Lilly’s Retatrutide Scores Triple Trial Triumph With 26% Weight Loss, But New Safety Signal Emerges

 

Analysts were hoping for a safety profile similar to what was achieved in Phase II but an abnormal sense of touch, called dysesthesia, has emerged in the late-stage TRIUMPH-4 trial.

Eli Lilly’s triple hormone receptor agonist cut body weight by 26.6% on a placebo-adjusted basis and reduced knee osteoarthritis pain by 75%, capping off a trio of wildly successful readouts for the pharma’s obesity pipeline.

BMO Capital Markets dubbed the Thursday readout for retatrutide “a true TRIUMPH”—playing off the trial’s name, TRIUMPH-4. The results are the strongest efficacy reported to date, the firm added.

“These results help to solidify retatrutide’s profile as an even higher efficacy next generation GLP-1+ asset,” BMO wrote Thursday morning.

Lilly’s shares breached the $1,000 mark in pre-market trading Thursday morning, rising 1.5% to $1,008 from $993.64 at close yesterday.

Previously, BMO had been anticipating weight loss of about 20%-23% for retatrutide, which targets GIP, GLP-1 and glucagon, along with 50% pain reduction. This would have put the drug ahead of rival Novo Nordisk’s semaglutide, which scored 42% pain reduction in a previous trial.

But retatrutide went even further, with patients in the Phase III TRIUMPH-4 trial on a 12 mg dose losing an average of 28.7% of their body weight at 68 weeks. The trial featured adults who have obesity or are overweight with knee osteoarthritis.

Adjusted for placebo, the weight loss was 26.6%, according to BMO. In addition to the pain achievement, retatrutide improved measures of physical function, reduced cardiovascular risk and blood pressure.

While the weight loss was significant, Lilly did report a safety signal called dysesthesia in 8.8% and 20.9% of patients on the 9 mg and 12 mg doses, respectively. In the placebo arm, just 0.7% of patients reported this side effect. Dysesthesia is an abnormal sense of touch that causes normal sensations to feel unusual or painful. It’s often a symptom of a larger diseases, such as multiple sclerosis or diabetes.

Analysts were hoping for a safety profile similar to what was achieved in Phase II. Dysesthesia was not reported in Lilly’s earlier mid-stage trial for retatrutide. BMO noted the dysesthesia signal and said the firm will be watching for it in readouts to come and for more detailed data from the TRIUMPH-4 trial. Lilly said that the dysesthesia events did not seem to lead to discontinuation.

Other safety events included typical gastrointestinal symptoms, with nausea rates of 43%, vomiting of 21% and diarrhea of 33%. Lilly reported discontinuation rates of 12.2% and 18.2% across the 9 mg and 12 mg arms, respectively, compared to 4% in the placebo group. Some of these discontinuations were due to “perceived excessive weight loss,” according to Lilly.

“Discontinuation rates appear to highlight the speed and strength of weight loss was excessive for some patients with lower BMI,” BMO wrote. The firm dubbed the discontinuation rates “overall acceptable,” with rates lower in patients with a higher BMI.

Retatrutide has been designed to be more effective than Lilly’s tirzepatide, which is approved for weight loss as Zepbound and diabetes as Mounjaro, Leerink Partners said in a note previewing the readout earlier this month.

It’s unclear if the adverse effect will impact retatrutide’s prospects. Lilly is positioning retatrutide as an option for patients seeking more significant weight loss than what can be achieved with the available GLP-1 therapies. TRIUMPH-4 was not powered to show the maximum weight loss possible, BMO noted.

The Cherry on Top

The retatrutide readout is just one of three major trials that Lilly has read out in recent months, all of them successful. In October, Lilly reported that the oral therapy orforglipron bested AstraZeneca’s Farxiga in a Phase III diabetes trial. The drug, which Lilly has already submitted to the FDA for review, also beat Novo’s semaglutide at reducing A1C in a head-to-head trial for type 2 diabetes in September.

