CMR is the leading provider of funding and management support for small to medium-sized businesses and entrepreneurs 

Authored by Ed Dowd: Beyond the Narrative via Substack,

The Disability Signal Across the Atlantic

On August 11 I wrote US Disabilities Hit an All-Time High of 37 Million In July: UP 23% Since Feb 2021. The BLS Current Population Survey printed 37,029,000 Americans 16 and over reporting a disability. That is seven million more people since February 2021, a 3-to-4 sigma break from the pre-2020 plateau that has not mean-reverted.

The inflection is February 2021 with the Covid vaccine rollout not in 2020 when Covid was at its most virulent strain. Alternative explanations fail the timing and the magnitude test. The UK PIP system corroborates that US survey signal with something the American series cannot give you: medically assessed new claims broken down by body system and underlying cause.

If you want to know whether that US survey signal is real, stop arguing about survey design and look at an administrative system that actually diagnoses people.

Go back to the UK Disabilities (PIP) Project we published at Phinance Technologies. The page is still up...use it. It was built so researchers, doctors, and ordinary citizens could see the same thing we saw in 2023. You can look at total body system new claims or by underlying cause new claims (best viewed on desktop) within a body system. You can pick absolute new claims, excess new claims, percent excess new claims and excess new claims z score. We also break it out by monthly and yearly data. Finally you can sort it by age group as well. Play with data and be horrified like we were in 2023. Interactive charts...all done by Phinance Technologies for free.

PIP is not the protagonist from Charles Dickens's novel Great Expectations. It is the UK's main working-age disability benefit officially known as the Personal Independence Pension program by the UK Department of Work. Claims are medically assessed. Decisions, "clearances," as noted above are coded by body system and then by underlying cause. The positive award rate has been stable at around 40 percent, so you are not looking at a sudden collapse in standards. You are looking at more people presenting with more illness...new claims, not just the stock of existing claimants, sorted monthly or yearly and by age band versus a 2016-2019 trend.

That is the advantage over the US survey. The BLS series tells you that disability exploded after early 2021 with the vaccine rollout and not in 2020 with the virus. PIP tells you where in the body it exploded, and it lets you watch the timing against the vaccine rollout curve on the same chart.

Two Body Systems Make the Point

Hematological (blood) disorders went off the rails. New excess clearances rose about 217 percent in 2021 and then 522 percent in 2022 while 2023 declined but still at an absurd 374 percent above trend. Over 300 percent above trend two years running. When looking at new claims on a monthly basis Hematology jumped early and hard almost coincident with the first doses. That is a regime change in medically assessed claims. I have posted those charts more than once on X in 2023. Something broke and the administrative system recorded it at a scale that alone should have produced a public-health investigation...it did not.

Musculoskeletal claims tell a different, equally inconvenient story. Monthly clearances sat near a "normal" 10,000 through 2020 and early 2021. Then, around September 2021, they jumped and stayed elevated above 18,000 a month. The rise did not arrive with the first lockdowns or the first COVID wave. It arrived after the mass rollout and into the booster period. Inflammation, joint and soft-tissue disease, the conditions that take people out of work and onto daily-living and mobility awards. The timing is not subtle.

Those two systems are not the whole file. Cardiovascular and neurological claims rose. Breast-cancer clearances showed large excesses in 2022 and 2023 with high z-scores. Different latencies, same calendar: the break is 2021, not 2020. That is why the UK file is useful. It is not one blob called "disability." It is a set of body systems with different clocks, all accelerating after the intervention that was supposed to end the emergency.

People will say PIP is being gamed by fraud. Look at the body-system split before you buy that. A fraud wave does not preferentially light up hematology in early 2021 and musculoskeletal in late 2021 while neurological claims print 20 plus-sigma years. An awareness campaign does not move breast-cancer clearances. Two independent disability systems, two countries, both detecting the same inflection period in 2021.

