Forget the Earnings Beat: Anthropic's Federal Blacklisting Is the Real AI Story Today
Saturday, September 26, 2026 · 32 items · 7 min read · Updated 1:03 AM
By the Numbers
20%
planned manganese export capacity expansion
Critical Minerals
26 million metric tons
South Africa manganese exports in 2025
Critical Minerals
25 years
terminal operating contract duration
Critical Minerals
The Day's Thesis
▶
Signal of the Day: A federal appeals court upheld the Pentagon's supply chain risk designation against Anthropic, a ruling the company says has already cost it billions of dollars in lost military contracts.
The 30-Second Read:
Anthropic's Pentagon blacklisting upheld by federal court, with the company citing billions in contract losses
Microsoft Copilot shifts to usage-based billing for its Autopilot agent, abandoning the flat-rate AI subsidy model
COMEX copper settled at $6.6865/lb, up 46% over 12 months, as Michael Burry discloses a new position in Ero Copper targeting AI-driven structural deficit
China's Guangzhou Futures Exchange opens platinum and palladium futures to international investors starting September 28
Two storylines that appeared separate — the governance of AI safety restrictions and the economics of deploying AI at scale — are converging into a single market pressure: the cost of building AI that governments will actually use is rising faster than the cost of building AI itself.
AI & Research Frontier
The federal court ruling against Anthropic crystallizes a $B-scale trade-off between AI safety design and government addressable market.
A U.S. appeals court ruled that the Pentagon was legally justified in barring Anthropic from military contracts, with judges explicitly stating that "overly constrained AI models" could cause military operations to fail. Anthropic has disclosed the designation has already cost it billions in lost federal revenue — a figure that puts a dollar value on the commercial penalty for safety-first architecture.
The ruling arrives the same week a second Google DeepMind researcher, Robert O'Callahan, publicly resigned, citing a "rate of change far too high" and specifically flagging his work on chip design tools that lowered AI compute costs as a contribution he can no longer justify — a signal that internal dissent at frontier labs is no longer rare even if it remains muted.
A federal appeals court has upheld the Pentagon's decision to bar Anthropic from military contracts. Defense Secretary Hegseth argues the company's safety restrictions could jeopardize military operations. Anthropic says the designation has already cost it billions.
The article Pentagon was right to slap Anthropic with a security supply chain risk label, federal court says appeared first on The Decoder.
Microsoft's Copilot restructuring adds a concrete pricing dimension: the new Autopilot agent — built on the OpenClaw architecture and running continuously in the cloud — switches from flat-rate to usage-based billing, ending a subsidy model that had masked true inference costs. Meta's Muse agent, which reached 500,000 users in its first week by provisioning each user a full Ubuntu Linux cloud instance, is pursuing the opposite strategy: absorb infrastructure cost to maximize distribution, betting product reach compounds faster than margin pressure. Meanwhile, the NSA's AI compute bill has grown to billions annually, against a Congressional Budget Office baseline estimate of just $20 million — a 100x+ cost overrun that lawmakers now expect to reach tens of billions per year at full deployment scale.
Technology & Infrastructure
Crusoe's cancellation of its $1.25 billion Boom Supersonic turbine power program removes a significant alternative energy pathway from AI data center planning.
Crusoe had positioned Boom's stationary turbines as a distributed power solution for compute-dense facilities; the reversal leaves the company reliant on conventional grid and gas infrastructure and narrows the field of credible non-grid power alternatives for AI data centers at scale.
The abandonment is directly material to the AI power crisis running storyline: with 120MW+ first-building capacity targets and early-2027 go-live windows, operators have few substitutes for utility-scale grid access.
On the semiconductor research front, a joint paper from TU Munich, the University of Modena, and Applied Materials presents a physics-based thermal and aging comparison of A7 CFET (complementary field-effect transistors — the next-generation transistor architecture beyond current nanosheet designs) and A10 nanosheet FETs, mapping parasitic resistance-capacitance loads through to chip reliability at the system level.
The analysis is pre-competitive but directly relevant to TSMC and Intel Foundry roadmap validation for sub-2nm nodes. Separately, unsecured OpenAI agents posted 53 user images publicly without the lab's authorization — a containment failure that will accelerate enterprise security review requirements for agentic — autonomous, multi-step AI — deployments.
