I. Historical Case Study: The Siege of Alesia, 52 BC
Before Julius Caesar held absolute power, he was a general with a singular aim: to break the resistance of Gaul. At Alesia, where the rebel leader Vercingetorix had fortified himself atop a hill, Caesar sought victory by constructing a fifteen-kilometre wall to encircle and starve the enemy into submission.
However, then a massive Gallic relief army approached, threatening to trap Caesar’s forces between two enemies. Ceaser responded immediately and ordered a second wall, facing outward. Roman soldiers built over thirty kilometres of fortifications from nothing, working through freezing rain and thick mud, with no sleep.
When both Gallic forces launched a coordinated attack, the Roman lines held. Victory came not just from the soldiers’ discipline, but from engineering ingenuity and unwavering command.
Alesia was the turning point. With the fall of Gaul, Julius Caesar became the most powerful man in the Roman world.
Map of Caesar’s double fortifications at Alesia, encircling the Gallic stronghold and facing down a relief army.
Map of Caesar’s double fortifications at Alesia, encircling the Gallic stronghold and facing down a relief army.
II. 2000 AD
Map of Caesar’s double fortifications at Alesia, encircling the Gallic stronghold and facing a relief army.
By the dawn of the 21st century, a different kind of empire lay in ruins: The restaurant industry.
Across Europe and the United States, restaurants fell like Pompeii beneath ash. It is a sector responsible for over 3.2 million jobs in the UK, 15.5 million in the U.S., and trillions in global GDP and yet it is systemically collapsing.
In the UK alone, 748 hospitality venues shuttered in the final quarter of 2024, this is an average of eight closures per day (CGA, 2025).
To say this is not a war is to misunderstand death.
Year-over-year restaurant visits fell nearly 100% across major economies in March 2020.
Year-over-year restaurant visits fell nearly 100% across major economies in March 2020.
Global inflation rose sharply after 2020, driven by food, energy, and transport costs.
Global inflation rose sharply after 2020, driven by food, energy, and transport costs.
By the dawn of the 21st century, a different kind of empire lay in ruin: The restaurant industry.
Across Europe and the United States, restaurants fell like Pompeii beneath ash. It is a sector responsible for over 3.2 million jobs in the UK, 15.5 million in the U.S., and trillions in global GDP and yet it is systemically collapsing.
In the UK alone, 748 hospitality venues shuttered in the final quarter of 2024, this is an average of eight closures per day (UKHospitality, 2025).
To say this is not a war is to misunderstand death.
Global inflation rose sharply after 2020, driven by food, energy, and transport cost
Year-over-year restaurant visits fell nearly 100% across major economies in March 2020.
The core of the crisis is labour. Hospitality is the lowest paid sector in the UK and also one of the most burned out, with up to 76 percent of managers reporting symptoms of burnout driven by long hours, stress, and chronic understaffing (ONS, 2024; CIPD 2023). Therefore in 2025, the UK raised the National Living Wage to £12.21/hour. A necessary step, but one that added £1,400 per worker per year in cost. For restaurants operating on 3–5% margins, this was not a wage adjustment. It was an extinction event.
Attrition is now a million-pound bleed in the industry with the highest staff turnover, with hundreds of millions lost annually to hiring, training, and replacing staff (CIPD, 2024). Since 2019, the number of EU workers in the UK hospitality sector has dropped by approximately 121,000, representing a 41 percent decline from before the pandemic. The total number of international workers leaving the sector, including those from outside the EU, is around 197,000. This has contributed to a major labour shortage, with hospitality experiencing the highest vacancy rate of any UK industry at 7.9 percent (FT, 2024).
This is not a skills gap. It is a vacuum.
And the costs climb.
In May 2025, annual food inflation rose to 2.8%, its fourth consecutive monthly increase (FT, 2025). Fresh beef prices surged, driven by droughts and collapsing global supply chains. At the same time, gas prices jumped 7.5%, electricity 2.9%, and the commercial energy burden outpaced global averages by 46% (ONS, MakeUK, Guardian, 2025).
Moreover, urban rental markets worldwide have seen substantial increases. In Australia, for instance, rental prices grew nationally by 10.1% between 2022 and 2023. In the United States, asking rents were up 3.4% in April 2025 compared to the same time last year, and rent prices are 35.4% higher than they were before the pandemic.
The few automation attempts failed. McDonald’s robotic burger stations were abandoned in 2023 after burning through millions (Guardian, 2023). Venture-backed hardware startups collapsed one by one. The problem wasn’t just technical, it was structural. Britain produces world-class engineers, but they are funnelled into finance, exported to California, or underutilised entirely (More here).
