A Japanese Nano-Cap for 3x FCF
The attractive business model and ownership structure is also appealing to activists.
Get a 90% discount off Bloomberg’s No.1 competitor. Get a 30-day free trial here
Silicone Studio (3907) is a technology and HR business founded in 1999.
They’re headquartered in Tokyo, Japan.
At today’s price the market is implying that the operating business is in terminal decline.
It also indicates that all the cash on the balance sheet is inaccessible or will be destroyed by management.
The reality of the business is starkly different.
It’s a profitable, capital-light technology provider, currently expanding its revenues into extremely durable markets.
Here are the valuation ratios:
NCAV Ratio = 1.5
TBV Ratio = 1.3
EV/FCF Ratio = 3.2
P/FCF Ratio = 9.4
The market cap is currently ¥1.88b and the enterprise value is ¥640m.
As always, I calculate the EV using all debt-like figures, including leases, and all cash and marketable securities.
The 5Y average FCF figure is ¥175m and last year it generated ¥286m.
I calculated FCF by starting with the OCF and then subtracted all forms of unavoidable business costs.
This usually includes items not regularly included in OCF, such as lease costs.
After researching the annual reports from the last few years, I believe it’s reasonable to expect an average annual FCF of roughly ¥200m.
The reason for this is that the company conducted a merger with one of its subsidiaries in 2024.
The purpose of this was specifically to optimise the efficiency of the operating business.
Lower costs and better margins.
They are also successfully expanding their expertise out from their traditional markets.
These improvements contributed to last year’s FCF figure being higher than the average.
It seems reasonable to expect this uplift to continue, even to a small degree, going forward.
Stock Price Decline
There are a few things not to like about the business.
First, it’s a typical Japanese cash-hoarder. The net cash represents 65% of the total market cap.
Next, management haven’t exactly been generous with their capital repayments.
The average yield over the last 5 years is practically 0%.
They have initiated a dividend going forward, but this offers a yield of roughly 1.4% at today’s market cap.
Another issue affecting the business is a severe lack of IT talent in Japan. This has been a headache as the business tries to compete.
The main reason the stock price has fallen almost 50% in the last 5 years, is that it was simply too expensive to begin with.
The actual business model is pretty sweet.
The market got too excited about it and, at one point, applied a PE multiple of 32.
The recent sell-off is just all the hot-air being released from the overinflated stock price.
Today, the price can now be considered cheap, relative to that sweet little business model it still owns.
The Business Model
Silicon Studio operates a highly efficient B2B model.
They have two core revenue segments.
The first one is called the Development Promotion and Support Business (DPSB).
DPSB generates roughly 62% of the revenue.
They basically supply the picks and shovels for digital content companies.
This sounds a bit vague, so here are a few examples to help illustrate exactly what they do:
They create advanced software tools (called “middleware”) that plug into game engines to make digital worlds look incredibly realistic.
For example, they offer their “Enlighten” software.
In the real world, light bounces off walls and softly lights up a room.
Calculating this indirect light in a video game is incredibly difficult, but “Enlighten” calculates it instantly, allowing game developers to easily create natural, beautiful lighting.
Another example is “YEBIS”.
This tool takes a flat computer-generated image and adds realistic camera effects.
This might be things like the dazzling glare of the sun, the soft blur of a background, or the slight distortion of a camera lens.
It basically makes a digital image look like a real photograph.
There’s also “Bone Dynamics”.
This tool adds realistic physics to digital characters so that things like clothing and hair sway naturally as they move, without glitching or passing through the character’s body.
The cool thing about this technology is that it isn’t limited to video game developers.
They use their high-end video game technology to help real-world industries, like automotive, manufacturing, and construction to solve practical problems.
One example of this is when a company needs to train a custom AI model for a specific environment.
For an Artificial Intelligence (AI) to learn how to drive a car or inspect a factory part for scratches, it needs to look at thousands of images.
Instead of forcing companies to take thousands of real photos, Silicon Studio generates massive amounts of fake, highly realistic 3DCG images for the AI to study.
They create fake roads for self-driving car simulations, fake human faces for driver-monitoring systems, and fake defects so factory robots can learn to spot them.
Another example is from the world of civil engineering.
They take “point cloud data” (measurements taken by 3D laser scanners in the real world) and turn it into highly detailed 3D virtual environments.
This allows construction companies to run safety simulations, review building plans, or check the progress of a construction site inside a virtual space.
Finally, they also provide and manage all the servers that the online gaming companies need.
When a popular online game launches, it needs massive servers to handle the web traffic.
Silicon Studio designs, builds, and operates server networks capable of handling tens of thousands of players all connecting at the exact same time without the game crashing.
They also provide support for this 24 hours a day, 365 days a year.
The second business segment is called the Human Resources Business.
This is responsible for 38% of revenues.
This is a highly specialised recruitment service.
Because Silicon Studio actually builds these technologies, they deeply understand exactly what skills are needed to succeed in the industry.