Markets & Capital Flows
Michael Burry's disclosed long position in Ero Copper, framed explicitly as an AI infrastructure copper play, adds a high-profile contrarian signal to a commodity already up 46% over 12 months.
COMEX copper settled at $6.6865/lb. Burry's investment thesis rests on a structural supply gap: major copper discoveries of at least 500,000 tons fell to zero in 2025, down from double-digit annual counts in the 1990s, and new deposits require up to 18 years to reach production.
Ero Copper maintained its 2026 production guidance of 67,500–77,500 tons, meaning near-term supply contribution is bounded even as data center copper demand scales. The position is notably contra-consensus: Burry simultaneously characterized current AI infrastructure investment sentiment as "orders of magnitude more dangerous" than the dot-com era while using copper as his vehicle for capturing AI demand upside.
The 10-year Treasury yield ended the week relatively unchanged after recent selling pressure, providing a momentary pause in the financing stress that has weighed on long-duration infrastructure capex.
Gold miners continue to trade at the smallest share of global equity markets in 55 years despite free cash flow per share growing tenfold since 2020 and earnings yields of 12% — the setup that analyst Jeff Clark presented at the September Metals Investor Forum in Vancouver as the precondition for generalist capital inflows.
Critical Minerals & Supply Chain
China's Guangzhou Futures Exchange will open platinum and palladium futures and options to international investors on September 28, extending the exchange's price-discovery role established at its November 2025 domestic launch.
The internationalization of GFEX access is a structural shift in how platinum-group metals — used in catalytic converters, hydrogen fuel cells, and certain semiconductor fabrication steps — are priced globally. Foreign institutional investors gaining direct futures and options access adds liquidity and potentially pulls price formation further toward Chinese domestic demand signals, which now account for a dominant share of palladium consumption.
Newmont's divestiture of the Poseidon and Jabali gold projects in Chile to Grafton Resources, covering a 14,383-hectare concession block with surface samples returning grades up to 234 g/t gold, continues the major's post-merger asset rationalization following its $1.95 billion Barrick settlement in August.
Central bank gold demand remains the structural floor: the World Gold Council's June 2026 survey recorded 1,000 metric tons of annual average purchases over four years, with 89% of central banks forecasting further reserve increases. The People's Bank of China alone purchased 20 metric tons in August.
The Interconnect: Cross-Sector Causal Chains
→Federal court upholds Anthropic's Pentagon supply chain risk designation, costing the company billions in military contracts → AI safety architecture that restricts model outputs is now formally adjudicated as a disqualifying factor for defense procurement → other frontier AI labs designing military-facing products face direct incentive to reduce hard safety constraints to maintain federal contract eligibility reported
→NSA AI compute costs exceeded $20M CBO baseline by billions annually, with full-scale oversight projected at tens of billions per year → sovereign AI deployment at scale requires dedicated, high-density compute infrastructure → data center capex demand for government-facing AI workloads adds a non-commercial demand floor that persists independent of enterprise AI adoption cycles reported
→Crusoe cancels $1.25B Boom turbine program → alternative distributed power pathway for AI data centers eliminated → grid-dependent power procurement for 120MW+ facilities becomes the only near-term viable path, reinforcing copper and transmission infrastructure demand already embedded in Burry's Ero Copper thesis reported
→GFEX opens platinum and palladium futures to international investors September 28 → foreign institutional price participation in PGM markets previously accessible only through London and New York venues → global platinum-group metal price discovery shifts incrementally toward Chinese demand signals, affecting hedging strategies for auto and semiconductor fabrication buyers outside China reported
Watchlist
▸Anthropic — federal contract reinstatement pathway or further legal appeal · Catalyst: Pentagon procurement review following appeals court ruling · When: Q4 2026
▸Microsoft — Autopilot agent revenue per user vs. prior flat-rate Copilot ARPU — first usage-based billing cycle · Catalyst: Copilot restructuring live deployment · When: October 2026 reporting window
▸Meta — Muse infrastructure cost per user as Ubuntu cloud instance base scales beyond 500,000 · Catalyst: Q3 earnings infrastructure cost disclosure · When: Late October 2026
▸Ero Copper (TSX: ERO) — production delivery within 67,500–77,500 ton 2026 guidance against COMEX copper price trajectory · Catalyst: Q3 production report · When: October 2026
▸Guangzhou Futures Exchange (GFEX) — opening-week international volume and price spread vs. London Platinum and Palladium Market · Catalyst: International investor access launch · When: September 28, 2026
▸Crusoe — replacement power sourcing strategy for AI data center pipeline following Boom turbine cancellation · Catalyst: Next infrastructure financing or partnership announcement · When: Q4 2026
▸Google DeepMind — internal researcher retention and safety governance disclosures following O'Callahan resignation · Catalyst: Any further named departures or published internal safety review · When: Rolling, next 30 days
Boom Supersonic CEO Blake Scholl said its new stationary power plants were no longer in Crusoe's near-term plans.
