(MENAFN- ValueWalk) >Hayden Capital's commentary for the third quarter ended September 30, 2025.

Table of Contents

Toggle

• Built By Bubbles
  • “Canal Mania”
  • “Railroad Boom”
  • “Dot-com Bubble”
• Portfolio Review
• Conclusion

Dear Partners and Friends,

The markets continued to power upwards, as AI excitement raged on. If I were to title the recent quarter, perhaps it'd be called:“Hot AI Summer: When Your GPUs Run Hotter Than Your IRR” (sorry, you can thank ChatGPT for that one).

However, skepticism has been seeping into the markets more recently, with the popular debate now being whether we're in an AI bubble or not (in my view, it's still too hard to know).

But regardless of the market's fickle behavior, one trend is clear. Whether participants in this surge of AI infrastructure investment become fabulously wealthy, or if they end up involuntary benefactors to the advancement of human progress – AI technology is here to stay.

Instead of debating“bubble or not”, I think our time is better spent contemplating what the world looks like after the current phase of AI infrastructure investment. And especially, how our portfolio can take advantage in the“deployment” stage of development.

For example, in a world where AI permeates everything, where will the biggest profit pools be? Will it be the foundational LLMs and chip-manufacturers powering the industry's backbone? Or the companies who own the customer relationships and utilize AI to enhance operational efficiency? Will it be the AI applications that enable this business transformation?

Or will these benefits ultimately pass onto the consumer, who utilize their newfound time & resources to spend on real-world leisure activities?

While the world throws money at the sector, fundamental data becomes distorted, and it can be hard to see past the fog4. But one safe bet, is that companies that meet existing human needs through new technology will inevitably gain significant value.

For example, online shopping evolved from physical stores, which replaced town markets, and which was an improvement from nomadic traders and bartering. In every stage, new business models that embraced“technology” to offer customers greater value, took profits from the old.

Human needs don't change – but the technology to meet those needs more efficiently, does.

Historically, we've made some of our best investments in companies that leveraged platform changes in technology. As such, we're eager to see which new businesses will emerge in coming years.

We gained +20.1% in Q3, outperforming the S&P 500's +8.1% and MSCI World's +7.5%. Since inception, we've returned +16.7% annualized after fees, ahead of the S&P 500's +13.5% and MSCI World's +10.2%.

Currently, our portfolio is allocated ~45% to Asia, ~40% to North America, ~12% to Latin America, with the rest in cash.

Built By Bubbles

While investors are distractedly arguing whether we're in an AI bubble or not, I want to posit something different:

Maybe... bubbles are actually beneficial for society?

Among financial circles, the term“bubble” carries a negative connotation. Immediately bringing to mind their inevitable and painful aftermaths – such as the dot-com crash of the early 2000s or the housing market collapse later that decade.

However, what if we should welcome bubbles as a necessary component of driving humanity's progress? History shows that bubbles are responsible for driving massive investments in new technologies and infrastructure, after all.

First, they lower the cost of capital for experimental ventures – precisely when the risk is highest. FOMO from speculative investors lowers the burden for unproven ideas to become reality, and paves the way for entirely new types of businesses in the bubble's aftermath.

Second, these speculative manias have incentivized and coordinated the capital, talent, parallel innovation, societal openness to new ideas, etc. necessary for technology“ecosystems” to develop (which symbiotically reinforce & create value for each other6).

**

Ben Thompson of Stratechery (one of my favorite writers), referencing economist Carlota Perez, argues that each major technological revolution – such as the industrial revolution, railways, electricity, and automobiles – was accompanied by a speculative bubble (LINK ).

These bubbles funded an“Installation Phase” of new infrastructure, laying the groundwork for a later“Deployment Period” of broader economic benefit. In this sense, bubbles are a feature, not a bug, of technological progress.

And from a societal perspective, the most beneficial bubbles are those that finance long-lasting physical assets, effectively gifting society with a“free” technological foundation on which new business models can be deployed7.

