🔥 Money & Machines |
Some weeks you report, and some weeks you just sit with it. Arti's piece in today’s Money & Machines edition is a little different. But, on the emerging tech side, we look at the most-talked about Emergent's case and how ARR is being rethought in AI. |
This week's piece on 'Money' — a bit unusual from my end |
I was in Mumbai recently. Met a very senior banker…we were talking about everything you and I have been tracking over past few months – the HDFC chairman exit, the arrest of Fino MD Rishi Gupta, the online gaming transactions mess, and the Haryana cases of IDFC Bank and Kotak Mahindra Bank. One case after another. |
And somewhere in the middle of all this, he said: "The standards of this society have fallen so much that till two-three years ago, when an employee was suspected of being involved in something shady, we used to still say 'innocent until proven guilty'. There was still some benefit of doubt where we used to think maybe he was caught in the system…maybe there was more to the story. But, that has changed now. It's the opposite…it's 'guilty until proven innocent'.” |
This line of him made me pause for a bit. |
Because we were talking about an industry that runs entirely on trust. Over a billion people have their life's savings in these banks – with a faith that the system and people are trust-worthy. |
Some bank officials get caught, even get arrested. But then somehow find their way out. "And what used to be absolutely unthinkable in banking until few years ago is these very people get hired again by other banks," he said. |
One may think that there are just few cases here and there and we are just being cynical for no reason, but the point is there have been too many instances where individuals inside the system have gamed it, made money, and somehow found their way back into it. |
This is the state of affairs. I genuinely don't know where we are headed. |
While writing this piece, I realized I couldn't stop at banks – at least there, the RBI is watching. In the startup ecosystem? The same pattern plays out, right in front of everyone's eyes, again and again. |
For instance, take Rahul Yadav, someone who has now been associated with multiple ventures, each surrounded by serious fraud allegations. |
Housing.com ended badly where he was pushed out after clashing with investors. Then came 4B Networks, where Info Edge – despite Yadav's highly publicized past – picked up 59% for Rs 288 crore. Within less than 18 months, the real-estate startup was in a freefall. Severe cash burn, employees unpaid for months, a forensic audit, and eventually an FIR over alleged fund misuse and suspicious related-party transactions. Info Edge wrote off the entire investment. |
The third entity, Kult – a beauty app with his wife at the helm, though by all accounts Yadav was present at every townhall and involved in every key decision. The company managed to raise $20 million in April last year – this is after Housing, 4B Networks, the FIR. And in less than a year, same story of unpaid salaries, vendor dues, fund misuse allegations – leading to the collapse of the this venture too. |
As if two episodes were not enough of warning. The money still came. I'm sure there are many more like Rahul Yadav in the ecosystem…quite easy guess. |
Sometimes, it is no longer about one founder making a comeback again and again, but it starts reflecting the ecosystem enabling it. |
I tweeted about this and few readers highlighted side arrangements. And, that brings me to my next point. We've all heard stories about such side dealings, incentives that don't always align with LP interests, relationships that go beyond boardrooms. You can't always prove them. But you hear enough versions of them to know they are not isolated whispers. |
I can say it – it's not just a hearsay. |
I remember working on a story a couple of years ago about a high-profile startup (not a fintech :)). The co-founder and CEO and a senior executive from the lead investor had an interesting arrangement. The VC on the company's board lived in an apartment owned by the founder. Another apartment in the same building – owned by the founder's sister – was home to the VC's sister. |
I sent both of them a media query and asked whether this constituted a conflict of interest; whether this information was disclosed to the board; and if there was any board approval. There were no clear answers. The founder insisted the flats were rented at market price and that an independent audit had benchmarked the rent. |
The founder requested to not publish the story. The VC also said the same thing in an indirect way too. |
Despite having property proofs, rent documents, ground reporting, source reporting – the story did not see the light of the day. To this day, I don't know why. |
But what I do know is every time the fund made a new investment into the company, the founder made good money in secondary transactions. Three times, if I remember correctly. Coincidence? Maybe. |
I believe governance isn't just about what is illegal…it is also about what 'feels' right when no one is watching. |
A fintech founder I once reported on (different story, different company) gave me a long lecture when I broke news of the acquisition of his startup. He was furious, he literally told me, "It's a private firm, private funding, unlike public listed companies…and it's nobody's business to report on us." Actually his real irritation was that his own employees found out through my news, not from him. |