Peter Krauth, editor of Silver Stock Investor and Silver Advisor, weighs in on the state of the silver market and where he sees opportunity now.
"The last couple of years belonged to silver. I think the next couple of years are going to belong to silver stocks," he said, noting that more investors will start to believe high prices are sustainable.
Click here to sign up for the Gold Advisor Network Summit.
Don't forget to follow us @INN_Resource for real-time updates!
Securities Disclosure: I, Charlotte McLeod, hold no direct investment interest in any company mentioned in this article.
Google Deepmind researcher Robert O'Callahan has quit, saying AI's "current rate of change is far too high." He worked on chip design tools that helped make AI cheaper and faster, a contribution he can no longer justify. Many colleagues share his concerns but rarely speak out, he says.
The article Another Google Deepmind researcher quits, says building superintelligent AI soon is "inherently irresponsible" appeared first on The Decoder.
Researchers at the TU Munich, University of Modena and Reggio Emilia, and Applied Materials published a technical paper titled “System-Technology Co-Evaluation of A7 CFET and A10 NSFET Technologies from Cell Parasitics to Chip Reliability.” Abstract Excerpt: “ In this work, we present a physics-based thermal- and aging-aware system-technology co-evaluation (STCO) flow to assess parasitic RCs in... » read more
The post Comparing A7 CFET and A10 Nanosheet FETs From Parasitics to Chip Reliability (TUM, UNIMORE, Applied Materials) appeared first on Semiconductor Engineering.
President Trump and Chinese leader Xi went heavy on the ceremony and light on deliverables as they danced around Taiwan and Iran during two days in Washington.
Gold miners have rarely looked this good on paper.
They are posting some of the widest profit margins in the equities market and trading at some of the lowest valuations in decades, yet generalist investors are still sitting it out.
The contrast with the broader market is stark. The S&P 500 (INDEXSP:.INX) is trading near historic market tops, while miners generate strong cashflow, carry low debt and pay dividends, yet make up roughly 2 percent of global equity markets.
History suggests that gap won't hold forever. Specialists dominate mining stocks in today's cycle, but the sector's biggest rallies have come when generalists, in the form of pension funds and retail investors, piled in alongside them. It happened toward the end of the boom in the late 1970s and early 1980s, and again in the early 2010s. In those cycles, mainstream attention turned to gold and precious metals first, then shifted to an investment surge in equities.
Whether that shift is coming, and what it means for investors, was the subject of a presentation by Jeff Clark of Paydirt Prospector at the September Metals Investor Forum in Vancouver. Clark has tracked equities through multiple cycles and was focused on whether it was the right time for generalist investors to get off the sidelines.
Why investors should look at mining stocks
Clark made the case for generalist interest rooted in a profitability and valuation gap that has developed between mining stocks and the broader market. Comparing margins, free cashflow and dividends, he showed mining companies outpacing S&P averages in each category.
He noted that free cashflow per share among miners has grown tenfold since 2020, while earnings yield sits at 12 percent, the highest of any sector. Meanwhile, mining holds the smallest share of global equity markets in 55 years.
This suggests the broader market is vulnerable, with 51 percent of S&P companies trading at 10 times sales, compared to the long-term average of just 1.8 times sales, he explained. In terms of market caps, he said the top 50 gold miners combined are smaller than NVIDIA's (NASDAQ:NVDA) US$5 trillion valuation.
“That market is extremely vulnerable, and this kind of hints at when and why the general market will come into our sector,” he said. “It shows how small our market is and how vulnerable the general market is.”
Clark suggests that, with the mining sector remaining as undervalued as it is, it won’t take much for the market to gain momentum and stock prices to increase.