For example, three of the most famous bubbles from the past two centuries – Britain's canal system in the early 1800s, railroad tracks in the late 1800s, and telecom fiber in the 1990s – illustrate this point. Each left behind infrastructure that remains in use decades or centuries later.

“Canal Mania”

In the late 18th century, British investors rapidly financed canal construction, resulting in over 4,000 miles of waterways by 1830. While many speculative canal companies failed and investors lost money when the bubble burst, the resulting network sharply reduced transport costs and supported the Industrial Revolution by enabling efficient movement of heavy and delicate goods.

Investors who bought at the peak in the 1790s suffered losses when the bubble burst, but Britain was left with infrastructure that boosted its economy for decades.

Canal transport became cheaper after the bubble burst because the collapse erased the original capital costs and forced competition among overbuilt routes. New owners could price at marginal cost instead of the much higher, original construction cost, turning canals into near-free public goods8.

This network of canals, combined with lower prices, created a brand new industry of bulk freight logistics. The efficiency of canals standardized freight rates, allowing factories to reliably source cheap raw materials and contributed significantly to the subsequent industrial boom.

“Railroad Boom”

By the late 1800s, railroads dominated U.S. equities and absorbed much of the nation's investment. Track mileage grew from about 30,000 miles in 1860 to nearly 200,000 by 1900, establishing the world's largest rail system.

Generous federal and local incentives – including land grants and bond guarantees – spurred a surge in railroad construction, often outpacing demand. Because the bubble led to overbuilding and intense competition, rates fell -80% between 1849 and 1910 (LINK ). This resulted in a quarter of railways going bankrupt (LINK ).

Despite the turmoil, the resulting nationwide rail network unified distant regions, lowered transport costs, accelerated industrialization, and created new population centers. By 1900, America's rail system surpassed the rest of the world combined, providing a key advantage for future economic growth.

Many new business models, such as nationwide mail order businesses and industrialized farming, emerged to utilize this new, cheaper infrastructure.

Companies like Montgomery Ward and Sears could now sell virtually everything (clothing, furniture, farm tools, even full pre-fab house kits) into every town served by a train depot, instead of relying on local walk-in customers (i.e. essentially Amazon with unlimited“shelf-space”, pre-internet). And farms became larger and specialized – benefiting from economies of scale, and thus lower unit costs – producing for national markets rather than just local subsistence.

“Dot-com Bubble”

The 1990s Internet boom is perhaps the most familiar example for investors today. Speculative fervor led to massive investments in anything“.com,” resulting in many startups crashing in 2000 but leaving behind lasting gains.

Telecom companies overbuilt fiber networks and data centers, much of which went unused after their bankruptcies. That“dark fiber” became the near-free backbone of the modern Internet, enabling broadband, e-commerce, and mobile technologies to scale cheaply.

The bubble also triggered a wave of“parallel experimentation,” with the development of new browsers, web technologies, search engines, and e-commerce models.

The internet frenzy brought nearly the entire U.S. population online by 2000, creating both the infrastructure and user base for the Internet economy that would dominate the next two decades. This provided the ingredients to form future consumer Internet giants, and prepared a generation of workers for web-based work (LINK ).

On the downside, overcapacity and subsequent investor wipeouts led to dramatic reductions in long-haul bandwidth prices (prices fell 90 - 95%). For instance, Global Crossing was famously sold for ~$250M post-bankruptcy, or just 1% of its original costs (LINK ).

However, these lower costs made new business models viable that previously would not have been under the old price. Cloud computing, video streaming, and software-as-a-service all became feasible and gained mainstream adoption in the years that followed.

**

All these historical examples follow a similar pattern: during the bubble, capital is misallocated and many investors lose money. But the surge of investment builds infrastructure and consumer knowledge & habits that endure long afterwards.

These“fixed assets” have large upfront costs, often wiped out during bankruptcies, but minimal marginal costs. The infrastructure, once“gifted” to society after the burst, allows for a plethora of new businesses to be built utilizing the lower pricing.

