I’ve automated 80% of my 2024 desk job. AI now saves me 600+ hrs per yr. On 30th June 2025, I shared my 10,000 hours of AI goal. I’m now 750+ hours into this journey. My set-up costs ~$12k per annum and saves ~$60k in notional labor costs i.e. a base case ROI of 5x on labor cost savings and upside benefit from faster & 24x7 work completion. I now have ~3.5 hrs of additional time per day to focus on outbound activities e.g. getting on the plane, meeting founders & experts F2F, going to events etc Here’s a quick summary of my work OS:

In 2019, Saurabh Pandey was helping his mother run her YouTube channel.. This experience showed him how difficult it was for small creators to find an audience, engage with fans and monetize their content. A year later, Saurabh left Flipkart to start EloElo. Raised a small ₹3 cr Seed round. Saurabh walked away from a lucrative leadership role where he was managing a ~₹2,000 crore turnover business for Flipkart. EloElo Group just raised an $80M Series C co-led by Hornbill Capital and NBCUniversal!! When I caught up with Saurabh last month, I realized that EloElo's business has evolved tremendously over the past 6 years: From a single social app in 2020 into a house of consumer apps by 2025: Voxy AI (voice companion), Master (micro learning) and Story TV (vertical microdrama). Story TV was the real breakout. It is one of India’s most downloaded apps & is the no.1 player in the micro-drama space. High volume content production was the genesis of EloElo Group’s 3rd evolution: They developed an internal AI-production engine to handle tasks such as multi-shot angles, character consistency, lip-syncing, voice modulation & stitching and editing. This engine allowed their team to test 1 Lakh creative assets per month This internal technology has been packaged into Dolphin AI, a product that creators, brands and studios can use to generate videos from prompts and scripts. They have an enduring competitive advantage: Story TV gives EloElo an understanding of what audiences want to watch. Dolphin helps the company produce that content more efficiently. But, Saurabh has a larger ambition: India can become an export hub for the global Orange Economy. There is a $15bn+ opportunity to offer a combination of creative + AI services to the world from India! Massive congratulations to Saurabh and the entire EloElo team! The possibility is endless & EloElo is right at the forefront of innovation 🙏

The Pixxel confusion stems from the fact that there are two different Indian Pvt Ltd which have similar names: 1. Spacepixxel Technologies Pvt Ltd (2018) - Incorporated in India on 8 June 2018. - Ownership: Awais Ahmed (49.95%), Kshitij Khandelwal (49.95%) and Pixxelspace India (0.10%). - FY25 income: ~₹73 lakh. - ~31 employees (March 2026 estimate). - This is the entity that signed the Ministry of Defence's 350th iDEX contract in June 2024 and received another iDEX grant in 2025. 2. Pixxelspace India Pvt Ltd (2019) - Incorporated on 27 March 2019. - Ownership: ~99.9999% held by US-incorporated Pixxel Space Technologies Inc. - FY25 revenue: ~₹39.3 crore. - ~186 employees (May 2026 estimate). - This is the entity associated with satellite operations and IN-SPACe authorisations. - It is also the lead entity in the consortium selected to build India's ₹1,200+ crore Earth-observation constellation. ➡️What does this mean? Pixxel has an Indian founder-controlled entity as well as a much larger Indian operating entity owned by its Delaware parent. The 1st co is focused on Indian defence development programmes, while the latter handles more global commercial projects Caveat: I don't have access to intercompany agreements establishing precisely which entity owns which IP, employs which engineers, or holds particular contractual rights (neither would you)

