A 59.8 MB file shipped by accident on March 31 exposed 512,000 lines of Anthropic’s Claude Code source. Within hours mirrors had spread across GitHub, security researchers were cataloguing 44 hidden feature flags, and two developers in San Juan had already started building. The leak dominated trade headlines for a daemon mode codenamed KAIROS and an internal feature that strips AI attribution from employee commits. The buried lede sat in a directory called /src/buddy.

What that directory contained was a complete pixel-pet system Anthropic had built but never turned on. By April 1, design engineer Agnel Nieves and his collaborator known as Peroni had a working public version live on the open web. The speed of that turnaround says something about how AI coding tools are being designed right now, and it is not about benchmarks.

What the Source Map Coughed Up

Version 2.1.88 of the @anthropic-ai/claude-code package on the npm registry landed in the early hours of March 31. Bundled with the binary was cli.js.map, a debugging artifact that referenced unobfuscated TypeScript hosted on Anthropic’s R2 cloud storage. Chaofan Shou, a security researcher, posted the find on X at roughly 8:23 UTC. The version was pulled about four hours later. By then the codebase was archived on multiple public repositories and the mirrors had collected tens of thousands of stars.

Anthropic called the incident a release packaging issue caused by human error rather than a security breach. The root cause was unglamorous: Claude Code is built with the Bun JavaScript runtime, which generates source maps by default. Adding *.map to .npmignore would have prevented the disclosure. Nobody did.

The exposure was wide. Roughly 513,000 lines of code across 1,906 files sat readable in plain text: the agent harness, prompt internals, model routing, and 44 feature flags gating unshipped functionality. Among them was KAIROS, a persistent background-agent mode referenced more than 150 times in the source, and a build flag called Undercover Mode that activates for Anthropic employees on public repositories. Undercover strips Co-Authored-By attribution from commits and forbids references to unreleased models.

Undercover Mode is dead-code-eliminated from public builds, so paying users never get it. The function it serves is straightforward: keep internal contributions to external open-source projects from carrying obvious AI fingerprints. A company built a feature to suppress that signal. The same release that exposed the feature exposed the feature’s existence.

The 18-Species Pet Hiding in src/buddy

Deeper in the leaked tree, alongside the daemon machinery, sat a complete companion-pet implementation. The Buddy system is a gacha mechanic in pixel art. Every Claude Code user is supposed to receive a unique creature generated deterministically from their account ID, summoned into the terminal with the /buddy command and capable of popping speech bubbles while a session runs.

The mechanics inside the leaked source were fully built:

• 18 species covering Duck, Goose, Jelly, Cat, Dragon, Octopus, Owl, Penguin, Turtle, Snail, Ghost, Axolotl, Capybara, Cactus, Robot, Rabbit, Mushroom, and Chonky Cat
• Five rarity tiers, weighted Common 60%, Uncommon 25%, Rare 10%, Epic 4%, Legendary 1%
• An independent 1% chance for a Shiny variant with an alternate palette
• Five stats on a 0 to 100 scale: Debugging, Patience, Chaos, Wisdom, Snark, with one peak and one dump stat per pet
• A two-layer architecture splitting deterministic visual traits (the "skeleton") from a one-time name and personality generated by the Claude model itself (the "soul")

None of this was a stub. The pet could observe a coding session, float hearts when petted, and persist across runs. It was sitting behind a feature flag waiting for an April 1 release window. The leak handed the design to the open internet five days early.

Two Builders, One Day, Shipped Live

By April 1, a public companion generator at claudebuddy.me was live. The app takes any name or Claude Code user ID, hashes it, and renders a pixel companion onto an HTML5 canvas with retro CRT animations and pop sounds. There is no login, no database, no server-side state at all.

The builders are Agnel Nieves, a design engineer with 15 years of product work, and his collaborator known on X as @peronif5. They ship under the name Basement, a small product studio with a hackathon cadence. In a thread on X, Nieves described the response to the leak in five words: "Shipped it."

