Act Globally, Eat Locally

February 1, 2017

By 2050, we’ll need to find a way to feed 9.7 billion people.

Current food production methods will not suffice. They’re too inefficient (it takes at least 9 calories of food production to produce 1 calorie of edible meat) and harmful to the environment (see Environmental impacts of farming). Many developed countries will even struggle to meet their Paris Climate Accord commitments.

Fortunately, some fascinating new innovations in agritech are emerging. Vertical farming, plant-based meats, lab-grown foods, drone-enhanced farming to name a few.

Japan is one of the most forward-thinking countries in this area. The government as well as a number of Japanese industrial groups are seeking to spur innovation in this domain. I’ve been solicited to help identify investment opportunities in agritech startups that are interested to explore the Japanese market.

Beyond finally learning how to distinguish a tasty tomato from a bland one, I’m still a neophyte investor in this domain and eager to learn more. Any relevant technology startups innovating in agriculture and interested in Japan are welcome to contact me.

Furthermore, Asia’s largest agritech summit will take place in Tokyo in May and is worth a look.

A day without AI is like a day without wine

January 23, 2017

As I mentioned in my predictions post, I believe 2017 will be a breakout year for deployment of Artificial Intelligence. By this I mean that AI companies will start to master the processes of expectation management and social engineering in order to bring AI across the chasm of consumer perception. Several VCs smarter than I am have made predictions along these lines, so I am by no means uniquely prescient. I also share the view that AI will become the new mobile, and accordingly have begun familiarizing myself with the variety of startups innovating around AI in Asia.

One such startup, for instance, has developed a clever application for deep learning to improve rice crop yields. By taking photos of their rice fields with their smartphones, farmers can obtain AI-driven analytics and disease diagnoses in real-time in order to optimize usage of pesticides, fertilizers, water, and energy on their crops.

This led me to thinking: would a similar application be appropriate for vineyards? For example, could a winemaker minimize pesticide usage based on an AI-informed recommendation of insect susceptibility? Or perhaps even the harvest date could be optimized? If this field is ripe for innovation, it would seem to me that a startup from one of France’s major wine regions would have some legitimacy in this area.

Happy to hear from anyone working on this.

Incidentally, my favorite venue for meeting AI entrepreneurs in Tokyo is the aptly-named Deus Ex Machina, a surf shop that doubles as a café (or vice versa). If you appreciate AI (or even just a well-balanced Gibraltar, albeit brewed by a human), Deus is worth a stop. If you bump into me there, I’m probably meeting with an AI entrepreneur, so please introduce yourself. However, please don’t “out” me as a VC since the staff of the Deus only know me as the California surfer dude.


More entertaining than nitrous oxide, but M&A was no laughing matter for AirGas

January 17, 2017

Even those who may not get excited about the stodgy industrial gases business but with a passing interest in the topic of M&A may find this story intriguing. It gives a glimpse into two hostile takeover defenses: the poison pill and a staggered board.

It’s also a tale that hits particularly close to home (if home is France) because the happy ending involves Paris-based CAC40 member Air Liquide successfully acquiring its American counterpart, AirGas.

Finally, as with any compelling story, there’s a victim-turned-hero: AirGas CEO Peter McCausland, and a villain: American giant Air Products (and arch-rival of both AirGas and Air Liquide).

The industrial gases sector is a textbook oligopoly. A handful of corporations dominate the boring but ridiculously profitable business of producing chemicals in gaseous state — largely oxygen, nitrogen, hydrogen, but also a basket of other gases like inerts or rare gases. Customers of industrial gases count almost every type of manufacturing business (semiconductors, automotive, food, consumer products, etc.) as well as service sectors like health care.

[An inside joke once shared with me by my VC fund’s LP, Air Liquide, went as follows. Q: How does the production of oxygen for the health care sector differ from that for the manufacturing sector? A: Slap a big “HEALTH CARE” sticker on the side of the oxygen tank and triple the price!]

Anyway, the industrial gases sector is slow-growth but predictable. Innovation usually takes the form of incremental cost savings initiatives. Profitability in this mature market follows general macroeconomic trends. The relatively low volatility and consistent dividend policies of this sector’s firms makes them a reliable bedrock of a long-term retirement portfolio. The only way these firms can unlock a step function in growth is via the rare acquisition of a peer.