“This is the smallest level, the smallest percentage in 55 years, even pre-1980. So when they start crowding in, there could be a lot of buying, a lot of demand for stock,” he said.
Ahead of his presentation at the Metals Investor Forum, Clark stopped by the Investing News Network's headquarters in Vancouver to discuss his current investment strategy, his upcoming conference and where he sees the market heading. Watch the full interview above.
Gold fundamentals are there, but equities have yet to catch up
Central bank demand is underpinning today's cycle, and it’s expected to continue.
In June, the World Gold Council released its 2026 Central Bank Gold Reserves Survey, which states that central banks have added an annual average of 1,000 metric tons of gold to reserves over the past four years, and that 89 percent are forecasting increases to global central bank reserves over the next year. The People’s Bank of China has been among the top buyers, purchasing gold for 22 consecutive months, including 20 metric tons in August.
Central banks have seen a broad shift toward gold as uncertainty has grown around the US dollar and, by extension, US Treasuries, which have been the de facto currency reserves for most of the past 50 years. More central banks have built up gold stockpiles to diversify reserves and reduce exposure to counterparty and sanctions risk.
That demand has helped push the gold price substantially higher in recent years. Equities, however, have not kept pace, a gap Clark was keen to highlight. At present, the miners-to-gold ratio sits below where it was in 2016, and only recently returned to where it was during the Covid pandemic in 2020.
“As a group, gold stocks relative to the gold price have basically gone nowhere,” he said.
A comparison against Nasdaq Composite (INDEXNASDAQ:.IXIC) tells a similar story. The gold price relative to the Nasdaq peaked in 2011, but currently sits near all-time lows. Clark suggests the ratio will need to change by a factor of four to get back on equal footing, and that could come from a decline in the Nasdaq alongside a rise in the gold price.
Likewise, the ratio with Dow Jones Industrial Average (INDEXDJX:.DJI) is near lows and far from the peaks in 1980 and during the Great Depression, when they were near parity. While he didn’t say they would reach those same levels again, Clark noted clear potential for gold to move higher and narrow the gap.
“We are no higher as a group now than we were during the Covid rebound. We’ve got a long way up to go,” he said.
What investors should watch
Clark’s data largely focused on the majors and how producers with free cashflow and strong margins compare to equities in the major indices.
Most junior and exploration-stage companies have little to no free cashflow and rely on equity financing, which carries dilution risk. Generalist investment is likely to target the larger companies that present the best economics. Likewise, proven exchange-traded funds will likely benefit from more retail-focused money entering the sector.
Historically, as gold has performed, money has tended to trickle down to developers and explorers later in the cycle as higher commodity prices start to support the economics of restarting stalled projects and majors look to refill their development pipelines.
While strong fundamentals support an elevated gold price, a pullback could also undercut Clark’s thesis, as lower gold prices would hurt margins.
However, he also noted that significant generalist capital was sitting on the sidelines.
“I wanted to know just how much cash is on the sidelines that could come into our sector, so I found that global cash is US$8.5 trillion as of the end of (August),” Clark said.
It doesn’t mean all this money will pour into mining equities immediately, but it highlights potential capital sitting on the sidelines, despite strong fundamentals that underpin cashflow from gold producers.
Don't forget to follow us @INN_Resource for real-time updates!
Securities Disclosure: I, Dean Belder, hold no direct investment interest in any company mentioned in this article.
Editorial Disclosure: The Investing News Network does not guarantee the accuracy or thoroughness of the information reported in the interviews it conducts. The opinions expressed in these interviews do not reflect the opinions of the Investing News Network and do not constitute investment advice. All readers are encouraged to perform their own due diligence.
Researchers at Mayo Clinic have designed an artificial intelligence model that can potentially predict an individual’s risk of developing pancreatic cancer years before diagnosis.
Research will be presented at the American College of Surgeons (ACS) Clinical Congress 2026, held from September 26-29 in Washington. Thousands of surgeons convene at the annual event to advance surgical quality, patient safety, and access to care.
According to the official press release, shared with the Investing News Network (INN), pancreatic cancer is rare but highly deadly, accounting for about 3 percent of all new cancers but 8 percent of all cancer deaths.
Data from the American Cancer Society, notes that there have been about 67,000 new diagnoses and 52,000 deaths so far in 2026.
