- To cut down on AI research process times, firms are investing in hardware – like coprocessors and GPUs.
- AI will help companies cut costs and improve revenues.
- VC’s don’t expect AI to slow down anytime soon, they are quite excited.
Imagine a time when computers start to beat humans at their own games, like world class grandmaster chess players, the best Go players or trivia champions. Well, this actually happened years ago.
With the power of artificial intelligence and big data, computers are able to solve complex scenarios like this much faster than humans. Watson and Deep Blue, developed by IBM, or AlphaGo, developed by Alphabet subsidiary DeepMind, were the systems behind these experiments and are likely the most well-known examples of artificial intelligence (AI). While not a small feat at their time, these experiments now represent a fraction of what artificial intelligence can accomplish today.
AI is coming of age, but its roots date back to 1666 when Gottfried Leibniz (a German philosopher, mathematician, and all around Renaissance man) theorized that all ideas are basically a combination of a small amount of concepts.¹ Similar to how humans and computers can recognize numbers. The number “8” for example is comprised of two little o’s stacked on top of each other vertically or how all of physical life is made up of the relatively small amount of elements in the periodic table. Breaking things down into its components, that being physical objects or numbers, is what Leibniz essentially theorized and how data scientists actually view and solve problems today.
This is relevant to AI because neural networks (one of the many AI research techniques) break down tasks into layers.² The computing power required to have a system run through these layers in order to analyze the possible combinations of the solution is massive. While the power required to train a program to read the number “8” is manageable, high-level problems like autonomous driving or AI driven medical research require a lot more resources. One of the fundamental reasons why AI is exploding of late is because of advances in microchips that divvy up these tasks better.
Stepping back, these chips are the MVP. Graphics Processing Unit (GPU) Accelerators allow data scientists to push through data sets at a much faster pace…so they can iterate on their program quickly. These chips are additions to your CPU (Central Processing Unit) and work to complement existing hardware, in most cases. AMD, Intel, and Nvidia are some of the key players in the chip market that have products specifically catered to computational tasks like this.
Nvidia, for example, suggests that their GPUs speed up deep learning training sessions 10-20x, which translates from weeks to days!³ These chips can be added to virtually wherever you want to compute; local workstations, data centers, and the cloud.⁴ While there is intense competition within the sector (coprocessors vs. CPUs vs. GPUs), these businesses have been a net winner for those investors seeking out hardware exposure with an AI flair. Continuing with the Nvidia example, their revenue growth has been astounding. The firm posted Q2 2017 revenue growth of 56% year-over-year⁵ versus the GPU industry’s global sales growth of 40%.⁶ Considering all the chips flying off the shelves around the world, there must be numerous user facing developments in the space.
AI Boosting the Bottom Line
AI touches end-users in a variety of ways, from Facebook being able to recognize users in newly posted pictures, to Amazon’s Echo Look + Style Check suggesting your optimal outfit. Now that there are all these apps, there must be a bigger strategy than creating a catchy app or Internet of Things (IOT) device that uses AI.
The thing is, the more engaging an app is, the more likely a user will share data with it to unlock its potential. Take the aforementioned Amazon Echo Look for example, at its core this piece of hardware is basically an internet connected personal photographer. But looking deeper, this device could drive some serious revenues. The app allows users to save outfits to your “Lookbook”, which Amazon will then analyze and suggest clothes you will probably like the next time you log in. To do so, the company said that the device uses “the best in machine learning and advice from fashion specialists” to optimize the user experience.
Thinking down the consumption chain, after a garment is ordered, the way goods from e-commerce retailers, like Amazon, will be getting to you will also be changing…to squeeze out more earnings. Just last month on the FedEx conference call, CEO Fred Smith stated that FedEx has, “some significant efforts underway in the evaluation of autonomous vehicles, but…none of these efforts are ready for prime time.” Autonomous vehicles like this will certainly require AI and vast amounts of research, but the main take away is that AI enables the potential for revenue expansion and cost contraction.⁷
Another indicator that AI is a hot trend and those in the space are being rewarded can be found in conference calls and YTD stock returns. Conference calls are essential storyline updates for analysts and investors, from those crafting the overall narrative…top corporate managers. These managers have a high level view of where the market is going. It’s interesting to note that the best performing tech stocks in the Nasdaq 100, (Micron, Nvidia, and Lam Research Corp – each up just about 80% YTD) all had managers that mentioned “Artificial Intelligence” numerous times in their latest conference calls, while the worst performing tech stocks in the index (Seagate, Qualcomm, and Akamai – down 10%, 19% and 24% respectively) did not mention AI at all. Akamai did mention machine learning once in the last call, but it was in reference to an immaterial startup they acquired.⁸
Another indicator of money flowing into the AI business comes straight from the wallets of big time investors, venture capital decision makers. In a recent survey of Silicon Valley VC’s, venture funds are “most excited” about AI when contrasted with the IOT, AR/VR, blockchain and e-commerce bots. When asked about AI, 58% of respondents said this was the most important trend for the next 5-10 years.⁹ The survey was courtesy of Upfront Ventures, an LA based VC firm, results displayed below.