Back in 2010, Air Products thought it saw some vulnerability in its U.S. rival and made a hostile takeover bid for AirGas at a value of approximately $5 billion. Remember, in 2010 the U.S. stock market was still early in its recovery from the financial crisis. AirGas chief executive Peter McCausland judged this valuation too low and resisted the takeover bid.

When Air Products persisted, raising their offer and upping the pressure, AirGas enacted a poison pill defense, which prevented parties from acquiring more than 15% without the approval of AirGas’ board of directors. Furthermore, AirGas had a staggered board, meaning that only three directors (out of 11) were eligible for election each year.

So Air Products escalated its battle on two fronts: a proxy fight and a lawsuit. Air Products lobbied AirGas shareholders to advance the annual meeting and replace the directors: three at a time but six within a four-month window. The lawsuit challenged the validity of the poison pill in a Delaware court. AirGas lost the first skirmish but prevailed in court in a controversial ruling, essentially granting the company more time to fend off the hostile bid.

In 2015, Air Products’ new CEO approached McCausland again with an aggressive stance to acquire AirGas by leveraging the support of some activist shareholders. The historical bad blood between the companies, reinforced by the new Air Products CEO’s menacing attitude, motivated McCausland and AirGas to seek a white knight.

They found their chevalier blanc in Air Liquide. Air Liquide offered and all-cash deal valuing AirGas at $13.4 billion (including its outstanding debt), more than double the highest hostile bid from Air Products.

It’s hard not to root for the AirGas CEO in this story. By resisting the hostile takeover, McCausland doubled the value of his company and found a friendlier acquirer.

Air Liquide proved far more lavish on price, paying twice as much as Air Products had offered. However, the complementary geographic coverage turned Air Liquide into a highly motivated strategic buyer. In contrast, the bad blood resulting from Air Products’ aggressive tactics impeded a constructive deal, probably at any price.

Reinforcing the rationale for this acquisition was the year-end merger of two other sectoral giants: Linde and Praxair, a transaction whose highest praise was “not the worst deal in the world” by the Motley Fool.

Finally there’s a moral to this story for all activist investors and meddlesome board directors: sometimes, even against your better judgment, management is right.


Experimenting vs Predicting

January 11, 2017

I am more inclined to invest in entrepreneurs who excel at experimenting rather than excel at predicting. I acknowledge the irony of writing this on the heels of posting a series of tech predictions for 2017.

Let me explain.

The act of making predictions is fun… and invaluable. It’s a prerequisite that entrepreneurs form a view of what’s going to happen in a particular market and build a company toward that vision (similar for VCs, by the way). Smart people are capable of thinking critically about their environment and accordingly, are good at making educated guesses about the future. While even the most brilliant minds are often wrong, they’re right often enough to make a difference.

The principal dangers with predictions are two-fold: i) wasting too much time ruminating so as to impede action (analysis paralysis); and ii) holding too steadfastly to our worldview while the market evolves. Changing our view means admitting we made an error. It’s a failure of sorts, and we’re wired to abhor being wrong. So we dismiss disconfirming evidence and embrace data that reinforces our view.

Meanwhile, as time progresses, the cost of reversing course increases. The stakes rise.

Experimentation, in contrast, has the expectation of failure baked in.

Adopting a mindset of experimentation is difficult because experimenting means admitting we don’t have all the answers. It means embarking on a series of paths where failing or reversing course will be necessary on multiple occasions. It means positioning ourselves to be wrong at some future point.

I submit however that experimentation is one of the most important habits of entrepreneurs, especially for those operating in a fast-paced sector like technology. Predicting the direction of change is one thing, but predicting the speed and specific form in which it arrives is impossible to do reliably.

By valuing experimentation, entrepreneurs are recognizing the complexity of social-technological-economic systems. Conducting a collection of small-scale experiments is the most efficient way to identify the best solution for the challenge at hand: be it related to product, pricing, positioning, business model, etc. Each failed experiment becomes an opportunity to learn, when the stakes are low, and enables the company to allocate more resources to the experiments that are working.

Entrepreneurs should espouse a culture of experimentation in their startups. In turn, their investors and boardmembers should foster an environment in which the founder feels safe to experiment and is not crucified for the inevitable missteps.


