Seeking Alpha by Adding Beta

Saturday, November 07, 2026

Diyas, Democrats & Data Centers...!

Happy Diwali and Naya Saal Mubarak! What a week to light the diyas: a blue wave, the first national referendum on AI data centers, the worst day for chips since the Covid crash, and market message to Mag7 - cash flow matters!

In my August post, The Reckoning, I argued the market had moved from blind faith to "show me the ROI" and that hyperscalers would soon adopt a "no capex increase" tune. Now voters have joined the CFOs in asking the same question. Here's my take...

1. Shubh Deepavali: The Original Year-End Review

The Significance: Light over darkness, and a fresh set of books.

Diwali falls on Sunday, November 8 this year, on the new-moon night of Kartik Amavasya. Whether it's Rama's return to Ayodhya, Krishna's victory over Narakasura, or Lakshmi visiting homes that are clean and lit, the message is the same: knowledge over ignorance, hope over fear, light over darkness.

What I have always loved as student of economics, technology and markets is that Diwali is also the original fiscal year-end. For generations of Indian traders (I come from trader family myself), Lakshmi Puja meant Chopda Pujan, closing the old ledgers and blessing new ones. Vikram Samvat 2083 starts with a clean slate, and Muhurat trading remains Dalal Street's symbolic first trade. Indian merchants were doing annual reviews long before earnings calls.

So let's close the books on a very eventful week. Given what chips did on Thursday, some corners of the market could use a few extra diyas!

2. The Blue Wave and the Data Center Referendum

The Shift: From "AI is inevitable" to "not in my backyard"

Democrats won both chambers on November 3. The Senate flipped with wins that sounded like fantasy a year ago: Talarico in Texas (first Democrat since 1988) and Andrews in South Carolina (first since 1998). The Iran war and affordability were the headline issues. The sleeper issue was data centers.

AI showed up on the ballot as a local issue for the first time: power bills, water, land, and campuses that employ only few people once the cranes leave.

Exit polls: Over half of voters called data centers a crucial or major factor (NBC).
Texas: Gina Hinojosa unseated Gov. Abbott on a data center moratorium, in the AI buildout's own backyard.
The pre-game: New York's moratorium in July, Abbott's own pause in August, and Monterey Park's permanent ban (86% support).
The irony: Hundreds of millions in pro-AI super PAC money may be the worst ROI of any AI spend this year!

I'm calling this the birth of NIMBYtes: Not In My Backyard, applied to bytes, and to who pays for the grid behind someone else's chatbot.

The market's verdict came in two acts. Wednesday brought the usual "uncertainty is over" rally (S&P +1.6%). Thursday read the fine print: S&P -4.4% to 7,354, Nasdaq -5.5% into correction, VIX up 70% to 28. It was the worst second session after a midterm since at least 1950, concentrated in chips, AI hardware and power equipment.

Don't over-read Washington: a national moratorium needs legislation, a seated Congress and a President who won't veto it. The real bite is at the state and county level: permits, interconnects, water and zoning. Slower than a bill, but harder to lobby away...

3. Mag7 Earnings: Beat, Raise... and Get Spanked

The Shift: From "how much are you spending?" to "how much cash is left after you spend it?"

On paper, Q3 was great: Azure +48%, Google Cloud doubled (+103%), AWS +39%. Yet all six reporters fell the next session. The market is grading free cash flow and capex discipline, not revenue.

The pattern is exactly the one I flagged in August: reward restraint, punish the burn.

Company	2026 capex plan	Capex signal for 2027	Free cash flow
Amazon	~$220B (kept)	"Meaningfully more"	-$42.4B (trailing 12 months)
Alphabet	$195-205B (kept)	"Increase significantly"	-$9.1B (Q3)
Meta	$135-145B (narrowed)	Not guided yet	-$2.1B (Q3)
Microsoft	~$170B (cut from ~$175B)	Capacity slips into 1H27	+$18.6B (Q3)
Tesla	~$20B (cut from >$25B)	Restraint	+$1.5B (Q3)

From the tape:

The cutters got a pass. Microsoft trimmed capex and stayed FCF positive: flat stock. Tesla cut capex a fifth: a mild -3.1%.
The spenders got spanked. Alphabet missed on both lines, burned $9B of FCF, and grew long-term debt from ~$22B to $125B in a year. The buildout now runs on bonds, not just cash flow.
The bill is landing on the consumer. Apple's new CEO called memory pricing "unprecedented" and Qualcomm's handset sales fell 26%. When AI eats the memory supply, your next phone pays for it.

Add it up: the Mag 7 spent about $647B on property and equipment over the last four quarters, up 82%, and the five that guide plan roughly $750B for 2026. That's nearly all of Big Tech's operating cash flow, leaving no cushion if growth, rates or politics turn.