AI’s Bottom Line
It doesn’t sound like AI is going to be a blip on the business radar. Unlike other VC fads like “daily deals”. According to Transparency Market Research, AI will be a $3T opportunity by 2024, growing at 36% CAGR.10 The fuel for such growth is also driven by the sheer amount of data creation, Goldman Sachs estimates that for the next 5 years, users will also be expanding on their data footprint by 36%.¹¹ This also implies that AI will likely become even more accurate with larger and more robust data sets, despite some firms doing more with less.
Geometric Intelligence, a recent acquisition by Uber, then appropriately renamed to Uber AI Labs, has its roots in smaller more efficient AI. The Geometric Intelligence team gained recognition in the MIT Technology Review after releasing XProp, a program that could read handwritten digits, “after seeing only around 150 examples of each digit, it would recognize only around 2 percent of new digits incorrectly,” while its competition needed to see 700 examples to reach the same error rate.¹²
Uber, the world’s most valuable private unicorn ($63B USD), clearly has a focus on AI with self-driving, mapping, carpooling and trucking goals. The company now has a seat at the AI table. Uber joins the exclusive ranks of Silicon Valley firms that have a dedicated AI lab…joining the likes of Google, Facebook Artificial Intelligence Researchers (FAIR), and Baidu’s Silicon Valley Artificial Intelligence Lab to name a few.
The best performing stocks in the technology market have been getting serious about AI for some time now, dedicating hundreds of employees to research. Consulting firm McKinsey & Company estimates that AI will drive $20-30B in spending, 90% going to R&D and 10% going to M&A. Considering the opportunity, there is reason capital is flowing into AI equities.¹³
1. Belaval, Yvon. Gottfried Wilhelm Leibniz. Encyclopedia Britannica, Inc. August 3, 2017.
2. Technopedia. What is the difference between artificial intelligence and neural networks. June 2, 2017
3. Huang, Jensen. Accelerating AI with GPU’s: A New Computing Model. January 12, 2016.
4. Nvidia Info Blog. Deep Learning Demystified.
5. Nvidia Corporation. MorningStar Financials.
6. Intel on the outside. The rise of artificial intelligence is creating new variety in the chip market,
and trouble for Intel. The Economist. February 25th, 2017.
7. FedEx’s (FDX) CEO Fred Smith on Q1 2018 Results – Earnings Call Transcript. Seeking Alpha.
8. Brief-Akamai buys Cyberfend in an all cash transaction. Reuters. December 19, 2016.
9. Suster, Mark. Upfront Ventures VC Survey 2017.
10. Global Artificial Intelligence Market: Market Expected to Cross Overall Valuation of US $3,-000 Billion by End of 2024, Observes TMR. September 13, 2017.
11. Artificial Intelligence: The APEX Technology of the Information Age. Goldman Sachs Research
Report. December 2016.
12. Simonite, Tom. Algorithms That Learn with Less Data Could Expand and AI’s Power. Intelligent Machines. MIT Technology Review. May 24, 2016.
13. McKinsey & Company Global Institute. Artificial Intelligence the Next Digital Frontier? Discussion
Paper June 2017.
Elite Wealth Management Team
Full Disclosures: http://elitewm.com/disclosures/
This article is not intended as investment advice. Elite Wealth Management or its subsidiaries may hold long or short positions in the companies mentioned through stocks, options or other securities. For a complete list of recommendations made within at least the past year, please contact us at (425) 828-4300 or firstname.lastname@example.org. Please Note: past performance is not indicative of future performance. Furthermore, we do not represent that the information contained herein is accurate or complete, and it should not be relied upon as such. Opinions expressed herein are subject to change without notice. Certain information contained herein (including any forward-looking statements and economic and market information) has been obtained from published sources and/or prepared by third parties and in certain cases has not been updated through the date hereof. While such sources are believed to be reliable, Elite Wealth Management does not assume any responsibility for the accuracy or completeness of such information. Elite Wealth Management does not undertake any obligation to update the information contained herein as of any future date.