The year was 1964. Sean Connery had arrived in Switzerland to shoot Goldfinger. Along with much of the cast and crew of the James Bond movie, he stayed in a quiet mountain retreat perched high above Lake Lucerne, where steep forested slopes gave way to sweeping views of the Alps.
Over the decades, the resort became accustomed to famous guests—movie stars, royalty, business leaders, and heads of state. That place is Bürgenstock.
This week, its most important visitors may have been the ones whose success is measured not by what happened but by what did not.
US Vice President JD Vance led an American delegation that met Iranian negotiators, alongside mediators from Pakistan and Qatar, to work through the technical details of a peace framework aimed at ending the conflict in West Asia.
The meeting did not end the crisis—much remains unresolved. But it produced a 60-day roadmap to negotiate the remaining technical issues, offering markets their clearest sign yet that diplomacy had regained some momentum.
Markets began to find their first evidence that diplomacy might translate into economic normalisation.
The Strait of Hormuz, through which roughly a fifth of the world’s oil trade passes, has begun to stir back to life.
Since the agreement was signed, at least 172 vessels have crossed the strait, according to maritime intelligence firm Kpler. Traffic remains well below pre-conflict levels, but it marks a meaningful improvement after weeks of disruption.
In the latest sign that confidence is gradually returning, around 20 million barrels of oil moved out of the Gulf in a single day this week, with flows approaching pre-conflict levels even as mine-clearance operations continue and many shipowners remain cautious.
Markets have responded just as quickly. Brent crude, which briefly traded above $93 a barrel earlier this month as traders priced in the possibility of a prolonged disruption, has since fallen below $75 – its lowest level since before the conflict began. The oil futures market has also shifted, with later-dated contracts now trading above near-term ones, suggesting traders are increasingly worried about excess supply rather than immediate shortages.
The return, however, is still tentative. More than 250 tankers and 440 cargo ships remain inside the Gulf, many still anchored. Shipping companies continue to navigate new permit requirements, conflicting messages from Iranian authorities, and the lingering risk posed by sea mines.
Markets, much like shipowners, are proceeding with caution. Equity benchmarks Nifty and Sensex rose for the third week in a row, posting their longest winning streak in seven months.
For India, where more than 85% of crude oil is imported, that repricing matters. Lower oil prices, if sustained, would ease pressure on inflation, government finances, and the current account, even if those benefits take time to feed through the economy. But some positive signs are already visible as the government increased LPG supplies to commercial and industrial users after imposing curbs soon after the war broke out.
For investors, the significance of Bürgenstock lies less in the symbolism of the venue than in what it represents. The talks did not eliminate geopolitical risk, but they reduced the probability of the most disruptive outcome.
Markets rarely wait for certainty. They respond when the balance of probabilities begins to shift.
For now, the conversations that took place high above Lake Lucerne matter less for what they promised than for what they persuaded markets to believe.
Cost of Intelligence
From war, let’s move on to another hot topic that is keeping investors busy—artificial intelligence.
For much of the AI boom, investors have focused on one question: who is building the smartest model?
Increasingly, a different question is beginning to matter: who can deliver that intelligence most efficiently?
Two announcements this week suggest the economics of AI may be entering a new phase.
OpenAI unveiled Jalapeño, its first custom inference chip, developed with Broadcom.
On the same day, Qualcomm and Meta announced a multi-generation agreement to develop data-centre CPUs for Meta’s AI infrastructure.
Different companies, different technologies, but both announcements point in the same direction.
The first phase of the AI race rewarded companies that built increasingly capable models. The next phase may increasingly reward those that can run those models more efficiently, more reliably and at lower cost.
To understand why, it helps to distinguish between two very different kinds of AI computing.
Training is the process of teaching a model. It requires enormous bursts of computing power over a relatively limited period. Inference is what happens afterwards. Every time someone asks ChatGPT a question, generates an image or uses an AI assistant, the model performs inference. As AI applications move from experimentation to everyday use, these inference workloads are growing rapidly.
That makes efficiency increasingly valuable. Even small improvements in cost, power consumption or response time can translate into meaningful savings when AI systems serve millions of users around the world.
That helps explain OpenAI’s latest move.
Rather than designing another training processor, the company chose to build an inference chip. Jalapeño has been developed with Broadcom as the first member of what OpenAI describes as a multi-generation family of AI accelerators. The company says the chip has been optimised around the networking, memory and computing requirements of large language models and is intended to complement, rather than replace, its existing hardware suppliers.
Meta’s agreement with Qualcomm reflects the same shift from another direction.
Qualcomm built its reputation in smartphone processors. Under the new multi-generation partnership, it will develop data-centre CPUs designed specifically for Meta’s expanding AI infrastructure. The agreement signals that custom computing hardware is becoming a larger part of how leading AI companies plan for the next decade of growth.