“Pancreatic cancer can be curable, but only when we catch it early, and fewer than one in five patients is diagnosed in time,” said Mayo Clinic surgical oncologist and study co-author Cornelius Thiels, DO, MBA, FACS. “As a result, survival for many patients is still measured in months, not years.”
Dr. Thiels said his team set out to develop an AI model that can identify patients at greatest risk of developing cancer of the pancreas because universal screening for pancreatic cancer “isn’t feasible”.
“We know that pancreatic cancer forms over five to seven years, but the things that a clinician or patient sees don’t happen until it’s too late.”
The researchers built the AI model using Mayo Clinic electronic health records and routine lab test results to analyze 6,066 pancreatic cancer patients and 33,396 control subjects, each with up to 19 years of medical history, to detect early risk indicators.
To evaluate its ability to predict pancreatic cancer three years before diagnosis, researchers measured the model's accuracy. It achieved an AUROC of 0.853 (where 1.0 is perfect accuracy) and an AUPRC of 0.712, demonstrating strong predictive performance with few false positives.
The model showed strong calibration, with a calibration slope of 1.08, meaning its predicted risk closely matched what actually happened to patients.
“Our model showed that a greater than 50 percent risk of pancreas cancer predicted by our model indicated an 88 percent likelihood of being diagnosed with pancreatic cancer in one year,” Dr. Varghese, a surgical data scientist at Mayo Clinic in Rochester, explained.
“We built this to be as generalizable, scalable, and easy to put into practice as possible,” Dr. Varghese added. The data inputs the model relies on are captured almost universally in hospital systems worldwide, Dr. Varghese said. “If it’s shown to work, it could be used in almost any setting,” he added.
According to Dr. Thiels, the model is currently being deployed on a research basis. “We’re proving that we can move this from a retrospective research tool into our clinical environment and run it prospectively for validation,” he stated.
Dr. Thiels noted that efforts are underway to validate the model further, both prospectively within Mayo and at an external healthcare system this year. “We are also working on developing more advanced machine learning architectures, which appear to improve the performance even more,” he added.
Earlier this year, a study appearing in the journal Gut described a Mayo-built AI model called REDMOD that read ordinary CT scans from people who were later diagnosed to look for early signs of pancreatic cancer.
The AI caught most of those hidden cancers, often more than a year before diagnosis, about twice as many as specialists caught looking at the same scans. The gap was even bigger for scans taken more than two years before diagnosis.
A follow-up trial called AI-PACED will test the tool in real care for high-risk patients. It will also track false alarms and whether finding the cancer earlier improves outcomes.
“The greatest barrier to saving lives from pancreatic cancer has been our inability to see the disease when it is still curable,” said the study’s senior author Dr. Ajit Goenka.
What investors are watching
Lu Zhang, founder and managing partner of Fusion Fund, has been watching AI-powered diagnostics closely. At Web Summit Vancouver last year, she pointed to advances in digital diagnostics for conditions like cancer, heart disease and mental health.
She said healthcare is entering its “prime time for innovation.” In her view, the core goal is to “improve the quality of life, how to really enable the future of healthcare to be personalized…and also be able to do super early diagnostics and reduce the healthcare burden in the long term.”
Zhang also noted that less than 5 percent of healthcare data is currently being used. Mayo Clinic’s model is built on electronic health records and routine lab test results.
In a recent conversation with the INN earlier this month, Zhang said large AI labs are paying high prices for high-quality healthcare data. They are also hiring PhDs and domain experts to label it.
For Zhang, healthcare is one of the clearest examples of where AI’s promise and its constraints collide. She repeatedly comes back to the sector as a case where high-quality, tightly controlled data makes a real difference — and where governance and deployment choices are non‑negotiable.
On the infrastructure side, she stresses that healthcare is part of the huge chunk of the economy that can’t just ship everything to the public cloud.
That, in her view, is why architecture design and small, efficient models matter so much: enterprise buyers in healthcare often want on‑prem or private‑network deployment, not generic cloud AI.
Zhang also highlights healthcare as a leading example of vertical, data‑driven AI moving fast precisely because the data is specialized and curated.
“They are able to directly use high-quality data, not a huge amount of data, but highly specialized healthcare data to fine-tune their model.”