They leverage this knowledge to act as a highly specialised recruitment agency for the entertainment and tech sectors.
They manage a database of approximately 30,000 specialised creators, designers, and engineers.
If a game company or a manufacturing firm desperately needs a 3D designer or a programmer, Silicon Studio finds the perfect match.
They either dispatch them as a temporary worker or place them as a permanent hire.
They receive a commission for each placement.
Revenues and customers
Over the last 5 years revenues have generally been growing.
In 2021 they sat at ¥3.99b, and last year they were ¥4.3b.
However, they dropped in 2024 and 2025 due to reduced appetite for work in the gaming sector.
This affected the HR division primarily.
Silicon Studio operates almost exclusively in Japan.
In 2025, Nintendo was the largest single client, representing 17% of total revenues.
Soleil is also a client.
However, over the last couple of years they have been pushing to get more clients from outside the gaming industry.
They successfully acquired Kawasaki Motors, Witz Corporation, and the Chuo University.
Market positioning and durability
As you can imagine, Silicone Studio is an elite technical provider.
Their programmers regularly publish research as tier-one academic conferences like Eurographics and SIGGRAPH.
This provides a pretty strong moat around the business.
The fact that a giant like Nintendo would rather pay Silicon Studio for their solutions than build them itself, is a testament to their expertise.
Essentially Silicon Studio is one of the go-to names for gaming and graphics solutions.
This makes the business pretty durable.
They don’t create video games themselves, they just provide the core technologies that all video game developers need.
They are also highly capital efficient.
In 2025, they generated ¥322m in operating cash flow but only spent ¥46m on tangible and intangible asset investments.
Their two main segments also reinforce each other’s value.
The company’s deep technical expertise in 3D graphics gives its Human Resources segment an unmatched advantage in matching highly specialised engineering talent.
Of course, the lack of skilled engineers is a threat to all this, as is the customer concentration with Nintendo.
Management’s Strategy
To counter these issues, management has defined a strategy to diversify away from video game clients.
They are adapting their high-end 3D graphics technologies for highly durable non-entertainment fields.
They target the automotive, civil engineering, aerospace, defence, and medical sectors with digital twin and simulation technologies.
Management has made significant, tangible progress towards these goals.
For example, they successfully launched ”BENZaiTEN” to provide artificial intelligence training data.
The acquisition of clients like Kawasaki Motors, Ltd. and Witz Corporation proves this industrial expansion is working.
To improve organisational efficiency, management also absorbed its wholly-owned subsidiary, Ignis Image Works, into the parent company on 1 December 2024.
This consolidation immediately improved profitability.
Following the merger, the company reported a 10.7% increase in operating profit in 2025, reaching ¥147m, despite the drop in top-line revenue.
The Operating Business
The primary feature of this business is its extreme capital efficiency.
Because the company provides software and human talent, it does not require expensive manufacturing plants, heavy machinery, or vast warehouses.
Once the company covers its fixed operating costs, almost all surplus operating cash drops directly to the bottom line as owner free cash flow.
The primary flaw is extreme working capital volatility.
Silicon undertakes large, complex contract development projects.
Revenue and cash collection are tied to specific project delivery milestones.
This creates massive year-over-year swings in accounts receivable, work-in-progress inventory, and customer advances.
For example, in 2025, operating cash flow was heavily boosted by a ¥314m decrease in inventory, but simultaneously dragged down by a ¥172m decrease in contract liabilities.
This lumpiness makes annual cash flow figures look erratic, even though the underlying business is stable.
Offer the last 5 years revenues have grown 8% in total.
OCF has been volatile.
During the same 5 year period, it has generated a low of ¥-29m, and a high of ¥527m.
Over the last five years, the business generated an average annual revenue of ¥4,153.4m and an average annual OCF of ¥241.6m.
Over the same period it converted 5.8% of its revenue into operating cash flow and 4.2% of revenues into extractable FCF, on average.
Again, this fluctuates quite a lot because of working capital swings.
For instance, in 2022, the business converted 12.5% of its revenue into OCF and 11.7% into FCF.
In 2025, it converted 7.5% of its revenue into OCF and 6.4% into FCF.
Overall, after the merger, the business is becoming structurally more profitable.
It also expenses its R&D costs rather than capitalising them.
This results in lower headline margins which artificially suppress the true earnings power of the business.
Another thing to understand is the Return on Equity.
In 2025, the company reported a Return on Equity (ROE) of 11.8%.
This headline figure is highly misleading.
It makes the business appear to be a mediocre, low-return asset to standard growth investors.
This is essentially distorted by all the excess cash on the balance sheet.
If we strip that out, the business has an adjusted ROE of around 100%.
Obviously, this is no compounder, but it is a business that could marmalade cash out to its owners for many years to come.
The Assets
The assets here provide some solid downside protection.
If the business did liquidate tomorrow, the theory is that we wouldn’t lose much money in the process.