Taken together, the two announcements suggest that AI companies are no longer thinking only about the software they build. They are increasingly investing in the infrastructure that delivers that software.
That has implications for investors. Much of the early excitement around AI centred on model developers and graphics processors. But as AI adoption broadens, value creation is also spreading across the infrastructure stack — custom silicon, networking, memory, data centres and power systems that enable AI to operate at scale.
This is not a sign that the race to build better models is ending. Rather, the basis of competition appears to be widening.
The companies developing frontier AI models are increasingly seeking greater control over the infrastructure beneath them—not simply to improve performance, but to manage costs, secure supply and scale their services more efficiently.
Every technology cycle eventually reaches a point where invention alone is no longer enough. AI may be approaching that moment.
The next phase of the industry may be defined not only by who builds the smartest models, but also by who can deliver that intelligence most efficiently. For investors, that broadens the opportunity—and the range of companies that could benefit from it.
Beyond Renewables
While India maybe lagging behind in the AI race, it is trying to catch up in another critical sector—nuclear power.
For decades, nuclear power occupied a unique place in India’s energy system. It was seen as strategically important, capable of producing large amounts of low-carbon electricity. Yet unlike thermal, renewable or transmission projects, the government kept it limited to state-owned companies until it allowed private companies late last year.
This week, Adani Group outlined its plans for the sector.
At its annual general meeting, the group announced plans to enter the nuclear power business through a new entity, Adani Atomic Energy, with a target of building 10 gigawatts of capacity by 2035. It said it has identified land for projects, describing nuclear energy as part of its strategy to meet India’s growing demand for “clean, round-the-clock power”.
The ambition is significant in the context of India’s existing nuclear programme. The country currently has around 8 GW of installed nuclear power capacity. The government aims to expand this capacity to 100 gigawatts by 2047. State-run Nuclear Power Corp of India, currently the sole operator of nuclear plants, aims to lift its capacity to 50 GW while NTPC, also state-owned, is aiming for 30 GW of nuclear capacity. If Adani Group meets its targets, it would be the largest private operator and the third-largest overall in the nuclear sector.
But why does Adani Group want to enter the atomic power sector? Part of the answer lies in how India’s electricity demand is evolving.
For much of the past decade, the conversation around the energy transition centred on replacing fossil fuels with renewable energy. Solar and wind have transformed the economics of clean electricity and now account for a growing share of new capacity additions. But India’s power system is also changing in another way.
Manufacturing, electrification, digital infrastructure and large data centres are increasing demand for electricity that is available throughout the day, not only when the sun shines or the wind blows. That does not diminish the importance of renewables. But it changes the role that different technologies are expected to play.
Nuclear power occupies a distinctive position in that mix. Unlike solar and wind, it can provide continuous low-carbon electricity. Unlike thermal power, it does so without ongoing carbon emissions during generation.
That helps explain why Adani described nuclear power as a source of round-the-clock electricity rather than simply another clean-energy investment.
The nuclear announcement was also part of a much wider investment plan. Alongside it, the group raised its thermal power target to 45 GW by FY32, expanded its data-centre ambitions to 3 GW by 2030 and announced new partnerships in aircraft manufacturing and hydropower.
Viewed separately, these announcements span different industries. Viewed together, they suggest Adani is planning for an energy system that will rely on multiple sources of generation rather than a single dominant technology. Renewables, thermal power, transmission, storage and nuclear each address different parts of the growing electricity demand.
Adani’s proposed entry into nuclear power suggests the next phase may be less about one technology replacing another and more about building an electricity system in which different sources work together to deliver reliable, low-carbon power as demand continues to grow.
Beneath the Bell
Investing has never been easier. A few taps on a smartphone are enough to buy shares, mutual fund units or derivatives contracts. Behind that simplicity, however, sits a financial system that has become steadily more complex.
Every additional investor, trading channel and algorithm places greater demands on the technology that keeps markets running.
Two consultation papers released by the Securities and Exchange Board of India this week suggest the regulator is preparing for exactly that future.
One seeks to expand Direct Market Access (DMA) – a facility that allows investors to place orders directly into an exchange’s trading system through a broker’s infrastructure.
Until now, DMA has largely been available only to institutional investors. SEBI has proposed removing that distinction, noting that advances in technology mean its benefits can now be extended to other categories of investors while retaining the necessary risk-management checks.
It has also proposed allowing exchanges to determine, from time to time, which categories of investors may use DMA in exchange-traded commodity derivatives, bringing those rules in line with other market segments.
The second proposal looks much deeper into the machinery of the market itself.
SEBI has proposed consolidating multiple technology and cyber-security circulars governing stock exchanges, clearing corporations and depositories into a single framework aligned with its Cyber Security and Cyber Resilience Framework.
