Weekly Research Briefing: A Most Important Handoff

To the index overlords who steer trillions of dollars in flows through the eligibility rules they set, adding a company like SpaceX is simply a matter of reflecting the market as it really is. Nasdaq Inc. changed its rules so SpaceX can join the Nasdaq 100 Index, a cohort of the largest non-financial companies listed on its exchange, in just 15 trading days, down from a three-month minimum. FTSE Russell adopted a similar approach, shortening the waiting time to five trading days, and S&P Dow Jones Indices is expected to decide soon following its own consultation on the subject, potentially opening the S&P 500 Index itself to welcome instant mega-caps like SpaceX…

Fast-track timelines vary by index, but if S&P follows its competitors, the resulting passive demand for SpaceX from funds tracking indexes would be nearly $20 billion, according to Bloomberg Intelligence analysts James Seyffart and Rob Du Boff’s estimates. With a $75 billion raise, that would be roughly a quarter of SpaceX’s offering.

SpaceX is planning an early release schedule for insiders whose shares are subject to so-called lock-up agreements preventing them from selling immediately, and if S&P’s proposal is adopted, at least 90% would be unlocked by the time SpaceX could enter the S&P 500, Seyffart and Du Boff wrote in a note. The shares sold by employees and early investors, likely giving them windfall profits, could be snapped up by index-tracking funds.

GOLDMAN DESK: “.. US IPO gross proceeds will total a record $225 billion in 2026. Our previous forecast was for $160 billion of IPO issuance. Our new estimates reflect a continued healthy macro .. and recent changes to the outlook for potential launches by some of the largest private companies.”

And then the S&P 500 impact is the biggest, just because it is the most-used index, with “about $11.8 trillion of passive equity funds tied to the S&P 500.” It is also float-weighted, but it is slow: SpaceX won’t be eligible for the S&P for at least six months, by which point a lot of SpaceX’s stock will probably no longer be locked up. If SpaceX’s float is $1 trillion in six months — if it has a $2 trillion market capitalization, and half of its stock is unrestricted — then it will be about 1.5% of the S&P 500, and S&P 500 funds might have to buy about $175 billion of its stock. So much more than the IPO raise, though much less than the float at that point.

Multiple studies, including Lynch and Mendenhall (1997) and Chen et al. (2004), have documented the tendency for stocks added to major indexes to experience significant price appreciation between the announcement date and the effective date of inclusion, with deletions declining correspondingly. Expanding on earlier work, Arnott, Brightman, Kalesnik, and Wu (2023) analyzed S&P 500 reconstitutions from October 1989 through June 2021 and found that index additions outperformed the broad market by an average of 499 bps during the grace period between announcement and implementation. They also showed that discretionary deletions lagged the market by 723 bps over the same window. These sizable performance gaps, as summarized in Exhibit 2, highlight the enduring market impact of rebalancing flows triggered by index changes.

SpaceX isn’t alone in its ambition to raise a whopping sum to help it win the AI race. OpenAI is preparing to file confidentially for an IPO in the coming weeks, people familiar with the matter have said. Bloomberg News reported that Anthropic PBC was considering its own public debut as soon as October of this year. And while they continue to scale up in the private sector, most of these AI-forward companies are landing eye-watering appraisals. OpenAI raised funds at an $852 billion valuation back in March. After Anthropic achieved a $380 billion mark earlier this year, it’s now in talks for a round at a more than $900 billion valuation. Both would likely exceed their private valuations in an IPO.

Companies are not supposed to show up on exchanges already the world’s most valuable businesses. New ventures raise money privately when they’re small and speculative, but established businesses have almost always turned to public markets when they needed sizable capital for growth. Everyone wins: Venture investors offload their shares to cash in on their early support, and new investors participate in potentially even bigger gains ahead. That’s the story of practically every public company trading today, including tech behemoths Microsoft Corp., Amazon.com Inc. and Nvidia Corp.

