The Berkshire of Software
What this covers
Savneet Singh, who is building a software-focused permanent holding company called Tera, explains why he buys ordinary software businesses rather than backing startups. The argument starts with arithmetic. SAP's renewal rate near 98% implies customer tenure of roughly fifty years, and a dentist who runs billing, scheduling and payroll in one system will not switch for a 25% discount. Software at one to three times recurring revenue looks expensive, he says, but equals about six times steady-state cash flow if margins reach 30 to 50%. He defends a holding company over a fund, plans heavy leverage and seller financing, and credits Mark Leonard-style Process-Based Acquisition, which needs no view of the future but works less well as capital grows.
The rest ranges widely. On venture, he found no repeatable pattern in his own angel wins and expects persona-driven models to be challenged by platforms like AngelList. On crypto, he is blunt: ten years in, no distributed-ledger application is mission-critical, and the returns came from speculation, not from disintermediation. He also expects governments with weak currencies to block parallel ones. His FinTech view is that the first wave only moved existing loans and insurance online, while the next will price risks and borrowers that were previously unpriceable. His forecast that better data could make life insurance affordable for people managing chronic conditions is the most contested.
Operating detail fills out the picture: paying outbound reps for each funnel step, tracking customer attributes to see what a top region's results really reflect, and the Culture Survives Founder Test, which Danaher passes. He closes on investing in the ecosystems around platforms, such as Slack add-ons and Shopify storefronts, rather than the platforms themselves.
Savneet argues that durable, low-churn software businesses bought at modest multiples through a permanent holding-company structure, run with culture and process discipline, compound better than venture bets or conventional public-market approaches, and that the next real FinTech value lies in new products rather than customer acquisition.
- Mission-critical software with high retention and switching costs produces visible recurring cash flow that compounds when cash is recycled rather than paid out
- Venture returns are not systematic and rely on brand or persona, while process-driven buying of unglamorous businesses is repeatable
- First-wave FinTech only moved existing products online; second-wave FinTech creates products for risks and borrowers that could not previously be priced
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
SAP's renewal rate is about 98%, or 99% excluding bankruptcies and mergers, which implies an average customer tenure near 50 years, creating an effectively permanent moat that lets the company raise prices every year and reinvest, yet the market rarely prices such software below about six times recurring revenue.
“You'd be shocked to find out its like 98%. If you remove bankruptcies and mergers, it's 99%. What that's implicitly saying is the average customer stays for 50 years.”
A dentist who has built billing, scheduling, payroll and practice management into one software system will almost never switch for a 25% discount, because the operational risk of migration outweighs the modest price saving, so switching costs protect pricing power.
“If someone comes to them and says, 'Hey. I'm going to give you a 25% discount if you take my product.' There's almost a 0% chance that dentist is going to go for that.”
Life insurance that rejects anyone with cancer or charges diabetics very high rates reflects a lack of data; better data could let insurers price chronic conditions fairly, conditioned on behavior, so that people who manage their conditions can get coverage at reasonable prices.
“Now you can say, 'Hey, we can price you at a relatively fair price provided you act in this type of behavior pattern, which allows us to price it that way.'”
Venture investing is not systematic: Savneet found no repeatable pattern among his angel successes, which he argues is why traditional allocators struggle with the asset class, and brand or persona-based VC models are being challenged as platforms like AngelList and blockchain let anyone become a VC.
“the single lesson I've learned from venture, which makes it so hard as an asset class is it's not systematic.”
Software trading at one to three times recurring revenue looks expensive but is cheap if steady-state cash-flow margins reach 30 to 50%, since buying at three times revenue equals roughly six times steady-state cash flow, especially where prices can be raised annually.
“If you believe the steady state end margins are 50%, you're really buying it at six times steady state cashflow.”
Structuring Tera as a permanent holding company rather than a fund lets cash from stable businesses be reinvested in better opportunities without dividends and taxes leaking out, and avoids the fee structure that rewards selling businesses, so it compounds capital over 20 to 50 years.
“It would be unfortunate to buy one company. Take that money. Dividend it out and then pay taxes on it and not be able to reinvest it.”
Buffett's question of finding a non-brand business that keeps margins for 30 years is harder today because brands mean less as commerce is disrupted, so consumers buy on Instagram and reviews rather than inherited family brands.
“a brand means a lot less today than it meant yesterday. The brand of my father is probably not the brand of me.”
Ten years into cryptocurrency development, no application using a distributed ledger is used regularly or is mission-critical, which Savneet finds alarming because the technology may have no real use case.
“There's still not an application that depends on a distributed ledger that anyone uses regularly or has any mission-critical parts of it.”