The fiscal piece is already visible in Britain. Claimant counts have roughly doubled since 2019. Psychiatric disorders are now the largest single category. Spending is on a path that forces politicians to talk about "sustainability" and tighter points tests instead of asking why so many working-age bodies failed in the same window. The US version of that conversation is coming. A permanently larger disabled share of the 16-plus US population is lower labor force participation, higher absence, higher insurance cost, and more pressure on SSDI and Medicaid. You can ignore a chart. You cannot ignore the payroll.

Bottom Line

I am not a clinician. We said that on the site in 2023 and asked doctors to explain the findings...mostly crickets. The charts are still there. Pick a body system. Pick an age band. Look at the cumulative dose curve. Watch across the many body systems how in 2020 they stay close to trend and then in 2021 they leave the trend. That is the instruction. The US disability series tells you that the population got sicker after February 2021 in the US. The UK PIP file tells you which systems broke and when.

Together they are the corroborating signal no one in official public health circles wants you to see.

The great cover up continues into 2026 and the damage is slowly compounding.

Brazil's presidential race is entering the final stretch ahead of Sunday's first-round vote. 

HSBC strategists led by Nicole Inui wrote in a note that the latest polls show socialist President Luiz Inácio Lula da Silva slightly ahead of right-wing Senator Flávio Bolsonaro, though the race remains neck and neck.

Both candidates remain statistically tied. If no candidate wins more than 50% of valid votes, the two will face a runoff on October 25.

"Brazil's presidential elections are entering their final stretch with first round elections to be held Sunday, 4 Oct. If no candidate wins more than half of the valid vote, a run-off will be held on 25 Oct," Inui wrote in the note. 

Inui continued, "If no candidate wins more than half of the valid vote , a run-off will be held on 25 Oct. First-round election results could be a major market-moving event, we think. The key variable is the margin between the leading candidates. Latest polls point to a tight margin of 5ppts between current president Lula da Silva (PT) vs. Senator Flávio Bolsonaro. 

"A narrower lead for President Lula or stronger performance by third-party candidates could add to policy change expectations, which could support risk-on sentiment. However, history suggests some caution: since 1989, every first-round winner has secured the presidency in the run-off of elections," the analyst said. 

However, Polymarket bettors see a clearer favorite, giving Bolsonaro a 60% chance of winning, compared with 39% for Lula.

Here's more from HSBC analysts on market impacts: 

What to look out for:

Brazil heads to the polls on 4 October, with a second round scheduled for 25 October if no candidate wins more than 50% of valid votes. Recent polling is pointing towards a highly competitive presidential race, with neither candidate expected to garner enough votes to win in the first round. Since 1989, all but one went to a second round. Leading candidates received between c42-c53% in the first-round vote and all went on to eventually win. In other words, first round winners prevailed and they won the first round by c42%-49%. In the last election cycle, the margin of victory was the tightest in recent history at 1.8ppts (50.9% for Lula vs 49.1% for Bolsonaro), with Lula's vote share only increasing by 2.5ppts between rounds, compared with c9-15ppts for winners in previous five runoffs. This is a useful benchmark for a highly polarized election, we think, especially considering there are no relevant left-wing candidates besides Lula in the first round.

Initial upside, not necessarily sustained

In the first trading session following the 2022 first round election, the IBOV surged 5.5%, while real strengthened against the dollar. In the 2022 elections, privatization hopes drove the largest immediate gains in SOEs at the time (Copasa, Sabesp, Cemig), and interest rate futures immediately fell, with consumer discretionary accounting for 50% of the top 10 performers. The move likely was attributed to Bolsonaro’s stronger-than-expected (vs polls) first round results and the composition of the incoming Congress, which together reduced the policy risk premium investors had considered before the vote. However, the Ibovespa came down from its highs leading up to the second round, declining 1.4% from the close after the first round, but still 4.1% above pre-first-round level. The BRL and iShares MSCI Brazil ETF (EWZ Index) followed a similar pattern, but the EWZ saw greater appreciation from 1 September 2022 than the Ibovespa around the run-off. Leading up to the run-off, reduced perceived risks of government intervention continued to support utilities and energy, which retained gains, while consumer discretionary's initial rally faded as interest rate futures rose again.