That model has been upended by private markets awash with capital and capable of feeding practically unlimited cash to the most sought-after businesses. The result is companies staying private longer and coming to market with ever higher valuations. Nvidia was valued at roughly $600 million when it went public in 1999. Five years later, Alphabet Inc. raised the stakes with a $23 billion valuation at IPO. Meta Platforms Inc. raised the bar again to $104 billion in 2012, which seemed like a fortune at the time. That’s loose change compared with this year’s blockbuster IPOs.

I don’t blame companies for sidestepping the rigors of public markets when private money is readily available, but it’s a disservice to public market investors. By debuting at valuations comparable to Big Tech, much of the value of SpaceX, OpenAI and Anthropic’s future growth will be captured by private investors at IPO, at the expense of public shareholders. And that’s best case. If expected growth doesn’t materialize, these IPOs will be a massive wealth transfer from ordinary investors to private equity — and a major blow to the appeal of public markets.

Nor should anyone really worry too much about the health of Nvidia — at least in terms of its survival. It has total debts of $8.5bn, cash of $10.6bn, and generated free cash flow of almost $100bn in the year to end-January, according to LSEG data. Its profits could plunge 90 per cent and it would still rank as one of the 100 most profitable companies in the world.

There’s a reason why five-year credit default swaps on Nvidia are trading at under 40 basis points, roughly the same price as it costs to insure against the US government defaulting (although that’s probably not the yardstick of unimpeachable creditworthiness that it once was).

Honeywell International Inc.-backed quantum computing company Quantinuum Inc. boosted the size of its initial public offering to as much as $1.46 billion, raising the number of shares offered and the price range.

The company is offering 26.5 million shares for $53 to $55 each, according to a filing Monday with the US Securities and Exchange Commission confirming an earlier Bloomberg News report. Quantinuum had previously offered about 21 million shares for $45 to $50 each.

At the top of the new range, the company would have a market value of $14.3 billion based on the outstanding shares listed in its filing.

The quantum computing firm’s IPO drew orders for a double-digit multiple of the number of shares available, people familiar with the matter said Friday. The company is on track to price after the market closes in New York on Wednesday and start trading the next day.

Quantinuum makes powerful quantum computers that are capable of solving complex tasks beyond the abilities of traditional processors that could make exponential leaps in computing.

The 19% gain in the S&P 500 over the last 9 weeks is the 16th biggest 9-week gain for the index since 1950. What tends to follow the biggest short-term advances? More upside and above-average returns. Strength begets strength.

@TheChartReport: The Russell 2000 $IWM is on track for its strongest start to a year since 1991.

We expect the beneficiaries of AI infrastructure investment will account for roughly half of S&P 500 EPS growth in 2026 and 2027. Semiconductors are the primary direct earnings beneficiaries of the AI investment boom, and consensus estimates show NVDA and MU together accounting for a third of S&P 500 EPS growth this year. A number of tech hardware, industrials, and utilities companies are also receiving large earnings boosts from the AI build out. Growing hyperscaler depreciation expense will partially offset the boost to S&P 500 earnings, with a larger impact in 2027 than in 2026.

“Each one of these [AI factories], at the one-gigawatt level, started at $20 to $30 billion. It’s now at $50 to $60 billion. And soon, it will be $80 to $100 billion per gigawatt. $100 billion into an AI factory. It must work the first time, and it must work right away. The cost of capital is incredible. The complexity is incredible.” - Nvidia CEO Jensen Huang

ASML CEO: Demand on AI is coming so strongly that we will be in a supply-limited market for quite a while; Sees global semiconductor market "tense" due to tight supply for the foreseeable future.

AMAT management said AI and “agentic” demand now give it “8 plus quarters” of visibility and support “continued strong growth” in WFE spend into 2027 and 2028.; says the better 2026 systems outlook is “not fundamentally a pull-in from 27 into 26”; instead, growth is being driven by a pipeline of greenfield projects that are now coming online.

An exchange traded fund investing in a narrow group of AI chipmakers has surged through $10bn in assets just 50 calendar days after its launch, in a further sign of investor excitement over the rapid build-out of huge data centres by tech hyperscalers.

The Roundhill Memory ETF, which has the stock market ticker DRAM, hit the figure on Friday after surging 87 per cent since launch on April 2 and after $8.6bn of inflows, in the quickest rise to a $10bn valuation for any ETF on record, according to data from Morningstar.

