My notes from attending Camp Hustle, a unique LP-GP gathering. Includes actionable tips for those raising Funds 1 and 2.
It was a great experience to attend Camp Hustle last week. I have been following the work of Elizabeth Yin (co-founder of Hustle Fund, a pre-seed fund that organized this event) on X for some time now. All this while, I kept seeing posts from the last 2 editions of the event.
This time, I was in town, in a relationship-building zone, and looking to add new and interesting LP-GP folks to my community. So, I landed up at the event (who would say no to spending 2.5 days in the idyllic surroundings of Los Gatos and Saratoga anyway!).
Honestly, I didn’t arrive with a specific agenda or expectations from the event. I just went in with an open mind and knowing the vibe of the Hustle Fund team, I instinctively knew this would likely be the best frame of mind for this gathering.
The event turned out to be a pleasant surprise. Now I know why the name has the word “Camp” in it – the entire event had an informal, outdoorsy, campy, yet energetic and authentic vibe to it. Everyone agreed to an informal social contract – no explicit pitching, no so-called networking and no shallow talk. Everyone bought into the idea of just getting to know a bunch of folks and really bringing their whole, authentic selves to the event.
While the free-flowing, candid conversations amidst nature were the highlight of the event, I did end up with some really actionable insights shared by the Hustle Fund team, other emerging managers as well as a few LPs. Sharing my notes below:
While interacting with potential LPs, focus on making them a “fan” of the fund first. That is the first step towards eventually converting to an LP.
One common mistake during fundraising as a first-time manager is chasing people too aggressively. The key is to put out your story and let people come to you.
A great way to engage potential LPs is to send out a monthly/ quarterly newsletter. Also, Virginie mentioned doing informal LP meets in the Spring and Fall, so folks stay connected with the fund.
The majority of potential LPs you meet today might eventually invest in Funds 2 and 3. So, it’s important to start building relationships from now.
I asked Charles a question on ways to increase conversion on warm intros that a GP gets via existing LPs. While intuitively one might expect a healthy conversion on this type of lead, Charles confirmed that in his fundraising experience, the conversion on these referrals was indeed lower than expected.
2/ Venture investing learnings from the Hustle Fund team
During an informal AMA, Hustle Fund co-founders Elizabeth, Eric, and Shiyan shared the following top venture investing learnings from their anti-portfolio:
Always bet on your friends.
Don’t penny-pinch on valuation (they passed on an initial round of one of the largest consumer Internet outcomes because of valuation).
“Good deals have legs” – when you like a founder, push as hard as possible to get into the deal. Don’t be afraid of being perceived as a pain, if it can help you get into the deal.
Don’t over-index on what a market or company looks like right now. Learn to imagine what the market or a company can become “over a period of time”.
3/ Tips from an institutional LP
Courtney McCrea (Co-founder of Recast Capital) is one of the most experienced institutional LPs out there. In a candid Bonfire Session, she shared some insightful tips for emerging managers:
During an LP pitch, don’t be afraid to talk about how great you are. In fact, spend the first 3 minutes in a pitch just talking about your unique superpowers.
If you are having trouble creating a unique narrative for why your fund is different, ask your portfolio founders why they picked you and how they would pitch you to their friends.
LPs look at who you co-invest with and who does follow-ons in your companies, as signals for the quality of your deal flow.
There are so many LPs out there who aren’t pitched very often. Try and focus on them to improve your odds. Don’t underestimate the amount of capital that is out there looking to be deployed.
4/ Other helpful convos
Matthew Stotts of Cerulean Ventures shared that outlining the 10-year vision and story of what the fund is looking to do, goes a long way in generating excitement as most LPs are looking not just for financial returns but also for impact in whatever their personal mission is.
To re-engage with potential LPs in your funnel, try going back to them when you have an interesting development or story to share from the portfolio (eg. “we invested in Company X at the pre-seed stage and now, 12 months later, they just cracked a $XMn ACV deal”). This could also be done with a new differentiated investment or interesting deal flow.
[Via Rahul Vohra, Founder and CEO of Superhuman] One of the best pieces of advice Rahul got while building Rapportive was to pick a strategy to go from Point A to Point B, never change your mind about it, and continue relentlessly executing it. The goal could be say reaching 1Mn users, $100k MRR, or any other metric. The key is to stick with it.
Hope these notes help emerging managers out there. Once again, thanks so much Team Hustle Fund for creating this unique event format. Am excited for the next one!
PS: For those in SE Asia, the next Camp Hustle is in Bali in September🏖️
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
Grit needs not just perseverance, but also passion for a long-term goal.
Loved this talk on “Leading with grit” by Angela Duckworth at the recent Norges Bank Investment Conference 2024. This is a particularly interesting topic for me given scouting gritty founders is one of the pillars of my venture strategy.
Some interesting insights from this talk:
1/ Definition of Grit
GRIT is sustained passion and perseverance for especially long-term goals.
Grit needs not just perseverance, but also passion for a long-term goal.
2/ Talent & IQ has NO correlation with Grit
Grit actually unlocks the latent talent of people.
3/ A nuance of the 10,000-hour rule
It’s not just the quantity of practice that makes a world-class expert, it’s also the quality. High-quality practice can be called “deliberate practice”. Low-quality practice will only take you to the plateau of “arrested development”.