To be sure of this we need to analyse the assets and confirm that they can indeed be converted into cash, in the real-world.
After going through everything and stripping out all the things that I don’t believe could be sold, here is what remains:
Cash = ¥1.59b
Net-Receivables = ¥616m
Income Tax Refund = ¥3m
Office furniture and PCs = ¥65m
Lease Deposits = ¥84m
Total liabilities = ¥948m
This gives us an adjusted NCAV of roughly ¥1.26b and an adjusted TBV of ¥1.40b.
From what I could tell, there are no hidden assets beyond what is stated on the balance sheet.
The biggest feature is the foundation these assets provide for the operating business.
Although there are clearly excess assets, this makes it a safe investment for someone buying the stock today.
The Shareholder Yield
This will be a short section.
There isn’t really any shareholder yield.
In 2023, and 2024 they did conduct buybacks but these were not to return value to shareholders.
They were actually conducted to buy the stock back from a long-time shareholder.
Now that transaction is complete, it is unlikely to be repeated in the same way a genuine buyback programme would be.
Crucially, the company does not issue new shares to fund ongoing employee bonuses.
Instead, it operates a “J-ESOP” trust.
The company pre-funded this trust with cash, and the trust purchased 25,000 shares of the company’s stock.
When employees earn points and trigger benefit pay-outs, they are granted shares directly from this existing trust pool.
This structure ensures that ongoing employee stock benefits do not continuously dilute outside shareholders.
This is pretty much the extent of their SBC activity.
Going forward, as mentioned earlier, we can realistically expect a total yield, paid in dividends, of roughly 1.4%.
This comes from the recent introduction of a dividend policy from management.
Anything else, would be a totally unexpected bonus.
The Ownership Structure
Uniquely, the company reports ownership and voting power net of treasury stock.
This means we don’t need to account for it, when calculating voting power, in the way I normally would.
In 2025, the top 10 shareholders controlled a combined 27.58% of the voting power.
The CEO holds the largest individual block, at 6.5%.
Core insiders and aligned shareholders control roughly 10% of the voting power in total.
Silicon also follows the Japanese tradition of cross-ownership.
If we assume that these are friendly shareholders, then the total amount of control between these and the insiders rises to a total of 21% of total voting rights.
Minority stock holders here are completely safe from an involuntary delisting at an absurd price.
The most compelling set-up here is for a potential activist.
With such a fragmented ownership, and a free float of roughly 70%, it would be possible to gradually build a controlling stake.
This could be used to clear away the excess capital and push it into something more valuable, like a massive buyback campaign.
What remains would be a highly efficient cash machine paying out most of its cash-flows as dividends.
The Risks
The primary risk is that the cash and value are never unlocked for a minority shareholder.
However, the ownership structure makes it likely that an activist will step in at some point to unlock it.
This becomes more true, the further the stock price declines.
The other risk is the Nintendo reliance.
Losing 17% of revenues probably wouldn’t kill the company, but it would severely damage the FCF for a few years, until it recovered.
This is driving management’s decision to expand and diversify the client base.
This is also something an activist investor could push for more aggressively.
Once the dependence falls below 10%, the business will be at far less risk from that worst case scenario.
There is also a case against the extreme concentration on the Japanese market.
The general demographic declines happening in Japan create a natural ceiling for entertainment growth.
The pivot to different sectors doesn’t necessarily protect Silicon from this.
The Investment Case
The investment case here is pretty simple.
We can buy a highly cash-generative business and get our money back in 3 years, if we include the excess cash on the balance sheet.
This is completely feasible with the current ownership structure.
The longer it remains cheap like this, the more likely it is that an activist will come along to unlock all that value.
This is also made more compelling by how attractive and niche the business model is.
I believe management will continue to expand and evolve the business in the right direction, but I don’t think it’s feasible to anticipate any major capital returns.
This gives us two main routes to making money.
The first, is through the business expansion gaining further traction and more non-gaming clients coming on board.
This is already happening and seems rational to continue over the next 2-3 years.
As that happens, revenues and cash-flows should stabilise, and the market can start pricing out the cyclical nature of the business.
The current pricing is now overly pessimistic, so even any type of positive update should re-rate the stock significantly.
The second is through an activist, competitor or industry player accumulating a controlling stake or pushing for a full private takeover.
Again, the ownership structure gives us a solid chance of receiving a fair offer, far above today’s price.
If it’s an activist, then riding on the coat-tails of special dividends or buybacks offers a nice route to making a market-beating return.
The price today is ¥686 per share.
My estimate of fair value, as things stand, is ¥1,142 per share.
I don’t own this stock yet because I already have some Japanese stocks that I consider better deep-value set ups.
You can view my own portfolio stocks with an annual subscription plan.
The annual plan also includes access to exclusive bonus content that you might also enjoy.
Get a 90% discount off Bloomberg’s No.1 competitor. Get a 30-day free trial here