Crypto's rise was driven by speculation tapped from low rates, inequality and anti-establishment sentiment, so proponents who were right on price were wrong on the disintermediation thesis, meaning the returns came from speculation rather than intellectual accuracy.
“That's a bear case where you're winning not because you were intellectually accurate. You're winning because of speculation.”
Both licensed maintenance and subscription software models eventually produce recurring revenue with 30 to 50 percent cash flow margins at steady state, which lets an investor estimate a mature company's economics early.
“In both those models, the end state recurring revenue generally has anywhere from 30 to 50% cashflow margins.”
Crypto's original promise of replacing weak currencies in countries like Kenya or Zimbabwe is shortsighted because governments with weak currencies have strong incentives to block parallel currencies that challenge their control, just as gold is not used as currency today.
“If you're a country that has a weak currency, you are not incentivized as someone running that country to have another parallel currency.”
Software's visible recurring revenue makes it a good candidate for debt, and Tera intends to use substantial leverage and seller financing because lowering the cost of capital is valuable when cash flows are predictable, unlike most large holding companies that use little debt.
“I actually think software is a beautiful business to use debt because you have great visibility into that revenue.”
Bitcoin's bull case is that a community will keep ascribing it store-of-value status, similar to gold, although Bitcoin's history is far shorter than gold's 5,000 years of cultural valuation, so its value rests on belief rather than utility.
“Just like gold, gold has very little utility but it has value because you think it has value.”
Lenders who find segments with no credit history can lock in attractive terms for themselves, because traditional capital will not enter markets lacking five years of data and is not interested in small checks such as $5 million, and the underwriting skill for these niches is not transferable elsewhere.
“If you find a new segment that has not had credit before, you as the debt capital can lock in amazing terms for yourself. Traditional capital won't come in because there's not five years of history.”
When Savneet considered entrepreneurship, his brother tested whether he truly wanted his fund manager's life by asking whether he would take the same job for a $35,000 salary, revealing that the attraction was the lifestyle and compensation rather than the work itself.
“Now, would you take that job if you're paid $35,000 a year?”
Paying outbound sales reps for each funnel step, such as calls, demos, in-person meetings and contracts, with rising amounts for later steps, prevents them from chasing only low-hanging fruit and rewards building a pipeline rather than waiting a year for a closed deal.
“Instead of just commissioning that person on the final step, which is signing the contract, you said, 'Okay. I'm going to compensate you on the quantity of calls then the number of demos.'”
Top-performing sales regions like New York may reflect the customer attributes they serve rather than the salesperson's skill, so tracking meeting-level customer attributes can reveal that a rep's real edge is selling to a specific buyer profile, which can then be exploited.
“Is that really the best sales person? Is that because they have that territory?”
The first wave of FinTech mostly acquired customers online for the same loans and insurance products banks already sold, so it produced VC success but few sustainable economic models; the second wave will be defined by genuinely new products such as lending to previously unbanked markets and insuring previously uninsurable risks.
“You've had the first wave of FinTech be primarily customer acquisition, acquisition businesses.”
Constellation's process-based buying means it does not need claimed software expertise or a view of the future; it simply checks retention and revenue thresholds, so it can hire many people to find deals, but this model works less well as capital grows large.
“By making it completely process-based, you can hire hundreds of people to keep finding deals for you.”
The simplest test of a great company culture is whether the business sustains after its creating leader leaves, and Danaher, on its third CEO with little change in approach, passes that test.
“The simplest way I define a company with a great culture is, will that company sustain after the leader that created that business leaves?”
When a founder says an industry like life insurance has not changed in a hundred years, Savneet treats it as evidence the business is excellent, since a model surviving a century implies durable economics and a reason incumbents have not been displaced.
“If something's lasted for a hundred years, that's the definition of an amazing business.”
A thought experiment asks whether an investor would decline a project built by the world's best engineers, and Savneet argues most people would still invest, showing that the price of exponential technology is hard to assess and that FOMO and greed override skepticism.
“It's hard to price the existential. It's hard to price the exponential. It's hard to figure out these things that you see such talent and it's that FOMO.”
Investors fixate on platform stocks like Airbnb, but the more mispriced opportunities sit in the ecosystems around them, such as Slack add-on tools, Shopify storefronts, and pet-focused sites that gain affiliate leverage from traffic.
“everyone always runs to the platform and doesn't look to the ecosystem... What are the derivatives of that that people aren't paying attention to?”
Rather than relying on a manager's individual skill, one should find the right pool of businesses to invest in, so that the investor can pick the horse rather than the jockey, which led Singh to the idea of a software holding company.
