Corporates raise money from capital markets and use that money to buy cryptos. Their share prices raise, and that makes them repeat the whole exercise. MicroStrategy, a software company, owns Bitcoin worth $64.8 billion. Companies vie with each other to build their own crypto stash. It seems there is a race to build crypto corporate treasury.

A coffee company whose shares had fallen 90 per cent in 2024 have doubled in value since it went in for a crypto buy in May. It holds 69 Bitcoin, and wants to add thousands more.

Most corporate treasurers are risk averse. They would not buy volatile tokens. They too are lured by crypto wave. There is investor demand for exposure to crypto vehicles, it makes sense for more players to hop aboard the bandwagon and offer supply to match.

There are risks. A crypto sell off could turn into a crash. These are overpriced tokens held by the weak and indebted companies. It could create a wave of forced selling. It should be remembered that in the last crash, Bitcoin fell more than 50 per cent.

Are we using the Bitcoin innovation which is blockchain based prudently? Or is it speculative frenzy to make ‘money from money’?

Hard-boiled candies costing around Re. 1 or even less are popular both with the children as well as adults. These candies are widely distributed through grocery stores and other small retail outlets.

Pulse candy was launched by Dharmpal Satyapal in kachcha aam or raw mango flavour a decade ago. Within a year of its launch, it became a Rs.100 crore brand. Last year ending in March 2025, it became a Rs.750 crore brand. It is growing at the rate of 15 per cent (against the industry average of 9 per cent). DS has a share of 19 per cent in the Rs.4000-crore hard-boiled candy market. Pulse brand is likely to reach the 1000-crore mark in a couple of years from now.

The competing brand Alpen Liebe has advertising and marketing muscle, but pulse does not. It focuses more on product innovation and consumer experience. Pulse did some clever packaging. It grew despite no marketing push and has become a case study at IIMA.

It bets on bold Indian flavours. The market is replete with sweeter Western flavours. It expanded the market from kids to adults. It finds flavours that appeal to adults. Its major revenue is derived from adults. Its growth is organic. The brand is an outlier. There was raw mango flavor in the market prior to Pulse. But it was launched as a combination of sweet and salt, mimicking raw mango that most of us savoured in childhood. There is salt in the center, which refines the candy experience. It has become its best-selling candy. It introduced other flavours also such as litchi, orange, guava, and pineapple. Pulse has also launched tamarind flavor that makes Indians nostalgic about their childhood.

Pulse is an indulgence at a modest price. It has remained non-preachy. It relies on buzz word and school-yard gossip. Its signature is its masala core. It could explore more extensions, say lollipops and chews.

Its competing products are Perfetti Van Melle’s Alpen Liebe, ITC Candyman, Parle Products Toffees and Ravalgaon. Perfetti is the largest competitor which earned a revenue of Rs.3500 crore. DS brand Pulse is around Rs.1000 crore, and the company wants it to reach Rs.5000 crore by 2029.

Though initially marketing was not the focus, the marketing spend has now reached 6-8 per cent of the annual revenue, on par with industry norm. Below-the-line promotion and distribution matter more in confectionery industry. Sampling and influencers have worked in favour of Pulse. It reaches 35 lac outlets in the country.

Indian agriculture has registered a growth of 4.6 per cent in FY 2025, and further growth of similar magnitude a decade hence can reach us to 5 per cent growth. It is vital for India to become a developed nation.

India’s agricultural growth is associated with crops and non-crops. The share of crop sector has declined, and yet it remains the largest contributor comprising cereals, pulses, oilseeds, other field crops, and horticulture. The non-crop sector consists of livestock (with milk-production remaining dominant), fishing, aquacultures. The rising share of non-crop sector is remarkable.

While free trade deals are negotiated, we have to bear in mind the interests of six million farmers who cultivate crop. To illustrate, US origin soyabean prices are around $ 390 a tone, whereas India’s MSP is $ 620 a tone. If there are cheaper imports at lower prices, our domestic cultivation will be affected. The difference in cost is on account of the large farm holdings in the US, that give them economies of scale. The average size of farm in US is 466 acres whereas in India it is less than 2-7 acres.

While negotiating a deal, we have to consider the India tariffs on agricultural goods of 396 per cent which is eight times the tariff the US charges. India’s tariffs must come down. There should be quotas set. However, 80 per cent of Indian agriculture is competitive. There are some exceptions — wheat and dairy items (skimmed milk powder).