Today, no founder acknowledge scrutiny…they may not say it in as many words but they get equally furious. |
Having said that, the logic of private money means private accountability doesn't hold anymore. |
Whether it is public depositors' money in a bank, or LP money pooled into a VC fund, or retail investors in listed companies – none of it exists in a vacuum, and somewhere it all connects. And once things start going bad, we all know it doesn't stay in just one corner. |
I've reached a point where I believe that no matter how much you write, or what you write, very little actually changes. There are stories I have heard that make me grind my teeth, yet you just can't write those stories without proofs. |
What is clear is that the scamming of the system – public or private, by individuals or by institutions – is far more widespread than what makes headlines. The difference between what gets reported and what actually happens is a very wide gap. |
And the ones gaming the system – most of them are just fine. |
I guess maybe we have all adjusted. Founders know how far they can stretch, investors know what to overlook, and institutions know what can be managed. |
And the rest of us – we make notes, discuss, and then move on. |
-by Arti Singh |
|
When ARR stops meaning recurring in an AI world |
Vibe coding startup Emergent, which raised $100 million within seven months of its launch from the likes of SoftBank, Khosla Ventures and Lightspeed, has found itself at the centre of a controversy over its high ARR claims. |
It began after an X user tweeted that media publications were working on investigative reports to debunk those numbers. |
Responding to the post, Hemant Mohapatra, partner at Lightspeed, weighed in with a response that sparked a broader debate on how ARR should be defined in the age of AI. |
“Emergent made more in accrued revenue in a WEEK than most startups make in a YEAR. This is actual, banked, cold, hard revenue,” he said, adding, “Token consumption is all that matters in AI.” |
It thus became clear that Emergent was using a different definition of ARR altogether. While SaaS companies typically report annual recurring revenue tied to contracted income, Emergent’s numbers appeared to reflect an annualised run rate based on recent consumption. |
To be sure, Emergent, founded by Indian-origin twin brothers Mukund Jha and Madhav Jha, is not the only vibe coding startup that has announced high ARR. In November last year, Stockholm-based vibe coding startup Lovable announced it had surpassed $200 million in ARR, just four months after a $100 million ARR milestone. |
It is not very clear if Lovable also follows the same method as Emergent to calculate ARR, but what has become evident is that there is no single way to define ARR in the AI ecosystem today. Different companies appear to be using different methodologies, often shaped by whether their revenue is driven by subscriptions, usage, or a mix of both. |
But as Mohapatra underlined, is token consumption all that matters in AI? |
Supporting Mohapatra, Pratyush Rai, founder of AI companies Thine and Merlin, argued that token usage should be seen as a proxy for distribution, with companies prioritising scale today even at the cost of margins, in anticipation that falling AI costs will make that usage highly profitable over time. |
“AI token costs are dropping around 100x over the next 2 years. If you're scaling token usage today, even at a loss, you're doing what Uber and WhatsApp did,” Rai wrote. |
Rahul Mathur, investor at DeVC, also supported Mohapatra, noting that so-called ARR in these companies is effectively an annualised run rate, closer to marketplace GMV than traditional SaaS revenue. In that sense, he said, the metric is less about contracted income and more about capturing usage at scale. |
"Therefore, is it alright to 12x your last month volume and report it as ARR? Absolutely, as long as you keep the ARR definition and methodology consistent across each reporting period,” he added. |
Mathur’s comparison of ARR to GMV is reminiscent of the debate during the early years of e-commerce, when companies leaned heavily on gross merchandise value as a measure of growth. While GMV captured scale, it was widely criticised by investors such as Bill Gurley as a vanity metric, as it reflected transaction volume rather than the revenue companies actually retained. |
Much like GMV drew criticism, Mohapatra’s argument that token consumption is all that matters in AI has also faced pushback. |
Ashish Sinha, founder of NextBigWhat, a platform covering India's startup ecosystem, technology, and product management, likened the focus on token consumption to the GMV obsession during the early days of e-commerce, questioning its usefulness as a standalone metric. “This metric is… meh,” he said, adding that usage can often be a function of a platform’s own architecture. |
Venture capital investors The Head and Tale spoke with also did not totally agree with Mohapatra. One of them stated, “Some investors just want to grab attention. Same ones who were reporting SaaS ARRs earlier and now for convenience want to talk about it differently.” |
For now, the debate remains unresolved. Vibe coding and AI businesses are still in their early stages, and the metrics around them are evolving just as quickly. In the end, the real test will not be how ARR is defined, but whether these companies can translate growth into the kind of exits that validate both the founders’ bets and investors’ conviction, as Flipkart once did. |
-by Joseph Rai |