Their build mirrors the leaked mechanic but reimplements every asset and every value. The species count is 12 rather than 18, hand-drawn from scratch in pixel art and given original names like Blobbit and Nebulynx. The four-stat block is renamed to Vibe, Chaos, Focus, Luck. The shiny rate is bumped to 5%. A one-command terminal install drops the buddy into the Claude Code statusline. The whole project is open source on GitHub.

The aesthetic decisions matter as much as the mechanics. The CRT animation, the pop sounds, the QR code generator for sharing your buddy: these are choices that read as fan tribute rather than clone. Anthropic has not commented on the public app. There is no obvious reason it would object.

Inside the Generation Math

The interesting part is how little code it takes to make a deterministic gacha feel like ownership. The Basement implementation runs entirely in the browser, calls no APIs, and resolves to the same buddy for the same name every time.

The Hash Function

The input string is lowercased, trimmed, and salted with a fixed prefix to reduce trivial collisions. The salted string is fed through DJB2, the classic Daniel J. Bernstein non-cryptographic hash that converts an arbitrary text input to a 32-bit integer in a single pass. DJB2 is not cryptographically secure, but cryptographic security is not the requirement here. The requirement is consistency: the same input must always yield the same integer.

The PRNG Roll

That integer seeds Mulberry32, a tiny pseudo-random number generator (PRNG, a function that produces a reproducible stream of numbers from a single seed). Mulberry32 fits in about ten lines of JavaScript and is fast enough to run in a single animation frame. Each subsequent random draw advances the generator state, so the species roll, the shiny check, and the four stat rolls all chain off the original hash. Change the input by one character and every roll changes.

The Rarity Weights

Rarity tiers are weighted with cumulative probability bands. The PRNG returns a float between 0 and 1; if the float falls below 0.6 the pet is Common, below 0.85 Uncommon, and so on up to a 0.97 threshold for Legendary. The shiny roll is independent and triggers 5% of the time in the Basement build. Stats are clamped between 1 and 99 with a rarity-keyed floor that ensures Legendary pets cannot roll worse than mid-range Commons.

The whole pipeline closes in under a kilobyte of logic. Everything else in the project is sprite work, sound design, and the CRT shader.

Basement’s Build vs Anthropic’s Leaked Original

The community version and the unshipped original are not identical. They are recognizably the same idea executed against different constraints: one inside a CLI binary, one inside a browser tab.

The mechanical resemblance is close enough that the Basement build reads as a respectful reimplementation rather than a copy. The personality layer is the most telling difference: Anthropic planned to call the model for the soul, which would have meant a per-pet API cost and a per-pet inference latency. Basement, with no model access budget, chose a curated text pool. The result feels nearly identical at the user level, costs nothing to run, and explains why a two-person team could ship in a day.

Why Whimsy Is the New Moat for Dev Tools

The Buddy story is a marker in a broader shift. GitHub Copilot positions itself as "your AI pair programmer." Cursor’s pitch is the AI code editor. Anthropic, sitting on the highest-scoring coding agent on most public benchmarks, also decided to build a virtual pet. OpenAI’s Codex CLI shipped its own pet system the same week, which is either coincidence or evidence of how thoroughly the major labs have read the same product memos.

Developer tools don’t have to be purely utilitarian. The best ones have personality. A pixel pet that lives next to your cursor won’t make you a better programmer, but it might make the work feel a little less solitary.

That framing came from Nieves on X. It tracks with how the category has evolved. Once base capability across coding agents converged inside a narrow benchmark band, the next axis of competition became attachment. A user who has named a pet, watched it level, and gotten a Shiny is a user who opens the terminal slightly faster the next morning. The retention math on whimsy is real even if it never shows up in a sales deck.