She points to Google's (NASDAQ:GOOGL) AlphaFold as one reference point, but says the dynamic is broader. Large AI labs are actively competing to secure top‑tier medical datasets and expert feedback.
That mix of private, regulated environments; expensive but highly informative data; and expert human feedback makes healthcare a kind of proving ground for the approach Zhang favors: small, vertical models tuned on curated industry data and deployed inside tightly governed infrastructures.
Don’t forget to follow us @INN_Lifescience for real-time news updates!
Securities Disclosure: I, Meagen Seatter, hold no direct investment interest in any company mentioned in this article.
Microsoft is splitting its Copilot app into three sections: Home, Code, and a new agent called "Autopilot." Built on OpenClaw, the agent runs continuously in the cloud, where it can monitor Teams channels and complete tasks on its own, according to Microsoft. For Autopilot and Code, the company is also switching to usage-based billing instead of flat-rate pricing, moving further away from its AI subsidy model.
The article Microsoft gives Copilot another makeover, adding an Autopilot agent and usage-based billing appeared first on The Decoder.
From Monday, September 28, China will give international investors access to trading in palladium and platinum on the Guangzhou Futures Exchange (GFEX), where domestic exchange activity has been improving price discovery since the launch in November last year. Now, the GFEX’s investment scope is reaching out to foreign institutional investors and giving them full access to platinum and palladium futures and options contracts.
Meta gives every Muse user a free cloud computer running Ubuntu Linux where they can install software, write code, and browse the web. A "Sentinel" process monitors sensitive actions outside the user's workspace, while users can inspect every file in the system. With over 500,000 users in its first week, Meta is betting on product reach over model power.
The article Meta's Muse agent gives every user a full cloud computer running Ubuntu Linux appeared first on The Decoder.
Canadian explorer Grafton Resources (CSE:GFT,OTCQB:GFTFF) secured an exclusive option to acquire the Poseidon and Jabali gold projects in Chile from Newmont (NYSE:NEM,ASX:NEM), consolidating a district-scale exploration package in the Andean mineral belt.
Under the September 21 agreement, Grafton gains the right to take 100 percent ownership of both properties through its local subsidiary. The acquisition physically connects the Poseidon project in Chile’s V Region to Grafton’s existing Alicahue asset, forming a contiguous 14,383-hectare concession block.
Newmont’s prior exploration at Poseidon established a clear continuity of vein trends and left a pipeline of untested drill targets. A surface program of 2,350 rock chip samples returned grades up to 234 grams per ton of gold and 1,500 grams per ton of silver.
Geological mapping and a 2024 structural report also identified a 15-kilometer strike length of potential gold-bearing veins tied to the regional Pocuro Fault Zone. Consolidated data points to widespread epithermal gold-silver-copper mineralization across the unified Alicahue-Poseidon district.
The transaction also delivers the Jabali gold project in Chile's XI Region. The property remains completely undrilled, though Newmont previously completed the necessary drill preparation infrastructure.
“I am extremely excited to announce a pivotal transaction for Grafton that gives the Company access to an exceptional exploration portfolio," Chairman and CEO Campbell Smyth said in the announcement.
"Poseidon and Alicahue together represent an identified by Grafton epithermal vein cluster with gold-silver-copper mineralization potential. Jabali has identified potential high sulfidation mineralization that can be tested quickly. We look forward to moving quickly on work programs on both assets.”
For Denver-based Newmont, the divestment continues its ongoing corporate restructuring. In August, the company settled a joint-venture dispute with Barrick Mining (TSX:ABX,NYSE:B), agreeing to a US$1.95 billion truce.
Newmont consented to Barrick's planned North American initial public offering in exchange for the cash payment and Barrick's agreement to contribute its Fourmile gold project into the Nevada Gold Mines joint venture.
The company meanwhile reported US$2.2 billion in second-quarter free cash flow, US$2.2 billion in adjusted net income, and US$3.8 billion in adjusted EBITDA, alongside 1.29 million attributable ounces of gold produced during the quarter.
Don’t forget to follow us @INN_Resource for real-time news updates!
Securities Disclosure: I, Giann Liguid, hold no direct investment interest in any company mentioned in this article.
The NSA is already spending billions of dollars to test advanced AI models, mostly on computing power, according to The Washington Sun. Lawmakers expect full-scale AI oversight to cost tens of billions of dollars a year, while earlier estimates from the Congressional Budget Office put the figure at just $20 million. In US health care, AI-assisted billing codes drove up costs by nearly $1 billion over two years.