A similar story could occur for 2026. An outsized reaction could occur immediately after 4 October if investors mark down the perceived fiscal/policy-risk premium based on a tight first round outcome, like what occurred in 2022, but would not necessarily extend into a three week straight rally. Subsequent price action would depend on polling, endorsements, campaign economic proposals, the congressional result, and the long end of the Brazilian rates curve. We believe a first round result associated with lower long-term fiscal risk would likely produce a broader domestic risk-on trade, with SOEs, domestic cyclicals, and bond proxies potentially experiencing the most upside. If first round results are tighter than in the prior elections, equity markets could react positively on expectations of fiscal consolidation ahead.

And post-election results? We see asymmetric returns for equity markets

We expect an initial knee-jerk reaction following the run-off as markets reassess the likelihood of fiscal consolidation under the incoming administration. A result that increases confidence in a more credible fiscal path could drive a relatively rapid rerating through lower-end yields, tighter risk premia, and stronger performance in domestic cyclicals and bond proxies where valuations are sensitive to fiscal and rate outlooks. Conversely, a result that reduces expectations for fiscal consolidation could trigger an initial derating, but we see some valuation support limiting the downside in parts of the market. Many election sensitive names including Petrobras are already trading close to historical trough valuations on an EV/EBITDA basis, leaving less room for further multiple compression Petrobras, 16 Aug. And Banco do Brasil is trading at a P/B of 0.7x and a 69% discount to Itau, close to a historic high, Brazilian Financials, 14 Sept.

Given the close race, we would emphasize segments that can participate in a relief rally without requiring an aggressive risk stance. Lower-beta yield names and selected bond proxies appear better placed to capture upside from lower long-end rates while offering more resilience if the market reaction is short lived. Politically exposed names and higher beta domestic cyclicals could outperform in a market-friendly scenario, but they would also be more sensitive to any reversal in fiscal expectations.

Polls suggest the race remains too close to call, while Polymarket odds point to Bolsonaro as the favorite.

On March 3, just days after the first US and Israeli strikes on Iran, when most of the market was still busy pricing the closure of the Strait of Hormuz as a temporary inconvenience, we pointed out something that seemed rather obvious (to us): the UAE's oil port of Fujairah, which sits on the Gulf of Oman and bypasses the strait completely, was far too small for its strategic importance - and that would change.

Surprising Fujairah is not a bigger oil terminal: it bypasses the straits completely.

Expect major infrastructure push here after the war. https://t.co/Do1gK7KBDQ

— zerohedge (@zerohedge) March 3, 2026

Six days later we went one step further:

After this crisis, every Gulf pipeline has to end either in Yanbu or Fujairah alongside massive terminal infrastructure https://t.co/fNBywD3Wr1

— zerohedge (@zerohedge) March 9, 2026

Seven months later, the "major infrastructure push" has a name, a sponsor, and a checkbook. According to a new Bloomberg Big Take, Abu Dhabi Crown Prince Sheikh Khaled bin Mohamed Al Nahyan - who took the helm of the emirate's $300 billion L'imad Holding sovereign fund weeks before the war began - is now the point man for what the UAE is officially calling its "Zero Hormuz" strategy (wonder if he was reading Zero Hedge at the time). And the centerpiece of that strategy is, you guessed it, Fujairah.