A few more predictions for 2017

January 7, 2017

I sincerely appreciated all the positive feedback on publishing my past predictions scorecard as well as on this collection of 2017 predictions from an exclusive all-women cast of investors. Thank you.

Some readers also prodded me to stake out my own set of tech forecasts for the year ahead. While I contend that the best insights came from my peers, here are a few more:

  • Mass consumers in the West will finally acknowledge definitively that Asia is capable of producing truly inventive new applications of consumer tech. My guess is that a breakout gadget like this “facial perfection” selfie camera in China will take off in the vanity-stricken cultures of most Western countries. With functionality embedded directly into the camera, this Casio automatically removes blemishes, slims the face, whitens the skin, and enlarges the eyes (perhaps a skin tanning function will emerge in the culture of sun worshippers).
  • The “down-with-the-banks” element to the fintech revolution will lose some of its luster. The saturation of fintech startups in Europe will witness consolidation, with the winners having already attained scale and found ways to partner with the banks rather than replace them; however, Japan’s fintech space will accelerate.
  • Agritech (or Agtech), in contrast, will ascend the curve of inflated expectations. “I can’t believe it’s not meat” substitutes and water-efficient avocados will attain widespread acceptance, taste great, and be less filling.
  • 2017 will be the year AI crosses the chasm of consumer perception. I cannot take credit for this prediction because a number of smarter VCs have made the same prognosis. However, I share the view that AI will become the new ‘mobile’. I also suspect we will see more surreptitious acts of AI, such as the recent stealth play of AlphaGo in the online Go community.
  • The invincibility of the app stores will soften, with the pendulum of centralized vs. distributed IT architectures swinging back. Streamable ‘apps’, HTML5 light versions, and integrations of mobile applications in other creative ways will emerge as a viable alternative to the tedious mobile app on-boarding process.
  • Kisenosato will win a sumo championship. He is the only ozeki to have never brought home the Emperor ‘s Cup despite being the winningest wrestler of 2016.

Happy and healthy new year to all. がんばってください。!

Forecasts for 2017 from Five Insightful Investors

December 28, 2016

With the first full year of Brexit implementation, French elections, the inauguration of President-elect Trump, and a start to the baseball season with the Cubs as the reigning world champs, the world could not feel more unpredictable. In 2017, entropy feels more palpable than ever before.

‘Tis therefore the season to invite some of my colleagues in VC to offer some predictions for the upcoming year. These talented individuals are much smarter than I am, so it’s an honor to solicit their wisdom during this period of chaos.

Claire Houry, Ventech, Paris

For 2017, I see e-tailers going offline, reinventing the offline world and merging the physical and digital world to create an immersive environment. We shall see augmented and virtual reality booming in marketing applications, data-driven indicators and online analytics entering the offline world and intelligent apps (virtual customer assistant) performing some of the functions of a human assistant. Be ready to see internet of things invading all types of products.

Geopolitical views: With the world’s biggest start-up campus to open in Paris in April 2017, I also see France getting high attention in the tech world. Ventech supports the project and will launch its ParisPOC program for its foreign portfolio companies, willing to expand into the French and the European market.


Anna Boffetta, Balderton Capital, London

i) The list of winners in machine learning will shift from including only those who can gather the biggest dataset for training to also include those who develop training methods that rely on fewer and fewer data points. In the past years we have made huge progress in Artificial Narrow Intelligence (automation, optimisation, prediction etc.) but, thanks to the increase in computational power, the next years will begin to build towards Artificial General Intelligence, where computers get closer to human-like intelligence.

ii) 2017 will see the rise of new interfaces. While today’s dominant human-computer interaction is done through the display and visualisation, mostly on mobile, new interfaces like voice, audio, messaging etc. will start seeping into everyday life. More data will be collected, structured and analysed by machines which will deliver the results to humans through conversational interfaces.


Allison Baum, Fresco Capital, Tokyo

i) Increase in cross-border M&A – The US startup ecosystem is ripe for consolidation and most corporates in Asia are currently cash rich but innovation poor. We are already starting to see the beginning of this trend with our own portfolio companies, where we are receiving inbound interest from international companies looking for new business lines, tech IP, or an entry point into the US.

ii) Edtech goes mainstream – Education technology has incorrectly been identified and relegated to a niche market. However, the combination of the economic and political turmoil we’re seeing in all corners of the world, and a rapidly accelerating rate of technological progress (self driving cars are very real now), jobs are going to be in focus. Tech related and tech enabled training will no longer be a nice to have, but instead a critical need for the health of individuals, corporations, governments, and society as a whole.