The consultation paper seeks to remove overlapping provisions while bringing together rules covering annual system audits, cyber security, business continuity planning, disaster recovery, capacity planning and algorithmic trading. It also proposes harmonising technology requirements across equity and commodity exchanges, introducing common capacity-planning standards and giving greater flexibility to vendors operating inside exchange co-location facilities.
Taken individually, both proposals appear administrative. Taken together, they tell a broader story. As markets become easier to access, they also become more dependent on the technology that supports every trade. Regulation, therefore, is evolving in two directions at once – widening access while strengthening the infrastructure beneath it.
That shift is easy to overlook because investors rarely interact with that infrastructure directly. Most never think about the systems that process orders, monitor capacity, recover from outages or defend exchanges against cyber threats. Yet those systems have become as essential to market confidence as trading rules themselves.
Some of SEBI’s proposals illustrate that changing mindset. The regulator has suggested common standards for capacity planning across market infrastructure institutions and proposed that exchanges and clearing corporations take immediate action when utilisation of critical IT infrastructure crosses prescribed thresholds. The emphasis is not merely on responding to technology failures but on ensuring systems remain resilient before they become stressed.
The proposal to simplify the rulebook is equally revealing. Rather than creating an entirely new layer of regulation, SEBI is attempting to consolidate existing requirements, eliminate duplication and align legacy rules with a more modern cyber-resilience framework. As markets become technologically more sophisticated, the regulator appears to be making the rules governing that technology easier to administer and supervise.
For investors, these proposals are unlikely to change tomorrow’s market direction. Their significance lies elsewhere. Indian capital markets are becoming larger, faster and more digital. As that evolution continues, trust will depend not only on the companies listed on an exchange or the rules governing trading, but increasingly on the resilience of the invisible systems that make every trade possible.
The strongest markets are often defined not only by how efficiently they allocate capital, but by how reliably the infrastructure beneath them continues to function.
Market wrap
India’s stock market benchmarks rose this week, as falling crude oil prices helped boost sentiment. A statement by the Reserve Bank of India Governor Sanjay Malhotra that it was premature to talk about rate hikes also buoyed sentiment.
The Nifty 50 rose 0.2% and the BSE Sensex climbed 0.4% in the holiday-shortened week—markets were closed on Friday. This is the third week of gains in a row and marks their longest winning streak in seven months.
In the broader market, however, the mid-cap index fell 1.2% while small-caps were little changed.
Only six of the 16 major sectors recorded gains this week. The pharma index, which is less exposed to crude oil and monsoon risks, rose 2.1%. The metals index posted a weekly loss of 4.4%, tracking weaker global prices.
Among Nifty stocks, IndiGo parent InterGlobe Aviation surged 8.5% as lower oil prices eased concerns for India’s biggest airline. Drugmakers Cipla and Dr Reddy’s Labs jumped over 6%. Sun Pharma, and hospital chains Max Healthcare and Apollo rose 2-3% each.
ICICI Bank was the top financial stock, rising 3.4%. Shriram Finance, Kotak Mahindra Bank, Axis Bank also ended higher.
At the other end of the spectrum, tech stocks continued to bleed. Infosys plunged 7.65%, HCL Tech slumped 5.3%, TCS lost 4.9% and Wipro shed 4.3%. Among metal stocks, Tata Steel melted 5.9% while Hindalco cooled by 5.5% and JSW Steel dropped 4.9%.
Other Headlines
- Meta to invest $900 million in CRED, names founder Kunal Shah as new WhatsApp CEO
- Adani Airports to invest Rs 20,000 crore in developing airport cities
- Govt restores 50% LPG supplies to commercial and industrial customers
- Amazon to invest additional $13 billion in India cloud, AI
- Enforcement Directorate flags forex breaches, missing records at Rajesh Exports
- Billionaire Gautam Adani’s lawyers urge US judge to dismiss criminal case after settlement
- Premature to talk about rate hikes, says RBI governor Sanjay Malhotra
- Bajaj Auto says ransomware attack hits systems
- Tata Electronics hit by cyberattack claiming to expose Apple, Tesla trade secrets
- Tata Motors Passenger Vehicle says aiming to almost double revenue, sales volume by 2031
- Govt to sell up to 2% stake in Indian Railway Finance Corp
- India’s infrastructure output growth slows to eight-month low of 0.5% in May vs 1.8% in April
- Sun Pharma to acquire Innovcare Lifesciences for Rs 271 crore
- That’s all for this week. Until next week, happy investing!
Interested in how we think about the markets?
Read more: Zen And The Art Of Investing
Watch here: Investing in International Markets
Start investing through a platform that brings goal planning and investing to your fingertips. Visit kuvera.in to discover Direct Plans and Fixed Deposits and start investing today. #MutualFundSahiHai #KuveraSabseSahiHai