Truist Securities estimates that Lilly’s weight loss trio of incretin medicines, Mounjaro, Zepbound and orforglipron, could reach $101 billion in peak sales worldwide. Retatrutide could be the cherry on top.

Lilly has six more late-stage studies in the TRIUMPH program underway for retatrutide which are expected to readout by the end of 2026. In total the program, which kicked off in 2023, has enrolled 5,800 people with obesity, obstructive sleep apnea and knee osteoarthritis.

https://www.biospace.com/drug-development/lillys-retatrutide-scores-triple-trial-triumph-with-26-weight-loss-but-new-safety-signal-emerges

Amgen Scores Uplizna Expansion Into Generalized Myasthenia Gravis

 

Generalized myasthenia gravis is Uplizna’s second new indication this year, after the FDA cleared the anti-CD19 antibody for IgG4-related disease in April.

On Thursday, the FDA approved a key expansion for Amgen’s autoimmune therapy Uplizna, allowing its use in patients with a certain type of generalized myasthenia gravis.

Specifically, the regulatory agency cleared Uplizna for adults positive for autoantibodies against the anti-acetylcholine receptor (AChR) and anti-muscle specific tyrosine kinase (MuSK). Uplizna is the first CD19-targeted B cell therapy approved for this indication, Amgen said in its press release, and allows long-term disease control with just two doses per year.

Data from the Phase III MINT study supported the regulatory verdict. A September 2024 readout showed significant improvements in activities of daily living at 26 weeks, as measured by the Myasthenia Gravis Activities of Daily Living (MG-ADL) scale, as compared with placebo. Uplizna’s benefit remained clinically meaningful and statistically significant in both AchR-positive and MuSK-positive subgroups.

In March, Amgen released follow-up data showing that Uplizna not only maintains but also deepens treatment response at 1 year. Placebo-adjusted improvement in MG-ADL was 2.8 points at one year, as compared with 1.8 points at 26 weeks. Scores on the Quantitative Myasthenia Gravis scale, which assesses functional performance, improved by 4.3 points at one year from a 2.5-point average improvement at 26 weeks.

Reacting to this 52-week readout, analysts at William Blair said in a March 12 note that these data “solidify a meaningful role for Uplizna in the treatment of myasthenia gravis” and support its “blockbuster potential.” The analysts also noted that the therapy’s “infrequent every-six-month dosing interval” will give it an edge over “weekly, monthly, or every-other-month dosing regimens with other therapies.”

Uplizna is an IgG1 monoclonal antibody that works by targeting CD19 on immature and mature B cells. Its specific mechanism of action in myasthenia gravis isn’t yet completely understood, though Uplizna could act by depleting B cells, which are key drivers of the disease’s pathology. Patients with myasthenia gravis suffer from muscle weakness, breathing problems and difficulty in swallowing and speech.

The FDA first approved Uplizna in 2020 for neuromyelitis optica spectrum disorder, a rare autoimmune and neuroinflammatory disease of the optic nerve, spinal cord and brain stem. The drug secured a label expansion in April this year for IgG4-related disease.

https://www.biospace.com/fda/amgen-scores-uplizna-expansion-into-generalized-myasthenia-gravis

FDA Plans Black Box Warning for COVID-19 Vaccines: Report

 

The FDA intends to place a black box label—its most serious warning—on COVID-19 vaccines, according to reporting by CNN. It is unclear if the warning would apply to Moderna’s, Pfizer/BioNTech’s and Novavax’s shots, or to all age groups.

The FDA may be gearing up to slap a black box warning on COVID-19 vaccines, according to two individuals who spoke with CNN on Friday.

A boxed warning, which alerts the public to the serious risks associated with a drug or medical device has been attached by the FDA to products including opioids for addiction, for example, and the acne medicine Accutane for birth defects. These risks should be weighed against the treatment’s benefits.