4/ “Dropping out” can be valuable in certain cases
If you start something and with some practice, realize that you aren’t enjoying it, or that the opportunity cost is too high, or things are just not working out for some reason, it’s perfectly acceptable to drop out.
Essentially, mindless grit should be avoided.
5/ Three elements of deliberate practice
(a) Decide on a small sub-skill to practice.
(b) Practice with 100% focus (the opposite of “multi-tasking”).
(c) Feedback and reflection.
While the best at this practice alone, they also work with a coach to show them the results of their practice and get feedback.
6/ The hierarchy of goals
Important to have a top-level goal that is long-term. But then, this goal should be broken down into a set of mid-level goals that in turn, are broken down into extremely tactical low-level goals with a daily/ weekly cadence.
Leading with grit means being extremely stubborn with the top-level goal but highly flexible and agile with low-level goals.
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
Whenever someone thinks of starting anything new, say a habit like working out or reading regularly, a side hustle, a passive income project, or even a full-time venture, they should expect to enter the “Funnel of Doing New Stuff”.
This past week, I launched an experiment of extending An Operator’s Blog into a podcast (I call them “jamming sessions”). Through these sessions, my idea is to discuss real operating questions and challenges faced by early-stage founders and folks in the venture ecosystem. Also, invite guests who are deep experts in their area but aren’t very visible.
The first episode is on “Doing 0-to-1 as a cybersecurity startup” where I went into brass tacks with two amazing venture-backed founders and engineering leaders in the domain – Buchi Reddy B (Founder and CEO of Levo.ai; ex-Traceable AI, ex-AppDynamics) and Ruchir Patwa (Co-founder and CEO of SydeLabs; ex-Google, ex-Mobile Premier League).
If there is one thing I learned from my past experience as a founder, it was to ship the MVP and put it in front of users as fast as possible so that the iterations can begin. Therefore, we recorded the episode on Apr 19 and I released a fully edited version to the public on Apr 23. Even before experimenting with this new format, I had resolved to ensure that whenever an episode gets recorded, it gets released within the next few days.
This idea also stems from my frustration wherein I was a guest on podcasts where the episodes were still not public even after a few months of recording. I thought this was just my experience but talking to others in my network, this apparently is quite common.
Given we live in an age where content is being thrown at us with high velocity from every direction, even the most insightful conversations have a relatively limited shelf life. An episode recorded 2 months back is likely to feel stale to listeners today. Then why are these part-time recorders hell-bent on maintaining a huge backlog of recorded material?
I was brainstorming this with my better half and an interesting idea developed during the conversation. Whenever someone thinks of starting anything new, say a habit like working out or reading regularly, a side hustle, a passive income project, or even a full-time venture, they should expect to enter the “Funnel of Doing New Stuff”.
This is how the funnel plays out in real life – almost everyone out there is constantly ideating about something new they want to do. Everyone wants to start posting more on LinkedIn, or write more, or hike more, or network more.
However, this is where the first stage of drop-offs happens. Very few people take the first step. The inertia of being busy with daily life kicks in for most people. Other times, it’s the fear of failure that stops folks. Or the potential public embarrassment in case things don’t work out.
Next, for the few brave hearts who take the first step, a new challenge awaits them. This is the challenge of staying consistent with this new thing. This is another stage of massive drop-offs, where people begin but don’t consistently execute and eventually give up.
I see a few psychological aspects at play in this stage of the funnel. People generally struggle with any new habit formation, in part because they are unaware of nudges and brain hacks one can use to make the process easier.
Also, we are dopamine-driven creatures wherein our brain naturally seeks excitement. And this excitement is easily found more in doing new things and getting into new experiences (eg. travel, adventure sports, trying new restaurants, social media, etc.) vs. repeating the same task. Hence, consistent repetition is always a mental and psychological challenge for most people.
Now, even for this next cohort of amazing souls who have both started and also stayed consistent, the game isn’t over yet. For any new initiative to translate into real outcomes (business, financial, or life), it’s crucial to constantly iterate and improve on the initial minimal offering (eg. my podcast MVP).
The initial phase of any new project is almost always internally driven – a hypothesis, belief system, or worldview. But for it to resonate with others – your users, customers, audience, partners, or even your own sensibilities, requires running a continuous feedback loop that includes perpetual learning and refinement until people start loving it. Then, this love is a currency that one can use to drive many types of outcomes.
This loop of continuous iteration and improvement isn’t natural to even the best talent, hence there are again drop-offs at this final stage of the funnel. Staying grounded in reality, listening to feedback, and having the humility to change or let go of stuff are all ingredients needed to succeed at this stage. Many don’t make the cut here.
This funnel idea also has an interesting implication for how one should view competition. At a macro level, most fields look cluttered and competitive because one tends to focus on the top of the funnel (the ideating mass) as competition. The reality is that the real competition is at the bottom of the funnel – the handful of highly driven people who started, stayed consistent, and also constantly improved with each rep*. That number is usually fairly small in whatever area or field you look at, and in my view, there is always room for more there.
*An analogy that people who grew up in India would understand – for a serious student attempting to crack the prestigious IIT JEE engg. entrance exam, the competition isn’t the 1Mn students who have registered to take the exam but only the ~100k or so who have put in adequate reps to prep for it.