And more importantly, bubbles coordinate society to direct resources – capital, people, ingenuity – at a singular problem. This unplanned, massive coordination allows for the simultaneous creation of infrastructure, applications, and services that feed off one another, accelerating adoption and innovation.

As Ben Thompson of Stratechery describes, the purpose of bubbles is to“install” physical capacity and“create cognitive capacity” for the future (LINK ).

“The fundamental utility of inflection bubbles comes from their role as coordinating mechanisms. When one group makes investments predicated on a particular vision of the future, it reduces the risk for others seeking to build parts of that vision.

For instance, the existence of internet service providers and search engines made e-commerce sites a better idea; e-commerce sites then encouraged more ad-dependent business models that could profit from directing consumers. Ad-dependent businesses then created more free content, which gave the ISPs a better product to sell. Each sector grew as part of a virtuous circle.”

Such rapid, society-wide engagement could only have occurred through the concentrated effort of speculative bubble.

**

So how does this relate to today? Since ChatGPT launched in 2022, AI stocks have driven roughly 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital spending, per JPMorgan9. It's obvious this concentrated,“coordination” mechanism is already taking place.

But does this mean we're in an AI bubble? I'm not so sure – some argue valuations like Nvidia at ~30x earnings are reasonable, while others say earnings are inflated due to excess data center building (making the P/E ratio look artificially low), which in turn is over-estimating future end-user demand that will never materialize.

One argument AI skeptics make, is that unlike canals or railroads, today's GPUs quickly become obsolete. Remember, the most beneficial bubbles create long-lasting, physical assets. Spending hundreds of billions on GPUs that will be obsolete in a few years, doesn't quite fit the bill.

But I'd argue what endures, is the technical expertise gained – building fabs, training more AI engineers, upgrading power systems, and nearly 100 million Americans familiar with ChatGPT (LINK ), all ready to embrace the next wave of AI applications.

In every case, the countries that permitted a speculative boom emerged with a head-start in the new technology and equally important, skilled workers & a educated consumer. While regions that avoided or suppressed these manias risked falling behind.

Even if the AI skeptics are right, society will still benefit in lasting ways.

(As a sidenote, this also makes me question whether China's recent approach to market speculation, such as trying to prematurely stamp out any signs of market frothiness, might ultimately place it at a disadvantage. Allowing investor greed to flourish can encourage the coordinated effort and lower cost of funding necessary to bring new technological shifts into reality. While investors may lose money and social unrest may be a risk, society benefits in the long run by being educated and primed to use the technology first, when more sustainable business models are built on top.

Perhaps America's government deserves more credit for its general acceptance of market speculation and its willingness to address consequences after bubbles burst. This might be the key factor in maintaining America's legacy of leading innovation.)

**

And if you've stayed on the sidelines during this installation phase (as we have), does the debate really matter?

For us, the conclusion is the same regardless – picking winners (and importantly, at an attractive IRR) simply seems too hard at this stage. Maybe others are better at it, and that's okay.

Luckily, we have the luxury to wait for a better entry point, when the outcome is more predictable.

I'd rather spend our time thinking about what the next“deployment” phase will look like – as that's when history shows most value is realized. For example, I think the investment case was clearer for Montgomery Ward or Sears, which utilized the railroads to become the first national retailers, rather than owning the railroads themselves. Or Procter & Gamble, who was the first to utilize radio advertising in 1923 (and even invented the“soap opera” genre), rather than investing in RCA / radio stations (LINK ).

As such, we're eagerly awaiting the new business models and applications, that will be built on top of today's infrastructure. And maybe at some point, we'll even be lucky enough to purchase assets at 1% of replacement costs. One can hope...

Portfolio Review

Timee (215A.T): Timee operates a short-term job marketplace in Japan. We've been investors for over a year, and it's our first investment in the country. I'm excited by its prospects, as the company showcases all the hallmarks of a leading two-sided, asset-light, marketplace. It's an archetype for one of the world's best business models, one where we've historically found some of our best investments.