The plans for a black box warning, first reported by CNN on Friday morning, are expected to be rolled out by the end of this year. It is unclear, however, whether they will apply to all COVID-19 vaccines or only mRNA vaccines—which have faced considerable scrutiny since Health Secretary Robert F. Kennedy Jr. took office in February, and since being politicized by the pandemic. Moderna and Pfizer/BioNTech make the two mRNA-based vaccines approved for use in the U.S., Comirnaty and Spikevax. The only other FDA-approved COVID vaccine is Novavax’s protein-based shot Nuvaxovid.

It also isn’t clear whether the warnings will apply to all age groups, CNN reported.

In a statement to the publication on Thursday, Department of Health and Human Services spokesperson Andrew Nixon said, “Unless the FDA announces it, any claim about what it will do is pure speculation.”

In an email to BioSpace Friday morning, Nixon reiterated this statement and added, “The FDA takes very seriously any death that is attributed to a regulated medical product.”

The black box initiative is being led by Center for Biologics Evaluation and Research head Vinay Prasad, according to one of the people who spoke with CNN. This follows Prasad’s claim in an internal FDA memo late last month that the deaths of 10 children could be linked to COVID-19 vaccinations. The alleged deaths were first flagged by acting Center for Drug Evaluation and Research director Tracy Beth Høeg, then a senior advisor to the agency.

The memo set off an eruption of protest from healthcare and ex-regulatory leaders. More than a dozen former FDA and CDC officials in an editorial published in The New England Journal of Medicine earlier this month decried the proposed vaccine policy changes laid out in Prasad’s memo.

“I think it’s irresponsible and unprofessional,” Paul Offit, director of the Vaccine Education Center at Children’s Hospital of Philadelphia, told BioSpace of the FDA’s reported black box warning plans, “because this agency, either Vinay Prasad or [FDA Commissioner] Marty Makary, have not in any way made clear why it is they think the COVID-19 vaccine killed children.”

In a statement to Endpoints News on Dec. 3, the FDA said it is still investigating “additional adverse event cases” and plans to make a report covering the reported 10 deaths publicly available by the end of December.

Included in these data, Offit said he would like to see autopsy data “showing that there’s severe damage to the heart muscle,” and that there are antibodies in the child’s bloodstream against SARS-CoV-2 spike protein. This, he said, would indicate that the child was vaccinated and not naturally infected with the virus that causes COVID-19.

The FDA is also expanding its investigation of deaths potentially related to COVID-19 across multiple age groups, an agency spokesperson told Bloomberg earlier this month.

Notably, the CDC on Thursday released a report showing that the 2024–2025 formulations of the COVID-19 vaccine were 76% effective at preventing emergency or urgent care visits in children aged 9 months through 4 years.

https://www.biospace.com/fda/fda-plans-black-box-warning-for-covid-19-vaccines-report

US FDA approves Innoviva's oral antibiotic for gonorrhea

 The FDA has approved Nuzolvence (zoliflodacin; Innoviva Specialty Therapeutics), a new, single-dose oral treatment indicated for uncomplicated urogenital gonorrhea in adults and children 12 years and older who weigh at least 77 pounds. The product is administered as a single oral dose and represents 1 of 2 newly approved treatment options for uncomplicated urogenital gonorrhea in nearly 2 decades, per the approval announcement.1

“These approvals mark a significant milestone for treatment options for patients with uncomplicated urogenital gonorrhea,” said Adam Sherwat, M.D., director of the Office of Infectious Diseases in the FDA’s Center for Drug Evaluation and Research (CDER).2

According to Innoviva Specialty Therapeutics, the approval reflects a collaboration with the Global Antibiotic Research and Development Partnership, which sponsored and led the pivotal clinical trial that formed the basis for regulatory review.

Pavel Raifeld, CEO of Innoviva, stated, “The FDA’s approval of Nuzolvence marks a pivotal moment for patients and the broader healthcare community managing gonorrhea infections. For the first time in decades, both patients and their health care providers will have a single-dose, oral treatment option for uncomplicated urogenital gonorrhea.”