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
As we enter the era of a new generation of critical technologies, from AI and AR/VR to EVs and robotics, the technical and entrepreneurial horsepower of immigrants will be more important than ever for Silicon Valley.
Yesterday, I had the opportunity to attend Entrepreneur First‘s first-ever demo day in San Francisco. Folks in the US and India might not be too familiar with EF – they are one of Europe’s top incubator programs, with a particularly strong presence in London and programs now in Paris, New York, and Bangalore.
Reid Hoffman at the EF Demo Day in SF
EF’s model is interesting – they operate in the -1 to 0 stage, spotting deeply technical founders, mostly in their early-to-mid 20s with many straight out of college, and help them identify and incubate a startup idea that aligns with their core technical skillsets and achievements.
And they are definitely spotting some outlier talent. Within this cohort, I saw everything from a Math Olympiad gold medalist, a Material Science PhD from Cambridge, and a 3rd year PhD dropout in Brain-Computer Interfaces to a Formula 1 aerodynamics engineer, someone who built systems for the US Department of Defense and another who worked on JP Morgan’s first AI systems.
This is what made the demo day super interesting for me. With the advent of AI, Europe is gaining prominence in the global tech scene courtesy of excellent technical universities and research institutions that produce some of the most cutting-edge research talent. A majority of EF cohort companies are in deeptech/ applied sciences and therefore, this demo day in a way, gave a glimpse into the future that leading AI research can potentially bring to life.
My 1 line takeaway from seeing these 32 companies pitch – the future is brighter, and full of “tech magic”, than we can probably imagine right now. Get a load of some of the ideas that are already in early productization:
1/ World’s first AI training processor using photons (directly taking on Nvidia).
2/ Optimizing farming 24×7 with low-cost swarms of Roomba-like robots that live in fields and spray everything from fertilizers to pesticides.
3/ AI platform that does automatic product placement within creator videos (a YouTuber can place everything from a Nike shoe to a Fiji bottle within a video in a matter of minutes).
4/ AI-powered real-time language translation that freelancers in non-English speaking nations can use to work with clients across geos.
5/ Exponentially simplifying going from a 3D render to a detailed pre-manufacturing drawing & design for any production process.
6/ Non-invasive neural links that can help soldiers in a hot zone communicate with each other without talking (telepathy brought to life?).
The raw intellect of these founders, combined with the product progress they appeared to have made in a short period, makes me think that many of these ideas are not that far away from commercialization.
What EF is smartly doing is relocating this entire batch to Silicon Valley, where the founders will live full-time, building product and raising capital. Seeing the ambition level of ideas the cohort is taking on, they definitely need the risk appetite and vision-backing mindset of the Bay Area. Can’t think of any other ecosystem in the world where such technically complex and capital-intensive ideas can be backed by a combination of talent, risk capital, institutional knowledge, and diverse networks.
Which brings me to another thought – how talented immigrants continue to move to the Bay Area to build the future. Imagine such unique outlier talent from places like Europe and India choosing to uproot themselves from their home countries, moving to the Valley, and offering their unique skills & knowledge to companies here. This makes me super-long on the Bay Area and clearly shows that the Silicon Valley immigration flywheel is still as strong as ever.
This macro trend is what also makes me equally excited about India’s emergence as a key supplier of founder talent for the world. And not just to the US, but also to regions like SE Asia, the Middle East, and Australia. I believe the Indian diaspora will make a defining impact on the global knowledge economy over the next 20 years. Combine this with the rise of a new generation of critical technologies (AI, EVs, AR/ VR, robotics, semiconductors, etc.), and this transforms into a generational opportunity that energizes me as a venture investor in the US-India corridor.
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
Life is cumulative. It unfolds in a geometric progression and therefore, the timing of your initial wins has a major impact on future outcomes down the road.
Met someone earlier this week who has had a bunch of major outcomes as an angel investor. This sparked my curiosity and I asked him about his backstory and career trajectory. Turns out these venture outcomes were a result of a series of parlays and things coming together over a decade-and-a-half:
He was early at a leading startup in the mobile supercycle, which was acquired by a BigTech in a marquee transaction. He continued at the BigTech for a few years (which in hindsight, was still relatively small at that stage), thus rapidly compounding both his net worth and network.
His spouse was an early engineer at a now-leading BigTech, then left to become an early engineer at a startup that was acquired in another OG marquee transaction. Through this journey, she also built a deep relationship with one of the OG Tier 1 venture firms in the Valley.
The couple used the capital acquired from this track record to start writing angel checks. Alumni of all the companies they worked at gave them access to some of the best deals.
The guy also went on to join a venture firm later, which further added to his creds and network.
Essentially, as a direct outcome of their early individual successes, this couple benefited from a self-compounding flywheel of relationships and capital. Pooling these assets as a married-team further magnified their impact. To their credit, in addition to being highly capable, they had the hustle, risk appetite, and foresight to keep taking shots at various opportunities that came along their way.
Btw, this story is not that uncommon in Silicon Valley. Though the extent of financial outcomes might vary, I know of many such stories where people have benefited from similar flywheels in their tech careers. In fact, this is one of the things that makes the Valley a unique place as there is an adequate density of talent, capital, networks, positive intent, and implicit trust within a small geographical region, which enables such flywheels to take shape in people’s lives. PS: I had written about this idea in my post ‘The Success Flywheel‘.