It is unfortunate to see Pixxel’s founder Awais being targeted by online trolls for his religious beliefs. I had to delete several a few such terrible comments on yesterday’s Breakdown video. Even more terrible given that our History tells us contributions from multiple religions have supported our DeepTech ambitions; Nuclear was spearheaded by Dr. Homi Bhabha who was a Zoroastrian; Missiles were sprearheaded by Dr Kalam who was a follower of Islam etc. Let me put Religion aside now that it should be addressed. There is a LOT of misinformation surrounding Pixxel; please find below some of these concerns addressed: 1️⃣Why does Pixxel have a US entity? Well, back in 2019 when it was founded, Pixxel was admitted to the Q3 2019 batch of Techstars Starburst Space accelerator which requires a Delaware holding company structure. This is why Pixxel’s holding company is Delaware incorporated Pixxel Space Technologies, Inc. Remember that in 2019, IN-SPACe didn’t exist & the future of India’s pvt Space industry was unclear until the 2023 Indian Space Policy was released. Having a foreign holding structure was quite the norm e.g. Groww, Meesho, Pine Labs etc have all flipped back to India before going public. Lots has changed in the intervening 7 years. Therefore, back in 2019, it was perfectly rational for Pixxel to flip to the Delaware holding structure. 2️⃣Is Pixxel an American company? This requires a nuanced understanding of history (which I shared above). So yes, the Pixxel “holding company” is US. But, if you look at the Principle of Materiality, Pixxel is very much Indian - Pixxel’s founders are Indian nationals (own ~25% of the US holding company), LinkedIn shows ~310 India based employees against 35 US based ones (latter includes their investors as well). Most US roles are front office (Govt liaison, Sales, Operations) whereas most Indian roles are R&D focused. This same Materiality test can be applied to Indian IT Services companies and you will reach the same conclusion: Just because your front office & revenue comes from the US, doesn’t make you a US company. Of course, you are welcome to your own interpretation of “American” in this context. As more Indian companies pursue global ambitions, not all of them will have the same entity structure but all of them will have large Indian R&D teams like Pixxel. 3️⃣“Govt of India has given Pixxel a ₹1,200 cr project award” NOT true. A consortium of Indian SpaceTech companies led by Pixxel won the tender process for a EO satellite constellation (estimated to cost ₹1.2k cr to build). Pixxel & its peers won because they bid at NIL value (this was confirmed by IN-SPACe chairman Pawan Goenka in Aug ‘25). In fact, Pixxel & its peers turned down the up to ₹350 cr of Govt support offered for this project. The L2 (losing) bid was led by another party that bid for ₹97 cr (Google the name, would explain a LOT about why today’s drama is taking place 😆) 4️⃣Has Govt of India given major grants to Pixxel? Once again, NOT true. Pixxel got its 1st iDEX grant on 25th July 2023; over 4 years after it was founded. The cumulative Govt funding support would be < ₹100 cr across a few iDEX SPARK, PRIME & DISC grants. This is would be < 5% of the total capital raised (even as far back as 2023 when the 1st grant was received). The early capital for Pixxel came from private Indian & US investors; the late stage money has come from overseas. 5️⃣“Pixxel has created the US entity to funnel IP from India to US” This misunderstanding is related to Pixxel Federal LLC. Under §117.9, only a US entity can engage in sensitive Govt contracts (technically, even Pixxel’s holding company could have met this criteria). However, to have a better chance at building a long term commercial relationship, it makes more sense to create a ring fenced entity with only US staff. Therefore, Pixxel Federal LLC is that ring fenced entity to cater to demand from US Govt entites e.g. the NRO which signed a contract with Pixxel. So, this required? Technically no. But, is it advisable? Of course! FYI, almost all major non-US Govt contractors follow such a structure e.g. BAE Systems of UK, Tales of France etc. FWIW, even US owned Capella Space created Capella Federal when it got a Govt contract Remember: In DeepTech, IP is never static & takes multiple forms. IP is built through novel ideas (easy to transfer), process understanding (hard to transfer) and personnel level tactic knowledge (impossible to transfer). As long as the R&D teams in India continue to chip away at hardware & system design, the “real” IP will always remain in India (BTW, this is also why it is so hard to beat global Aerospace giants despite them having local JVs) The sad part here is that the same clowns who cry about lack of Defense & Space exports fail to realize that Pixxel is exactly what DeepTech exports to foreign Govts are going to look like: Our IP first companies will have to hire local staff, create a ring fenced local entity and potentially co-build additional IP for the buyer. For me, Pixxel is a celebration of the fact that 2 young BITS graduates could (with a small amount of capital) create globally optimal technology from India (with a local R&D team comprising young talent including their own BITS batchmates). The NRO contract is the best global validation we could get. Our DeepTech IP export opportunity will look exactly like this. This is the only way we will be able to build global Aerospace & Defense (A&D) prime contractors from India for the world. The last mile always requires local JVs, some IP transfer & doing a bit extra to make your customer feel comfortable.