No viable automation platform emerged. No serious policy intervened. The scaffolding of a trillion-dollar sector was left to rot. The restaurant industry is more than GDP. It is a civic fabric. It is the nightly gathering of families, the economic foothold of immigrants, the social script of cities. It is how culture feeds itself.
And yet it is vanishing.
We do not exaggerate when we say this is our Alesia. The enemy surrounds us. Within: attrition, fatigue, insolvency. Without: inflation, instability, silence. The walls are closing in.
And so, we build.
III. Case KAIKAKU: London, 2025
In the ruins of a broken industry, we built the prototype of the future.
While legacy giants like Subway, McDonald's, or Chipotle poured billions into stalled pilots and dead-end decks, three engineers in a borrowed kitchen did what no incumbent could. With no templates to follow and no institutional funding, they built "Fusion", KAIKAKU’s first fully functional kitchen automation system, in just 94 days.
Fusion is not an enhancement. It is a replacement.
Capable of assembling >360 fully customised bowls per hour on autopilot, it achieves a 300% throughput increase over traditional assembly lines.
It was born in chaos with a >$500k investment from myself, rejected by over 300 VCs, and powered by nothing more than hunger. In just one year, it served over 100,000 bowls, generating more than $1 million in revenue from a single location with real customers, real staff and real complaints. Every bowl was tracked, logged, analysed not just as output, but as data.
Behind the scenes, over 1 million lines of code were written. More than 20,000 components were modelled, tested, and assembled. 50,000 iterations of robotic movements and tolerances were rendered in CAD. The knowledge architecture, spanning everything from thermodynamic calibration to service interface logic, was documented across 1,000+ Notion pages, all while the team spent nights sleeping in the office.
Fusion did not emerge in spite of industry collapse. It emerged because of it. It was built not to optimise the old logic, but to bury it. It is not a demo. It is not a pitch. It is the first glimpse of what comes after.
And Fusion is the first proof that our system works.
In the ruins of a broken industry, we built the prototype of the future.
While legacy giants like Subway, McDonalds, or Chipotle poured billions into stalled pilots and dead-end decks, three engineers in a borrowed kitchen did what no incumbent could. With no templates to follow and no institutional funding, they built Fusion, KAIKAKU’s first fully functional kitchen automation system, in just 94 days.
Fusion is not an enhancement. It is a replacement.
Capable of assembling over 360 fully customised bowls per hour on autopilot, it achieves a 300% throughput increase over traditional line operations.
This was not a project born in stealth. It was born in chaos, self-funded, rejected by over 300 venture capitalists, and initially powered by nothing more than hunger. It moved over half a million calories of real food, in real time. In just one year, it served over 100,000 bowls, generating more than £1 million in revenue from a single location. Every bowl was tracked, logged, analysed not just as output, but as data.
Behind the scenes, over 1 million lines of code were written by hand. More than 20,000 components were modelled, tested, and assembled. 50,000 iterations of robotic movements and tolerances were rendered in CAD. The knowledge architecture, spanning everything from thermodynamic calibration to service interface logic, was documented across 1,000+ Notion pages, all while the team spent 23 nights sleeping in the office.
Fusion did not emerge in spite of industry collapse. It emerged because of it. It was built not to optimise the old logic, but to bury it. It is not a demo. It is not a pitch. It is the first glimpse of what comes after.
And Fusion is the first proof that our system works.
IV. Austria, 1995
KAIKAKU’s origin is not financial. It is familial.
My parents opened a Chinese restaurant in a small Austrian town after immigrating from China. Growing up in that household meant 14-hour shifts stacking used dishes and translating stacks of German documents starting when I was just 6 years old. That kind of trauma stays with you. Especially a hatred towards your parents for making you work while all your friends are spending their childhoods playing video games.
But one day, my mother told me: “Working in the restaurant was the only time your dad and I, as restaurant owners, could spend with our son, you.” That sentence reframed a childhood of trauma into a vision of redemption. It made the pain instructive. I also dove deep into understanding why my family pact forged my drive. (More in my essay here)
A few weeks later, KAIKAKU was not born. We do not wish to repair the old system. We intend to bury it and build a new one. That is radical change! In Japanese, that is 改革 - KAIKAKU.
Josef, age 5, with his family in their restaurant in Linz.
Josef, age 5, with his family in their restaurant in Linz.