The Cost Side

Whimsy is also cheap to build relative to model improvement. The Anthropic source files for Buddy are a small fraction of the codebase. The personality generation rides on infrastructure that already exists. A pixel art commission and a few hundred lines of deterministic glue produces something users will photograph and post. The same dollars spent on a 0.3 point improvement on the SWE-bench Verified benchmark do not generate the same posts.

The Risk Side

The downside is that whimsy ages fast and reads as gimmicky when the underlying tool is shaky. Tamagotchi mechanics have a half-life. The labs that win the personality layer will be the ones that update it without making last quarter’s pet look obsolete. Anthropic has built a model where the soul layer is generative, which is the closest thing to a hedge against that aging problem. Basement’s pool is fixed, which is fine for a fan project and a problem for a product.

The leak that mattered was supposed to be KAIROS, the always-on agent that suggests where Anthropic’s product is going. The leak that will be quoted is the one that revealed an AI lab spending engineering time on a pixel cat. The first one shows the roadmap. The second one shows the culture. For the people choosing which coding agent to live inside ten hours a day, the culture is starting to count more.

Buddy was scheduled to ship on April 1. It shipped on April 1 anyway. Just not from Anthropic.

Anthropic is in talks to raise at least $30 billion at a pre-money valuation north of $900 billion, a price that would push the Claude maker above OpenAI’s $852 billion mark from March. Bloomberg first reported the figure on April 29, and the Financial Times has since named Dragoneer, Greenoaks, Sequoia Capital and Altimeter Capital as lead investors, each writing checks of $2 billion or more.

The number is the headline. The underlying question is what these investors are actually paying for, because the company they are buying at $900 billion looks very little like the company OpenAI sold to its own backers eight weeks earlier.

Two Numbers, Two Different Companies

OpenAI’s close on March 31 was a consumer story. OpenAI’s official funding announcement framed the round around scale: $122 billion in committed capital at a post-money valuation of $852 billion, paired with growth figures that lean heavily on ChatGPT’s reach. The company says enterprise is closing in, making up more than 40% of revenue and on track to reach parity with consumer by the end of 2026, but the brand and the cash flow still come mostly from a chatbot used by hundreds of millions of people every week.

Anthropic’s valuation is built on the opposite arithmetic. Claude holds approximately 3.5% of the global generative AI chatbot market compared to ChatGPT’s roughly 60%, but in enterprise AI Claude holds an estimated 29% market share, and by mid-2025 Anthropic’s enterprise revenue had surpassed OpenAI’s. The two companies are competing in the same industry and trading at the same order of magnitude, but they are not selling the same thing.

The cleanest way to read the gap is by where the revenue concentrates:

• $30 billion annualized run rate at Anthropic, per the company’s February disclosure, with one source telling TechCrunch the real figure is closer to $40 billion.
• ~$24 billion implied annualized rate for OpenAI at $2 billion a month.
• $2.5 billion of Anthropic’s run rate sits inside Claude Code alone, a single command-line product launched in May 2025.
• 900 million weekly active users for ChatGPT against tens of millions for Claude on the consumer side.

The Coding Lane Holding the Valuation Up

If you strip Anthropic down to one engine driving the $900 billion math, it is Claude Code. Claude Code was made available to the general public in May 2025, and its run-rate revenue has grown to over $2.5 billion, more than doubling since the beginning of 2026. Business subscriptions to the product have quadrupled since the start of the year, and enterprise use has grown to represent over half of all Claude Code revenue.

The product also wins where word of mouth lives. JetBrains’ January wave of its AI Pulse survey, which polled more than 10,000 professional developers worldwide, found that 57% of developers had heard of Claude Code in January 2026, up from 31% in spring 2025, and 18% currently use it at work, a sixfold rise from roughly 3% the previous spring. Its CSAT (customer satisfaction) sits at 91%, and its NPS (net promoter score, a recommendation index on a scale from minus 100 to plus 100) is 54, both category-leading numbers.