Rising rates and limited treatment alternatives

Uncomplicated urogenital gonorrhea is among the most common sexually transmitted bacterial infections worldwide, with more than 82 million cases estimated annually. In the United States, more than 500,000 cases are reported each year, and the highest rates are observed in sexually active young adults.3

Without timely treatment, gonorrhea may lead to reproductive or systemic complications. The approval announcement emphasized that “Gonorrhea is the second most common sexually transmitted bacterial infection worldwide,” and noted persistent transmission among sexually active populations.

Antimicrobial resistance and evolving therapeutic need

The emergence of antimicrobial resistance has contributed to the diminishing effectiveness of established therapeutic options. The announcement stated that the World Health Organization has identified antimicrobial resistance as one of the most critical global health threats, driven partly by widespread resistance among Neisseria gonorrhoeae strains.

Edward Hook, MD, of the University of Alabama at Birmingham, reported, “The decades-long absence of new gonorrhea treatments, combined with rising global antibiotic resistance, has created significant challenges in managing this common but potentially serious sexually transmitted infection.”

Some patients require alternative approaches because of medication allergies or limited access to injectable therapy. Hook added, “A new antibiotic that does not require injection and can be used for patients who are allergic to penicillin or related drugs meets 2 important unmet needs in the treatment of gonorrhea.”

Clinical trial evidence informing approval

The FDA approval was supported by multinational clinical data evaluating the safety and efficacy of a single oral dose of zoliflodacin compared with the current standard injectable combination therapy. The approval press release stated, “In the pivotal phase 3 study, Nuzolvence demonstrated non-inferiority compared to the current standard injectable therapy, including in infections caused by drug-resistant strains, while offering the convenience of a single oral dose.”4

Previous results published in The Lancet also demonstrated non-inferiority of oral zoliflodacin relative to ceftriaxone with azithromycin among adolescent and adult participants with uncomplicated urogenital infection, with similar microbiologic outcomes at extragenital sites. Those findings were part of the largest clinical trial conducted to date for a new treatment targeting Neisseria gonorrhoeae infection.4

Mechanism of action and indication

Nuzolvence is a bacterial type II topoisomerase inhibitor with a mechanism distinct from currently approved antibiotics. The product was approved for adults and pediatric patients 12 years and older who weigh at least 35 kg for the treatment of uncomplicated urogenital gonorrhea due to Neisseria gonorrhoeae, including strains resistant to first-line therapies.

According to the announcement, Nuzolvence “demonstrated activity against drug-resistant Neisseria gonorrhoeae” and offers a single-dose oral approach at a time when injectable therapy remains the only recommended first-line option in the United States.

Implications for clinical practice

Approval of a single-dose oral therapy may expand access to treatment and provide an alternative for individuals unable to receive injectable agents. For clinicians, the availability of an oral medication may facilitate timely management in settings where in-office administration of intramuscular therapy is not feasible.

Hook noted, “In the pivotal phase 3 study, Nuzolvence demonstrated non-inferiority compared to the current standard injectable therapy,” while emphasizing the broader clinical need associated with emerging resistance and limited therapeutic innovation.

Continued implementation and post-approval experience will determine how Nuzolvence is incorporated into treatment strategies for uncomplicated urogenital gonorrhea and its role in addressing antimicrobial resistance trends.

https://www.contemporaryobgyn.net/view/fda-approves-zoliflodacin-nuzolvence-for-uncomplicated-gonorrhea

Intel has tested chipmaking tools from firm with sanctioned China unit, sources say

 Intel (INTC) has tested chipmaking tools this year from a toolmaker with deep roots in China and two overseas units that were targeted by U.S. sanctions, Reuters reported.

Intel (INTC), which fended off calls for its CEO’s resignation from President Donald Trump in August over his alleged ties to China, got the tools from ACM Research (ACMR), a Fremont, California-based producer of chipmaking equipment, the report said.