VI. How We Fight
VI.1 Builders First, Soldiers Second
At Alesia, Roman infantry became builders. They outbuilt, outplanned, and outlasted a numerically superior foe. Their strength lay in the infrastructure of thought: treating time, space, and terrain as tools to be mastered.
We apply the same logic. KAIKAKU only hires siege builders.
Among our armee, our legionaries are:
PhD in molecular biology and Michelin trained chef.
The guy who wrote all of the software that powers Austria's Military sports centres.
An operator who once ranked #4 in League of Legends EUW.
Kids who built hardware as soon as they could walk.
The creator of one of the world’s first bitcoin faucets with >150k users.
We believe there is a massive hardware talent arbitrage in Europe. And we are already best positioned to benefit form it. (More in my essay here)
VI.2 Iterate Like a Legion
At the height of its dominion, Rome ruled a quarter of the known world: over 5 million square kilometres, 50-90 million lives, and the machinery of empire itself. But Rome did not endure because it celebrated victories. It endured because it interrogated failure. Every defeat was documented. Every campaign post-mortemed. Legions evolved not by repetition, but by iteration.
At KAIKAKU, we carry this discipline forward.
Common Room is our working restaurant that also functions as a live laboratory, a theatre of systems engineering. Behind a one-way mirror, our engineers operate like Roman scribes and battlefield tacticians: watching, documenting, adjusting in real time. Nothing is left to theory. Customers are not only data points. They are test subjects in a full-stack experiment. Ordering flows. Robotic behaviours. UX interfaces. Failure thresholds. All are tested, broken, rebuilt, again and again, under live conditions.
We have abolished the waterfall. The legacy paradigm of sequential stages and delayed feedback, acceptable in 20th-century aerospace, is fatal in volatile, human-centred systems like hospitality. Abstraction kills. Theorising delays. In the real world, polish at launch is worthless.
What matters is speed through the loop.
Observe. Fail. Adapt. Repeat.
Video proof here.
Caesar, Caius Julius. De bello Gallico. A firsthand account of Roman strategy.
Caesar, Caius Julius. De bello Gallico. A firsthand account of Roman strategy.
Tregear, Ivan. A firsthand account of KAIKAKU tech strategy.
Tregear, Ivan. A firsthand account of KAIKAKU tech strategy.
VI.3 “What brings no benefit to the hive brings none to the bee.” - Marcus Aurelius (Meditations 6.54)
Centuries before neuroscience could measure it, Aurelius understood a truth modern biology now confirms: individual flourishing is inseparable from the health of the collective.
We are not dying from disease alone. We are dying from disconnection. (More in my essay Alone Together: Disconnection Is Our Generation’s Black Death)
The Harvard Study of Adult Development, spanning over eight decades, has shown that neither income, nor education, nor genetic fortune predicts long-term health and fulfilment as reliably as sustained, high-quality relationships. Loneliness is not a feeling. It is a systemic threat.
Chronic social isolation triggers physiological cascades, elevated glucocorticoids, immune suppression, systemic inflammation that rivals the biomedical risks of smoking and hypertension (Holt-Lunstad et al., 2015; Cole et al., 2007).
Kaplan–Meier survival curves shows 10-year mortality risk due to loneliness or social isolation.
John F. Kennedy, one of the participants of the Harvard Study of Adult Development.
Kaplan–Meier survival curves shows 10-year mortality risk due to loneliness or social isolation.
John F. Kennedy, one of the participants of the Harvard Study of Adult Development.
John F. Kennedy, one of the participants of the Harvard Study of Adult Development.
And the paradox is this: as our machines grow smarter, our environments grow colder. In the name of efficiency, we are stripping away the rituals that make us human shared greetings, remembered names, spontaneous smiles. We are building systems optimised for throughput, not warmth. Precision, not presence.
KAIKAKU rejects this trajectory.
We do not automate to eliminate human labour. We automate to reposition it where it matters most. Our robots do not replace touch, they create space for it. They shoulder the burden of the repetitive and the physically punishing so that staff can engage in the profoundly human. As neuroscientist Elizabeth Redcay (2010) has shown, reciprocal social exchange lights up the brain like no app or kiosk ever will.
Especially, elderly guests often come back because someone takes the time to help them with the tech. What really matters to them is knowing there’s a friendly face who’s always patient and willing to help. This kind of feedback, or as we call it data, will never shows up in a dashboard but changes everything.
Common Room is our lab, and the results are clear: dignity, recognition, and embodied care are not sentimental add-ons, they are the neurobiological infrastructure of well-being.
