Stock’s in the spotlight with lots of queries. Reposting teaser solution, but not updated since. Company had uneventful quarter, but stock surged in mid-Feb due to renewed A.I. enthusiasm. Recent chip innovations are still awaiting concrete updates.
NVIDIA has consistently been one of the top-performing big-cap stocks globally during the last seven to eight years, delivering a staggering 1,300% return in only in the last 5 years, 12,000% during the last decade. However, securing these profits would have involved weathering a 60% drop in 2019-2022, a difficult hurdle for many investors. That’s the claim: that we can expect similar gains from whatever hidden shares Kohl is promoting.
The so-called “urgent announcement” and “one-time opportunity” are likely exaggerated, especially considering this ad relies on data from a year ago, and I originally discussed this offer in January….However, time will tell. Investment teasers often create a sense of urgency with deadlines and catalysts to push people toward a subscription, the short-term pressure is more of a marketing tactic than a meaningful indicator.
Alright, the stock being teased? Let’s examine the clues in the ad Here’s how clues start to reveal themselves…
That’s simply image of typical office building in an industrial park in California., though we’ll find out later if it aligns with our clues.
As for the bold claims about potential profit, they mostly echo the usual ‘AI is going to be huge’ headlines that have been circulating over last year.
It’s likely Alexander Wang, a co-founder with Scale artificial intelligence. I believe he overtook Austin Russell of Luminar Technologies as the ‘youngest billionaire’ when his stake in Scale AI reached that valuation by mid-2023, only six years following his exit from MIT. Both leveraged the AI boom and an influx in venture capital to achieve ‘paper billionaire’ standing at the age of 25 to 26, although I believe Wang…hit that milestone several months earlier than Russell. Russell reached that mark a couple of years ago, soon after leaving Stanford. The takeaway here is, naturally, ‘leave a top 10 school if you want to become a billionaire early’… though this route could also lead to a prosperous career in selling used cars or fixing photocopiers, so make your decision carefully.
So what is the major opportunity? Kohl claims AI will revolutionize the world, he believes it’s facing a challenge that will leave it ‘incomplete’…
It seems that Kohl is referring to the bottleneck caused by the infrastructure and data processing overheating. This is a topic that was also brought up by Karim Rahemtulla in an AI presentation several months ago. Rahemtulla emphasized the importance for data centers adopting liquid cooling technology to manage the heat generated by high-end NVIDIA processors that power applications like ChatGPT. However, it appears that Kohl is approaching it from a different angle.
And some details…
Before we dive into that, it’s worth noting that by that time. NVIDIA had spent several years investing in AI technology.The company had long been a dominant and consistently profitable player in the high-end graphics chip market, primarily serving the gaming industry. While NVIDIA’s growth accelerated in the years that followed—As their autonomous vehicle chip technology advanced, a booming gaming sector, and increasing demand for data center acceleration—the company was already well-established, with a market cap nearing 10 billion dollars, a decade back, 10 billion dollars was still considered a substantial figure. Broadcom (AVGO) was slightly larger than NVIDIA, while Intel (INTC)…), the dominant force in the semiconductor industry, held a market capital of approximately $120 billion. Achieving NVIDIA-like returns doesn’t necessarily require starting with penny stock, though the prospect is certainly intriguing.
That has typically been the progression In semiconductor technology—new technologies and software are first powered by cutting-edge processors advanced chips. Once the demand is well-defined, the market reaches a scale that justifies custom silicon development. ASICs are far more cost-effective and more optimized than high-performance chips initially used for breakthroughs. A similar pattern played out in cryptocurrency mining: it began with standard CPUs, then shifted to NVIDIA’s graphics processors (GPUs) as competition intensified (since crypto mining is essentially a race for faster calculations), and eventually, as the market matured, companies developed specialized chips and equipment for mining, and those ended up dominating a significant portion of the market.
Will AI follow the same trajectory? It’s hard to say for certain, but likely—eventually. With billions flowing into AI development, custom ASICs will likely be funded, especially given the diverse range of artificial intelligence computation. The market could fragment as various hardware solutions emerge to tackle different tasks.
We’re not at the stage of custom AI ASICs yet, it seems he’s referring to a chip firm with an alternative architecture designed to tackle memory constraints and handle the high data throughput demands of AI…
And what is the ” 1 time opportunity” promoting?
Any other clues? The company’s shares recently “soared 512%,” yet insiders didn’t sell, retaining 31% ownership in the company.
Plus, it’s currently trading for “a 2 of bucks.”
This one wasn’t too hard to piece together… and it triggered a distant memory from the archives of Your go-to source for uncovering stock teasers neighborhood Gumshoe. This stock was originally hyped over ten several years back defunct investment newsletter. It appears that Kohl is likely pointing to GSI Technology —a company Lou Basenese was pushing in 2011. A small-cap stock valued at $200 million, renowned for its innovative memory chip architecture.
GSI Technology is still in the SRAM chip business, but while its revenue peaked at around $100 million in 2011, the company has been on a slow downward trajectory ever since. Its valuation has dropped significantly from that $200 million level and now sits well below valued at $152 million that Kohl hinted at—currently hovering around $90 million.
I tend to be wary of these small, niche tech firms that make big promises but have little to show for them. Maybe I’m being overly critical of GSIT, but if we examine its financial track record over the past two decades, the outlook isn’t exactly promising—revenue (purple) has remained flat, while both operating cash flow, net income have consistently been in negative territory.