This story also highlights the importance of something I think about a lot, even from reflecting on my own career – there is a massive advantage to putting points on the board early on in life.
Life is cumulative. It unfolds in a geometric progression and therefore, the timing of your initial wins has a major impact on future outcomes down the road.
Mark Spitznagel, famous tail-risk trader and Taleb’s Partner at Universa, talks about this concept in the context of financial portfolio management and risk mitigation in his book ‘Safe Haven‘.
We are not a casino, or a portfolio of our distribution of possible simultaneous returns. Rather, we are one wager compounded through time. We only get one chance, and, if we shine a bright light on that, we will avoid many mistakes—start thinking about the right things, with a better internal valuation metric: making sure this chance maximizes its chance.
The idea is simple but powerful – having early wins enables the player to parlay the fruits of that win into the next opportunity while also having a long enough time runway for significant geometric compounding.
Being in the right zipcode like Silicon Valley in tech or NY in finance, also provides a large enough sample set of opportunities for continuous parlaying as well as high rates of compounding given the inherent leverage in these ecosystems.
This idea also makes the case for why students try so hard to get into Ivy Leagues, or why VCs try their best to get into prominent logos early in their track records. It also frames the competitive advantage folks get by starting as a fresh undergrad Analyst at Goldman, engineers who joined Google in the mid-2000s straight out of college, or those in their 20s joining OpenAI right now. The difference in getting these early wins starts showing up a decade later when the slope of the curve of these folks is markedly steeper than those who didn’t.
Of course, logging early wins isn’t by itself a sufficient driver or a definite leading signal of holistic success later in life. Everyone has their own unique journey and has to walk their own path. Still, given the sheer leverage these early points provide, it’s worthwhile to have this at the back of your head while executing your career strategy.
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
In Q4 2014, top 3 Indian OEMs Micromax, Intex, and Lava together had >30% smartphone market share. In Q4 2023, there are now ZERO Indian companies in the top 5 smartphone vendors!
There is a critical lesson here that India needs to learn as the AI supercycle dawns on it.
It has been fascinating to see videos showcasing Xiaomi’s SU7 EV sedan emerge over the last few days. The interior UX, in particular, looks fantastic. With BYD having entered India in 2021, and Xiaomi’s well-known expertise in product design and global distribution, China has emerged as a powerful player in the global EV space. Of course, with its prowess in hardware manufacturing, it was never a question of ‘if’, only ‘when’!
In parallel, the AI supercycle in the US is having positive ripple effects in hardware and robotics. Apple is expected to make significant progress in the next couple of years on the 1st gen Vision Pro. Meta is focused on rapidly improving the Quest 3, while it also collaborates on the Ray-Ban smart glasses. Zuck also recently spoke about developing a neural interface in the form of a wristband, which can read the signals your brain sends to hands and arms.
US startups are already going gun-ho on imagining new form factors to deliver AI capabilities to users. These include AI pins, AI necklaces, and desktop robots that turn your smartphone into a whimsical companion.
With massive investments going into the chips and infra layer in the AI cycle (read my post: ‘AI Musings #6: The Bull Run Is Just Beginning (90s Telecom Boom Vibes)‘), the resulting drop in compute prices and democratized access to it is providing tailwinds to hardware verticals whose adoption has been previously blocked by uncomfortable form factors.
With all this hardware action in the US and China, I believe it’s a matter of national importance that India now steps up its game in key verticals of consumer hardware, and avoids ceding this territory to overseas players.
In the mobile supercycle, I had a ringside view into how in smartphones, homegrown Indian companies like Micromax, Karbonn, Intex, and Lava didn’t end up investing in domestic manufacturing capabilities, playing the short-sighted game of buying stock hardware from Chinese OEMs/ ODMs, and just putting a sticker and packaging on it.
Further, even with a solid local software talent pool, these companies also under-invested in software, building only basic wrappers on top of stock Android. Essentially, Indian smartphone OEMs ignored both these key areas of competitive differentiation, diverting capital away from capex & R&D, and towards brand building and customer acquisition.
This strategy was found wanting when Chinese OEMs like Xiaomi and OnePlus aggressively entered the Indian market. Given their deep manufacturing and digital expertise in China, these companies had strong differentiation in both hardware and software. Plus, they had deep pockets to outspend Indian OEMs on marketing. Therefore, despite entering the Indian market much later and starting well-behind Samsung and Indian OEMs, Chinese smartphone companies managed to build strong local brands in a fraction of the time.
The end result is this – in Q4 2014, top 3 Indian OEMs Micromax, Intex, and Lava together had >30% smartphone market share. In Q4 2023, there are now ZERO Indian companies in the top 5 smartphone vendors!
Given the ambitious goals we are setting for the Indian economy over the next decade, and the national importance of critical technologies like AI, automotive, energy, space and defense, it’s important India doesn’t repeat the strategic mistakes of the mobile supercycle. It’s imperative that indigenous hardware capabilities get built during this next AI supercycle (and other constituent subcycles in AR/VR, EVs, space etc.).