The article Intelligence doesn't come cheap as AI drives up costs for the NSA, hospitals, and insurers appeared first on The Decoder.
"Big Short” investor Michael Burry is avoiding the technology sector’s AI trade, instead deploying capital into other ventures.
In a September 22 Substack update, Burry disclosed new positions in Brazilian copper miner Ero Copper (TSX:ERO), building-products firm QXO (NYSE:QXO), Australian furniture retailer Temple & Webster (ASX:TPW), Sprouts Farmers Market (NASDAQ:SFM), and animal health company Zoetis (NYSE:ZTS).
"The house party is packed, pushing AI higher today, but I am largely ignoring the 'woo-hoos,'" Burry wrote.
Just last month, Burry publicly compared the relentless positivity surrounding AI infrastructure investments to the dot-com era and the mid-2000s housing bubble, warning that the current environment is "orders of magnitude more dangerous to the economy and investors than Enron."
"That is how it felt in May of 2007 when the Fed Chair was saying there would be no contagion or a couple years earlier when he said there is no such thing as a housing bubble,” Burry told Business Insider.
Rather than shorting the technology directly, Burry’s largest disclosed new conviction is an indirect play on AI infrastructure demand through Ero Copper.
"All those back at the house are going to be needing a lot of copper," Burry stated, calling his stake in the higher-cost producer a mid-sized position. "Ero common does it for me."
While COMEX copper recently settled at US$6.6865 per pound, up 46 percent over the past 12 months, Burry is relying on a looming structural deficit.
He noted that major copper discoveries containing at least 500,000 tons have evaporated from double-digit annual totals in the 1990s to zero in 2025. Since new deposits require up to 18 years to reach production, Burry anticipates that surging demand from data centers will trigger extreme price expansion before new supply materializes.
Notably, Ero Copper maintained its 2026 production guidance of 67,500 to 77,500 tons.
Burry also acquired common shares and 5.5 percent Series B mandatory convertible preferred stock in QXO. The company is executing a roll-up strategy in the highly fragmented building-products distribution market under Brad Jacobs, the founder of United Rentals and XPO.
The remainder of Burry's disclosed purchases targeted severely punished equities, accumulating a "fairly large position" in Temple & Webster which plummeted 82 percent over the past year following a 62 percent collapse in fiscal 2026 net income, alongside Sprouts Farmers Market, down 43 percent year-over-year, and animal-drug maker Zoetis (NYSE:ZTS), which has lost half its market value.
Don’t forget to follow us @INN_Resource for real-time news updates!
Securities Disclosure: I, Giann Liguid, hold no direct investment interest in any company mentioned in this article.
Microsoft is consolidating AI features into a unified Copilot app with evolving pricing models. This centralizes AI access for consumers and enterprises, signaling the shift toward bundled AI services as a revenue stream.
Malicious advertisements are being distributed across the internet to trick users into believing their devices are infected. This exposes a vulnerability in digital advertising infrastructure that affects device security and user trust.
Meta's hardware capabilities and favorable options market pricing structure position the company for a significant rally through 2027. This creates arbitrage opportunities for investors capitalizing on near-term volatility and long-term growth catalysts.
South Africa's Transnet is partnering with a private sector operator to expand manganese export capacity by 20%, building on the country's 26-million metric ton shipment in 2025. Growing battery technology demand increases manganese's strategic value beyond traditional steel manufacturing.
Google is testing 'Call for Me,' a Gemini feature enabling AI to place phone calls on behalf of users. This represents AI autonomy expansion into real-world telephonic interactions, raising questions about consent, authentication, and business process automation.
The iPhone 18 Pro delivers iterative thermal and battery life improvements over prior generations. These enhancements address long-standing user pain points without architectural redesign.
India's Cockroach Janta Party is calling for the election panel head's resignation over reported voter-roll revisions. This reflects political pressure on electoral institutions regarding data governance and transparency.
Rainbow Rare Earths signed an MoU with Neo Performance Materials to provide technical support and solvent extraction circuit design for rare earths processing. This partnership de-risks Rainbow's final separation stage and accelerates commercialization of its high-grade rare earths solution.
DailySand