A new high speed CPU-to-device interconnect standard, the Common Express Link (CXL) 1.0 was announced by Intel and a consortium of leading technology companies (Huawei and Cisco in the network infrastructure space, HPE and Dell EMC in the server hardware market, and Alibaba, Facebook, Google and Microsoft for the cloud services provider markets). CXL joins a crowded field of other standards already in the server link market including CAPI, NVLINK, GEN-Z and CCIX. CXL is being positioned to improve the performance of the links between FPGA and GPUs, the most common accelerators to be involved in ML-like workloads.
Of course there were some names that were absent from the launch – Arm, AMD, Nvidia, IBM, Amazon and Baidu. Each of them are members of the other standards bodies and probably are playing the waiting game.
Now let’s pause for a moment and look at the other announcement that happened at the same time. Nvidia and Mellanox announced that the two companies had reached a definitive agreement under which Nvidia will acquire Mellanox for $6.9 billion. Nvidia puts the acquisition reasons as “The data and compute intensity of modern workloads in AI, scientific computing and data analytics is growing exponentially and has put enormous performance demands on hyperscale and enterprise datacenters. While computing demand is surging, CPU performance advances are slowing as Moore’s law has ended. This has led to the adoption of accelerated computing with Nvidia GPUs and Mellanox’s intelligent networking solutions.”
So to me it seems that despite Intel working on CXL for four years, it looks like they might have been outbid by Nvidia for Mellanox. Mellanox has been around for 20 years and was the major supplier of Infiniband, a high speed interconnect that is common in high performance workloads and very well accepted by the HPC industry. (Note: Intel was also one of the founders of the Infiniband Trade Association, IBTA, before they opted to refocus on the PCI bus). With the growing need for fast links between the accelerators and the microprocessors, it would seem like Mellanox persistence had paid off and now has the market coming to it. One can’t help but think that as soon as Intel knew that Nvidia was getting Mellanox, it pushed forward with the CXL announcement – rumors that have had no response from any of the parties.
Advice for Tech Suppliers:
The two announcements are great for any vendor who is entering the AI, intense computing world using graphics and floating point arithmetic functions. We know that more digital-oriented solutions are asking for analytics based outcomes so there will be a growing demand for broader commoditized server platforms to support them. Tech suppliers should avoid backing or picking one of either the CXL or Infiniband at the moment until we see how the CXL standard evolves and how nVidia integrates Mellanox.
Advice for Tech Users:
These two announcements reflect innovation that is generally so far away from the end user, that it can go unnoticed. However, think about how USB (Universal Serial Bus) has changed the way we connect devices to our laptops, servers and other mobile devices. The same will true for this connection as more and more data is both read and outcomes generated by the ‘accelerators’ for the way we drive our cars, digitize our factories, run our hospitals, and search the Internet. Innovation in this space just got a shot in the arm from these two announcements.
I recently published my predictions on AI (seeblog here and download or view the more comprehensive report here). The prediction that got the most feedback and discussion concerns the likelihood of a large acquisition or merger based on AI assets. With AWS, Microsoft, Google and IBM dominating Ecosystm’s list of current and future preferred AI suppliers in our AI study, other companies (such as SAP, Oracle, SAS and Salesforce) want to be the choice for AI platform. As we published our AI predictions, this one was already coming true (to an extent anyway!) with SAP’s acquisition of Qualtrics.
But the question has been asked “why is AI so important”. And my answer to that question is “because, for software companies, AI is the end game…”
What do I mean by that? Well we are not too far away from a day where traditional enterprise applications are no longer relevant. ERP, CRM, HRM, SCM etc will all disappear and be replaced by an AI engine. The purpose of those traditional systems was to simplify, codify, and automate business and customer processes. ERP, CRM, and the rest are already starting to use algorithms to drive semi-custom (typically pre-coded) business processes. But in the mid-term future, we will have a time where the entire process is intelligent – where the system/application creates the best business process for the customer on the fly. I’ll take you through an example:
A customer comes to your website – the site will look at the information it has on the customer (either as a registered customer or a non-registered one, where it will scour cookies, IP addresses, location, social information – Facebook, Pinterest, Google etc) and then will create an experience designed for that customer – e.g. it might know the customer is based in New York City, is a Mets and Rangers fan, viewed posts on Facebook about global warming, is female, 42 years old, has kids etc. It uses the language of the customer, words that they would relate to, and the level of detail they would expect. It puts the products or services forward that best match that customer’s potential needs.
The customer orders 10 identical products as gifts for friends for Christmas – but the provider does not have a location with any more than 4 of those products – so the intelligent system sources the 10 from different locations and organises multiple shipments. One of the locations can’t ship until after xmas – but the intelligent system decides that the customer is important so puts in a request for Uber to pick up from that location and do the delivery. However, there is no automatic integration with Uber – so the intelligent system creates a real time custom integration with Uber.