4. Semageddon: Before the SaaSpocalypse Is Even Over

The Shift: From "capex is going up" to "capex is going up more slowly or peak capex is here" — and for chip stocks, that's the whole ballgame.

Software has rallied off its April SaaSpocalypse lows, but the scars aren't healed. And before that story is even finished, the baton has passed to the other side of the AI trade. Welcome to Semageddon.

Here is my hypothesis in one line: semiconductor stocks are priced on the second derivative of hyperscaler capex, not the level. Chip revenue tracks capex spend; chip valuations track capex growth. Capex doesn't need to fall for chips to crack; growth only needs to peak.

That's what Thursday started pricing:

The peak moved forward. Dan Ives, usually the loudest bull, now sees capex growth peaking in 2027, not 2028-29, with 10-20% of 800-1,000 planned US data centers cancelled or delayed.
The carnage: SOX fell 12% on Nov 5, its worst day since March 2020, and sits ~32% below last month's peak. Micron and AMD fell ~17% each, despite AMD's data center revenue nearly doubling.
Suppliers executed, spenders acquitted: From Oct 28 to Nov 5, twelve AI hardware suppliers fell ~32% on average while Microsoft, Amazon and Meta were flat. Fewer new data centers make the existing capacity more valuable: the customer's relief is the supplier's pain.
It went global: TSMC -10%, Advantest -12%, Kospi circuit breakers, Nikkei in a bear market.

The history rhymes, loudly. In 2000, carriers were still raising capex when Cisco peaked; the SOX went on to lose ~80% by 2002. The fiber got used, but its sellers never got their multiples back. Shovel stocks discount the end of the gold rush before the miners stop digging.

Now, the case against Semageddon. It is fair to note that this is, so far, a valuation event and not an earnings event. Nvidia guides ~$108B, memory is scarce, and 2027 visibility is intact. Nobody is cancelling orders; markets are just removing 2028-29 growth. And Tokenomics cuts both ways: fewer GPUs per token, but far more tokens.

My take: we are probably closer to peak capex growth than peak capex. That means a regime change in semis from "buy every dip" to "pick the bottleneck": memory, power, optics and custom silicon over pure capacity plays. Market is fearing "cuts" in capex in later part of 2027 and pricing the possibility now...

5. What to Expect for the Rest of 2026

The Shift: From a narrative market to a calendar market. The next eight weeks will be decided by four dates, not by headlines.

My New Year's Day range for the S&P was 6,000 to 8,000. We touched 7,819 on October 6 and closed Thursday at 7,354, about 6% off the top. Friday's bounce shows dip-buyers are still here, but they're waiting on these dates:

Date	Event	Why it matters
Nov 10	October CPI	Sets the tone for the Fed; oil near $97 isn't helping
Nov 13	October PPI	Pipeline inflation after the Iran-war oil spike
Nov 17	Nvidia Q3 FY27	The Semageddon referendum; guide is ~$108B
Nov-Dec	Lame-duck Congress	Spending deadlines; any AI or energy bills
Dec 8-9	FOMC	Futures price ~70% odds of a rate hike

And then there is the real test of AI - Anthropic IPO. If that goes well, it may give short-term respite till the first earnings report.

Note the Fed line: under Warsh, the market debates not how many cuts but whether we get a hike. Warshnomics, meet NIMBYtes: costly capital plus local resistance is tough on a debt-funded buildout.

My scenarios for year-end:

Range trade into Santa (55%): S&P chops between the 200-day (~7,320) and the election-day close (~7,570). Nvidia beats, but the market wants 2027 visibility. Leadership rotates to AI spenders, software, healthcare and financials.
Relief rally (25%): Strong Nvidia guide, cooler CPI, fading hike odds. S&P retests 7,800; semis recover half the drawdown.
Semageddon spreads (20%): A sustained break of the 200-day, more Qualcomm-style guide cuts, and moving state bills take the S&P to ~7,040, a 10% correction. The 2018 midterm path is the closest precedent.

Longer term, this week fits my August call of a 1929-style reckoning around 2028-29. Bubbles don't end when spending stops; they end when the growth stops, and shovel sellers get repriced first. This may have been the opening act.

The Bottom Line

The AI trade has a new stakeholder. On November 3, voters and utility customers joined the CEOs and CFOs, asking the same question: show me the return, and show me who pays for it.

That doesn't kill AI. It slows the pace of the buildout, and for its suppliers pace is everything. The SaaSpocalypse hit incumbents first; Semageddon hits suppliers first.

Investor takeaways:

Watch the second derivative. Capex growth drives chip multiples; watch Nvidia on Nov 17 and January's 2027 capex guides.
Own the bottlenecks, not the capacity. Pricing power over pure volume.
Spenders over suppliers, for now. A slower buildout eases the hyperscalers' FCF squeeze and helps software.
Map the politics locally. States and permits, not just Congress.
Respect the Fed. A December hike is live. Keep dry powder; no margin for this dip.
Keep the long view. Close the ledgers like a Diwali merchant and start Samvat 2083 clean.