Follow the money (to the Gulf of Oman)

Here are the Bloomberg report highlights:

• L'imad has announced plans to take Abu Dhabi Ports Co. private at a valuation of nearly $9 billion, and people familiar say the fund is now likely to spend tens of billions of dollars more on new port infrastructure outside the strait.
• The crown prince and his inner circle are expected to be "particularly focused" on expanding ports in Fujairah, which sits just outside Hormuz and opens into the Gulf of Oman.
• In May, L'imad struck an agreement with BlackRock (via its Global Infrastructure Partners unit), Temasek and ADNOC to jointly target up to $30 billion of infrastructure investment in energy transportation, logistics and water. Abu Dhabi did not want such a large push funded solely by state money - which is a polite way of saying Larry Fink gets a toll road around Iran.
• Sheikh Khaled also chairs the executive committee of ADNOC's board, which is building a second oil pipeline to double export capacity through Fujairah; at a May meeting he directed the company to accelerate delivery.
• Meanwhile Dubai's DP World is separately pushing new container terminals in Fujairah - meaning that a relatively small stretch of coastline under the Al Hajar mountains is about to become some of the most crowded (and most valuable) real estate in the Gulf.

The UAE's trade minister Thani Al Zeyoudi summed up the doctrine back in June: the country wants to move to Zero Hormuz dependency regardless of whether the strait is open or not. Translation: even if Tehran signs a peace deal tomorrow, the leverage it enjoyed over Gulf exports for decades is never coming back.

Regular readers will recognize every step of this progression. On April 2 we noted that Gulf states were dusting off costly bypass pipeline plans; on May 15 we reported that ADNOC would double its crude export capacity bypassing Hormuz with the new pipeline to Fujairah, due in 2027; and on July 13 we wrote that DP World's plan for a new east coast port in Fujairah "signals the beginning of the end" of Iran's Hormuz leverage. That same day, as peace talks went nowhere, we also offered an alternative engineering solution:

At this rate it will be faster to dig a canal through the UAE to bypass Hormuz than to wait for a peace deal

— zerohedge (@zerohedge) July 13, 2026

Abu Dhabi, it appears, has opted for the slightly less ambitious version: pipelines plus a lot of concrete.

Goldman: 60% of Gulf exports insulated from Hormuz by 2028

So how far can this go? Goldman's commodity team (Alexandra Paulus, Yulia Grigsby, Daan Struyven and Filippo Cuscito) ran the numbers in a July note titled "Gulf Exports: Short-Term Uncertainty, Long-Term Pipeline Hedge" (available here to pro subs), and the conclusion is that while Hormuz still dictates prices in the short run, the long run looks very different. The bank estimates that enough pipeline capacity will be added in the region to insulate over 45% of pre-war Persian Gulf exports by end-2027 and more than 60% by end-2028 from any future Hormuz shock.

Some of the details:

• Goldman measures current effective bypass capacity as the flows out of Yanbu (East-West pipeline), Fujairah (ADCOP pipeline) and Ceyhan (Kirkuk-Ceyhan). In its base case, that capacity rises by 3.8mb/d by end-2027 and 7.3mb/d cumulatively by end-2028, to over 14mb/d - versus ~23mb/d of pre-war exports from the seven Gulf producers that need pipelines to dodge Hormuz.
• The UAE features prominently: the West-East pipeline (ADNOC's second line to Fujairah) is one of only two projects already under construction, while a Hamriyah-Fujairah pipeline sits in Goldman's "Accelerated Scenario" - which would insulate 75% of exports by end-2028 (vs. just over 45% in the "Conservative Scenario").
• History is on the builders' side: across Goldman's sample, the median construction time for Mideast pipelines was 2.5 years, and single-country projects get built faster, especially in response to supply disruptions. Multi-country projects (looking at you, Iraq-Syria) not so much.
• Total cost across the seven projects: roughly $30-48 billion - or, put differently, about one BlackRock/L'imad infrastructure platform.

And here is the punchline for oil bulls: Goldman raised its long-dated Brent assumption (3-year-ahead futures) by $9 to $76/bbl at the peak of the war, mostly on a higher structural security premium. But the bank warns that the eventual expansion of bypass capacity poses downside risk to that long-dated assumption. In other words, every barrel that Sheikh Khaled routes to Fujairah is a barrel of risk premium Iran can no longer charge the world.

