Alarum Stock: Data Collection Is Key For Enabling Gen AI (NASDAQ:ALAR)

In addition to providing IT security solutions, Israel-based Alarum Technologies (NASDAQ:NASDAQ:ALAR) has diversified into data collection solutions. This is at a time when massive amounts of data are needed to train artificial intelligence models.

Thus, it is unsurprising that its shares delivered above 1000% upside during the last year as shown in the Orange chart below. It has even outperformed Nvidia (NASDAQ:NVDA) and Microsoft (NASDAQ:MSFT) by a hefty margin and this thesis aims to show that this should continue.

At the time of writing it had dipped to about $33.58, constituting a buying opportunity based on its ability to have rapidly developed three products for data collection while at the same time boosting sales from its NetNut IP proxy solution by double-digits. It is also making progress on the profitability front.

First, I provide an overview of where it fits in the AI story.

AI Needs Data Collection

After an enterprise has purchased Nvidia’s accelerator GPUs and hosted them in its data center or subscribed to Microsoft’s intelligent Azure cloud to consume AI in subscription mode, the next step is to build AI models, termed the training phase. Now, the problem is all the data required is not in one place and while some of it may lie in corporate databases others are located on faraway websites.

This implies the need to collect data from different places on the internet. Moreover, since AI models developed tend to be dynamic, the collection is a continuing process to keep the information up-to-date. For this purpose, the company introduced its AI Data collector product line in February.

Looking deeper, this is not just a passive tool for navigating the web and collecting data but more of an instruction-driven product to access websites to extract specific parameters to train a model. Additionally, it automatically translates the unstructured website data into a structured format before feeding them to clients’ AI algorithms thereby facilitating the data analysis part while reducing cost since this is normally a manual process involving several man-days of work.

Now, the first version of the AI Data Collector should be launched by the end of 2024, but in the meantime, the company is generating revenues mostly through NetNut. This is Alarum’s subsidiary operating as an IPPN (IP Proxy Network) infrastructure provider for enabling enterprises to access the internet through a secured proxy network. Talking adoption, given the rapidly evolving and tough cybersecurity landscape, IPPN got rapid traction from 2018 to 2019 with customers coming primarily from the eCommerce, data collection, ad verification, and social media verticals as shown below.

Thus, $8.4 million of sales were generated in the first quarter of 2024 (Q1), or a 47% YoY surge, with its track record showing revenues accelerating by 78% CAGR over five years. Therefore, far from being a startup developing a new technology, this is a business already growing at double digits, and, more importantly, is building the AI collector solution as a revenue diversifier while leveraging its market positioning as an IP proxy.

Valuing as a Data Collector for AI Model Training

In this context, considering data collection is key to successfully deploying AI projects and with the cost of training a single innovative model can rise to $30 million depending on the size of the data sets, it is not unrealistic for clients to spend money on Alarum’s data collector especially as it comes with automation features which reduce labor-intensive tasks at a time of scarcity in data engineers.

Moreover, till its AI data collector hit the market, the company launched SERP (Search Engine Results Page) Scraper in October last year. This is an API (application program interface) or piece of software code that enables the collection of real-time structured data from global search engines depending on how it is customized for corporate needs.

Moreover, the fact that SERP is already getting traction shows that Alarum has penetrated the data collection and labeling market, valued at around $2.57 billion in 2024, and is expected to grow by 18% CAGR by 2034. Interestingly, this estimate only includes previous flavors of AI like machine learning, not Gen AI. Moreover, there are different players already providing services and Alarum aims to play a disruptor role versus conventional web-based data collection methods.

Therefore, assuming Alarum manages to sell ten data collectors worth $1 million each next year, this totals $10 million, which would only represent 0.4% (15/2570) of the $2.57 billion market. However, when added to the $45.90 million of projected revenues for FY-2025, this could more than double the YoY growth from 25% to 52% as charted below. The incremental revenues should reduce the forward price-to-sales ratio of 4.05x to 3.33x (45.9/55.9 x 4.05), or a decrease of 18%.

Conversely, this means a higher share price and I have a target is $39.6, obtained after incrementing the current share price of $33.58 by 18%.

Noteworthily, this is a long-term target and also depends on the product being available by the end of this year or the beginning of 2025. Noteworthily, as a revolutionary product with no code (no need for software coding), it will have to be beta-tested before being deployed on clients’ premises.

Therefore, in the meantime with an RSI of above 65, the stock could be impacted by volatility especially if the trailblazer of AI, Nvidia were to suffer from significant outflows as was the case recently as per the introductory chart.

Deserves Better Valuation as an AI Enabler with Three Products for Data Collection

On the other hand, the stock remains well above the 100-day average which means upside potential. To further support the bullish thesis, Alarum’s forward P/S of 5.06x exceeds the IT sector median by 77% but its forward revenue growth outshines by 428% or more than five times the percentage by which it is overpriced signifying it deserves its higher valuations.

Second, its P/S is less than one-quarter of Nvidia’s forward price to sales of 25.81x. Therefore, Alarum may not necessarily be overpriced from the perspective that both companies are Gen AI enablers in addition to belonging to the IT sector, the semiconductor giant with its accelerated computing, and the Israeli company for the data collection part. Thus, it may be because of its lower multiples relative to Nvidia, that Alarum is outperforming as per the introductory chart.

Pursuing further, in addition to SERP and AI Data Collector, as I mentioned earlier, the company also launched Website Unblocker in January this year as pictured below.

Now, as its name suggests, the Unblocker tool unblocks access to sites that have been restricted as the problem encountered during data collection is not all admins are allowing AI crawlers to scan their websites to extract the data. The reason is they want to protect their intellectual property or monetize the ad traffic instead of letting the internet link become congested by AI crawlers fishing for data.

Therefore, just having data collection technology is insufficient and with its background in the IP proxy world relying on a network built for years and composed of servers located throughout the world, Alarum has positioned itself as a key player in collecting data through three products to feed to AI models for training purposes.

A Demonstrated ability to Monetize and progress on Profitability but not without Risks

Looking at the monetization part, the company has been able to upsell and cross-sell its products to its existing customer base as with Website Unblocker launched in January 2024. This is reflected in the NRR or Net Retention Rate of 1.66 (166%) which takes into account revenue expansion from existing customers minus churn. Now, since an NRR of 130% is considered exceptional, 160% means excellent customer satisfaction and potential for organic growth. I also verified that innovation is attracting new customers.

Furthermore, the company is making progress on the profitability front after gross profits increased by 75% to $6.5 million mainly because it has been scaling down its less profitable Cyberkick consumer business. This also resulted in a reduction in operating expenses by 6% YoY despite bearing more costs related to scaling the enterprise business. Thinking aloud, this strategy to focus on enterprise business puts it on a strong footing to continue on its profitability journey after breaking even in the March quarter of 2023. Moreover, its cash position advanced to 15.1 million from $10.9 million while debt amounts to $1.9 million.

On the other hand, the company has been selling equities as shown in the chart below. Well, now that it is profitable and has started generating positive cash flow from operations (of $4.7 million) in FY-2023, there is less need for equity issuance, unless it opts for an acquisition.

Finally, this thesis has made the case for an investment in Alarum with the potential for an 18% upside due to opportunities in data collection for AI. This may represent a moderate target. However, it is justified in a context where companies are spending billions of dollars on Nvidia’s GPUs, meaning there should be no shortage of competitors that are likely to emerge including existing providers reorienting their product line. Thus, in case it faces competitive pressure, Alarum may have to spend more on sales and marketing which could then impact profits forcing it to raise capital through equity issuance thereby diluting the shares.