There is always some give and take in negotiations. We should decide what we can export to the US in return for allowing them access to Indian markets. We can promote high value horticultural products such as bananas, mangoes, grapes, promegranates. In return, we can lower duties on walnuts, apples, cranberries and blueberries which we hardly produce or produce in negligible quantities.

Onboarding farmer unions is a recipe for successful trade negotiations.

AI affects industries, but it is too early to gauge its impact, both economic and social. AI has a very broad scope as a tool. It is being applied in education, healthcare, finance and across a variety of other fields. There is rapid progress in the application of technology. To illustrate, generating images from text seemed like magic. These days it is commonplace. It creates both opportunity and pressure. Innovation alone is not enough.

There are certain concerns about the use of AI. AI can spread disinformation. There could be situations where AI is not in sync with social objectives. In the long term, AI affects jobs and economic systems. As a society, we must be prepared to meet what is coming.

Knowledge, in the past, was limited by access. AI gives easy access to knowledge in the areas such as health and finance. It is a positive development. It raises the issues of shift of economic systems, and relevance of skills.

While using AI models, it is to be seen that the model uses data to train itself represents all viewpoints, languages and culture. Countries could produce their own country-specific data. Such a model reflects local realities.

We do require more reasoning capacity from models. They should think in context. They should not restrict to predicting the next token. The techniques used today can grow. These should cope with more complex problems. The models require high quality expert data.

AI can take over repetitive and simple tasks. People should focus on skilled work. It is not necessary to have breakthroughs in model design. What is needed is thoughtful model development.

Current models are designed to answer questions. They are not expected to raise questions or be skeptical. In future they should be more conversational. They, should learn to say, ‘I don’t know.’

Jane Street, a 24-story tower southwest of New Delhi, housed at least half a dozen high-speed trading firms. Jane street is a blue-glass building with a roof-top helipad. It was the center of a trading boom that made India the world’s biggest equity derivatives market by volume.

SEBI barred JS entities from operations in India. In 2024, JS India’s revenue exceeded $2.3 billion. They reaped a profit of Rs.36,502 crore between 2023 and 2025. This includes Rs.43,289 crore profits from Index options. Rs.7687crore were the losses in cash and futures. In January 2024, Jane profited Rs. 734.93 crore from Bank Nifty.

The other competing firms, JS peers, Citadel Securities earned $344 million in 2024, Optiver BV earned modest profits of around $300 million in 2024 from options trading in India. The other competing firms in algo-trading and high-speed trading include IMC Trading, Flow Traders, Tower Research, Two sigma, Jump Trading, Goldman Sachs, Hudson River and DRW.

The supporters of high-speed trading firms say these firms play a vital role as market makers. They facilitate price discovery by narrowing the bid-ask spread on trades, or the difference between the highest price a buyer is willing to pay and the lowest a seller is willing to accept.

Critics argue that they leverage their technological advantages — having their own server in data centers owned by stock exchanges. They front-run ooders by retail and other investors. They use sophisticated algorithms to profit from small price fluctuations, or benefit from erratic trading behaviour among retail speculators.

India presents a unique opportunity considering its market structure. The options market has high retail participation. The trade sizes are relatively small. There are bets possible with as small an amount as 12 cents. Options buyers get the right to buy or sell an asset. (say stock or index) at a specific price on or before a certain date. They enable traders to bet on the direction of stocks (that too for a fraction of the cost of buying and selling the actual securities).

Speed traders account for about 60 per cent of the trading volume in India’s equity derivatives market. There is a $ 3 trillion trade in notional value daily. It constitutes 40 per cent of stock trading in the county.

The spectacular success for global firms is at the expense of small investors. In the last three years, they lost $21 billion from futures and options.

SEBI seized Rs.4840 crore from the US-trading giant citing it as illegal gains. Jane Street has disputed SEBI’s findings. It is a watershed moment. Jane Street may appeal against the order to Appellate Tribunal and Supreme Court. SEBI officers are confident about the order and treat this as an open-and-shut case.

Several international players are reconsidering their India plans. It puts a damper on trading.

Indian officials are skeptic about outsized profits of global players and sophisticated traders. It puts high-frequency traders on watch. Though welcome, these players will have to play by the rules.

In India, firms make many rapid-fire traders each day. They pocket small amounts of money by exploiting market anomalies. There is little risk. These modest gains accumulate into substantial profits.

The US Senate passed the Genius Act, to regulate specifically stablecoins, marking a new phase for crypto eco-system. It adds clarity and confidence in cryptos. Cryptos started as competition to fiat currency but have now evolved into a widely held asset class. The evolution has been influenced by institutional moves, regularity recognition and investor behavior.