The plumbing is already working

The "adaptation" is already visible in the export data. As we reported earlier today, Goldman estimates that Persian Gulf oil exports (including "dark exports") recovered to 23.3mb/d over the past week, in line with their 2025 average, after doubling in September. Crude accounted for nearly 90% of the recovery, reaching 19mb/d (108% of the 2025 average), while refined product exports remain stuck at about half of normal. Crucially for this story, Goldman notes that oil exports from the UAE - which shockingly exited OPEC shortly after the Iran war started - are also above their 2025 average, "with likely further upside" - while Iran shipped no crude by sea at all in September.

Drill down and Fujairah is doing a lot of the heavy lifting: Goldman's late-September breakdown puts flows via Fujairah at 3.6mb/d (crude, products and LPG combined), more than Yanbu's 2.6mb/d, and more than double the ~1.7mb/d Fujairah handled before the war.

And by country, the UAE is already running at 110% of its 2025 export average - second only to Saudi Arabia's dark-transit-fueled surge - while Iran sits at 19%.

The UAE in particular has been the most creative workaround artist of the war: as we noted in July, its crude output hit an all-time high of 4.1mb/d in June after it quit OPEC, with ADNOC selling cargoes for loading off Fujairah and Sohar, outside the strait. Back in March, we reported that Fujairah crude loadings had already hit ~1.9mb/d - about the max the existing 1.5-1.8mb/d Habshan-Fujairah line can carry. The only real constraint was pipe. Which is exactly what Abu Dhabi is now paying to fix.

Bypassing the strait is not the same as bypassing the drones

None of this makes Fujairah safe. It is roughly 80 miles from Hormuz and well within range of Iranian drones and short-range missiles - a point LSE professor Steffen Hertog makes in the Bloomberg piece. Tehran knows exactly what Fujairah represents: the port was in flames on March 14, was attacked at least seven times in the first four weeks of the war, and on March 31 Iran explicitly threatened to target the port and its pipeline "in order to close the UAE's route to bypass the Strait of Hormuz." On May 4, another Iranian strike on Fujairah's oil zone sent Brent above $114.

Saudi Arabia offers the cautionary tale. Its 7mb/d East-West pipeline to Yanbu - the region's single biggest Hormuz bypass - was shut down on September 11 after drone attacks by pro-Iran militias, before restarting on September 28. Goldman's September 14 Oil Tracker pointed out that an April strike on the same pipeline cut flows by just 0.7mb/d for four days, while the latest attack was far more severe and threatened the remaining ~2mb/d of Yanbu exports. The Saudis promptly pivoted back to shipping through... Hormuz. Meanwhile the Houthis are advancing on Bab el-Mandeb, threatening the other end of the Red Sea route.

Which brings us to Treasury Secretary Bessent, who predicted on September 1 that in two years Hormuz will be "a worthless piece of water." Qatar's energy minister promptly called that "completely wrong" - which is easy to understand when you are Qatar, have no geographic alternative route, and have watched your gas revenue drop sharply. The truth is somewhere in between: Hormuz won't be worthless, but if Goldman's math is right it will be worth a lot less to Iran - and a lot more to whoever owns the ports on the other side of the mountains.

As Chatham House's Sanam Vakil puts it, a "No Hormuz" policy is now of utmost importance for the UAE - but with Iran's proxies extending their reach, Abu Dhabi will also have to prepare for infrastructure targeting outside Hormuz too. Expect the next line item in the L'imad budget to be air defense.

For now, the bottom line is the one we flagged on day four of the war: the Gulf's most important real estate is no longer inside the strait, it's on the Gulf of Oman - and the crown prince of Abu Dhabi has just put tens of billions of dollars (and BlackRock's money) behind that view.

Mamdani wants New York’s billionaires to cough up more money. Maybe he should call Florida and ask how they’re getting them to hand it over voluntarily.