The Industry Backdrop

Japan is in the early stages of a chronic labor shortage, one that's predicted to get worse over the next few decades. A declining birth rate, and increasing elderly population, means Japan will be missing -3.4M workers by 2030, rising to -11M by 2040 (~3% and ~10% of Japan's current adult population).

Most of these shortages are in Timee's key categories – with 1.6M“missing workers” across logistics, food, and retail today11. It's extremely difficult to find qualified part-time or full-time service workers in Japan – and it will only grow more difficult over time.

In 2018, the Japanese government began promoting“働き方改革” or“work-style reform”, to address these issues. This initiative encouraged side-jobs / multiple jobs, and normalized the practice.

As an example of how bad it's become, about 25% of job postings on Timee's app explicitly mention the desire to hire someone who can become a full-time employee, and the company estimates another 25% would like to.

This use of a short-term employment platform to source long-term employees demonstrates the desperation of businesses, and the extent they must expand their recruitment efforts beyond traditional channels such as job boards, referrals, and offline advertisements.

How Timee Works

Timee is an online two-sided marketplace that connects businesses needing immediate staff with individuals seeking short-term work. It was founded in 2017 by Ryo Ogawa, and pioneered a new labor category in Japan – instant, on-demand staffing.

When a business urgently needs a cashier, warehouse worker, waitress, or a variety of other service jobs to be filled, they can post on Timee's platform and have someone start as soon as the same day.

Through Timee's app, companies can post shifts as short as one hour – for example, a restaurant or retail shop can fill a sudden staffing gap – and workers can sign up in real time, without interviews or resumes.

Instead, trust is built upon a“rating” system and work history – with employees earning“badges” after showing excellence in a particular area. For example, a worker who receives repeat positive reviews after working in“dish-washing”,“cooking” or“shelf-stocking” will receive a badge in that specialty. Future employers have greater confidence in the worker's qualifications, and can even mark certain jobs as eligible to only badge-holders.

Think of it as a hybrid between Google Maps Reviews + Duolingo Badges, for individual workers. The system is proven to work, with >90% of positive reviews and just a 0.1%“no-show” rate.

Importantly for workers, the payment is processed quickly after the shift (often same-day). In contrast, part-time jobs are conventionally paid the following month in Japan – weeks after the job finished. In this way, Timee addresses a crucial pain point of temp work – especially as households face rising inflation and strained budgets.

For background, Japan is facing structural inflation for the first time in decades – rising to >3% vs. <1% pre-Covid. And many workers are now resorting to side-jobs to make ends meet.

By paying instantly, this helps workers meet immediate cash flow / household budget needs. For example, Timee claims ~21% or ~2.5M of their registered workers have full-time jobs.

Arguably without Timee, historically these individuals wouldn't have worked a side-job. In this way, Timee encouraged many people with full-time jobs to enter the part-time labor pool – single-handedly expanding the market.

Japan continues to face significant structural challenges related to population decline and labor shortages. Because the supply of service workers and the overall "labor pool" remains limited, society needs to efficiently utilize the pool of existing workers. Timee's platform addresses these issues by optimizing and expanding access to the available labor hours.

By offering businesses greater convenience and workers increased flexibility along with rapid compensation, Timee has demonstrated substantial product-market fit within Japan's tight labor environment.

In just seven years, the company has attracted approximately 12 million registered workers, with around 1.8 million actively engaging with the platform. Moreover, nearly 200,000 businesses regularly post opportunities on Timee's platform (with a total ~392,000 registered businesses).

Most Timee workers take on >8 jobs per month (~55% of workers), making its worker population highly engaged. And in a country of ~100M adults, Timee has already become a significant part of Japan's social fabric – equating to ~2% of the population using the service.

Timee initially focused on three core verticals – logistics (~43%), retail (~26%), and food service (~17%) – which account for ~86% of jobs. The company is now expanding into other industry verticals, such as hotels and elderly care, which currently comprise another ~14% of revenues.