Wishing you and your family a very Happy Diwali and a prosperous New Year. May the year bring more light to your portfolio and fewer diyas needed for the SOX!

/Shyam

Not investment advice; just my views. Do your own research.

Claude was used to do research and organize

Sources

Friday, August 07, 2026

The Reckoning: Tokenomics, SaaSocalypse and Odyssey to Adoption!

A lot has happened since my last update. Between summer distractions and a stubborn case of writer's block, I skipped a few posts I meant to share. But as the saying goes, better late than never! Here is the long-awaited update.

The tech ecosystem has officially entered its “reckoning” phase. If the last two years were defined by unconstrained optimism, massive capital deployments, and a collective fear of missing out, today's reality is dictated by unit economics, margin calls, and questions about return of investments. From corporate boardrooms to hedge fund desks, the blind faith in raw optimism is giving way to hard accountability.

Here is a breakdown of the structural shifts reshaping the current landscape.

1. From TokenMaxxing to Tokenomics

The Shift: Moving from metric-less AI consumption to rigorous optimization and ROI.

Earlier this year, tech culture birthed a bizarre internal performance metric: "TokenMaxxing." Spurred by remarks from tech leaders - that software engineers should be judged by how many hundreds of thousands of dollars in tokens they consume—enterprises began tracking raw AI consumption as a proxy for productivity. Meta employees famously built internal leaderboards ranking token usage, and companies gave digital badges to workers burning the most API bandwidth and tokens

The result was predictable: unconstrained agentic loops, massive prompt slop, and astronomical monthly token bills.

The industry has hit a wall, shifting violently from TokenMaxxing to strict Tokenomics.

Corporate Throttling: Companies like Microsoft, Amazon, Meta, and Uber have aggressively reined in internal AI budgets. Microsoft issued internal directives stating that maximizing raw token usage is no longer a corporate objective.
The ROI Crisis: Financial teams are utilizing specialized cost-intelligence platforms to audit where tokens are going. Spending millions on frontier models to have an agent loop infinitely in a code repository without shipping a feature is no longer tolerated.
The New Playbook: As stated by Uber's CTO, the goal has evolved into treating efficiency as an engineering problem rather than a budget restriction—leveraging advanced prompt caching, open-weight models, and smaller, specialized default models to do more with less. Army of startups are born to go after Tokenomics aspects of AI..

2. The End of the SaaSocalypse

The Shift: The stabilization and painful evolution of software-as-a-service models.

For eighteen months, the "SaaSocalypse"—the narrative that generative AI and agentic workflows would entirely obliterate seat-based SaaS software—terrorized public tech valuations. Critics argued that if an AI agent can do the work of ten people, companies would slash their Salesforce, ServiceNow, or Workday seat counts to zero, starving legacy software giants of revenue.

We are now witnessing the End of the SaaSocalypse, not because legacy SaaS won, but because the alternative proved too chaotic to deploy natively.

The Failure of DIY Agent Orchestration: Enterprises that tried to build their own agent layers from scratch using raw LLM APIs ran into massive deployment, security, and cost hurdles including unpredictability..
The Return of the Work Graph: Legacy SaaS companies successfully defended their territory by embedding AI directly into where the data already lives. It turns out enterprise buyers prefer paying a predictable premium to an established vendor over managing an unconstrained, multi-million dollar token budget for a custom-built bot.
The Value Shift: Seat-based pricing is steadily morphing into usage-based or outcome-based pricing, signaling that SaaS isn't dying; it is simply adapting to the cognitive era.

3. The "Situational Awareness" Blowup

The Shift: The spectacular downfall of the market's most levered AI macro trade.

Leopold Aschenbrenner’s Situational Awareness manifesto predicted a rapid, trillion-dollar industrial mobilization toward AGI, a narrative he parlayed into a massive hedge fund. However, operating with staggering leverage, the fund proved highly vulnerable when semiconductor and cloud computing stocks pulled back.

A chain reaction of margin calls forced the fund to rapidly liquidate its public tech holdings. While the fund preserved its private startup equity, this public blowup marked the exact moment the market stopped trading AI on pure narrative and began demanding actual cash flow. The Situational Awareness blowup echoes the 1998 collapse of Long-Term Capital Management (LTCM) as both firms relied on massive leverage to back high-conviction, concentrated theses that ultimately triggered severe margin calls when markets turned. LTCM had 25-to-1 leverage and was started by two Nobel Laureates and former Fed Vice-Chairman. But markets continued the momentum and eventually crashed in 2000-2001...if similar timelines hold, I am predicting at repeat of correction in 2028-29 (which would coincide with 100 years of famous 1929 crash)

4. Warshnomics: Monetary Policy Meets the Tech Reset

The Shift: The end of the "easy money" Fed era and its immediate impact on tech valuations.