Two of ACM's units, based in Shanghai and South Korea, were among several firms barred last year from receiving U.S. technology over claims they have supported the Chinese government’s efforts at harnessing commercial technology for military use and making advanced chips or chipmaking tools. ACM denies the allegations.

The two so-called wet etch tools, used for removing material from the silicon wafers that are transformed into semiconductors, were tested for possible use in Intel's most advanced chipmaking process, known as 14A. That process is due for an initial launch in 2027, the report added.

Reuters could not determine if Intel had made a decision to add the tools to the advanced chipmaking process and has no evidence that the company violated any U.S. regulations.

ACM, according to the report, said it could not comment on "specific customer engagements," but can confirm that "ACMR’s U.S. team has sold and delivered multiple tools from our Asian operations to domestic customers."

It also said it has disclosed the shipment of three tools to a “major U.S.-based semiconductor manufacturer,” which are being tested and some of which have met performance standards, the report added.

https://www.msn.com/en-us/money/other/intel-has-tested-chipmaking-tools-from-firm-with-sanctioned-china-unit-report/ar-AA1Sd0sE

Arcus, Gilead to Discontinue Phase 3 Trial of Potential Cancer Treatment

 

Arcus Biosciences and Gilead Sciences are discontinuing a Phase 3 trial of an experimental cancer treatment after an interim review showed the therapy was unlikely to help patients live longer.

Arcus said Friday that the treatment, STAR-221, failed to improve overall survival for patients with advanced gastric and esophageal cancers compared with currently available therapies.

An independent data-monitoring committee recommended that the companies stop the study due to futility.

Shares of Arcus fell 11% to $22.48 in premarket trading. Gilead's stock edged down 0.3% to $122.88.

The companies said the drug's safety profile was consistent with existing treatments and no new safety issues emerged.

Arcus and Gilead said they are working with trial investigators on next steps for patients and plan to conduct a fuller analysis of the data.

"We are disappointed with this outcome," said Arcus Chief Medical Officer Richard Markus, adding that the company remains focused on developing new treatments for cancer and immune-related diseases.

https://www.marketscreener.com/news/arcus-gilead-to-discontinue-phase-3-trial-of-potential-cancer-treatment-update-ce7d50d8d880f22d

'American hyperscalers: bearish catalysts are piling up'

Viewed through an industrial and financial lens, several bearish catalysts are indeed starting to pile up: AI cloud profitability, energy constraints, the scale of CapEx, ecosystem circularity, creative accounting on depreciation and rising debt. However, none of these factors calls into question the considerable technological progress that AI represents. However, their combination materially shifts the risk/return profile of these tech companies, which are tipping from an historically asset-light model into a heavy-industry logic with, in the process, accounting and narratives that are at times more opportunistic than convincing.

Here the aim is not to be alarmist, but rather to lay out the facts. Because at this stage of the cycle, I sincerely think it will be the most down-to-earth constraints - electricity, hardware obsolescence, industrial execution, and accounting - that will dictate stock returns.

The AI cloud business: a growth engine, but murky profitability

At MarketScreener, our architecture is hybrid. Our platform relies mainly on servers that we own, although we also rent infrastructure - notably for AI projects, where we need those famous GPUs.

For example, a server with a V100 GPU (16 GB of VRAM) rents for around €400 or €500 per month. Given that these chips date from 2017 and cost about €10,000 bare, the investment is certainly paid back for cloud providers. However, today, hardly anyone rents these V100s, as they do not support the latest software required to run the best LLMs efficiently. You are more likely to use an L40 GPU, launched in 2022, or an A100.

And that is the crux: both hardware and software development have accelerated. New chips are arriving faster and faster - Nvidia is on roughly an annual cadence - and the open-source community is continuously innovating on the software layer. The upshot is that infrastructure bought today may be technically OK in three years but commercially obsolete versus the efficiency gains of new generations.