Mapped against its rivals, the share picture looks like this:

Sources: JetBrains AI Pulse (January 2026), Microsoft FY26 disclosures, vendor figures. The Pragmatic Engineer’s February survey of roughly 900 senior engineers tells a similar story: when professional developers pick the tool they like most, the gap is not close.

That concentration matters because coding is not a niche inside enterprise AI. It is the category. Menlo Ventures data shows coding now accounts for 51% of all generative AI enterprise usage, by a wide margin the highest-value use case in the market, and within that segment Anthropic holds 42 to 54% market share. A $900 billion price tag for Anthropic is, in large part, a price tag for owning that lane.

Where the $30 Billion Comes From This Time

The investor list on this round is what makes it unusual. Anthropic’s previous $30 billion Series G in February closed at a $380 billion post-money valuation, led by GIC and Coatue, with co-leads including D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ and MGX. The current round, if it closes as reported, brings in a different anchor group.

Per the Financial Times reporting underpinning the $900 billion figure, the lead investors stepping in are:

• Dragoneer Investment Group, returning from the February round as a co-lead.
• Greenoaks Capital Partners, the San Francisco crossover firm.
• Sequoia Capital, one of Anthropic’s earliest backers.
• Altimeter Capital, the public-and-private hybrid fund run by Brad Gerstner.

Each of those firms is in for $2 billion or more, according to the FT. Amazon and Google, both of whom anchored the February round, are not expected to participate this time. Some early backers, particularly those who invested in 2024 or earlier, are skipping this round and are instead waiting to potentially cash out during Anthropic’s anticipated IPO later this year, as the company raises what is likely to be its last private round before going public to fund its massive computing needs. The capital push comes as the San Francisco-based company explores a potential initial public offering as early as October.

Compute Is the Other Half of the Story

If Claude Code is the revenue engine, GPU capacity is the constraint. Anthropic spent the past nine months locking in compute deals on a scale that would have been unthinkable for any private company outside the AI lab cohort. The most striking of them landed on May 6, when the company agreed to take over an entire xAI-built supercomputer.

The SpaceXAI compute partnership announcement spells out what Anthropic is getting. Colossus 1 features over 220,000 NVIDIA GPUs, including dense deployments of H100, H200, and next-generation GB200 accelerators, delivering extreme parallel performance for large language models, multimodal systems, scientific simulations, and generative AI at frontier scale. The agreement gives Anthropic more than 300 megawatts of new capacity and access to all of those GPUs within the month.

That deal sits inside a much larger build-out:

• Amazon: up to $25 billion in investment with 5 gigawatts of compute capacity for training and deploying Claude models.
• Google + Broadcom: a separate 5-gigawatt agreement starting next year, plus a Google commitment of up to $40 billion tied to performance milestones.
• Microsoft + Nvidia: a November 2025 partnership in which Anthropic agreed to buy $30 billion in Azure compute capacity.
• SpaceXAI Colossus 1: the entire Memphis cluster, 220,000+ GPUs and 300 megawatts, transferred from xAI’s own Grok training to Anthropic within May.

The compute story also explains why Elon Musk, who has spent two years calling Anthropic “woke” and “misanthropic,” agreed to lease his largest GPU cluster to a direct rival. Musk wrote on X that he was ok leasing Colossus 1 to Anthropic, as SpaceXAI had already moved training to Colossus 2. The economics of AI infrastructure now bend toward the highest-bidder lessee, even when the lessee builds the model your own model is supposed to crush.

OpenAI Owns the Consumer Lane Anthropic Doesn’t

The contrarian read on $900 billion is that it does not yet buy what OpenAI’s $852 billion buys. Anthropic has under 8% of consumer AI traffic, against ChatGPT’s share that, while down from roughly 80% a year ago, still sits north of 50%. A consumer brand at ChatGPT’s scale generates a flywheel that an enterprise-first business cannot replicate just by adding GPU capacity.