As tech companies struggle with internal token economics, they face an equally harsh macro environment driven by Federal Reserve Chairman Kevin Warsh. The arrival of "Warshnomics" represents a fundamental restructuring of US monetary policy that could have impact on high-multiple tech landscape.

The Death of Forward Guidance: Under Chairman Warsh, the Fed has aggressively dismantled the predictable, slow-moving "forward guidance" era. Instead, the central bank operates with data-dependent agility, keeping markets on edge and eliminating the safety net tech investors used to rely on.
Uncompromising 2% Target: Warshnomics places a hawkish focus on a strict 2% inflation mandate and an aggressive reduction of the Fed's balance sheet.
The Valuation Squeeze: Cheap capital is no longer available exactly when trillions of $ are needed for AI infrastructure buildout. Markets are asking for return on investment and discipline in capex spends and rewarding companies like Microsoft and punishing companies like Google and SpaceX just based on their Capex spending plans. That means many of these companies would adopt "no capex increase" tune over next couple of quarters.

The Bottom Line

The era of easy money, narrative-driven leverage, and vanity metrics in tech is over. Whether you are an engineer optimizing prompt caches to protect corporate margin, a SaaS vendor rewriting contract terms, or a growth investor navigating the new realities of Warshnomics, the directive is simple: Show me the return on investment (Capex, Tokens consumed or margin loans)

History reminds us that groundbreaking technology cannot bypass the grueling 10-to-20-year timeline required for widespread enterprise and societal adoption. We witnessed this long, evolutionary crawl with the personal computer, the internet, the smartphone, and cloud computing—none of them transformed global infrastructure overnight. Like Odysseus enduring his brutal ten-year journey back to Ithaca, the tech ecosystem is realizing that the shortcut to a magical AI destination (including AGI) may take at least a decade...personally my yardstick is simple...Cure of Cancer expedited with help of AI..and I am really hoping that with trillions of $ investment going in, we can see that happening in next decade - preferably before 10th anniversary of ChatGPT!

Sunday, May 24, 2026

Rockets, Relics & Roaring Markets: The $4 Trillion Crossroads of 1927 and 1999

Happy (almost) Summer! After watching Kevin Warsh get sworn in at a White House ceremony two days ago, tracking three S-1 filings that could collectively hoover up more capital than every U.S. IPO since 2022 combined, and watching 26-year-old stock charts finally break to new highs — it felt like the right moment to ask the uncomfortable question out loud.

Are we at a party that ends gracefully, or one that ends with the furniture on fire?

The market is simultaneously flashing the neon signs of 1999 and the orchestral excess of 1927. Most commentators reach for the dot-com playbook. I think the original Roaring Twenties is the better map. Here's why...

Assembly Lines to AI Clusters

Ford's River Rouge complex was the largest industrial facility on earth in the 1920s — raw iron in one end, a Model T out the other. Steel, rubber, and oil became the picks-and-shovels of the age. GE and Westinghouse were electrifying factories and homes. The infrastructure buildout was the story, not just the product.

Today, Nvidia's GPU farms are the River Rouge of the AI era. Corning lays the fiber, Micron supplies the memory, the power grid strains under the load. AI data center capex is projected at nearly $1 trillion this year alone. Jensen Huang recently called it "the single largest infrastructure buildout in human history." That is exactly what people said about the electrical grid in 1925. They were right. The infrastructure cycle always pulls in the old industrial players — and then eventually outruns them.

Swampland to Starships

By 1925, Florida lots were changing hands multiple times a day. Buyers bid on land described only in pamphlets, using "binder contracts" — options on future options — to lever up without ever taking delivery. The boom peaked in 1926. The bust came before 1929. And nobody connected the two until it was too late.

SpaceX, OpenAI, and Anthropic are being valued at $900 billion to $2 trillion on businesses that are real, growing — and largely unprofitable at scale. The SpaceX S-1 hit May 20th: $1.75–2 trillion target valuation. OpenAI targets $1 trillion in September. Anthropic follows in October with over $1 trillion Combined: $3.75 trillion pricing in a five-month window — roughly the GDP of France.

The Florida binder contracts are today's IPO allocations. The pamphlets are pitch decks. The lots are orbital launch rights and foundation model APIs. The underlying assets are real. The arithmetic, at these prices, requires a great deal of imagination. Every era has an asset class where imagination outruns math — until it doesn't.

Ticker Tape to Twitter Feeds

The 1920s gave ordinary Americans real-time price data for the first time. Bucket shops let small investors speculate on margin without owning the underlying. By 1929, nine million Americans — in a population of 120 million — held stocks. Speculation had been fully democratized. That was both the triumph and the trap.