From what I am hearing about all the foundational work already happening in semiconductors, automotive, space, and general manufacturing, this is totally doable if the Indian public and private sectors can come together, backed by an encouraging policy stance from the govt. We are already seeing early greenshoots of this in space tech, where ISRO is actively collaborating with Indian spacetech startups.
I want to throw out a challenge for Indian founders, asking them to be more courageous in picking tough hardware problems to solve. There is enough global capital available that is positive on India and in its chase for alpha, will be ready to back this courage.
Further, this is also the right time for Indian conglomerates like Reliance and Mahindra to step up in a meaningful way and drive INR capex in these critical sectors. In parallel, the govt’s policies should ensure that while hardware manufacturing attracts investments from all over the world, the strength of homegrown players also get built up in this supercycle.
China is a good example of leveraging foreign capital to strengthen its domestic manufacturing and digital capabilities. Technology indigenization is critical in this age of fickle geo-politics where everything from trade to currencies are being weaponized.
This current generation of Indian founders also have the benefit of home-grown role models like Sachin Bansal and Binny Bansal, who stood up to US and Chinese competition in eCommerce, ultimately ensuring a homegrown & enduring market leader like Flipkart continues to thrive to this day.
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
The current AI landscape is giving strong mid-90s telecom boom vibes. Studying that cycle suggests that we might only be at the beginning stages of a multi-year bull cycle, wherein the ongoing massive investments in chips, infra and foundational models will ultimately enable the rise of enduring AI applications.
The last few days have been hectic in the world of AI. Introspecting on how things are unfolding at present in public, private, and commercial markets, where AI is right now is giving me major mid-90s telecom boom vibes.
Hyperscalers continue to raise huge $$$
1/ Microsoft did a sort of acqui-hire of Inflection AI for $650Mn via an interesting licensing deal structure. Given that Inflection was one of the high-flying foundational model startups and had raised $1.3Bn from Microsoft and Nvidia (cash + cloud credits) at a valuation of $4Bn in June last year, it’s unclear whether this is a good or bad outcome for employees and investors. Although, going by this tweet from Reid Hoffman where he mentions “good future upside”, looks like shareholders got some sort of equity package from this deal.
2/ Amazon concluded its initially committed $4Bn investment in Anthropic by investing the second tranche of $2.75Bn. Am assuming like other hyper scaler deals, this is a mix of cash and cloud credits. As per CNBC, this tranche was done at the first tranche’s valuation of $18.4Bn. The press release from Amazon hinted at deep integrations between both platforms, with Anthropic using AWS as its primary cloud provider internally for product development while also offering the latest generations of Claude-3 foundational models to AWS customers via Amazon Bedrock (a fully managed service for LLMs).
3/ As per The Information, Canadian pension fund PSP Investments is about to co-lead a fresh round of financing in Cohere at a ~$5Bn valuation. The company’s last round was at a ~$2.1Bn valuation last year, and given it’s reportedly at a $22Mn ARR currently, this is a rich revenue multiple to pay. Based on my conversations with BigTech AI operators, Cohere is significantly lagging Anthropic in terms of foundational model capabilities.
The current phase of AI seems to be like the beginning of the telecom boom in the mid-90s
Personally, I am finding it hard to predict which of the current foundational model hyperscalers and AI-first application companies will survive. Further, given the inflated valuations these deals are being done at, barring logo grabbing, I don’t see how investors can make outsized venture returns in these deals.
In parallel, while meeting super-early AI companies in categories like dev tools, security, and deep domain applications, I am struggling to see a clear right-to-win for a majority of them. Given data and distribution advantages of incumbent products both in Enterprise and Consumer, it’s unclear which seemingly-white spaces are actually viable startup opportunities.
However, amidst these struggles as a venture investor, I am feeling good about one hypothesis – all this capital going into infra and foundational model companies is actually building capacity for the next generation of enduring AI products to be built. This is quite similar to the role that in hindsight, the telecom boom of the 90s ended up playing for the adoption of the Internet.
As the Telecommunications Act of 1996 opened up the telecom sector to competition, a host of new entrants came in to become ISPs. They were followed by companies like Cisco, Ciena, Lucent, Nortel, and others who were desperate to sell networking equipment to these telecom companies.
Comparing this to today’s AI landscape, cloud providers seem to be similar to telecom companies, while semiconductor companies selling chips to these cloud providers are like the networking equipment companies eg. Cisco.
Also, during this boom, telecom capex was unlike ever seen before. As per the earlier cited post, just in the year 2000, capital spending by publicly traded telecom service providers was at an astonishing ~$120Bn (~$213Bn in today’s dollar terms).
This telecom boom capex is one of the largest capital bases ever built in such a short amount of time. I can see the same vibes in the amount of dollars going into AI chips, infra, and foundational models today.
Btw, one more learning from the telecom boom is how these flywheels become even stronger as the adoption of new tech starts reflecting in productivity gains. Here are some interesting excerpts on this from the Fabricated Knowledge post:
As this telecom boom was unfolding, LTCM blew up and therefore, the Fed ended up cutting rates in 1998 to avoid negative ripple effects. This is like adding tons of gasoline to a raging fire, ultimately leading up to a massive dotcom bubble.
If I play out the AI cycle like the 90s telecom boom, we might only be at the beginning stages of a multi-year bull cycle, similar to say 1995-96 (perhaps the launch of ChatGPT is similar to the Netscape IPO?).