It is getting hard to imagine replaying the V100 scenario: renting a chip bought for €10,000 at €500 per month for more than three years. And again, we are talking bare chips, not a full, operational, cooled, cabled AI rack, etc. Nor are we talking energy costs: an AI rack running 24/7 pulls on the order of 100 kW. Far from negligible.

Two things can therefore happen - and we are already starting to see it.

1. The open-source community can adapt software to let the latest models run on prior-generation chips, extending their economic life. Pretty cool.
2. Hyperscalers will raise prices. And that is even necessary if they want to avoid ending up with a ROIC below 5%. Typically, OVH's CEO anticipates 5%-10% cloud price hikes by mid-2026, with internal server costs up 15%-25%, notably due to DRAM/SSD pressure and AI hardware.

And you know what? I think a good chunk of AI cloud demand will not follow if prices jump too abruptly. Simply because monetizing an AI project is not that simple. Prototyping is very easy - and I am well placed to say so - but putting a reliable, scalable, and profitable system into production is another story.

For several months our goal has been to make MarketScreener's search bar intelligent, and I can tell you: it is easier said than done if you do not want to spend €0.10 every time a user hits Enter. And if we dare to contemplate this kind of project, it is precisely because GPU rental pricing still looks affordable to us. But if the price doubles or triples, I sincerely think a large part of demand will not absorb those increases.

The physical bottleneck

Goldman Sachs estimates data center electricity demand will rise by about +165% by 2030, driven overwhelmingly by AI. 

Supply of dispatchable power is not growing anywhere near as fast. The question is no longer even "at what price?" but "will there be enough, in the right place, at the right time?"

In Northern Virginia, the world's top hub, new projects face waits of up to seven years to connect to the grid. And it is not a local exception: US and European grid operators report lengthening queues everywhere AI wants to set up.

Tens of billions of dollars of capacity are being built... that cannot run because they are not hooked up (shortages of transformers, cooling equipment, specialized labor, etc.). And the irony is that these assets depreciate while waiting to be plugged in.

That is the real opportunity cost of AI: not just capex, but industrial downtime.

It is such a mess that hyperscalers are signing or financing SMR nuclear projects (Oklo, X-Energy, etc.). But do not kid yourself: meaningful arrivals there are also more likely post-2030.

This energy bottleneck has two direct effects on hyperscalers:

• Unplanned additional CapEx: they must fund not only data centers but sometimes their own energy solutions.
• ROI delay: potential AI revenues are there, demand is there (for now, and at this price), but deliverable capacity is constrained, so monetization is pushed out.

Investments on a scale never handled by private companies

Make no mistake: we are in an unprecedented historical regime, with hundreds of billions of CapEx per year concentrated in a handful of companies.

I hardly dare cite exact figures for fear of being out of date. From what I recall, we are talking about over $600bn in annual AI investments by 2026-2027.

According to the Wall Street Journal, US AI investment may have accounted for half of the country's GDP growth over the first six months of the year.

Who can execute such investment programs without destroying value?

The players themselves acknowledge it: none of these companies has ever run a $50bn industrial project - and they are now launching a dozen simultaneously. At this scale, the smallest logistics error becomes a money pit, and delays cost fortunes.

Funding these investments is not trivial either. A true shift in capital structure is underway at tech companies we knew as extremely solid financially.

In 2025, more than $120bn of debt was issued by these hyperscalers, a sharp increase versus prior years, and the projected dynamic for 2026-2027 is even stronger.

Oracle is a good example: net debt is now above $80bn and leverage above 3.

Even if, since the end of the shutdown and the publication of macro figures, sentiment seems to have eased in the US, this step-change matters: less leeway if rates remain high, more sensitivity to the economic cycle, and therefore greater dependence on AI's commercial success.

Ecosystem circularity

One point repeatedly raised by analysts around the globe concerns the circularity of economic flows. This incestuous flow pattern leads us to think that part of the growth reported by actors in this ecosystem is not truly synonymous with net value creation.