The broad consumer reach of ChatGPT creates a powerful distribution channel into the workplace, where demand is rapidly shifting from basic model access to intelligent systems that reshape how businesses operate.

That is OpenAI’s own language from its March funding post, and it captures the part of the franchise Anthropic does not have. The consumer install base feeds enterprise upsell. OpenAI says its business side now makes up 40% of revenue (up from around 30% last year) and is on track to reach parity with consumer by the end of 2026. The lanes that Anthropic and OpenAI are pricing at $900 billion and $852 billion respectively are not symmetric; one feeds into the other.

The Anthropic counter is that the enterprise channel is increasingly the more valuable end of the funnel anyway. By the company’s own count on its Series G announcement page, two years ago a dozen customers spent over $1 million with Anthropic on an annualized basis; today that number exceeds 500, and eight of the Fortune 10 are now Claude customers. The number of customers spending over $100,000 annually on Claude has grown 7x in the past year.

The bull case for the $900 billion price tag is that those large-account economics scale faster than chatbot ad revenue. The bear case is that owning the developer lane in 2026 does not yet make you a consumer-AI platform in the way the headlines suggest, and the next OpenAI model release could narrow the coding gap in a single quarter.

The IPO Window That Matters

The financing round is, at this point, a pricing event for an IPO that has not been filed yet. TechCrunch sources have described the current raise as likely Anthropic’s last private round before going public, and the company is reportedly targeting a listing as early as October. OpenAI is sitting on a similar fork, with most analysts now expecting a filing window late this year or early next.

The mechanics inside the next six months will tell more about which valuation holds than any private-market term sheet. Raising significant funds now is viewed as essential for securing the massive computing power required to sustain Anthropic’s growth trajectory, but the same compute commitments that justify a $900 billion price tag also harden the burn. Public-market investors will see the bills the private ones got to look past.

If Claude Code’s $2.5 billion run rate compounds through the back half of the year, and if enterprise revenue keeps doubling, $900 billion will look conservative by the time the prospectus lands. If the lane consolidates and Cursor and OpenAI’s Codex close the satisfaction gap, the IPO will price at a discount to this round, and the investors writing the $2 billion checks today will get to find out what an enterprise-AI valuation looks like when retail does the marking.

Malaysia’s initial public offering market has cleared roughly US$1.2 billion in the first four months of 2026, already within touching distance of the US$1.4 billion total for all of last year, according to Bloomberg-compiled data. A May 20 listing of integrated-circuit designer SkyeChip Bhd and a fourth-quarter real estate investment trust (REIT) carve-out from IOI Properties Group Bhd would lift full-year proceeds to roughly US$1.8 billion, the most Bursa Malaysia has cleared in 13 years.

Capital that would normally route through Bangkok or Manila is concentrating in Kuala Lumpur. Bursa is, almost by default, becoming Southeast Asia’s listed proxy for the artificial intelligence (AI) memory cycle, with state-linked funds and provident schemes absorbing institutional tranches at premium subscription multiples.

The Numbers That Frame the 13-Year Mark

The four-month tally already eclipses three of the last four full years on Bursa. All of 2025 closed at US$1.4 billion. 2024 sat below US$1 billion on a softer pipeline. 2022 produced less than US$700 million. The 2013 benchmark, by comparison, was lifted by the Westports Holdings listing and a stack of state-linked privatisations.

Sunway Healthcare Holdings Bhd carried most of the early-2026 proceeds. The hospital group priced its initial offering at RM1.45, raised RM2.86 billion on March 18, and saw its shares jump more than 38% on debut. That single deal accounts for more than half of the year-to-date raise.

The forward calendar is what pushes the math toward the US$1.8 billion mark. SkyeChip will price its 400-million-share base at 88 sen, raising RM352 million, while the IOI REIT is targeting RM1.98 billion in offered units at an indicative 90 sen each. Two further names, Creador-backed pharmacy chain Big Caring Group and convenience operator KK Mart Retail Bhd, have filed draft prospectuses.