Investments into the buildout of AI infra could run into trillions of dollars. In parallel, it seems the public markets have already started pricing in some of the future promises of AI. Going by the telecom boom, this pricing-in of future expectations could significantly accelerate for several years from hereon, driving stocks of both the telecom service equivalents (Cloud providers) as well as the networking equipment equivalents (Nvidia and perhaps any new entrants into chip manufacturing?).
That all this is happening in a higher-interest rate environment is a critical point. If for any reason (economic, geopolitical, or otherwise) the Fed starts cutting rates (which they are publicly saying they will), this could provide a major kicker into an already accelerating bull market.
So, we can reasonably posit an oncoming AI bull market for the next few (at least 3-5) years. Ultimately, like all bull markets, it will transform into a bubble, which will then peak and eventually crash. If you look at the Internet wave, the massive telecom capex of the 90s ultimately enabled the rise of enduring Web 1.0 companies like Google and Facebook, but only after the dotcom crash. Hence, I tweeted this yesterday:
What do you think?
Bonus Section: Commentary On Sequoia Capital’s AI Ascent 2024
As I was trying to make sense of recent AI funding developments, I chanced upon the just-released videos from Sequoia Capital’s AI Ascent 2024. I found these points from the keynote particularly interesting:
1/ If we draw parallels with the Cloud wave, in 2010, the entire global software TAM was ~$350Bn, of which Cloud was a tiny ~$6Bn sliver. Cut to 2023, the global TAM has grown to ~650Bn but more importantly, Cloud has grown to a ~$400Bn large piece of this pie (~40% CAGR over 15 years).
The starting pie for AI is not just software products, but also services that can be automated. So the hypothesis is that the starting pie for AI is ~$10Tn.
2/ The “Why Now” for AI is really strong, wherein a set of additive waves, starting from semiconductors in the 60s to Cloud and Mobile in the 2000s, has brought us to this stage. The ingredients to take AI from research to commercial applications are all there today.
Personally, I feel Sam Altman created the ‘iPod’ moment for AI by taking the power of AI to everyday users via a step-change ChatGPT product. In parallel, Jensen Huang should get shared credits for this catalytic moment given Nvidia’s rapid progress on giving chips more power at smaller form factors and hopefully over the next few years, making compute considerably cheaper and easier to access.
3/ In the last Cloud and Mobile transition waves, a host of new categories were created and new leaders were born in each of them. For AI, most of the major categories – (a) Infra, (b) Security, (c) Data, (d) Developer, and (e) Apps, are open right now. Hence, a massive opportunity for new category leaders to be created.
Interestingly, by depicting the white spaces this way, Sequoia also seems to hint that the Infra category is likely to be dominated by BigTech incumbents in chips and cloud. I am also reading the sub-text that Sequoia, in a way, views hyperscalers like Anthropic to be embedded within the existing cloud ecosystems (and hence, no separate logos depicted).
4/ Sequoia estimates that Generative AI companies are clocking in ~$3Bn in annual revenues in aggregate at present. As a comparison, SaaS took 10 years to get to this aggregate revenue scale as an industry, something that AI has achieved in almost the first year out of the gate.
5/ One of the early signals that AI is a real transformative wave is the sheer traction that the early products are getting across both Enterprise and Consumer.
6/ Over the last year, a majority of the capital has gone into the foundational model companies. In the Web 1.0 wave, the Application companies that came later in the cycle (eg. Google) captured the most value. The current uneven distribution of funding indicates that the Applications layer in AI hasn’t even gotten out of the stables yet.
7/ The usage numbers of AI-first products are still way behind incumbents. Eg. the median DAU/MAU ratio of AI-first products is a mere ~14%, compared to ~51% for incumbent products. This indicates that AI adoption is still in its infancy.
It’s encouraging to see that the ability of foundational models is on a continuous upward trend. At some point, this will translate into product capabilities that meet the expectations of users, which will then eventually reflect in better usage and retention numbers.
8/ I loved this slide that showed how when the iPhone was launched, the first generation of apps were either gimmicky or basic utilities. It wasn’t until a few years later that companies learned how to harness the capabilities of the iPhone to build enduring products.
Reasoning by analogy, we should expect that it will take a few more years (though perhaps a smaller number than previous waves?) for enduring AI applications to emerge.
9/ Sequoia is calling AI primarily a “productivity revolution”, similar to farm mechanization. At a macro level, this should bring down the costs of doing any task or delivering services, creating a strong deflationary force in areas like education and healthcare that have historically seen a perpetual rise in costs.
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
With millions in the bank post a seed round, Founders often face the challenge of maintaining disciplined execution. Excess capital often ends up slowing progress towards real PMF. I share a brain hack to counter this.
Was in a working session with a founder recently. The company is going after a huge market opportunity and has raised a low single-digit Mn seed round from some very good VCs.
Issues with too much early capital
The issue though is that, like most category-creating seed startups, the precise customer persona and pain points to be solved are not obvious right now. After just a few months of execution, it’s quite clear that the company will need a grinding customer discovery process, with long and deep engagements with early design partners.