In 2020, MicroStrategy, a listed company, held Bitcoin on its balance sheet as a treasury asset. This was the first listed company to do so. Square and Mass Mutual followed suit. By 2021, large investors evinced interest. The Securities and Exchange Commission approved ETFs (exchange-traded funds) in 2024. It welcomed traditional institutions and retail investors to access Bitcoin through regulated channels, on the lines of gold ETFs.

In 2025, ETFs drive inflows. Global banks have started developing structured products tied to Bitcoin. There is murmur about Bitcoin reserves.

There is a shift in policy — from restriction to integration. Between 2018-19, regulatory pushback was a patchwork. In 2020, Bitcoin was treated as properly attracting capital gains.

In 2021, El Salvador made Bitcoin legal tender (since then revoked). India introduced a capital gains tax on cryptos. It tantamounted to recognition without endorsement. In 2025, President Trump supported stablecoin legislation. India has prepared a consultation paper.

In 2018, there was crypto winter because of subdued investor sentiment. By 2020, investor started to treat Bitcoin as a store of value, on the lines of digital gold. It is also a hedge against inflation.

The distribution mechanism of exchanges improved between 2022 and 2023. In 2024 and 2025, the spot ETFs became normal. Bitcoin entered mainstream portfolios. Bitcoin is no longer purely speculative. It is treated as a long duration asset.

India permits ownership and trading of Bitcoin. RBI is skeptical about systemic risks. The taxes introduced have become deterrents. India can come out with a consultation paper in 2025. Industry favours prudence, rather than prohibition. Informally, a Bitcoin reserve is proposed in India. India is cautious and is in the wait and watch mode.

Gold has a history of last 5000 years to maintain its value. Bitcoin requires a supportive infrastructure. There is a physical constraint on gold supply whereas Bitcoins scarcity depends entirely on code. Gold has industrial application and decorative value. To be a true store of value, it needs to lower its volatility.

Bitcoin adoption could be dependent on business cycles. It can reverse rapidly. Bitcoin’s cryptographic security could be challenged by advances in quantum computing. Central-bank digital currencies could reduce demand for decentralized cryptos. Extreme volatility does not make cryptos dependable.

It is to be seen how Bitcoin fits into broader financial system. Bitcoin may not replace fiat currency but it is here to stay.

Stablecoins are a type of cryptocurrency that could maintain a stable value. These are typically pegged to a reserve asset such as fiat currency, say the US dollar a commodity, say gold or through algorithmic mechanisms. The algorithmic mechanism uses smart contracts to control the supply and maintain peg without actual collateral.

In short, stablecoins are blockchain-issued cryptocurrencies designed to mimic the dollar. These are potential contenders in payment system.

Still there is skepticism about this mode of payment. The German regulator BaFin knows what havoc Wirecard caused in Germany. Wirecard AG was a celebrated Fintech company in Germany which collapsed in 2020, after it was discovered that 1.9 billion Euros were missing from its account. Since then, the regulator has become extra-cautious.

A racy technology promises how we pay and disrupt the old system of payment firms. In the past decade, the world is moving towards being cashless. There is easy shopping online. It facilitates fintech companies. There are electronic shuffles of money — A to B. But it also exposes us to online hacks, heists and fraud. Fraud rates are generally low, say 0.1 per cent of card spend. Regulators still face pushbacks since they have to prioritize between safety and speed.

Stablecoins too have similar tradeoffs. By 2030, this market could soar to $ 1.3 trillion. It has speed, low-cost and 24-hour access in cross border payments. Yet faster and cheaper do not necessarily mean safer. Stablecoins’ philosophy is ‘be your own bank’. It also means victims of fraud are less protected. There is risk of stolen funds too. The majority of illicit activity on the blockchain is through stablecoins. Criminals too want to reduce costs and maximize profits.

Stripe Inc, a payment company, is launching stablecoin-funded accounts in over 100 countries, and is partnering with Visa Inc.

Crypto world is ‘Wild West’. Some countries have become cautious, whereas some approve the digital currency as if a meal is ordered at McDonald’s. Tether Holdings is the biggest issuer of stablecoins. It has circulation of $150 billion. It is based in ELSalvador and has yet to be fully audited.

Let there not be another Wirecard to give us a wake-up call.

Some companies are leaders in building AI models such as OpenAI, Google DeepMind, Anthropic and others. These companies earn their revenues from multiple revenue streams — product offering, partnerships, licensing and cloud infrastructure.