Why Capital Is Crowding Into Kuala Lumpur

Two regional comparisons explain the rotation. Thailand’s political turbulence, household debt, and tighter fundraising rules have weighed on the SET; growth there is forecast at 1.9% for 2026 by AMRO’s Thailand macro surveillance work. The Philippine economy slipped to the slowest in major ASEAN, with Q1 2026 GDP growth coming in below Vietnam, Indonesia, Lao PDR and Malaysia.

Malaysia recorded 5.3% Q1 growth. The ringgit has been steady. The political backdrop has remained intact since the unity government formed in late 2022. For a regional fund manager building an ASEAN book this year, that is the lowest-friction venue available, and Deloitte’s Southeast Asia IPO rebound assessment shows the regional volume increasingly concentrating in Malaysia, Indonesia and Vietnam.

Specific Malaysian policy moves are also showing up in earnings. The National Energy Transition Roadmap and the National Semiconductor Strategy have moved from announcement into delivery, and listed companies in the affected sectors are reporting margin uplift to their books for the first time.

What that adds up to, on the ground:

• A liquid local pension and provident-fund bid that anchors institutional tranches at price
• A retail base that has been waiting since the post-COVID listing freeze and is currently subscribing the chip tranche at 95 times
• Lower competition from Thailand and the Philippines for the Southeast Asia AI-exposure trade

Inside SkyeChip’s Cornerstone Book

The cornerstone book counted 22 investors who took 155 million shares, equivalent to 58.6% of the institutional offering and 8.6% of the enlarged share capital. Named participants include Khazanah Nasional Bhd (through Pantai Feringgi Ventures), the Employees Provident Fund Board, Lembaga Tabung Haji, Lembaga Tabung Angkatan Tentera, Great Eastern Life Assurance, and JPMorgan Asset Management’s Singapore arm. Those 155 million shares carry a six-month moratorium on disposal under Bursa’s listing rules, which keeps the majority of the institutional tranche off-screen until mid-November. The retail tranche tracked the same way. The public-offer book was oversubscribed by 95.03 times, the heaviest retail demand on a Bursa Main Market issue since PetroChemical (M) Bhd’s 2010 listing, which cleared 73 times. Oppstar Bhd, the 2023 ACE Market chip-design comparable, cleared 77.

I expect the market to continue to be quite vibrant to be able to raise sizable amounts in the near term.

That came from Raymond Chooi, regional head of equity capital markets at Maybank Investment Bank, in remarks reported by Bloomberg this month. Maybank is also a joint advisor on the IOI Properties REIT, which makes the comment as much positioning as forecast. The RM352 million raise is heavily weighted to research: RM211.5 million, or 60%, lands in R&D, with the rest split across facilities expansion, licensing, tools and working capital. The company designs silicon intellectual property for high-performance computing and counts an HBM3E (high-bandwidth memory generation 3 extended, the specification used in current NVIDIA AI equity stacks and the wider accelerator order book) interface among its commercial product lines.

The Q4 Swing Factor From IOI Properties

IOI Properties Group Bhd announced in April it would carve out RM7.58 billion of retail and hospitality property, anchored by IOI City Mall and W Kuala Lumpur, into a real estate investment trust. The structure offers up to 2.2 billion units to the public at an indicative 90 sen each, raising as much as RM1.98 billion, per the IOI Properties REIT carve-out announcement. IOI Properties retains 60% post-listing. Maybank Investment Bank and AmInvestment Bank are joint advisors; DBS Bank acts as global coordinator.

Net gearing at the parent is forecast to drop to roughly 69%, from 89.6% at end-2025, on a combination of REIT proceeds and recent land disposals. That makes the carve-out as much a balance-sheet event for IOI Properties as a capital-raising one for unit holders.

For Bursa’s full-year tally, the arithmetic is direct. Take the year-to-date US$1.2 billion. Add the chip-design IPO’s roughly US$83 million. Layer the REIT’s roughly US$465 million at the top end of the indicative range, and the year clears around US$1.75 billion without any further pipeline.