I'd be careful to draw any conclusion from Deedy's work below for a few reasons: (1) Prof Ilya Strebulaev's June 2026 research shows an elite MBA (GSB, HBS etc) has much higher odds of success ($bn club). Deedy has grouped all MBAs together (which causes this issue). Prof's research is free to read online (search for it) (2) YC admission is a major confounding variable here; YC is a known hunting ground for most T1s esp. given the historic high unicorn hit rate (not sure if he has tested for this) (3) Interactions (eg undergrad in IIT and MBA at GSB) can create an attribution mess. Eg Arvind from Perplexity is IIT Madras, UC Berkeley PhD and ex OpenAI (so lots of signals including the PhD negative one) The messiest interaction would be YC alumni (since it cuts across all categories below) Ultimately, being T1 funded is not the goal. Becoming a large & profitable biz is. However, no doubt, having a T1 on your cap table changes the odds of success (I've seen it first hand as a founder) Deedy has done some good work here but like all statistical work, I'd always be careful since VC is a fat tails business where Outliers & pattern breakers make you the most money 😇

"Choices" is a must read to understand India's historic foreign policy decisions. Former NSA Shivshankar Menon lays down key moments in our foreign policy (China War, Pakistan, Sri Lanka uprising, becoming a Nuclear Weapon state etc) Mr Menon was in the room when decisions were made; therefore, the book covers some high pressure moments (eg a key Bush x Manmohan meeting before signing of the Nuclear treaty in Nov '08) The best line in the book is: "Politics is a process without clear mathematical solutions or distinctions between right & wrong, true & false and black & white" Hence, the title is very aptly chosen as "Choices" (often difficult & irreversible in politics where everything falls in the court of public opinion in minutes)

📍Old Cataract, Aswan This 127 yr old hotel has a personal story: My great-grandfather visited Egypt on work in 1948 (for cotton trading) but couldn't afford to have a meal at this hotel during their stay in Aswan The hotel is like a museum; was built by Thomas Cook (the guy whom the travel agency is named after). Most rooms are named after someone & have a story. This one wasn't a surprise for mom but we got a suite at the hotel This same room hosted a few royals, Prime Ministers and other famous people over the years (the view is fabulous) Hands down this is the best hotel & hospitality I've witnessed!

Our shipyard workers cost less than China’s. We still paid 2x more for the same work. An old Maritime India Vision benchmark: India: $3 to 4 an hour, 150 to 180 hours, labour bill about $620 China: $5 to 6 an hour, 50 to 60 hours, labour bill about $300 Cheap labour is useless if you need 3x as much of it. Add loans at 10 to 12% (China: 2 to 5%) and 95% imported engines, and the problem is clear before the money even arrives. Now ₹69,725 crore is going in, with new fine print: no support below 30% Indian content, full support at 40%. Can it work this time? We found 3 tests that will tell us. Full Breakdown below.

Sudhanshu & team turned down a life changing acquisition offer right after their Seed round led by Info Edge (back in Nov '24) This team has stuck together and gradually found PMF across various distribution channels in the past year (you've seen it on X and on YouTube) With 3 Lakh DAU, they are on a very good trajectory now! They already have usage from South Korea, US and Japan; three very exciting & deep profit pool markets for Consumer Tech It is wonderful to see a global edu game experience being built from India for the world! Very well deserved!

China built about 950x more ships than India in 2024. 39.1 million gross tonnes vs 40,923. India pays foreign ships $75 billion a year to carry its own oil, gas, coal and fertiliser. We plan 437 new vessels worth ₹2.2 lakh crore. We tried this in 2002. Share of global orders rose 60x, to 1%. By 2014, ABG, Bharati and Reliance Naval (ex-Pipavav) were in debt or insolvency. Now ₹69,725 crore. Same bet, or a real industry?

📍Step Pyramid of Djoser This is the oldest Pyramid, built 4700+ yrs ago. Also, the 1st major stone construction! Went down the shaft to see the tomb. One item off my bucket list

📍Mariott Mena House, Cairo Mom has wanted to visited Egypt for 20+ yrs. Something always got in the way ('08 crisis, '11 uprising, '20 COVID) Managed to land a big surprise for her with this hotel & room. The Pyramid of Khufu ("Great Pyramid of Giza") is visible from the room! Great way to start our longest family holiday ever!

Finally, @Breakdownbyaeos makes its debut on Twtr! In an attempt to cut down my camera-shyness, I did a video with them a couple months ago (a solid 2.5 years after my first conversation with @waiting4_asi about this). Much to learn ofc, but lfg!