Small chip companies often lose money when they first start out, and that’s what happened with this company. However, this company has been around for 20 years and while they haven’t burnt pouring nine-figure amounts into operations, they have still struggled to make a profit. They have managed to stay afloat by consistently generating revenue, they haven’t continuously issued a large number of new stocks just to stay afloat.
Recently, they initiated the development of new products for the military and satellite sectors, as well as AI. This has been costly, and they’ve had to issue more shares to raise money. While this hasn’t been too overwhelming so far, their share price has been falling, which become a problem. If they have to keep selling shares at progressively declining prices just to stay in business, it could be difficult for them to recover. Biotech stocks experience this constantly, but the same risk applies to any small, unprofitable company.
The stock price of GSI Technology (GSIT) reached $9 earlier this year, which was a significant increase from its previous value of $1.45. However, this increase was fleeting and did not help the company raise any significant funds. Currently, the stock price of GSIT is around $2.50, which is far from the teased market capital of $152 million by Kohl. The advertisement from Kohl’s side is widely circulated at the moment, and this version of it is the first time I have seen it. It is possible that the ad was written over the summer when the stock price increased. According to the company’s last quarter report, they expected a cash outflow of $14 million in 2023. If things do not improve, they foresee having sufficient funds to operate for “a minimum of a few years.”. The company believes that its most recent AI chip design will help it reach the major leagues
With SRAM purchases significantly reduced, they present themselves now:
They are currently working on Gemini II, a product with larger volume, and designing the Gemini III to enter the market for data centers. However, during the last conference call in October, they did not seem to consider the timeframe for this project to be particularly urgent.
Will the Gemini chipset play a significant role in the evolution of AI, especially in Edge AI? Hard to say—maybe. But it’s still early days, plenty of chipmakers are developing AI processors for different use cases, including low-power options such as GSI’s Gemini. Perhaps their memory integration is a game-changer, but I can’t say for sure—That is not the type of debate I’m qualified to judge while it comes to determining the best chip architecture. I’m neither a data analyst nor a semiconductor engineer, so in cases like this, I rely on real-world adoption. That means watching for orders, sales, revenue figures, as well as GSI Technology isn’t there yet.
Well, the process will likely unfold gradually—chip development is a slow and iterative journey, often much slower than investors expect given all the headline buzz. The chip fabrication step is a significant milestone for chip designers, marking the transition from a digital blueprint to a physical design, allowing for the creation of photomasks and the begin of actual manufacturing. However, production doesn’t start immediately—it still has to pass through fabrication, testing, packaging.
Right now, they’ve manufactured the initial Gemini II processors, mounted them on boards, and begun initial testing. This phase will take time as they internally verify performance, troubleshoot errors, and refine design challenges. The following step would likely involve getting key customers to test the chips in real-world conditions. Design tweaks before mass production wouldn’t be surprising—this could take some months or even years.
Six months back, they projected benchmarking data would be available “around summertime,” with earnings generated by Gemini II will be recognized in their fiscal year 2025, starting in April 2024. However, as of their last earnings call (July 2024), their goal remained to showcase Gemini II’s capabilities by December’s end, with hopes of securing orders and drawing hyperscalers to co-invest in Gemini III.
Keith Kohl suggested a potential “parabolic” moment for the Gemini II this fall. However, delays are common in this industry. The tapeout itself arrived concerning a year later than expected, and the chip still isn’t finalized for partner testing. As the company put it in their latest update:
They remain optimistic that their high-speed, low-power chip will attract strong demand in certain AI applications, particularly in Edge AI processing. The goal is to market Gemini II processors and computing systems for these use cases, while the still-theoretical Gemini III could eventually target more complex AI workloads, such as AI-driven language models akin to ChatGPT. However, bringing Gemini III to life would require substantial funding and a strategic partner to scale production. For now, much of this remains speculative, with only Gemini I being commercially available.
Investors shouldn’t expect a sudden influx if purchases or profits in the next earnings report, likely due in the last days of October. Revenue is projected to continue its downward trend, with Nokia (their largest customer) and defense-related contracts making up over half of total sales. Most of this revenue still comes from traditional SRAM chips and government research grants, rather than commercial AI deployments. Defense and satellite clients, drawn to the company’s radiation-hardened memory products, continue to be a core part of their customer base. Although Gemini I was built to withstand radiation, its adoption has not yet led to substantial revenue growth.
In essence, this is a small semiconductor company with an AI chip vision but no clear breakthrough in large-scale sales. Their pivot toward AI began in 2019, leveraging technology acquired in 2015, while Gemini I launched in 2020, Gemini II is still in its early testing phase, with potential customer evaluations slated for late 2024. Progress has been slow, and while updates on Gemini II’s development may arrive before year-end, the company remains a cash-draining operation.
Whether Gemini II proves to be a game-changer or if Gemini III ever comes to fruition is uncertain. Investors will need patience, as the company’s AI ambitions have yet to materialize into real revenue traction. For now, it’s an intriguing concept, but the path forward remains unclear.
What’s your take? Do you see Gemini II or III making a dent in the AI chip market? Are there other semiconductor firms with a more defined route to success? Share your thoughts in the comments below.