Big Caring Group, the pharmacy chain targeting a valuation of up to RM20 billion, and the KK Mart Retail convenience-store carve-out, would each push the headline past US$1.8 billion if either prices before December.

The pacing matters. A late-November or early-December REIT window puts the deal into year-end tax-loss flows and the closing of institutional books. The earlier the REIT moves, the more cushion the year-end number carries against any AI-cycle wobble at the chip name.

What the Oppstar Parallel Tells Listing-Day Bidders

Comparable hunting takes you to March 2023, when Oppstar Bhd opened its ACE Market debut at 63 sen and closed first session at RM2.43, a 286% premium. That remains the largest listing-day move from a Malaysian chip-design name, in a Penang cluster that MIDA’s chip-startup tracking work has been mapping since 2022. Oppstar’s retail tranche cleared 77 times oversubscribed at the time, just short of the current chip listing’s 95 times.

Three years on, Oppstar trades roughly 50% below that listing-day close. The pattern from the 2023 cohort, including Cape EMS Bhd, suggests the day-one pop is real but is also when the institutional cornerstone tranche begins its earliest secondary-market exit window.

Two reads follow for SkyeChip bidders. The first is that listing-day demand on a Malaysian chip-design name has historically been hot enough to break the indicative range; bookbuilds above 70 times oversubscription have all paid on day one in this cohort. The pattern parallels the pre-IPO demand pulse Cathie Wood has flagged on SpaceX in the US private-listing calendar. The second is that price discovery comes later, once the cornerstone moratorium starts unwinding and the first quarterly earnings update is on the tape.

The cornerstone moratorium clears around November 20, six months after the listing. The first quarterly earnings update will land before then. Those two events, not the opening bell, are when the listed chip-design name becomes a real price-discovery story for buyers who arrive at 88 sen.

Memory Cycle, Rate Path and FX Will Decide the Final Tally

Three conditions can move the post-2013 claim from in-reach to missed.

1. The AI memory cycle turns. The chip-design issuer’s R&D allocation rides on HBM3E and successor specifications staying in NVIDIA, AMD and hyperscaler order books. A pullback in Samsung or SK hynix HBM3E capacity utilisation, or a price reset in DRAM (dynamic random-access memory), would compress forward earnings before the November cornerstone moratorium clears.
2. The REIT prices into a tougher window. Bursa’s REIT segment has been priced off a Bank Negara Malaysia overnight policy rate of 3.00% for most of the year. A surprise hike, or a sharper US Federal Reserve path, would pressure unit pricing on the IOI Properties carve-out and risk the deal being trimmed or pushed into 2027, the way Airtel Africa’s listing has slipped its earlier timetable.
3. FX translation breaks the dollar headline. The 2026 league table is denominated in dollars in the Bloomberg dataset everyone is citing. A weaker ringgit shrinks the dollar headline even when the ringgit raise lands on plan; a stronger ringgit does the reverse.

The cornerstone moratorium expires in mid-November. The IOI REIT is targeted to price before December. The first quarterly update from the chip issuer lands in late August. The order in which those three events arrive will determine whether Bursa closes 2026 against its post-2013 high, or with a story about how the AI memory cycle outran the listing pipeline.

If memory pricing softens before November, the listed chip name reprices first, the REIT books a discount, and the year ends a few hundred million dollars short of the May headline. If the cornerstone unwind lands into a still-hot HBM order book and the REIT prices into a friendly retail market, the post-2013 line clears with room to spare, and the back-half pipeline of Big Caring and KK Mart starts being priced into a more confident 2027 window than anyone is currently underwriting.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. Securities markets carry risk; valuations, allocations and listing-day outcomes can move sharply in either direction. Consult a licensed financial professional before acting on any of the information above. Figures are accurate as of publication on May 18, 2026.