Today: zero-commission trading, zero day options, options on phones, Reddit-driven price action, and retail IPO allocations have extended the party to a new generation. The mechanics differ; the psychology is identical — easy access, leverage, and the primal fear of missing the decade's defining trade. Democratizing speculation is always celebrated right up until the margin calls arrive.

The Ghost Fleet & the Change of Guard

Woven through all three themes is an eerie subplot: the infrastructure names of the last technology boom are back at record highs. The companies that defined the dot-com era — Intel up nearly 200% in 2026, Cisco crossing century mark first time ever, Corning finally breaking its year-2000 peak, Micron at an $800 billion market cap — are all riding the AI infrastructure wave. Intel now trades at 108x forward earnings, richer than the most celebrated networking names were at their dot-com apex. Same ships, new cargo delivering the goods for the AI infrastructure

The Fed parallel is equally striking. In 1999, Alan Greenspan had spent three years warning of "irrational exuberance" — then watched the market double anyway. He finally raised rates six times between June 1999 and May 2000. The hikes were what ultimately punctured the bubble. Greenspan had the political independence to act. The market paid the price on his schedule.

Kevin Warsh just took the oath in the East Room two days ago, with the President watching. He inherits 3.8% inflation (three-year high, driven by the Iran war oil shock), real wages falling, and a political principal with a well-documented preference for rate cuts regardless of the data. The inversion from 1999 is the critical distinction: Greenspan had independence but waited too long. Warsh may lack the runway to use it at all. If he cuts while inflation sits at 3.8%, bond markets will reprice violently and every long-duration AI growth name will feel it. If he holds, the political pressure will be unlike anything the institution has faced in a generation. There are no clean exits from this inheritance.

1927 or 1999? Both. Simultaneously.

The three-layer reality:

The Real Thing. AI productivity gains are structural and compounding. The infrastructure buildout will run for years similar to internet infrastructure buildout because need for tokens would be exponential (When the dot-com bubble burst in 2000, global internet traffic averaged roughly 84 petabytes per month. Five years later, in 2005, monthly global traffic had grown significantly to about 2,426 petabytes (2.42 exabytes), representing roughly a 29-fold increase)

The Froth. Infrastructure names at 100x+ earnings, trillion-dollar IPOs priced on terminal-value imagination, retail speculation at historic participation rates. That's where 1999 lives.

The Structural Trap. A new Fed chair under political pressure, 3.8% inflation, a war that markets have priced as resolved when it isn't, and three capital-market whales surfacing simultaneously. That's where 1927 lives — and it's the layer most investors aren't pricing.

The 1927 lesson that the 1999 frame misses: the crash doesn't require fraud. It requires excess capital chasing a genuine thesis faster than that thesis can deliver returns. The 1929 bubble wasn't built on delusion. It was built on extrapolation.

What's the Play?

Mega-IPOs: entry price over first-day pop. SpaceX is the most defensible — real cash flow, near-monopoly in commercial launch. Size for the five-year hold. The first-day pop is for insiders. The five-year hold is for investors.
The Ghost Fleet: trade, don't marry. Infrastructure names at 100x earnings are real businesses priced for perfection. Know whether you're buying the thesis or the multiple.
Warsh Watch is the underpriced risk. The Fed chair transition is the single most important variable for the second half of 2026. Watch his first two moves carefully — they will telegraph whether he has the independence to act or the political constraints that make 1999's ending look orderly by comparison.
Cash is not trash. Three trillion-dollar IPOs competing for institutional capital creates dislocations in everything else. The patient investor who bought Amazon at $7 in late 2000 didn't need a time machine. Just dry powder and conviction.

Final Thought

We came into 2026 asking whether this was 1998 or 2000. The better map is a century older — assembly lines becoming AI clusters, swampland becoming starships, bucket shops becoming 24 hour trading on mobile apps. And a new Fed chair, like Greenspan before him, inheriting a market that has already decided what it wants to believe.

History doesn't repeat. But sometimes it rhymes so loudly you can hear it from the balcony.

The question for 2026–2027 isn't whether this ends. The question is whether you're positioned with enough real exposure to capture a genuine technological transformation — and enough intellectual honesty to know that the price you pay determines everything.

Enjoy the summer. It might be a memorable one.

/Shyam

Personal views and market observations, not investment advice. Do your own diligence.

Tuesday, April 14, 2026

The AI Siege: From SaaSpocalypse to BugArmageddon

Happy Spring! After navigating "Ides of March", it’s time to tackle the story that has every SaaS executive reaching for antacids and every cybersecurity CISO simultaneously terrified and quietly pumping their fist

This saga has everything: a mythology-named AI model too dangerous to release, a real-time startup mortality tracker called "Death by Clawd," and an unannounced emergency meeting at the U.S. Treasury

You genuinely cannot make this stuff up

. And we are not even halfway through April...