Adizah Tejani, Former Level 39 Fintech Accelerator, now at, London

Over the last few years, finance has undergone change as technology shifts the landscape on what is possible. With the up and coming PSD2 changes across Europe, I think the move to platform banking driven by an API economy will start to become clearer in 2017. Banks and technology companies will continue to adjust through collaboration to navigate this change.


Anja König, Novartis Venture Fund, Basel

I like to quote Yogi Berra, “It’s tough to make predictions, especially about the future.” With this caveat in mind, one doesn’t have to be a wizard to predict that it will be easier for European companies and academic institutions to recruit top talent from the US.



All the best for 2017.


My prediction scorecard

December 26, 2016

I ruffled a few feathers back in February when I wrote how I think HTML5 games will become trendy again. Nine months later, Facebook launches Instant Games, offering HTML5 games directly inside the chat feed. Perhaps a bit premature to brag, but I’d like to suggest that I really nailed this one.


In contrast, earlier I had predicted that the sense of touch would become more deeply integrated into our smartphone UXs, i.e. kind of haptic feedback 2.0. This one didn’t materialize, at least not yet.


I had also predicted that the UK would remain in Europe and that Hillary Clinton would be elected President of the U.S. I’m going to try to resist making political predictions in the future.


Coming up next, some astute predictions about technology and innovation from some truly brilliant and globally-minded VCs.

Giving thanks for 2016

December 20, 2016

Around this time every year I make a habit to reflect on the countless benefits in my professional life for which I am thankful. Out-ranking these items are a multitude of blessings on the personal side, but I’ll restrict this list to the business ones.

2016 proved a tumultuous year on the geopolitical front, making this exercise of appreciative stock-taking all the more important in my opinion. Here are five things for which I was grateful in 2016:

  1. For starters, the management teams of my various portfolio companies, both in Europe and Japan. A few experienced breakout years, others struggled, yet all approached the challenges they faced with dignity and diligence.
  2. The people that have worked for me, both current and former. I would not have accomplished so much without your collaboration.
  3. The luxury of my work allowing me to shuttle between my two favorite cities of the world: Paris and Tokyo, and the opportunity to interact with inspiring individuals in each.
  4. The progress which I’ve had a front row seat to witness of the innovation ecosystems of France and Japan. The former has sped from laughable to laudable in the 15 years I’ve invested in it. The latter is at an earlier stage, and I’m enthusiastic about its prospects in several sectors, such as fintech, AI, agricultural technologies.
  5. Pertaining to Japan, I’m also grateful for the professionals who have guided me as I stumble my way into a new investment ecosystem. The fellow VCs, the talented entrepreneurs, the savvy journalists and friendly bankers, the captains of industry who have hospitably welcomed me into their circles. Thank you.

Happy holidays to all.

France’s tech incubators and accelerators [infographic]

December 14, 2016

In complement to my post last week about Station F, which proclaims to not be an incubator but rather a ‘startup campus’, I figured it would be helpful to publish an infographic of the numerous incubator/accelerator options for startups in France.

If the chart below appears cluttered, that’s because tech incubators/accelerators are beyond the saturation level in France. Too many? Probably. However, I generally believe the best policy is to let a thousand flowers bloom and then garden appropriately.

Special thanks to my intern Wen-Chiao for assembling this RUDEscape of France’s tech incubators and accelerators.


Station F now accepting applications

December 9, 2016

Last summer I publicly weighed in — part tongue-in-cheek without being asked but part seriously — on what I thought Station F’s big hairy audacious goal should be: namely, become the startup hub of Europe.

Calling itself the world’s largest startup campus (love the ambition), Station F is now accepting applications from startups worldwide. This strikes me as an incredible opportunity to get in on the ground floor of a project that could become monumental. Backed by Xavier Niel and ably managed by the talented Roxanne Varza, Station F aims to house an entire startup ecosystem under one big roof.

