance stuff—gaming, 5G, video—but it’s not always the leanest on power for, say, a smart doorbell that’s awake 24/7. Akida shines there, using spiking neural networks to process only what matters, keeping power under a milliwatt. Edge Impulse already supports Akida, so my acquisition means I’ve inherited that compatibility. Why not leverage it? I could integrate Akida IP into my portfolio—maybe not for phones, where Snapdragon rules, but for IoT or wearables where power’s a dealbreaker.
Strategically, BrainChip’s not a direct threat—they’re not churning out chips to rival me; they’re licensing IP. I could buy that IP, partner with them, or even acquire them outright if the price is right. Their neuromorphic tech could be a secret weapon for my Dragonwing line, aimed at industrial and embedded AI. Imagine a Dragonwing chip with Akida baked in—low-power AI at the edge, paired with Edge Impulse’s tools. That’s a combo that could crush competitors like Nvidia in niche markets.
But there’s a flip side. I’m Qualcomm—I don’t like relying on outsiders long-term. If Akida’s tech is that good, I might rather reverse-engineer the concept or build my own neuromorphic solution in-house. I’ve got the R&D muscle, and Edge Impulse gives me the software to optimize it. Why split profits with BrainChip if I can own the whole stack? Plus, integrating Akida might confuse my branding—Snapdragon’s the star, and I don’t want some quirky brain-chip stealing its thunder.
So, here’s my play as Qualcomm: I’d keep BrainChip in the mix for now. License their Akida IP for specific low-power use cases—think smart sensors, wearables, or edge boxes—while I study it. Edge Impulse already bridges us, so it’s low risk. Meanwhile, I’d task my engineers with figuring out if I can replicate or beat Akida’s efficiency with my own twist. If BrainChip’s open to a cheap buyout, I’d snap them up to lock in the tech and talent. But if they’re stubborn or redundant, I’d phase them out once I’ve got my own neuromorphic ace up my sleeve.
Bottom line? BrainChip’s got a role as long as they bring something Snapdragon can’t—ultra-low power edge AI. But as Qualcomm, I’m not here to play nice forever—I’m here to win. So, they’d better fit my plans or get ready to be sidelined. What do you think—should I keep them close or cut them loose?
www.forbes.com