“what if you could just find the right pool to fish in so that you could pick the horse and not the jockey?”
Sikhism, which emerged in India over 500 years ago, rejects caste and gender hierarchy; most Sikh men take the surname Singh (lion or king) and women Kaur (princess or queen), meant to signal equality by removing surname-based caste markers.
“Most Sikhs have the last name Singh. Most women had the last name Kaur. Singh means king or a lion. Kaur means princess or queen.”
Because Tera is not competing as the highest bidder but as a value-added long-term operator, it is the only party at the table in its deals, which Singh says allows it to pay less than competitive auction buyers.
“Every single deal we do, we're the only party at the table. There's no one competing on the other end because it's very much a partnership.”
Tera's biggest sourcing differentiator is arriving with a detailed operational plan for sales, marketing, and culture rather than an offer price, which shifts the conversation from price to value creation and often wins the deal.
“the single biggest differentiator we have when we go sit down with a company is we come in with an incredibly detailed operational plan.”
Robo-advisors at scale become registered firms whose marketing must pass counsel review, so they converge on the legacy incumbent mindset, and their founders end up as missionaries who are unlikely to be the companies that disrupt product creation.
“When you have billions of dollars and you're registered and every marketing communication has to be approved by council and then by an outside counsel, you start looking more and more like the traditional incumbents.”
Many software founders do not want to sell to private equity and instead want a long-term partner that will keep their company's culture and employees intact, creating an opportunity for a Berkshire-style buyer that never sells.
“We want to buy your business and we never want to sell it. We want to reinvest in it.”
Tera's cold outreach achieves a 20 to 25% open rate, against roughly 5% for private equity firms, because outreach is data-tested and offers a deep, specific analysis of each target's business rather than a generic call request.
“Our open rate on a cold email is 20 to 25%. When we interview our friends at work in private equity, it's closer to 5%.”
The two key tests for a Tera target are whether the product is mission-critical and whether retention is very high, because those two features imply most of the business's quality and predictable recurring revenue.
“The key thing we look for in a business: 'is the product solution or tool mission-critical?'. Is it something that you can't live without? The second is, is the retention rate very high? If you can figure that part out, you've pretty much figured out the business.”
Process without culture becomes process for the wrong reasons, so durable investing or operating advantage requires combining a strong underlying culture with rigorous process.
“It's the combination of culture and process. We are obsessed with process. If you don't have the culture underneath it, it becomes process for the wrong reasons.”
At 3G, founders flying economy means employees feel pride in frugality because the culture rewards cost-saving, and this norm continues after founders depart, which is why its culture lasts for generations.
“When you have an organization like 3G where the founders fly economy, when you're flying and you're uncomfortable flying from New York to San Francisco on an economy flight, you feel pretty proud.”
Tera deliberately targets software markets around $400 million total addressable market, where large vendors like SAP, Oracle or Workday will not bother to compete, so the company avoids head-to-head competition with bigger players.
“We're not looking for the $10 billion market that SAP, Oracle or Workday is trying to go after. We're going for the $400 million TAM where it's not really worth the time the big guy to come in there.”
Gold is the oldest investment currency in the world yet is not traded electronically anywhere, a gap in the market that motivated GBI's platform for buying, trading, and storing precious metals with less counterparty risk.
“We were surprised that gold is the oldest asset investment currency in the world yet is not electronically traded anywhere.”
Small, cash-flowing online businesses can be bought for about three times earnings and are relatively sticky if well managed, and a marina directory bought as a test grew revenue about 50% by raising stale ad rates and making the site mobile-friendly.
“you can find lots and lots of small cash flowing websites online that are interestingly if you handle the right way, relatively sticky businesses that you can buy for three times earnings.”
Tera positions itself as the 3G of sales and marketing, seeking to grow acquired companies through centralized sales and marketing playbooks rather than cost-cutting, using scale across the portfolio to create economies of scale.
“We're the 3G of sales and marketing. We're not cost cutting but affiliate marketing.”
A portfolio company's New York region appeared top-performing, but collecting attribute data on each prospect revealed that a rep's close rate depended on prospect characteristics such as education and gender, so reps could be reassigned to the segments they closed best.
“Did you know that your close rate on the Ivy League-educated guy that lives in New York is 90% but your close rate on women is 2%?”
Companies that have taken venture money tend to develop poor financial hygiene, assuming endless capital and throwing money at growth problems, which can later create talent flight when options lose value, making them less attractive as acquisition targets.
“What we found though is the problem with taking venture money, you develop really bad hygiene. You assume that there's endless money.”