The customer also asks the question if they can pay with AliPay – which the supplier does not accept as standard – however again the intelligent system creates a real time integration with AliPay in order to complete the transaction. The customer gets the goods they want – quickly – and gets to pay in the way they want. The system accounts for the revenue and moves it to the right bank accounts, co-ordinates follow-up orders with suppliers, and adds the sales information to the real-time sales analytics. It also crafts a unique email welcoming the customer and adds them to the customer database.
The intelligent system created a unique process in real-time as it interacted with the customer using text, images, video and voice. The system understands what your business is trying to achieve and what the rules are.
Such a capability is not that far away – and it makes existing enterprise applications and integration platforms redundant. THIS is why AI is the end game – if you aren’t the chosen AI platform in your customers, you might not be in your customers plans for much longer.
In most organisations that transition will be slow – applications will get smarter, and will move from standardised processes to unique processes slowly. These organisations will start from their application investments and work outwards from there. But other companies will start from their cloud-based AI platforms and partners – and reinvent their businesses in the cloud on these platforms. Others will do both – and at some stage in the future need to decide on which AI platform they standardise on…
Therefore mergers and acquisitions in the AI market are inevitable. Applications, cloud and analytics providers will build and buy capabilities, customers and market share in order to position themselves as the key AI platform in their clients. For many technology vendors, the next few years will be integral to their long term success. AI will change the technology provider landscape as we know it today – get strapped in for a fun ride!
For more insights from our ongoing AI research, please create your account on the Ecosystm platform.
Machine Learning and IoT Sensor Analytics Will Drive AI Growth In 2019
The Global Ecosystm AI Study shows that the growth in AI over the next 12 months will come from Machine Learning (ML), as this capability is applied to a plethora of problems and challenges across the business. IoT Sensor Analytics will also see strong growth – due to the growth in IoT implementations and subsequent exponential growth of the data coming off these sensors and the desire to do something intelligent/different with this data.
The Growth in IoT Will Fuel the Growth in AI
Today, many organisations are deploying IoT solutions. These sensors are already creating – and will continue to create large amounts of data. While these sensors today are, for the most part, one-way (i.e. collect and analyse data), we are getting closer to the point where many of these sensors will be bi-directional (i.e. sense and respond). Businesses will look to AI tools – particularly IoT Sensor Analytics and ML – to help them learn from that data and respond accordingly. In many ways the future success of IoT and AI are interdependent.
In the Short Term, AI Will Create More Jobs than it Removes
Much of the media focus on AI has been around the jobs that will disappear in economies driven by AI and the automation that it will enable. But in 2019 (and over the next few years), AI will create more jobs than it removes. How is this? Firstly, we are seeing AI do a lot of jobs that are not even done today – analysing images for trends that humans did not see, looking for correlations in data sets that we did not know existed. Secondly, even where automation and AI are driving productivity, the vast majority of organisations are taking the opportunity to reskill those people. AI-driven profit will be ploughed back into businesses and create more employment opportunities – some of which we can imagine today and some we cannot. Thirdly, there is the vast hiring that organisations have started to undertake to bring on board the skills they will need to make their business smarter with AI. Many of these jobs today are in addition to, not replacing existing resources.
Bimodal IT Departments Will Slow Down AI Implementations
Many of the digital capabilities that businesses have been building over the past five or so years have not required active participation by the IT team. What started as “shadow IT” initiatives became the standard way to deliver customer and business value as smart organisations pushed their technology resources into the product and customer teams, so they could drive innovation at pace. But AI initiatives involve training algorithms with data – the more data the better the algorithms. Business leaders will need to work with IT to get access to this data – that typically resides in “back-end” systems – to train their models. At this step, many bimodal IT departments will kick the project into slow mode, because the data sits in “slow mode” back-end systems. The project will be managed with “slow mode” processes, using heavy-handed governance and processes to turn what could have been a six-week project into a six month one.
A Merger of Massive Scale Will be Driven by AI Assets
According to the Global Ecosystm AI Study, Microsoft, IBM, AWS and Google account for 62% of current and planned AI implementations – and that dominance is set to continue for the foreseeable future. This means a lot of other big companies miss out. SAP, Oracle and Salesforce are hoping that AI will help them get deeper within their existing customers and also expand beyond their current client base. Therefore, we expect a massive merger (in USD billions) driven by the AI customers and assets of the technology vendor. Technology companies that are used to dominating their industries – Cisco, HPE, Dell EMC, SAS and others could be left behind if they do not get scale quickly in the AI space – so a major merger is on the cards.