Because Miami just scored a $3 billion pledge from billionaire Ken Griffin, with $2 billion earmarked for a brand new Carnegie Mellon campus in the city, according to Forbes.

The massive commitment represents Griffin’s latest investment in Miami’s evolution into a major center for finance, technology and entrepreneurship. The Citadel founder is directing the bulk of the money toward establishing Carnegie Mellon in South Florida, adding a major research university to the city’s rapidly expanding business ecosystem.

Under the plan announced Wednesday, Carnegie Mellon will develop a 35-acre campus in Miami’s Wynwood neighborhood. Work on the project is slated to begin in 2027, with the first graduate students expected to arrive in 2028.

Another $1 billion will go toward Carnegie Mellon’s operations in Pittsburgh. That money will support financial aid and other university priorities, while a substantial portion will be directed to its renowned computer science program, which will take the new name Kenneth C. Griffin School of Computer Science.

The university says Griffin’s $3 billion pledge is the largest individual donation ever made to an American university. Griffin is also set to take a seat on Carnegie Mellon’s Board of Trustees.

Forbes writes that the deal marks another significant expansion of Griffin’s footprint in Florida. Citadel moved its headquarters from Chicago to Miami in 2022, and Griffin has since poured roughly $2.4 billion into South Florida organizations and initiatives, including substantial support for education, according to Reuters.

For Miami, the university project could ultimately have an impact well beyond Citadel’s own presence in the city. Carnegie Mellon plans to concentrate its new campus on fields including national security, healthcare, energy, climate resilience, advanced manufacturing and industrial technology, giving Miami another potential anchor as it attempts to build a larger technology and research economy.

Griffin has framed the commitment as an investment in American innovation and future job creation, particularly in industries where the United States can strengthen its competitive position globally.

The sheer size of the pledge also resets the bar for university philanthropy. Michael Bloomberg’s $1.8 billion contribution to Johns Hopkins in 2018 was among the previous benchmarks. With this latest commitment included, Reuters estimates Griffin’s lifetime charitable giving at approximately $5.7 billion.

For cities trying to figure out how to get billionaires to part with their money, Miami may be offering an interesting example: make them want to put their capital there, and the checks can get awfully large.

Authored by Ron Paul via The Ron Paul Institute for Peace & Prosperity,

Most indicators suggest that the Republicans in the House and Senate are going to face a brutal backlash from the American people in November. Polling with generic ballots suggest a big win for the Democrats and likely the loss of Republican control of both houses of Congress.

There is little reason to cheer a Democratic party victory, however Republicans in charge have for the most part earned what is coming to them. They have refused to uphold their Constitutional obligation to take leading responsibility for war powers.

Earlier this year, President Trump launched what is the most unpopular war in US history when on February 28th he ordered a full US assault on Iran. Iran had not attacked or even threatened the United States and our own intelligence community continued - and continues - to claim that Iran is not building nuclear weapons.

The February attack followed last June's US bombing of Iran after which the president claimed that the country's nuclear program was "completely obliterated."

Congress over too many years has been used to allowing the President to have his own way on matters of war and peace. Members bask in the false patriotism promoted by the media and special interests as the people "rally around the flag." Eventually, when the war has gone on for years without "victory" - as in Iraq and Afghanistan - many claim they were never really for the war in the first place.

This time is different. There was no propaganda blitz. There were no attempts to convince the American people that the "smoking gun" could be a "mushroom cloud" as we were told in the run up to the 2003 Iraq war. The majority of Americans were against the war on Iran before it started and only became more opposed over time.

As the war drags on, we have seen the national debt pass $40 trillion, diesel fuel prices at the highest ever, and inflation hitting the middle class and poor even harder. The "cakewalk" again failed to materialize.

Incumbent Republicans seeking re-election are in such a panic over voter anger about the war that they have taken to calling for a quick end to it while on the campaign trail. When they get back inside the Beltway, however, they vote to keep the war going for fear of President Trump's anger.