Long-term vision at compounders is operational rather than rhetorical: compensation tied to five-year value creation, and recruiting processes that wait for the right person, such as a company that spent six months on one hire, so the culture embeds patience.
“The company literally waited six months for that hire because they needed to find the right person.”
Software revenue quality is like an apartment building: a building full of doctors and lawyers gives far more predictable rent than one full of tourists, so investors should pay more for sticky, high-retention, deeply integrated recurring revenue.
“You're always going to pay more for Building A because there's great visibility in that revenue.”
Constellation Software's advantage is extreme acquisition discipline: a process-driven buyer that will walk away over a few cents, buying maintenance-revenue software companies, pushing margins to around 40%, and reinvesting cash flow into the next acquisition.
“they are the most disciplined buyer I have ever, ever seen. If it's one penny above that ... If you're 25 cents above that price, you won't go. These guys are by the penny.”
Across a dozen-plus compounding companies studied, the uniform traits were long-term vision, tax efficiency, a culture that outlasts founders, and adding operational value, even though industries and operators varied widely.
“The three things we observed were a long-term vision, very strongly focused on being tax-efficient and culture.”
Even when customer industries are cyclical, such as an energy company during a 40% oil price drop, companies rarely cut software that runs core functions like accounts payable or AML and KYC compliance, so cyclicality reduces growth rather than the recurring base.
“if you're an energy company and oil prices are down 40%, are you going to take away your accounts payable software? Probably not”
Most target software companies, averaging about 100 employees with 80 to 85 engineers and only two salespeople, are culturally run by engineers who reward adding features, so Tera must shift them to ROI and sales thinking through collaboration rather than coercion.
“they are culturally run by engineers. The reward for that culture, that organization was let's create more widgets on a product.”
Software serving industries with very low failure rates, such as dental practices with about a 1% failure rate, is durable because its customer base persists, whereas marketing technology suffers high churn because the underlying technology changes too fast.
“Dentists have a 1% failure rate. The lowest failure rate of any small business in America.”
Airbnb is already the largest hotel company in the world yet most people over 35 have never used it, and as the next generation that grew up in the shared economy gains wealth and travels more, the company is only at the tip of the iceberg of demand.
“Airbnb is already the biggest hotel company in the world.”
Tera's first hire worked for free because they wanted to learn, and the firm delays portfolio scaling until it builds the operating muscle to know what works, since building culture early is harder when buying many companies at once.
“If we were trying to buy 10 companies and create a portfolio, it would be harder and harder to create that type of culture day one.”
The Barcelona business became a distressed buyer of condos that moves faster than local brokers, buys at replacement cost, and locks in fixed-rate 20-year mortgages below 2.5%, so a broker's cash-flow listing can be flipped the next day to roughly two and a half times its value as a short-term rental.
“We lock in interest rates at less than 2.5% for 20-year fixed. We are buying at replacement costs.”
Properties with a Barcelona tourist license sold for only about 10% more than unlicensed properties while yielding two to three times as much as rentals, so the legal short-term rental permit was underpriced by the market and created the opportunity Savneet's group exploited.
“What we discovered was a property with a tourist license only sold for 10% more than a property without one. Yet the yield was two to 3X the price.”
The next wave of venture capital will be more domain-specific, with investors who have a particular operating skill set, such as customer acquisition, backing companies that can add concrete value, replacing investors whose edge is a famous brand or personal status.
“I bet you what happens to the next wave of venture capital will be a lot more domain-specific venture capital.”
Berkshire's playbook of buying high fixed-cost, reinvestment-heavy businesses like railroads and utilities will be harder to replicate because the economy is shifting toward asset-light businesses, so Berkshire tomorrow will differ from yesteryear.
“Berkshire tomorrow will be very different than the one of yesteryear.”
Centuries of prejudice and persecution against Sikhs created a cultural resilience and sense of duty to people who fought before, which shapes how members of the community behave today.
“it's gone through lots and lots of prejudice, lots and lots of genocide over time. What that did to the community... it created this idea of resilience.”
Savneet's two core life lessons are that circumstance and luck shape outcomes, most importantly his birth in America, and that people must take action and enter the arena, because those who never try fail to build anything valuable even if they fail outright.
“I often say like, 'I've learned two real lessons in life.' One of those lessons is the power of circumstance.”
The largest crypto business, Coinbase, is effectively a modern bank that charges more than a traditional bank, which is ironic given that Satoshi Nakamoto intended cryptocurrency to escape bankers and Wall Street intermediaries.