The company also introduced adjacent services such as“Timee Career Plus”, a placement service to convert proven gig workers into full-time hires (launched February 2024). As mentioned above, many Timee merchants are searching for full-time employees too, and this helps Timee monetize worker transitions and deepen its value proposition to job seekers.

Timee makes money by taking a commission (~30%) on wages paid for each shift arranged. While seemingly high, this is lower than traditional temp staffing agencies, where the cut is typically ~35-40%. This“pay per shift” model also appeals to employers because they pay only for actual labor received, as opposed to upfront job listing fees or fixed temp contracts.

Operationally, Timee runs a high-touch model to ensure quality and liquidity. Of ~1,200 full-time employees, some 61% (about 741 staff) are in sales and customer success. This large on-the-ground team actively onboards client companies, supports them in using the platform, and is a key differentiator for Timee's service.

Many Japanese businesses are culturally hesitant to change. So having a dedicated account manager that guides them through the process and introduces them to platform features, significantly increases customer satisfaction & engagement.

Field salespeople encourage multi-site employers to list all locations, then focus on filling shifts at the sites with the biggest labor gaps. The heavy investment in customer success helps new clients integrate spot workers smoothly (i.e. training managers on using last-minute staff) and contributes to Timee's superior fill rate (~86%) on posted jobs.

By comparison, rivals struggle with fill rates and no-shows, giving Timee a reputation for reliability that further reinforces its network effects.

The Numbers

Timee has grew revenues at a +78% cagr, over the last three years. Operating margins expanded from ~2% to ~22% over that period, showing significant operating leverage in-line with a rapidly scaling asset-light marketplace model. The company believes they can get to 30 - 40% operating margins over time, which seems achievable (incremental margins were ~40% last quarter).

Crucially, Timee's unit economics are favorable and highly resilient. The company can adjust marketing spend in line with demand – for example, management can scale back worker recruitment ads when job postings slow, which keeps operating profit on target even if revenues are soft.

This variable cost structure (worker marketing is discretionary) means margins can hold or even improve during revenue slowdowns. Timee's contribution margins are high because each filled shift yields a commission, while the marginal cost of serving an extra shift is low (mainly payment processing and a share of support costs).

I believe the most important KPI for this business, is having a high fill-rate (what % of job postings are filled). This is because a high fill rate:

Discourages employers from trying other platforms, since Timee meets their needs adequately13,

High customer satisfaction encourages employers to open up jobs for other locations / roles, which

Encourages more workers to use Timee as their platform of choice, given the higher density of jobs available. For these minimum wage jobs, applicants are not willing to travel far to attend a job. So job density across both time and geography is crucial. For example, you need jobs available within 1 kilometer of a worker, and also available when they're free on Tuesday at 2pm.

Because the workers actively engage / work on the Timee platform, they build a rating that locks them into the platform. Or else the workers will need to rebuild their positive review history (and implied trustworthiness) over again.

All these dynamics encourage a“winner-take-most” model. The platform with the highest density of jobs & workers will have a natural, compounding advantage (“flywheel effect”), that makes it hard for competitors to attack.

And crucially, we just received the best“proof” of this yet, with news of Mercari Hallo shutting down next month.

Mercari, one of Japan's best known tech companies, saw Timee's success and launched its own competitor in March 2024. They quickly amassed 11 million registered users a year later (LINK ). But while they had a similar registered user base as Timee, what they didn't disclose is that the fill-rate was less than half of Timee's.

Through our own sources & analysis, we found that only ~40% of Mercari's jobs were being filled. And this is despite them coming to market at 0% take-rates (i.e. giving the service for free, while Timee charges 30%).

Why would an employer who needs a worker tomorrow, post on Mercari Hallo, if there's only a 40% chance they'll find a suitable worker? It's clear that they wouldn't – and that the lost sales from an unfilled job far outweighed Timee's 30% cut. Timee saw evidence of this - some customers were lured by Mercari's“free” service, only to come back months later.