The scheme is fairly simple:

1. Hyperscalers invest massively in data centers.
2. A large part of AI demand comes from... AI players themselves: start-ups, labs, platforms, model publishers.
3. These players fund their cloud consumption via capital raises in which hyperscalers and semiconductor giants are often shareholders, partners, or exclusive suppliers.
4. Cloud revenues are thus inflated in part by a system where the supplier indirectly finances its own customer.

None of this is catastrophic, but it is worth noting that this circularity enables these players to post spectacular growth even though end demand, from the general public or traditional enterprises, has yet to prove that it is keeping pace.

Why is that fragile?

• Because the loop depends on the cost of capital. As long as money is plentiful and valuations remain high, start-ups can consume cloud at a loss. If market appetite slows, the loop contracts quickly.
• Because revenues are correlated. If one link cuts spend (for example, an AI platform slowing its growth), the hyperscaler sees its cloud revenues slow, reducing its ability to reinvest, which weighs on GPU demand... etc.
• Because end value is not yet guaranteed. The ecosystem is spending enormously today on a promise, but if monetization of use cases takes longer than expected, the whole edifice becomes very sensitive to a regime change.

Creative accounting

Some will call it fraud, others creative accounting. To me it is the most pragmatic point here: hyperscalers have started to assume the useful life of their servers and chips has increased, and are therefore lengthening depreciation schedules even as the product life cycle of AI hardware is shortening. A worrying divergence, suggesting they are booking earnings today at the expense of tomorrow.

Useful life of servers (depreciation schedule) for the three main hyperscalers. The coordinated shift from 3/4 years to 6 years is easy to observe over time. Source: "Why AI factories bend, but don't break, useful life assumptions", SiliconAngle.

In 2023, Alphabet extended estimated useful life for servers from 4 to 6 years and some networking gear from 5 to 6 years. Same for Microsoft.

In 2025, Meta extended to 5.5 years (previously 4-5 years). The same year, Oracle also began depreciating hardware over 6 years.

If you do not grasp the financial mechanism, a quick example:

Suppose a hyperscaler buys $100bn of servers/AI GPUs. If the company deems the useful life of this hardware to be 3 years, it books around $33bn per year of depreciation (assuming straight-line). If it stretches to 6 years, it books only $17bn per year.

Result: +$16bn of operating profit per year in the short term, without cash changing by a cent. It is just income statement timing, because in four years the carrying value of these assets on the balance sheet will be disconnected from economic reality, and the company may be forced to impair the residual value in one go - surprising many investors.

Michael Burry estimates the scale of under-depreciation of these companies' AI assets will reach around $180bn by 2028. According to him, this accounting choice boosted Oracle's profits by 26.9% and Meta's by 20.8%.

Burry is not the only one going out on a limb. The Economist runs with "the USD 4 trillion accounting puzzle at the heart of the AI cloud".

By their estimates, if these assets were depreciated over three years instead of the longer lives now used by companies, annual pre-tax profits would fall by about 8%. And if depreciation truly matched Nvidia's imposed pace (which is extreme and does not really make sense to me - just look at iPhones), the implicit shock to market value could reach $4 trillion.

We can debate the word "fraud," because it is not easy to estimate the useful life of equipment subject to more or less annual innovation. But the economic mechanism is undeniable: pushing the bill into the future while the tech accelerates. That divergence is odd. However, as I said earlier, we are indeed seeing an effort by the open-source community to build software that lets the latest models run on older-generation chips. And it is entirely possible that, in practice, the newest generations of chips prove very resilient, especially in the final phase of their life, for simpler applications (and not training) with clients like us.

In the end, perhaps these warning signs have prevented the AI bet from getting even more congested than it already is. That is Bank of America's somewhat bold theory, which thinks skepticism benefits the daring by leaving other investors on the sidelines. Needless to say, the US investment bank's semiconductor research team has no doubt about the sector's momentum, despite the noise.

 https://www.marketscreener.com/news/american-hyperscalers-bearish-catalysts-are-piling-up-ce7d50d8db8bf224