SaaSpocalypse: The First Shoe Drops

The trouble started well before Mythos. In early February, roughly $285 billion in software market cap evaporated in 48 hours

. Stalwarts like Salesforce, Adobe, and ServiceNow—names that had compounded investor wealth for a decade—suddenly dropped 25-35%

. A Jefferies trader coined the term that stuck: the SaaSpocalypse

The structural logic is elegant in its brutality. The SaaS golden model was built on one premise: charge per seat, watch headcount grow, and profit forever

. Then the AI agent arrived. The agent doesn't have a headcount; As Gartner analyst noted, if AI agents do the work of 100 seats independent of humans, tying licenses to humans simply doesn't make sense anymore

. The per-seat model is starting to look like a horse-and-buggy pricing strategy in the age of autonomous vehicles.

BugArmageddon: The Second (Bigger) Shoe

Just when investors thought the pain was priced in, Anthropic dropped Claude Mythos on April 9th and simultaneously announced they would not release it publicly

. It’s the equivalent of unveiling a car that does 400 mph and saying: "Impressive, right? It stays in the garage"

. and that was right decision to make sure that companies which are part of "Project Glasswing" (equivalent to "J.P. Morgan’s Library" of the AI era - more on this analogy later). It is a closed-door, high-stakes defensive perimeter established to prevent a total collapse of digital credibility.

The reason? Mythos can identify and exploit vulnerabilities at a scale that makes traditional security testing look like a candlelight audit in a hurricane

. In demos, it found a crash vulnerability in OpenBSD that had evaded hackers for 27 years

The market reaction was catastrophic

. Many software companies dropped 6-8% that day

. Meanwhile, Powell and Bessent quietly summoned Wall Street CEOs to the Treasury—no press release, just dread in a room full of people whose institutions run on software Mythos can dissect like a grad student with a library book. It reminded of similar meetings during depth of financial crisis of 2008 - but this time was not about liquidity crisis but potential of exposure to vulnerabilities!

The SaaSpocalypse was about AI replacing software workflows. BugArmageddon is about AI dismantling the security assumptions those workflows are built on

What’s the Play?

Markets are answering a healthy question with the precision of a sledgehammer. Here’s how I’m thinking about the fallout:

Trim the pure workflow SaaS: If your product's moat is "doing the task"—drag-and-drop UI or basic automation—you are exposed. Someone can screenshot your UI, hand it to Claude Code, and have a replica in days. The per-seat expansion story is structurally capped, and valuations haven't fully adjusted.
The cybersecurity contrarian setup: If Mythos-class models proliferate (and in a tech race for AI supremacy, they will), demand for elite security infrastructure doesn't collapse—it explodes. The same AI that finds 27-year-old bugs makes every enterprise desperate for zero-trust architecture and AI-assisted threat detection. This short-term pain may be setting up one of the decade's better long-term plays.
Don't abandon "system of record" SaaS: Salesforce, ServiceNow type SAAS companies are not just workflow; they are 10-20 years of "system of records". That doesn't get one-shotted overnight. These names would do ok as long as they evolve quickly with their own Agentic workflows based on system records to help their customers (and of course take care of BugArmageddon in their own software)

Final Thought

Mythos didn’t create the new "Y2K" type problem — it just turned the lights on and created AI speed urgency. We came into 2026 asking if this was 1998 or 2000. I have been thinking about historical analogies.

The dot-com comparison is the obvious one — and it’s partially right. We’re in a
period where transformational technology is genuinely real, but where the valuations had
priced in a frictionless utopia that reality doesn’t deliver on schedule. Some of 1999’s
casualties (Pets.com, Webvan) were genuinely worthless. Others — Amazon, Google — just
needed time.

There is one lesser but similar analogy: 1907 and the Bankers’ Panic. The panic wasn’t caused by bad fundamentals. It was caused by a credibility crisis — the realization that the financial system had grown complex and interconnected in ways that no one fully understood, and that a shock in one corner could propagate unpredictably. J.P. Morgan famously corralled the country’s major bankers into his library and essentially strong-armed a private sector bailout.

The Mythos moment feels similar. This isn’t fundamentally about bad software companies. It’s about the market suddenly confronting the fact that the software infrastructure of the global economy may have systemic vulnerabilities that we didn’t know existed — and that a single AI model can find them faster than the industry can patch them. That’s a credibility crisis, not an earnings crisis. The question isn’t whether software companies survive. Of course they will. The question is: at what multiple, with what margin structure...Markets are still trying to figure that out.

That's all for now...next week would be another interesting week...in meantime, I will enjoy my weekend in Mexico city exploring some history and excellent food!