“it's a little funny to me. A little bit ironic that the biggest business in crypto is Coinbase. It's the modern version of a bank.”
Singh's father deliberately involved him and his brother in major decisions from a young age, such as choosing flight times and memorizing highway exits, which built confidence in uncomfortable situations.
“One of the great things my dad did at a really young age was he involved us in big decisions.”
Tera's sourcing edge comes from presenting as operators who built businesses, not as capital, which resonates with founders more than a private equity pitch because it signals a commitment to their legacy and employees.
“We're coming there as like, 'Listen. We're some operators. We started these businesses. We'd love to help you build your business,' and they respond.”
Savneet no longer makes angel investments because he could not find a repeatable pattern, except possibly for deals sourced from his network of people he considers far smarter than himself.
“It's purely because I just couldn't find the pattern.”
Singh's father offered him and his brother the choice to drop out of high school for a tennis academy, which instilled that they could do anything they wanted and taught hard work and commitment.
“The thing that impacted me the most from that was the sheer fact of my dad giving me the option to do that.”
Founders of small software companies, often aged around 65, respond better to operators who arrive in ordinary clothes with backpacks than to young private equity deal-makers in suits and flying first class, because relatability and sincerity build trust that wins deals.
“I think that relatability really helps us. We walk in the door and we come in with backpacks and we say, 'Here's exactly how we think your business. It's not just a pile of money.'”
After the 2008 crisis, investors would move toward esoteric hard assets and toward assets with less counterparty risk, so owning assets directly rather than through intermediary shares was an attractive model.
“The idea was after the crisis, there'd be a movement into one esoteric assets but two, assets with less counterparty risk.”
Two years in sell-side investment banking provides exposure to the full Wall Street value chain, from IPO process to private equity and capital markets, and surrounds young people with higher-horsepower thinkers than college does.
“The second thing that was super useful was you get exposure to all parts of Wall Street.”
Losing his mother, experiencing a serious illness, and watching Lehman and AIG collapse within his first year at a hedge fund made Singh want to build something of his own rather than follow a conventional path.
“All those experiences at once make you think about what you really want to do with your life.”
A family friend's invitation to spend a weekend together, months after his mother died, was the first time the family had fun again and marked the start of their recovery.
“It was the first weekend that we, as a family had fun since my mom passed away.”
Sikh values of service and giving back are reflected in unusually high Sikh political representation in minority countries, with Savneet citing Sikhs holding 20% of the Canadian cabinet.
“20% of the Canadian cabinet is Sikhs.”
Six of the top ten performing private equity funds have a very high focus on software, which suggests the software opportunity has already been recognized by sophisticated capital.
“six of the top 10 performing private equity funds have a very high focus on software.”
Worldwide there are roughly 100,000 software companies in Tera's target size range, and the pool is growing because venture-backed firms that top out at 10 to 20 million dollars in revenue become acquirable boring businesses.
“We think worldwide, there's probably 100,000 software companies that are in the target range and size that we're looking for.”
Crypto's most durable contribution may be the sense that financial systems can be better, since wire delays and high international transfer costs motivate innovation even if the blockchain itself is not the answer.
“Why do I have to wait two days for a wire? Why does it cost so much to send money to a different country?”
Tera has given little thought to diversification across industries, instead focusing on the most defensible software businesses first while building operational musculature, and they will consider sector diversification later.
“I'd be lying if I said we've given it more than a 2% thought.”
Savneet's first angel investment was Uber in 2010 or 2011, which he describes as complete luck because he knew Travis Kalanick through a college roommate.
“I always joke the first deal I ever got into was Uber. It was complete luck.”
After visiting the Berkshire Hathaway meeting with friends, Savneet concluded that Buffett's contentment exceeded that of fund managers and executives, prompting the goal of recreating that life by building a holding company of powerful assets.
“Warren Buffett is really, really content. He's way happier than our investors or the people we look up to or our bosses.”
The hostile reaction to Singh's crypto bear article, including hate mail and social media attacks, suggests the community is emotional and therefore less logical than needed for long-term success.
“The quantity of hate mail and Twitter hate and medium hate, it was mind-blowing.”
Savneet has given portfolio diversification little thought, prioritizing finding highly defensible businesses with cash flow stacks and fixable inefficiencies over sector allocation across industrials, energy or telecom.
“I'd be lying if I said we've given it more than a 2% thought.”
As teenagers, Singh and his brother started selling koi fish online and then reselling baseball cards on eBay, which taught them to capitalize on the early internet boom through small entrepreneurial ventures.
“We took a bunch of pictures of the fish and put it up on Craigslist and started selling fish on the internet.”