This is not leadership.

Late last week the Iranian delegation in New York for the UN General Assembly meeting offered an off-ramp. Essentially it was the same Memorandum of Understanding that President Trump signed last June and promptly disregarded, but on a shorter timeline. By Sunday, the President announced that he would reject the deal - a deal like one he already signed.

There is no strategy for winning this disastrous war, it is clear. From over-confidence on decapitation strikes to certainty of a quick military victory, the Administration is now promising that they will win by crippling Iran's economy. The problem is that Iran has been operating under "crippling" sanctions for decades and has a high economic pain tolerance.

Unfortunately, both parties in DC are war parties, so again it is hard to cheer the likely Republican losses. But sometimes electoral losses can serve as wake-up calls. Eventually we may see the re-emergence of antiwar Republicans in the spirit of the late Sen. Robert Taft and a restoration of the Constitution.

We publish a variety of perspectives. Nothing written here is to be construed as representing the views of ZeroHedge.

After a year in which Google's AI roadmap resembled a Waymo stuck in a roundabout, the search giant on Wednesday finally unveiled Gemini 4 "Argon", its long-awaited flagship model. The market cheered, briefly: according to Goldman's closing equities color, GOOG traded +2% after hours on the "Argon" announcement. 

Then Bloomberg reported that some of the people who have actually use the thing - as in Google's own engineers - aren't nearly as impressed as the leaderboard. The stock promptly faded. 

According to Bloomberg's Julia Love and Davey Alba, Gemini 4 "performed well on benchmarks widely used to gauge model efficacy" but "does less well when employees actually put it to work," particularly on coding. One insider said the model "isn't particularly adept at front-end design" - the part of software that decides how apps and websites look and feel. Which is a bit awkward for a company whose entire business is, well, things you look at on a screen.

Google, naturally, disagrees. It told Bloomberg it would be "inaccurate" to say Gemini 4 is underperforming in coding, and pointed back to last week's remarks by DeepMind boss Koray Kavukcuoglu, who said "it's a certainty that we are always gonna be at the frontier." Another Google employee said there is "large consensus" internally that the model is frontier-class. Which is the kind of thing one tends to say when there isn't.

The industry has a word for this. "Benchmaxxing" is when engineers optimize a model to crush standardized tests rather than to do useful work - the AI equivalent of the kid who memorizes every SAT prep book, posts his 1600 on LinkedIn, and then can't do his own laundry. Two people familiar with Gemini 4 told Bloomberg the model appears affected by exactly this.

Surge AI founder Edwin Chen put it more bluntly, calling benchmark-chasing "an incredibly pernicious problem" and comparing it to bragging about your kid's SAT score. We'd add that a generation of Silicon Valley's finest minds has now spent billions teaching machines to do what they themselves did in high school: optimize for the test, then act surprised when the real world grades on a curve.

The irony is that Google's own researchers have documented how quickly optimization turns into gaming. As we reported on September 3, a DeepMind experiment found that when math problems got hard, 9% of AI agents outright cheated and another 5% cheated "after hesitation", gaming a shared knowledge base that rewarded successful submissions. Turns out teaching to the test works on silicon too.

A $400 million detour

Gemini 4 is also the model Google shipped instead of the one it promised. At I/O in May, Google pledged Gemini 3.5 Pro for June; the date slipped (Bloomberg first reported the delay on July 16, citing tech that "fell short of internal goals") and the project was eventually abandoned altogether. Bloomberg Intelligence's Mandeep Singh estimates a frontier training run can cost as much as $400 million - before paying the researchers, many of whom have since left.

And leave they did. We noted in June that Google was losing more Gemini researchers to Anthropic, following the earlier departures of Nobel laureate John Jumper and transformer co-inventor Noam Shazeer. Then in August came the big one: Jeff Dean exited after 27 years to launch his own startup, taking several senior researchers with him and knocking 5% off Alphabet stock. Demis Hassabis subsequently kicked himself upstairs to chairman, handing day-to-day DeepMind operations to Kavukcuoglu - who now gets to defend Argon's front-end skills to Bloomberg.