It's a clear proof-point that capital and cheap prices, are not enough to overcome superior job liquidity & density. And perhaps Timee has reached a“tipping point”, making it difficult for competitors to catch up.

These advantages have allowed Timee to gain 70 - 80% market share in the spot-work market14. They place 18M jobs per year versus other online competitors at a mere ~2.1M combined.

Despite its recent success though, Timee still has significant room to expand. The company's current ~12 million registered workers are only a fraction of Japan's workforce. And many industries (like healthcare, event set-up, etc.) remain largely untapped for spot work.

Timee's own TAM analysis indicates its barely penetrated in its core verticals, implying a long growth runway as more businesses adopt flexible staffing. Additionally, they are exploring new revenue streams – from full-time job placement fees to potentially fintech (e.g., earned wage access, insurance) – which could boost monetization per user.

International expansion is a further upside optionality (i.e. exporting its platform to other labor-challenged markets in Asia), though not in the immediate plan15. Overall, the structural trends (aging population, normalizing side-job work) support growth for decades, and Timee is well-placed to be the central infrastructure enabling that in Japan.

With competitors defeated, I expect Timee will continue compounding revenues in the low-20%'s over the next few years. Combined with a gradual ~1.7% annual improvement in operating margins, this would bring operating margins from ~22% today to over 30% within 5 years. If our forecast is correct, this means the company would be making ~¥93BN in 2030, with ~¥29BN in operating profits.

The stock is currently trading at an ¥138BN enterprise value. The company is free cash flow generative, has return on equity over 40%, and over ¥12BN of net cash on its balance sheet – metrics that reflect Timee's asset-light, high margin business model.

Thus at current prices, Timee is trading ~17x 2025 EV/EBIT, and just ~4.7x EV/EBIT on 2030 numbers. Comparatively, rivals like Recruit trade at ~19x 2025 EV/EBIT, have lower operating margins at ~16%, and are projected to experience much slower growth of ~4.4% per year over the same time period.

Based on fundamental growth alone, I expect Timee shareholders may realize >3x on their investments over the next few years. But it's also possible that the market will assign a premium valuation to shares, given the superior margin and growth profile. If so, this provides an additional“kicker” to our price targets, and investment returns.

We began purchasing shares at ~¥1230 last fall, and have continued to increase the position to an average price of ~¥1450. Shares have taken a dip over the last few months due to temporary growth headwinds (which we'll discuss in subsequent updates).

In light of Japan's significant labor challenges, I think this presents an opportunity for groundbreaking companies like Timee. The company is scaling quickly, has proven high unit economics, and successfully fended off much better financed competition.

We are enthusiastic about the investment, and are excited to watch it lead the way in Japan's spot-work sector over the coming years.

Conclusion

On the operational front, we successfully launched our new investment vehicle last month. The vehicle uses the same strategy as our separate accounts, but just offered in a different format.

This is important, however, because it allows us to accept partners from the European Union, Hong Kong, Japan, India, Australia, and New Zealand. We welcome our global partners who have already joined us, and look forward to strengthening these relationships in years to come.

Additionally, if you're interested in joining us, please reach out. I'm always happy to meet like-minded partners, and hopefully welcoming you to the Hayden“family”.

**

On a similar note, I'm also back on the road early next year, with planned stops in the UK, Switzerland, Germany, Spain, and possibly more. I'd be delighted to grab coffee while I'm in town, so please just reach out & let us know.

Lastly, we finally started a Substack page, and will also post more frequently on Linkedin going forward. Traditionally we've engaged with the broader investment community on X (formerly Twitter). But as the community has changed over the years and migrated elsewhere, I thought we'd give these other platforms a chance too. Please follow us on Substack or Linkedin, if you use those platforms.

I wish you and your family a wonderful holiday season. See you in the New Year.

Sincerely,

Fred Liu, CFA

Managing Partner

[email protected]

Read more hedge fund letters here

MENAFN28112025005205011743ID1110411897