If I weren’t based in Tokyo now I’d seriously consider setting up an office for our fund inside Station F, as other smart VCs like Ventech, Kima, and Daphni are doing.

My advice (again free and unsolicited): if you’re an ambitious entrepreneur in the market for a European office base, consider applying to Station F’s on-site residency program now.


Dash Camp Japan 2017

December 4, 2016

Pour la quatrième année consécutive, on m’a accordé des places pour inviter une délégation française de 3 startups privilégiées au prestigieux Dash Camp au Japon.

Dash Camp représente la conférence de référence en Asie qui facilite les échanges entre les entrepreneurs innovants et les dirigeants des grands groupes tech japonais. Elle est pertinente pour ceux qui souhaitent être sur le radar des acquéreurs japonais, chercher des partenaires commerciaux au Japon, ou bien attirer les investisseurs asiatiques qui ont un appétit pour l’Europe.

L’édition 2017 de Dash Camp aura lieu le 16 & 17 mars 2017, dans un cadre propice à Fukuoka, en dehors de Tokyo.



  • startups tech de moins de 8 ans d’existence
  • ambitions au-delà de la France pour le marché japonais (démarche commerciale, partenariat, m&a)
  • domaines d’intérêt: agri-tech, objets connectés, fintech, adtech, sharing economy

ようこそ !

Plus d’informations sur Dash Camp 2017

FB Messenger Instant Games: A new era of friendly play (except for iOS/Android)

November 30, 2016

Quick primer for the millenials in the audience who weren’t yet born: computing used to be performed by these (relatively) powerful, monstrous machines called mainframes. Mainframes were programmed and accessed via dumb terminals which resembled not much more than a keyboard, screen, and wired connection to the mainframe, kind of like on those portrayed on that old retro-futuristic TV show Max Headroom (oh nevermind, you were still in diapers).

Then one day, the personal computer was born. The aforementioned dumb terminals became intelligent, with their own memory, storage space, processing power, etc. Also thanks to Moore’s Law, mainframes shrunk in size while simultaneously becoming more powerful. Somewhere along the way we started calling the mainframes servers, and increasingly created software applications that would harness both the distributed PC resources of robust clients and the centralized server resources. The era of client/server computing had dawned. (not to mention the age of client/server consultants, err I mean advisors for strategic transformation of pretty much any enterprise that fit into cool new categories like ‘commediatainment’).

Unbeknownst to those of us without sufficient security clearance, some wizards at DARPA were staying up late building something that Al Gore would later “reinvent” as The Internet, which was really just a bunch of tubes, until a really smart dude named Tim created the world wide web and then another smart dude named Marc added a graphical interface allowing regular people to navigate the whole internet landscape. The point is, the web swung the pendulum back toward the centralized configuration of concentrating most of the computing processes into powerful, centrally-located servers, which were in turn accessed by relatively dumb terminals (now called thin clients) in the form of web browsers.

Trends like ASPs, SaaS, and eventually Cloud computing (which were essentially fancy buzzwords for the same basic concept) further reinforced these centripetal forces (as opposed to centrifugal forces) on the centralization of computing. Client/server became old school, and the paradigm of the thin client was back in vogue.

However, as you millenials know all too well, the equivalent of client/server has been alive and well since 2008 when a super smart dude named Steve released the iPhone and opened the App Store. The pendulum had swung back yet again, this time with the client machines manifesting themselves in the form of smartphones.

It’s no secret that the computing resources in our pocket are more powerful than those original mainframes. We’ve become so accustomed to finding a mobile app for everything that thought leaders began debating whether the mobile web was in irreversible decline.

I personally couldn’t accept this thesis that the universal cycle of life omnipresent in IT systems had been permanently knocked out of kilter. Not because I’m particularly insightful but more because I’m a fan of Herman Hesse, Star Wars, and Eastern belief systems that acknowledge the transcendental concept of duality.

Yesterday, Facebook confirmed recent rumors by launching Instant Games inside FB Messenger. One billion Messenger users can now play HTML5 games directly inside their chat feed with their friends. No need to hunt among the thousands of native games in the app stores, no need to suffer app download fatigue.

Granted, this is a relatively modest development for the moment, restricted to casual, snackable games. Yet something tells me this initiative may foreshadow potentially monumental implications for the app stores.


Visit Us On TwitterCheck Our Feed