As I shared yesterday, today I’m “unlocking” some premium content from behind the Brownstone Research paywall. Below, I’ll share three of my favorite companies leveraged to the adoption of artificial intelligence.
I’ve included my full analysis on each stock including a recommended stop loss and a profit-taking level. If readers decide to act based on this analysis, remember to keep a close eye on these levels.
Like I said yesterday, there are essentially two categories of AI investments: innovators and adaptors. Both categories present great potential as the AI megatrend takes shape in the years ahead.
And the first company is…
Advanced Micro Devices (AMD)
Next Earnings Announcement: August 2, 2023
Revenue Guidance: $5 billion – $5.6 billion, next quarter
Operating Income: $40 million – $1.2 billion, quarterly
Nvidia is – rightfully − known as the gold-standard for AI training. But for a market opportunity as big as artificial intelligence, second place isn’t a bad position to be in. That’s where AMD finds itself.
AMD doesn’t have to dethrone Nvidia to have a great AI business. There’s enough demand for alternatives, especially considering Nvidia’s steep margins and desire to lock customers into its ecosystem.
AMD is responding to Nvidia with the MI300 later this year. This is an innovation in chip design, architecture, and capabilities. The Accelerated Processing Unit, or APU, combines a GPU, CPU, and memory on one chip.
The MI300 design has a few advantages in AI workloads given the fact it reduces the space footprint and energy to run tasks. That saves on the overall cost to start and maintain a data center.
Given it takes several years of research and development to bring a GPU to market − it’s likely that Nvidia and AMD could have the market largely to themselves through 2024.
Estimates have Nvidia capturing 60-70% of the AI server market today. The companies that don’t use Nvidia largely rely on custom silicon designs that aren’t available for sale.
The potential buyer for the MI300 is still narrow. Industry estimates are for 70,000 MI300’s to ship in 2023 with 40,000 of those to go in the El Capitan high-performance computer.
The El Capitan computer is a U.S. government contract which likely carries lower gross margins but gives AMD a good way to demonstrate performance. By comparison, Nvidia is likely to ship 400,000+ H100 GPUs in a single quarter.
However, I suspect that Amazon could be a potential large buyer of MI300 for its data centers. That would be significant to AMD from a revenue perspective. Given that Amazon has its own AI ambitions, the rumor likely isn’t far-fetched.
AMD generates a $5B – $7B quarterly revenue run rate across all products. MI300 adoption from the likes of Amazon or Meta could dramatically increase this amount. That would drive the share price even higher.
And it’s worth mentioning that AMD trades at a relative bargain compared to NVDA. The price-to-sales ratio for Nvidia hovers around 41 today. In other words, investors are pricing the stock at 41 years of current sales. Maybe Nvidia can grow into that valuation. But on the flipside, any slight misstep will send the stock tumbling.
Meanwhile, AMD trades at a P/S of 7.7. That’s in the middle of its long-term historical range. And it’s significantly cheaper than the 12.3 ratio the company carried in late 2021. We like what AMD is up to as an innovator in this space.
Nvidia may be the king of AI chips, but there’s still plenty of money to be made if AMD can solidify its position in second place.
Target price: $157
Stop loss: $87.49
Next Earnings Announcement: September 15, 2023
Revenue Guidance: $19.3 billion (10% growth), full year
Operating Income: $1.5 billion – $1.6 billion, quarterly
I first started seeing text-to-image generators like Midjourney in 2022, which can create life-like images with a simple text prompt. And I was immediately concerned about Adobe’s core business.
Generative AI like Midjourney uses natural language − or written instructions − to manipulate and create images. For example, you can simply type “show me a cat eating cake in the woods” and an AI generated image is created.
Other use cases allow users to edit or alter an image such as changing the hair color or background of an image −simply by instructing AI. The results are extraordinary.
An AI-Generated Image
Just a short time ago this required more sophisticated software like Adobe’s photoshop. Initially, I wasn’t certain how quickly Adobe would adjust its legacy software to stay relevant.
To Adobe’s credit, it has responded quickly rolling out a beta version of Adobe Firefly in March. This is essentially a Midjourney-like product that generates and edits images based on natural language input text. Then in May, Adobe rolled out Generative Fill on Photoshop, adding some editing functionality using natural language.
These solutions help solve two issues for Adobe.