API and Subscription Services

ChatGPT has paid tiers, say available for use at $ 20 a month. Thus, one major source of revenue is through subscriptions.

Companies make access available to their AI models through APIs. Later, businesses integrate these to their products and pay per usage (tokens, queries etc.)

Enterprise and B2B Services

Model making companies offer tailored or customized solutions to large enterprises, say OpenAI’s service to PwC.

Clients fine-tune base models or proprietary data for specific tasks and pay the model companies.

Cloud Infrastructure and Partnerships

Cloud providers such as Azure, Google and AWS invest in AI companies and then resell AI capabilities through their cloud platforms. Azure gets exclusive rights to host and sell OpenAI services against Microsoft’s investment in OpenAI.

Licensing

Companies license their models for use in third party applications, e.g. integration of ChatGPT in Microsoft Word, Excel as Co-pilot for which Microsoft has to pay OpenAI.

Advertising

This is not a common source for foundation models. Models that offer AI search, say Gemini or Perplexity can monetize through ads or sponsored results on the lines of traditional search engines.

Grants, Investments and Strategic Tie-ups

Venture capital firms, governments tech grants schemes offer funds to foundation model building companies. They are later monetized.

To illustrate, ChatGPT plus receives @ $20 a month from 10 million users (10 million x $20 =200 million per month. It equals a yearly revenue of $2.4 billion). It earns from Microsoft via Azure usage a share of user revenue. It earns for API access from business using GPT. PwC has made available ChatGPT to 1 lac employees, and OpenAI earns tens to hundreds of millions per client per year. Fine-tuned GPT models for legal, medical and financial usage provide earnings to OpenAI.

There is heavy math behind LLMs so that they generate text shockingly fast.

These run on GPUs or TPUs built to do massive parallel operations. It boils down to linear algebra, mainly rallel matrix operations. It boils down to linear algebra, mainly matrix multiplications and additions. They perform thousands of these operations at once. The model does it in parallel, and not step-by-step as in CPU.

ently. The transformer architecture is designed for speed. Instead of processing language word by word (like RNNs), they use self-attention, allowing them to look at the entire context at once. The design generates output (inference) faster and more efficiently.

LLMs generate text one token at a time but the computation for each token is highly optimized. There are frameworks such as PyTorch, TensorFlow, JAY, CUDA backends that make the model run extremely efficiently.

While generating multiple tokens in a sequence (such as a full sentence), models cache the internal activations, (such as key and value vectors in attention layers) from previous steps. It is called KV caching and massively reduces computation time in generation.

Some models are optimized further by quantization and pruning, thus reducing memory and increasing speed with minimal loss of accuracy.

A modern LLM such GPT4, runs on a GPU such as A100, which can generate hundreds of tokens per second (though it involves billions of parameters and trillions of multiplications). It is thus a triumph of smart architecture, parallel computing and engineering optimization.

Axiom -4 (Ax-4) mission is the beginning of the Indian human space flight. In this mission, a commercial flight has been used operated by Axiom-4, a Houston-based private company. There is collaboration between NASA, ISRO, ESA and SpaceX. Shubhanshu Shukla pilots this mission. He is the second in India to go into space. The space craft docked into International Space Station (ISS) on Thursday, 26th June 2025. Shukla is the first Indian to visit NASA’s orbiting laboratory. Between Rakesh Sharma’s space flight aboard Soyuz in 1984 and Shukla’s AX-4, there is a gap of 41 years. Rakesh Sharma’s mission was symbolic, but Shukla’s mission will lay the foundation for the future of humanity. It’s a 14-day mission.

AX-4 will conduct a series of experiments. Some of the results will facilitate ISRO’s future manned flights into space. It will help India set up its own space station.

There are seven experiments related to health and growth of crops in gravity-less environment. There are experiments related to microorganisms (cyanobacteria which produce energy through photosynthesis on the lines of plants). Microorganism study may give clues to deep space exploration and long-term space habitation.

There will be an experiment on germination and growth of sprouts. One experiment will be on seeds and how they are influenced. One experiment on microalgae will be to study oxygen generation system, and act as a food source.

ISRO will study muscle behaviors by studying muscle degradation on account of ageing and natural causes and the weight of a person. These two causes can be isolated into space, since weight factor is eliminated. We can focus on muscle changes purely due to natural reasons.

AX-4 is a giant leap forward and for India. Its scientific objectives will benefit the whole of humanity.