You will be spoiled for choice when looking for stocks to take advantage of the booming artificial intelligence (AI) market. Almost all the well-known tech giants -- including NVIDIA, Intel, Amazon, Alphabet, and many others -- are betting big on this fast-growing field, as they don't want to miss out on an opportunity that could be worth a total of almost $60 billion by 2025.
But these aren't the only ways to take advantage of this space. Lesser-known stocks like Xilinx (XLNX), Ciena (CIEN 1.27%), and CEVA (CEVA 0.96%) could win big from the AI revolution. Here's how.
Xilinx makes programmable logic devices used across several growth segments such as data centers, automotive, and industrial. One class of its programmable logic devices, known as field-programmable gate arrays (FPGAs), has been gaining tremendous traction of late because it can train deep-learning models for artificial intelligence applications in data centers.
It is commonly believed that graphics processing units (GPUs) from vendors like NVIDIA are best suited for training deep-learning models in the cloud since they have massive computing power. FPGAs, however, are considered better than GPUs for certain applications.
For instance, Xilinx's FPGAs deliver more performance for each watt of power consumption as compared to GPUs. This makes them ideal for deployment in large-scale data centers where energy savings are important. Moreover, the programmability of FPGAs allows them to be optimized for carrying out different varieties of workloads, something that isn't possible for a GPU.
Because Xilinx's FPGAs can accelerate the workload in a data center while consuming less power, they are being adopted by big cloud-computing players. Huawei, for instance, recently announced a cloud server platform for North America based on Xilinx's FPGAs. Huawei claims that this FPGA-enabled platform can speed up the computation of intensive cloud applications by 10 to 100 times for machine learning and video processing, among others.
This isn't Xilinx's only win in the cloud computing space -- Alibaba and Amazon have already selected its FPGAs to accelerate their cloud services. Moreover, FPGA adoption in data centers is expected to exceed the growth of GPUs and central processing units (CPUs), according to Allied Market Research.
Xilinx is in the running to take advantage of the rapidly growing AI chip market that could be worth $11.1 billion in 2023 as compared to just $661 million in 2016. More importantly, Xilinx could carve out a sizeable share of this market since it has established a clear lead over its nearest rival -- Intel -- in the FPGA space with a 59% market share.
Ciena sells fiber-optic communication equipment. Its clients include telecom carriers and cable companies, which use Ciena's equipment to enhance connectivity speeds. The company operates in a tough industry where the competition is stiff and business is driven by the spending budgets of its clients.
For instance, Ciena's rival Infinera has been experiencing curtailed demand from its clients for its multiplexing equipment, which allows carriers to enhance network capacity without laying additional optical fiber networks. Instead, optical fiber networking clients have diverted their spending toward data center interconnection (DCI) equipment, which allows two or more data centers to share resources and workloads so they can be scaled up as required.
Demand for this DCI equipment is expected to increase at an annual pace of 10% over the next six years to an anticipated $6.4 billion in 2023. This isn't surprising as DCI will play an important role in tackling fast-growing data traffic and meeting increased cloud computing demand.
Ciena is looking to take the lead in this market with its WaveLogic Ai modem that's aimed at making networks more intelligent and flexible as needed. According to Ciena, WaveLogic Ai is an intelligent, autonomous, and self-driven networking chip that can be used by network operators and cloud service providers of all sizes to scale channel capacity while controlling costs.
A cloud service provider can deploy the WaveLogic Ai chip for both short-reach metro connections, which are used for short-distance data transmission, as well as submarine networks that cover thousands of miles. The intelligent chip will automatically decide the capacity required to transmit the data through the optical network, scaling the network as needed so that the remaining network capacity doesn't go waste, and can be used by either Ciena or third parties.
Ciena expects that sales of this WaveLogic Ai chip will grow substantially in 2018, and it should help the company maintain its leading position in the growing DCI market.
CEVA is an intellectual property (IP) vendor known for licensing its digital signal processors (DSPs) to larger chipmakers. The company licenses DSPs for several applications like imaging and computer vision, connectivity, audio, voice, speech, AI, and deep learning. Within the AI and deep-learning space, CEVA is focusing on fast-growing markets like automotive, augmented reality, and smart homes, among others.
The company also licenses its deep neural network frameworks to chipmakers for enabling AI applications in their devices. Apple, for instance, is reportedly using CEVA technology to power the much-hyped Neural Engine inside the new iPhone's A11 Bionic processor, as reported by Charlie Demerjian of the technology news site SemiAccurate.
The financial gains of this supposed partnership will only become evident when CEVA reports its next set of quarterly results on Feb. 6. If the company has indeed licensed its intellectual property to Apple, its royalty revenue from audio, imaging, and vision products should increase, as pointed out by Fool.com's Ashraf Eassa.
CEVA's potential win at Apple could be a sign of bigger things to come for the company. It's likely that Apple has decided not to use Qualcomm's DSPs because of its feud with the chipmaker, so CEVA could be a long-term beneficiary. More importantly, CEVA has a pretty huge list of licensees that use its designs. The company includes the likes of Cypress, Intel, Broadcom, among many others, as its publicly announced licensees.
The semiconductor IP market could be worth $6 billion in 2023, according to Markets and Markets. CEVA provides chips for the AI space, so it has put itself in a strong position to take advantage of this multibillion-dollar opportunity.
CEVA has generated $87 million in revenue over the past four quarters, and it is growing at a terrific pace. Last quarter, its revenue jumped almost 35% year over year, boosting its earnings by 72.7%. So investors should definitely watch this chip licensing play.
The Foolish takeaway
Xilinx, Ciena, and CEVA are taking advantage of AI in different ways. More importantly, they are being successful to different degrees with the help of this emerging technology. Investors shouldn't expect quick gains and should remain patient, as the AI chip market is still in its early phases, but these three companies are well positioned to take advantage.