Al Chip discoveries and development is one topic that has hit headlines this year countless times. Top websites like Google News, MIT Robotics, IBM News, TechCrunch and the most tantalizing of all Sanvada, have tried their best to deliver updates on everything around artificial intelligence.
However, there is one thing that has been going on behind the curtains that’s worth bringing to light, which is: startups have popped up across all fronts -as a result of the enormous demand for Al chips. In fact, it’s astonishing how AI has grown from being a theory to real-world application in the 21st century.
Industries with Higher Demand for AI Chips
The gaming industry tops in demand for AI chips, it wants the higher-level trainable software to make games more interactive. Data centers are also seeking neural networks that can help them arrange, store, process and avail information faster. The healthcare industry is also focused on optimizing patient records using AI, as well as robot doctors that would show higher precision and quicker diagnosis. While on the other hand, neural networks are being made more trainable by the memristors chip – which is intended to power fast-learning neural networks.
As of now, AI, in the area of deep learning is being used to helping Google users organize their photos, training digital assistants like Apple’s Siri, Amazon’s Alexa, Microsoft’s Cortana and the most eye-popping of all being helping self-driving cars see the world exactly like a human driver could see.
Interesting, right? This is happening right now in case it’s news to you. Where? Right here on earth! Just to tip you, Google, Telsa, Uber and other high profile car aficionados are pushing for large quantity production of autonomous cars, after successful tests -for public consumption.
As in, the demand for the AI chip to power things is beyond what a few companies can handle. Nvidia a firm that stocks and sells chips has had the best business ever this years but come 2018 which is a few hours from now, demands will hike higher. That is, all fields have their hopes on AI’s smart chips to be able to achieve their product goals.
Al Chip Startups Explosion
An array of startups saw the market need and have started working on their own version of hardware that can house top AI chip systems hoped to power future devices. What’s astonishing is that these startups have received enormous amounts of funding. Some of these startups have never shipped a product, but it’s so far open that they don’t need to worry about raising financing.
Let’s name them
Cerebras Systems started the trend by receiving a whopping $25 million from Benchmar Capital, but that was in December last year, meaning it’s possible that more funding was done in the cause of this year. Nonetheless, Graphcore made it clear that it got $ 50 million plus from Sequoia Capital as financing in November. Shortly after, Atomico pumped in another $30 million into Graphcore’s account making it a total of 80 million as financing in 2017 alone.
What sounds amazing is that as of now, both Cerebras Systems and Graphcore don’t have any splashy products compared to Nvidia. Well, before we proceed to other startups, it’s fair to mention that Nvidia prides of being the boss in AI chip production –exhibiting a company value of about $900 million.
Let’s proceed: Alibaba is on record for having financed Cambricon Technology –where unconfirmed memes say the funds raised the startup’s worth to $ 1 billion. Intel Capital, on the other hand, steered a $100 million investment into Horizon Robotics, while ThinkForce a growing startup managed to raise 68 million at the beginning of this month. Mythic, got its fair share of 9.3 million in financing, while Grog another startup interested in hardware development got $10 million from Social+Capital. Further details of these financings can be found on TechCrunch and Forbes.
Established AI innovators are also not doing nothing. Google is working on its TPU next-gen hardware after it’s done with the software. Apple is also on the verge of announcing how far they are with its GPU for the long-awaited next-gen iPhone devices. Siri, as it expounds in smartness may also require newer hardware -while Intel stated that the new Nervana Neural Network would be on shipment to clients anytime soon.
Top companies, as well as investors at Silicon Valley, are looking to buy faster chips in order to reduce machine training time and expenses, with the recently discovered memristors chip (which greatly condenses machine training time) being a great example.
Firms are also investing in AI made sensors, lower-powered chips and so forth, as in, technologies that can meander into the internet of things to truly fulfill the promise of super efficiency, and you know what, all these will require more startups -to meet the market demand for AI chip and their housing hardware.