In my eyes, this is all great. As an investor looking for non-incremental startups, I precisely expect this and in fact, get excited by it. Sharing an excerpt from my post ‘Building…one at a time‘ on my learnings as a founder on the 0-to-1 stage:
When the absolute user numbers weren’t met, my morale as a founder would get hit with each iteration. In hindsight, hitting numbers shouldn’t have been the goal at all. The ideal 0-to-1 mindset is like that of a scientist, with curiosity being the core driving emotion, backed by an iterative product development approach. The target outcome of this approach should be to gather insights that help refine the hypothesis.
Similar to how scientists drive their research process one experiment at a time, I have realized that building any new product or service from grounds-up requires moving one “unit” at a time. It’s up to you to decide what that unit should be – acquisition, activation, frequency of use, revenue or even just getting qualitative feedback!
The challenge is when a company has raised significant capital relative to its stage. While this de-risks the company from a runway perspective and opens up many options in each execution track, having money sitting in the bank often puts undue pressure on the founders to use that capital.
In my experience, this pressure starts manifesting in many ways at an operating level:
1/ While the seed stage needs founders to be directly talking to customers and building product, capital often creates a tendency to do premature functional hiring and delegating core aspects of PMF progress to new employees.
2/ Even as a seed startup is still figuring out the customer persona and pain points that it needs to solve, excess capital drives founders to invest in GTM even before the company knows what product needs to be taken to market. This could involve unnecessary paid marketing, attending events vs talking to customers, building PR rather than product etc.
3/ Excess capital can often create an environment where the team starts to feel victorious even before any material progress towards PMF. The mindset shifts from ‘doing things that don’t scale‘ to ‘doing fake work’ via mindless reps.
Ultimately, this creates a massive risk of founders not being honest to themselves about execution and learnings, while also setting wrong expectations with their Board/ investors. Most investors aren’t builders anyway, and given their primary concern is the next round markup, often push startups to increase burn and “show numbers” prematurely. Unless the founder can push back with a high-conviction execution philosophy that they believe the company needs at this stage (I espouse founder-led, lean, frugal tiger teams doing things that don’t scale), this Board pressure will create a negative flywheel.
Only founders who are honest with themselves about where the startup really stands can then push back on investors with the best model they believe is needed to make progress at this specific stage.
Drip-feeding as a brain hack
So, how can a founder create this disciplined, frugal, ‘doing things that don’t scale’ mindset even with millions sitting in the bank? During this working session I mentioned at the beginning of the post, I blurted out a brain hack:
“What if we just virtually ring-fence the funds, maybe even create a CD or something, and give ourselves say only $500k (the standard YC deal amount) or something similar for the next 6-12 months to execute? In a way, we use this artificial scarcity to discipline ourselves, and drip-feed execution till a certain set of milestones are reached.”
It’s almost treating raised capital like a 401k account – there to save your a** in the long run but not accessible day-to-day. It’s what HNIs do with trust funds – even with a large pool of capital, the kids still get drip-fed for their own good.
A similar spirit is reflected in grandma’s age-old wisdom that advises folks to minimize easily accessible funds in bank accounts and instead, lock them up in CDs. Adding that extra layer of friction itself acts as a nudge to avoid impulsive spending.
OG public market investors like Nick Sleep and Guy Spier have openly shared how they use behavioral nudges like keeping the Bloomberg terminal in an uncomfortable location or only placing Buy/Sell orders when the market is closed, to avoid unnecessary noise and the tendency to frequently trade at the expense of compounding returns.
This idea of drip-feeding immediately resonated with the founder and in fact, she encouraged me to blog about it. Hence this post! Am eager to see how the results of this execution nudge pan out. Will share the learnings on that soon.
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
Leveraging rare moments in time like Bitcoin requires having a prior mental model for how to behave when coming across an asymmetric option.
It’s crazy that in hindsight, Bitcoin was one of those super-rare, asymmetric-upside options that should have been a no-brainer to buy, esp. at the <$1k levels.
That’s why most tech HNIs, at least anecdotally from my network, ended up being early adopter buyers of Crypto. They had the right channels of early intel that informed them of why it was worth having at least some exposure to it.
Funnily enough, having these proprietary sources of info didn’t matter much given the long price runway that especially Bitcoin has shown. The dealer showed all the cards, multiple times over several years. Hell, you could have just read this 2011 post from Fred Wilson, trusted the OG who has gotten it right multiple times in tech, and bought maybe just a few thousand dollars of BTC. Do you know what price you would have entered at when this post was written? $2.75!!
And yet, few people bought any Bitcoin over these years, fewer ended up HODLing and even fewer ended up doubling down. Why do you think that is? I believe it’s because people don’t have a prior mental model for how to behave when coming across an asymmetric option.
Conviction comes from having a mesh of these mental models already in place, especially those that are drawn from experiential learning and therefore, become much more deep-rooted than those imbibed from mere academic study.
I don’t blame folks for not knowing what to do with Bitcoin. It is one of those once-in-a-generation movements and therefore, by definition, entire cohorts would have lived their lives without seeing anything similar to it before.
This is where experience becomes important. Ironically, even though Bitcoin is referred to as a Gen Z asset class, the people who have made real money off it are the grey-haired (or no-haired!) Michael Saylor, Mike Novogratz, and Bill Miller. Interestingly, both Saylor and Novogratz are 59 years old while Miller is almost 74!