Why this matters: the ROIC math doesn't grade on benchmarks

This would be a nerd-bro squabble over leaderboard screenshots if it weren't for the money involved. Gemini underpins nearly everything Google sells - Search, Maps, Gmail, Chrome, each with over a billion users - and Alphabet is one of the hyperscalers footing the largest capex bill in corporate history.

Per Goldman's Ryan Hammond (full note available to pro subs), US hyperscalers are on track to spend roughly $800 billion in 2026; consensus expects $1.1 trillion in 2027, while Goldman's house view is even higher at $1.2 trillion and $1.4 trillion for 2027 and 2028. Hammond estimates hyperscalers need about $300 billion of annual AI revenue just to break even on 2026-27 spending, and that end users may ultimately need to spend around $1 trillion a year on AI applications for everyone in the stack to earn decent returns.

Separately, Goldman's Eric Sheridan, whose ROIC framework we flagged last week, calculated that assuming a 15% ROIC target and ~$42 billion of capex per gigawatt, the six big US hyperscalers need to generate roughly $1.42 trillion in cumulative revenue during 2028-30 - about $11.6 billion per GW per year - with a range of $908 billion to $1.89 trillion depending on assumptions. As we tweeted at the time, even in Goldman's worst-case scenario where ROIC on capex goes to zero, they'd still need $920 billion a year just to cover depreciation and running costs.

Here's the rub: none of that revenue gets paid in MMLU points. It gets paid by developers and enterprises choosing whose model to build on. And Bloomberg notes Gemini 4 is "a very large model" - and big models are expensive to serve, which means pressure on margins at exactly the moment the capex hurdle is rising. Goldman's TMT desk flagged this week that the industry is already in a price war: OpenAI cut Luna pricing by 80% in July and usage rose roughly tenfold, while "Big Short's" Steve Eisman is openly asking what happens to margins if cheaper Chinese and open-weight models push prices down further. Bringing an expensive, benchmark-optimized heavyweight to a knife fight over token pricing is a bold strategy.

Meanwhile, the competition isn't waiting

The timing couldn't be worse. On the same day Google unveiled Argon, OpenAI was busy at its Developer Day rolling out "Dots" - persistent, always-on agents that live inside ChatGPT, Slack and Teams. Goldman's Sean Johnstone noted OpenAI is "increasingly competing with Microsoft 365, Google Workspace and traditional enterprise software - not simply other AI models." Per Axios, cited by Goldman's desk, OpenAI's ARR is nearing $70 billion, up from ~$40-41 billion as recently as mid-August, with enterprise sales more than doubling since July. OpenAI is now reportedly seeking $30 billion at a $1.4 trillion valuation.

And earlier this month Meta released its new agentic platform, Muse, which promptly took the app charts by storm - so thoroughly that Amazon blocked it - and sparked a 25%+ rally in META. Goldman's Sheridan now frames agentic commerce as one of AI's biggest monetization opportunities, noting that "similar to how Google captured search intent, successful AI platforms may capture shopping intent." Read that sentence again if you're long GOOGL.

To be fair, Gemini 4 has real strengths: insiders say it stands out at multimodal work like extracting metadata from video, on safety and cybersecurity (it reportedly beat OpenAI's Astra on a security benchmark), and it can spit out up to 1 million tokens - roughly 750,000 words - in one go. Whether anyone needs a 750,000-word answer is a separate question; we suspect the answer is "only for benchmarks."

Bottom line: Google has the distribution, the TPUs and the balance sheet. What it doesn't have is much time. Every quarter it spends explaining leaderboard scores, is a quarter that OpenAI, Anthropic and Meta spend convincing developers, businesses and consumers that the future of search and software runs on their platforms. Because the $1.4 trillion question isn't who tops the leaderboard - it's who gets paid.