/Shyam

Saturday, March 14, 2026

Statues, Seeds, and Straits: Navigating the Ides of March

March is always a month of transitions. We move from the scripted drama of Hollywood to the unscripted chaos of the basketball courts with NCAA March Madness and continued chaos in geopolitics due to surprise late Feb events. Here is my take on Oscar predictions and other events..

Statues: And the Oscar goes to....

Tomorrow night, the Dolby Theatre becomes the center of the universe. This year’s race is a fascinating battle between the "Record-Breaker" and the "Auteur's Epic." Ryan Coogler’s Sinners walked in with a historic 16 nominations, but Paul Thomas Anderson’s One Battle After Another has been cleaning up the precursors.

Here are my predictions for the top 10 categories:

Best Picture: One Battle After Another (Surprise: Sinners)
Best Director: Paul Thomas Anderson (One Battle After Another)
Best Actor: Michael B. Jordan (Sinners) - even though close second, Timothée Chalamet (Marty Supreme) may have to wait to win his first Oscar some more time!
Best Actress: Jessie Buckley (Hamnet)
Best Supporting Actor: Sean Penn (One Battle After Another) - this win should put Sean Penn in esteemed company of 2nd most Oscar winners (Meryl Streep, Jack Nicholson, Daniel Day Lewis, Ingrid Bergman and Walter Brennan)
Best Supporting Actress: Amy Madigan (Weapons)
Best Original Screenplay: Sinners (Ryan Coogler)
Best Adapted Screenplay: One Battle After Another (Paul Thomas Andersen)
Best Animated Feature: K-Pop: Demon Hunters
Best International Feature: Sentimental Value (Norway)

GTC 2026: The "Vera Rubin" Era and the Gigawatt Factory

Just as the Oscars wrap up, the "Woodstock of AI" begins. This week’s NVIDIA GTC in San Jose isn't just about faster chips; it’s about the architectural shift to Agentic AI and "Physical AI." While everyone is eyeing the Vera Rubin platform—which promises a 10x reduction in inference costs over Blackwell— watch out for the networking and energy play. Jensen is talking about "buildouts measured in gigawatts," and for those of us in the networking world, that means the AI Factory is the new unit of scale. Look for big news on the Spectrum-X photonics side and how BlueField-4 will handle the storage-heavy reasoning required for autonomous agents. If Blackwell was the engine, Rubin is the entire grid.

Seeds & Straits: The Geopolitical Full-Court Press

Selection Sunday is one day away, and while the Seeds (Duke, Florida, Michigan, and Arizona) are nearly settled on the top line, the real "Madness" is happening in the maritime chokepoints.

As we enter the Ides of March, the Strait of Hormuz has become the ultimate "bracket buster." Following the military escalations in late February, the effectively closed strait has sent oil traffic plunging by 70%. We are seeing roughly 20 million barrels a day—20% of global energy—sitting on the sidelines. Brent crude has already kissed $126 per barrel, and with the "Ides" historically being a time for settling debts, the global economy is footing a massive bill. One "unforced error" in these narrow waters can bust the entire global stability bracket. The geopolitical game is moving faster than the world and markets can keep up with it...

The markets are currently "shooting first and asking questions later." We’ve seen a massive rotation as the AI narrative shifts from pure excitement to "show me the productivity."

The AI Rotation: We are moving from the "infrastructure" phase (chips and servers) to the "implementation" phase (software and productivity). Watch for the laggards in the SaaS space to finally catch a bid if they can convince markets that they hold the key to "final mile" of enterprise productivity unleashed by AI and holds trove of "system of records" in enterprise (aka they control the "straits" to unleash the power of AI in enterprises)
The "Real Asset" Hedge: With geopolitical tensions refusing to simmer down, hard assets such as Gold and Energy remain the "anti-fragile" plays. I still believe "Stree-Dhan" (the wisdom of holding gold) will outperform the paper-pushers this quarter.
Volatility Rebound: The VIX has been suppressed for too long. Much like a 12-point underdog making a run in the final two minutes, expect a spike in volatility as we approach the next Fed meeting.

Final Thought: Whether you are predicting on a film, a team, or a stock, remember that March is the month where "momentum" is a fickle friend. Keep your positions tight and your eyes on the Ides.

Happy "Pi" Day!

/Shyam

Sunday, February 01, 2026

Stree-Dhan vs. Oracle of Omaha!

Happy February! After another brief hibernation from the blog—partly to digest the early year volatility and partly to observe the shifting sands of global liquidity—it’s time to look at some fascinating disconnects in the market.

Lately, I’ve been thinking about the "Unbeatable Asset Class." No, I’m not talking about the S&P 500 or Nvidia. I’m talking about a collective force that has quietly outperformed the "Oracle of Omaha" for over two decades.