This is because, over 4 decades of working and investing, these gentlemen have seen enough human behavior in the real world, as well as put skin in the game by taking multiple explosive-payoff bets one after the other, to recognize how the system works and how to leverage these waves to their benefit.
50% of my networth is in Bitcoin.
Bill Miller (born 1950)
In this fascinating interview, Bill Miller talks about how Roosevelt confiscated everyone’s physical gold in the US in 1933 and that’s the mental model that Bill uses to view Bitcoin as digital gold that can’t be confiscated due to the Internet (see my post ‘Bitcoin ETFs and The Challenges of Digital Gold‘). He then nullifies the argument used by the likes of Warren Buffett that Bitcoin has no intrinsic value, by saying that what intrinsic value does a rare baseball card or a Picasso painting have? They still sell for millions as their supply is scarce and people ascribe value to them.
I am actually a Bitcoin observer. I am observing its trajectory as a new technology and comparing it to things like the printing press, or the steam engine, or the railroads, or the automobile, or electricity. And it seems to be following a well-understood path to adoption of any new technology.
The benefit of age and living through multiple cycles is that one can fit the arc of a new tech wave within a very long historical view of how things have evolved in the past and leading up to this point, as Bill does above. It’s how Millennials like myself will likely use the lived experiences of GFC’08, ZIRP, the pandemic, and the peak of 2021 as mental models for decisions going forward.
Therefore, let’s bookmark this post as a note to self: the next time we encounter an asymmetric option, strongly consider taking a swing at it (after due consideration, of course!).
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
The imagination of the best founders will kill any TAM analysis. Rather than over-fixating on market size, I suggest a couple of other elements that are more useful for evaluation at the seed stage.
As a freshly minted VC in 2011, one of the first things I was taught to look for while evaluating new deals was a large total addressable market (TAM). It was pretty much a necessary, though not sufficient, condition for generating outlier venture returns.
After many years of venture investing now, I would agree that startups need to go after large markets to be VC-fundable. However, I have also seen a common VC fallacy, especially at the seed stage, where a startup gets prematurely filtered out simply because the immediately visible market doesn’t seem large enough.
In my experience, the best founders either create new markets or expand to adjacent markets over time. So the TAM keeps growing. I remember when Peyush Bansal of Lenskart (the Warby Parker of India) was pitching for Series A, investors thought that the Indian eyewear market wasn’t large enough. In hindsight, everyone underestimated 3 things: (1) the overall market would grow at a much faster rate than expected, (2) in addition to online, Lenskart would also go offline and (3) it would expand the market by launching private labels, as well as increase its gross margins by in-house manufacturing. Lenskart’s last valuation: $4.5Bn!
Same with FirstCry, where most investors viewed the TAM as mostly Diapers because that was the majority of its GMV in the early stages. FirstCry is IPO-bound, likely at $3-4Bn valuation!
In the enterprise space, I remember Amagi’s initial flagship offering was the ability to insert vernacular TV ads in national programming. Investors pretty much checked out when they sized up this niche. What they underestimated was Amagi’s ability to go global, expand its product offering to an end-to-end cloud platform, and serve the rising OTT market. Amagi’s last valuation: $1.4Bn!
The entire thread is super-insightful, wherein PG is repeatedly trying to explain to Fred that the eventual market will include hotel accommodations and therefore, will be large enough.
In fact, one of the strategies that is highly recommended for seed-stage startups is to identify a wedge in the market, usually a very sharp pain point or job-to-be-done, and focus on solving that in a differentiated way. By definition, these early wedges often create an illusion that the addressable market is limited to just that.
This is where the role of imagination comes in. Seed investors should spend adequate time with the founders to understand the eventual end-state of the world they are imagining. If this end-state is large enough and ambitious enough, usually that’s a leading signal that the company will continue to expand its TAM.
Turning the tables around, I also believe that founders should spend time crafting a strong narrative around this end-state and paint a picture that investors find easier to buy into. If this can include even a broad outline of what the path from the present wedge to the eventual end-state potentially looks like, even better.
I hesitate to call this picture a “vision” as this word is just thrown around a lot and frankly, is now considered faff. Instead, trying to visualize what the world will look like with your product in it, is much more tangible and real.
So, if not TAM, what aspects should seed investors evaluate instead? I recommend the following two:
1/ Founder-market fit – this is critical for the startup to go from the wedge to the end-state, and should be in place even at the earliest stages of company building.
2/ Competitive differentiation/ right to win – as I mentioned in my post ‘How To Differentiate As An AI Applications Startup?‘, for a startup to be viable, it’s not enough to just build cool tech. It has to be able to create significant competitive differentiation, especially against incumbent solutions. I call it “non-incrementality”.
A strong hypothesis around competitive differentiation, and eventual right to win, should exist in the founders’ strategy from Day 0. The edge will get gradually built out over time, but both the intent to create it as well as the building blocks for it need to exist from the earliest stages.
Looking back on the many startups I have seen or invested in across geographies, one thing I have learned is that if a startup remains sub-scale, in most cases it tends to be due to founder motivation, quality of execution, and team/culture issues, rather than the available market being small.
Subscribe
to my weekly newsletter where in addition to my long-form posts, I will also share a weekly recap of all my social posts & writings, what I loved to read & watch that week + other useful insights & analysis exclusively for my subscribers.