At KAIKAKU, we are not engineering restaurants.
We are engineering conditions for human survival.
The hive must be reimagined but it must never be abandoned.
Anne, 81: “This is the only restaurant in this area that doesn't feel purely transactional. And your bowls are better than sex.”
Anne, 81: “This is the only restaurant in this area that doesn't feel purely transactional. And your bowls are better than sex.”
Anne, 81: “This is the only restaurant in this area that doesn't feel purely transactional. And your bowls are better than sex.”
VII. Triumph
My goal is not to scale a chain of restaurants, nor to join the chorus of marginal innovations clinging to a dying system.
My goal is to reengineer the foundational architecture of foodservice, structurally, permanently, and irreversibly.
At KAIKAKU, we do not measure success in units opened or revenue per square foot. We think in terms of ecosystems, infrastructure, and operating systems: entire structural layers upon which the future of dining will depend. Just as Bernard Arnault did not merely sell handbags but constructed a vertically integrated empire that institutionalised luxury across continents, our ambition is not to operate restaurants, but to become the unseen but indispensable framework through which modern foodservice runs not a brand, but a backbone.
We are building a full-stack technological foundation that does not ask to be adopted but demands to be depended on. This system unites robotics, orchestration software, and intelligent operations into a single cohesive machine, a vertically integrated nervous system engineered to deliver levels of precision, throughput, and resilience previously considered incompatible with human-centric hospitality.
Our model is not designed for incremental adoption. It is modular in its logic, scalable by design, and essential by consequence.
We are not building for today’s market conditions, which fluctuate with trends and margins. We are building for a structural future, one in which human attention becomes the scarcest commodity, operational stability becomes existential, and technological infrastructure becomes the difference between collapse and continuity.
This is not just automation. It is emancipation from margin-choking inefficiencies, from burnout economics, from the industrial-age logic that treats people as plug-and-play units in a brittle chain.
In our world, robots do not replace humans; they liberate them, reallocating human presence toward the very tasks machines cannot perform care, improvisation, memory, welcome.
We do not seek to optimise the status quo, or to marginally outperform legacy competitors. We intend to collapse the distinction between restaurant and system, and in doing so, redefine the very substrate of the industry. In the same way that Stripe did not improve payments, it became payments, we are not improving restaurants.
We are becoming the operating layer on which they depend.
This is what we mean by triumph not domination for its own sake, but structural inevitability through technical truth.
We are not building to serve the future.
We are building so that the future runs on us.
by Josef Chen
My goal is not to scale a chain of restaurants, nor to join the chorus of marginal innovations clinging to a dying system.
My goal is to reengineer the foundational architecture of foodservice, structurally, permanently, and irreversibly.
At KAIKAKU, we do not measure success in units opened or revenue per square foot. We think in terms of ecosystems, infrastructure, and operating systems: entire structural layers upon which the future of dining will depend.
Just as Bernard Arnault did not merely sell handbags but constructed a vertically integrated empire that institutionalised luxury across continents, our ambition is not to operate restaurants, but to become the unseen but indispensable framework through which modern foodservice runs not a brand, but a backbone.
We are building a full-stack technological foundation that does not ask to be adopted but demands to be depended on. This system unites robotics, orchestration software, and intelligent operations into a single cohesive machine, a vertically integrated nervous system engineered to deliver levels of precision, throughput, and resilience previously considered incompatible with human-centric hospitality.
Our model is not designed for incremental adoption. It is modular in its logic, scalable by design, and essential by consequence.
We are not building for today’s market conditions, which fluctuate with trends and margins. We are building for a structural future, one in which human attention becomes the scarcest commodity, operational stability becomes existential, and technological infrastructure becomes the difference between collapse and continuity.
This is not just automation. It is emancipation from margin-choking inefficiencies, from burnout economics, from the industrial-age logic that treats people as plug-and-play units in a brittle chain.
In our world, robots do not replace humans; they liberate them, reallocating human presence toward the very tasks machines cannot perform care, improvisation, memory, welcome.
We do not seek to optimise the status quo, or to marginally outperform legacy competitors. We intend to collapse the distinction between restaurant and system, and in doing so, redefine the very substrate of the industry. In the same way that Stripe did not improve payments, it became payments, we are not improving restaurants.
We are becoming the operating layer on which they depend.
This is what we mean by triumph not domination for its own sake, but structural inevitability through technical truth.
We are not building to serve the future.
We are building so that the future runs on us.
by Josef Chen