1. The Golden Saree: Indian Women vs. Warren Buffett

If you look at the performance of Berkshire Hathaway (BRK-B) since the launch of the GLD ETF (the first gold ETF) in late 2004, you’ll find a startling reality. While Buffett is the gold standard of value investing, the "Gold Standard" itself—specifically in the hands of Indian households—has been a formidable rival.

Data shows that since the inception of the GLD ETF in November 2004, the total return on Gold has actually surpassed Berkshire Hathaway. In USD terms, Gold is up roughly 932% compared to Berkshire’s 744%. But for the Indian woman, the "Alpha" is even higher. Because she buys in Rupee (INR), she has benefited from a "Double Engine" of growth: the rising global price of gold and the steady 3–4% annual depreciation of the Rupee against the Dollar.

For decades, Buffett dismissed gold as a "non-productive" asset that just sits in a vault. But for millions of Indian families, Stree-Dhan has been the ultimate hedge, outperforming the smartest man in Omaha by simply buying gold on every festival and wedding in the family and holding it forever. My mom used to advise exactly this since I was young (which unfortunately I did not act on but fortunately as every Indian woman my wife did).

2. The "Warsh Effect"

Moving from real assets to macro policy, we have to talk about the Warsh Effect. With Kevin Warsh’s influence looming large over the fiscal intersection, markets are bracing for a "hawkishly pro-growth" stance.

The Warsh Effect is essentially the market’s anticipation of Rule-Based Monetary Policy.

Stocks: We are seeing a "Bad News is Bad News" regime. The market wants growth, but it’s terrified of a bond vigilante comeback and by appointing a credible candidate, Trump gave markets assurance that Fed would remain independent and balanced in its approach to inflation and employment
Cryptos: Liquidity came out of cryptos to chase hot rally in gold and silver. With Warsh effect and potential shrinking of Fed's balance sheet, it would be interesting how cryptos behave in new normal. Unfortunately extreme leveraging in cryptos and washout events like in Oct and maybe this weekend decide prices of cryptos more than macro factors.
Commodities: A tug-of-war is underway. A stronger dollar usually hurts commodities, but if the Warsh doctrine successfully re-industrializes the US, industrial metals like copper and silver should do ok. And as far as gold goes, it has almost become new "reserve currency" of the world.

3. Hot Memory and Storage: What's next?

Finally, the "Nuts and Bolts" of the AI revolution. 2024/2025 was about the "Brain" (GPUs). 2026 is the year of Memory.

Stocks like Micron (MU) and SK Hynix are hitting record highs because AI models are "memory-bound." You can’t have a super-intelligent AI if it has the short-term memory of a goldfish. HBM (High Bandwidth Memory) and Flash memory (for storage) are now the most valuable real estate in tech. No wonder analysts are predicting Sandisk to reach 4 digits followed by 1:10 split.

What’s next? We are approaching the "Cyclical Peak" anxiety. Usually, when everyone realizes memory is essential, the industry overbuilds and prices crash. However, the "AI needs" feels permanent. With Micron alone committing over $100 Billion in expanding capacity, eventually enough supply would come online to have usual cyclical patterns - maybe in 2028. As adjacent players, one should look at storage infrastructure players to understand how their business model evolves in this "hot memory" world

And the Grammy goes to...

I am no expert in western songs and very rarely listen to latest albums (I am into old classics and Hindi songs)....But that should not prevent me in enjoying a good show and best way to enjoy the show is to make predictions (even if I have no clue) and check how many were correct...

So here are my predictions with help from AI and various articles I read

Here is my take for the Big Four + Pop Album:

Album of the Year: Bad Bunny’s DeBÍ TiRAR MáS FOToS. With his Super Bowl halftime show looming next week, the momentum is undeniable. This would be historic since this would be first full non-English album to win.
Record of the Year: Kendrick Lamar & SZA - "Luther". Kendrick leads the field with 9 nominations, and this track has the "cultural resonance" the Academy loves.
Song of the Year: "Golden" (from KPop Demon Hunters). This one I know since I was the movie and it’s been 12 weeks at #1 for a reason.
Best New Artist: This is the most wide-open field in years. Olivia Dean has the "Grammy Soul," but Sombr has the algorithm. I’m betting on Olivia Dean for the win, but watch out for Leon Thomas
Best Pop Vocal Album: I’m betting on Lady Gaga’s MAYHEM. However, watch out for Sabrina Carpenter. Man's Best Friend is a masterclass in pop engineering, and she has momentum.

Bottom Line: Whether it’s the gold in an Indian locker or the HBM chips in a Blackwell server, the theme for 2026 is Tangibility. In a world of deepfakes and AI hallucinations, investors are soul-searching for things that actually exist—and can touch!

Stay invested, stay curious.

/Shyam