DRAM hit the milestone in less than one-tenth of the time taken by the iShares AI Innovation and Tech Active ETF, known as BAI, which reached $10bn in assets last month, 556 days after its launch — at the time, the quickest rise to that value for any thematic ETF.

“DRAM is a textbook case of an ETF issuer catching lightning in a bottle, pairing a perfect ticker symbol with impeccable timing and an investment thesis tailor-made for the current market,” said Nate Geraci, president of NovaDius Wealth Management…

“I would just caution folks. Three stocks make up most of the fund and they are all related. It’s not like they are different parts of the AI supply chain,” Rosenbluth said.

@RenMacLLC: Healthcare Equipment hasn't been this abandoned since the late 1980s. Our 64-year SERM model is in the 5th percentile bullish zone - a set-up that suggests it's time to start turning over rocks.

Exxon Mobil warned Thursday that oil inventories will fall to record low levels in coming weeks, forcing prices to spike and curbing demand.

“We’re approaching unheard of inventory levels,” said Exxon Senior Vice President Neil Chapman at a conference hosted by Bernstein in New York.

“I mean really, really low levels,” Chapman warned. “You can debate whether that’s going to hit, those really low levels, in two weeks or three weeks. Once you get to that point, then you’ll see price shoot up.”

The price of physical Brent oil cargoes will spike to $150 to $160 per barrel when inventories hit all-time lows in coming weeks, the executive said. “When the price gets to a certain level, demand destruction brings it back into balance,” he said…

Iran’s closure of the strait has cost the market more than a billion barrels so far, the largest oil supply disruption in history, according to the International Energy Agency. Oil stockpiles have mitigated the impact so far, but that “can’t last forever,” Chapman said…

Oil industry executives have warned for two months that the crude futures market is not reflecting the scale of the disruption triggered by the war in the Middle East.

“I don’t know, whether it’s two to three weeks or three to four weeks,” Chapman said. “What I’m really saying is, once you get to the minimum inventory levels and all-time low inventory levels, there’s only one way to go. That’s the situation.”

Berkshire Hathaway agreed on Sunday to buy home builder Taylor Morrison Home Corp. for $6.8 billion in cash…

Berkshire has long been involved in the housing industry, owning a large residential real-estate broker, and holding stakes in home builders in recent years.

Berkshire will pay $72.50 per share for the Scottsdale, Ariz.-based home developer, a 24% premium to Taylor Morrison’s closing stock price of $58.50 on Friday.

“This is an exciting transaction for Berkshire and reinforces our long-standing commitment to U.S. housing. Homeownership remains central to the American dream, and this investment expands our ability to serve that market,” Abel said in a statement.

Equity Residential (EQR) and AvalonBay Communities (AVB) have agreed to an all-stock merger to form a housing rental ‌company with an enterprise value of $69 billion, the firms said on Thursday, strengthening their presence in key U.S. markets.

The deal, which would create the largest publicly traded U.S. apartment real estate investment trust by market value, will also strengthen the capacity to invest in AI technology.

The companies said they have a 95% overlap in ⁠markets where they own rental properties, a concentration expected to improve margins by enabling neighborhood-based operations, centralized services and lower cost-to-serve.

The transaction is expected to generate $175 million in gross synergies by the end of 18 months after completion, driven by reduced corporate overhead costs and property management expenses, the companies said.

“If you are a mathematician,” one of the world’s leading mathematicians recently wrote, “you may want to make sure you are sitting down before reading further.”

And you’ll definitely need to sit down if you’re not a mathematician.

Because a famous math problem that stumped humans for the better part of a century has finally been toppled—by AI.

Not long ago, the most advanced AI models couldn’t do basic math. By last year, they were performing at gold-medal levels at the International Mathematical Olympiad. Now they are solving classic problems in combinatorial geometry using algebraic number theory. In no time at all, artificial intelligence has gone from stupid to frighteningly smart.

But even mathematicians were astonished when OpenAI announced that one of its models resolved a puzzle known as the unit distance problem without the help of any humans scribbling a bunch of equations on chalkboards.