First, it’s addressing competition so that Adobe still maintains industry leading capabilities. Competition is still going to chew into some of Adobe’s market share − however the company will retain the enterprise accounts. Most companies won’t risk running afoul of copyright laws or having buggy programs from young start-ups.
Second is the fact many Adobe products have a steep learning curve. The advent of features that use natural language to perform tasks will help Adobe gain new users and lower the learning curve.
The large advantage Adobe has over competition is the large stock images it owns through an acquisition in 2014 − and continuing to invest in the image database. Enterprise customers are going to be cautious using generative AI images without the proper copyrights and permissions.
Additionally, Adobe is working with specific enterprise clients on a version of Firefly that uses custom generative AI models tailored to the specific company. Here’s what Adobe CEO Shantanu Narayen said on its second quarter conference call about this product:
The product is not yet available, but the number of customers who want to sign on in terms of wanting to use it has been one of the most successful launches that we’ve had.
This supports the thesis that Adobe is likely going to gain enterprise clients because open-source or alternative generative AI carries legal risks these organizations can’t take. Additionally, if Adobe is able to give white-glove type support to large clients creating custom generative AI solutions, it further gives the company a competitive edge.
Adobe reported second quarter earnings on June 15. Revenues topped expectations and grew 9.8% year over year.
Guidance for the third quarter and the full-year on the revenue side were in-line with expectations of roughly 10% revenue growth for the full-year.
Some concern was expressed on the conference call by analysts that the investment into generative AI features would chew into operating profits. However, management at Adobe indicated they would use discretion when it came to spending.
My projection is Adobe will comfortably meet the ~ 10% revenue growth target with room for upside if enterprise adoption of generative AI tools accelerates.
Target price: $573
Stop loss: $331.16
Next Earnings Announcement: August 2, 2023
Revenue Guidance: Gross Bookings $33 billion – $34 billion, yearly
Adjusted EBITDA: $800 million – $850 million, yearly
Uber’s ability to increase revenue and profits is partially limited due to competition domestically with Lyft. Both companies are constantly competing on price, driver incentives, and perks.
Ideally, Uber continues to squeeze Lyft operationally − which will force Lyft to make adjustments to pricing since the company is not close to being profitable.
The tailwind of Lyft having to eventually generate a profit will benefit Uber since the company has a more diversified business and has already achieved adjusted profitability.
In terms of Uber’s exposure to artificial intelligence, I view it as an adopter. Specifically, the opportunity is to adopt self-driving technology into its cars.
But while self-driving technology is an opportunity, it’s also a risk. Uber attempted its own self-driving R&D only to have lawsuits and tragic accidents mar its development. Uber sold its self-driving unit in 2020 to Aurora. This has improved the profitability of Uber as the R&D spend was a drag on earnings.
But here’s what’s interesting…
While Uber may have sold its self-driving unit to Aurora, the former received a 26% stake in the latter. In other words, Uber still has a connection to self-driving technology through its partial ownership of Aurora. And Uber has two seats on the Aurora board. Aurora, for its part, inked deals with Toyota and Denso to develop self-driving Toyota Sienna minivans.
I’m reading the tea leaves a bit here, but if the Aurora/Toyota/Denso initiative is successful in the years ahead, an obvious way to commercialize this technology would be to begin offering self-driving ride hailing services. And who better to partner with to accomplish this feat than Uber?
In the shorter term, Uber’s greatest asset is the CEO Dara Khosrowshahi. He moonlighted as an Uber driver for several months to learn more about the company. He found that many of the complaints by drivers were valid and made a number of enhancements to improve driver retention and morale.
The changes he made, including tip-baiting, and allowing drivers to choose a geographic area to work were well received by drivers. These changes are credited with helping Uber increase market share against Lyft.
Above all, Uber’s business is scaling. In the most recent quarter Uber grew revenue 29% while sales and marketing spend was flat. An impressive feat, which highlights operating efficiencies and the Uber brand.
Overall Uber fits in well with the idea that “Gold Medal” stocks perform well. Gold Medal stocks is my term for market leaders. Like Google in search, Apple with smartphones − investing in the clear market winner is a great investment. Uber is well on its way to dominating ride-sharing while also creating meaningful business models around food delivery and freight.
Target Price: $46.54
Stop Loss: $29.02
Editor, The Bleeding Edge