September 12, 2026: Friday Bounce Not Enough

Last week, Dow Jones lost 1.6%, S&P 500 was down 0.8%, Nasdaq composite fell 0.7%. Despite a Friday bounce, markets remained weak as Treasury yields and oil prices rose. Nvidia is in talks to invest up to US$10 billion in Anthropic’s IPO, which could raise as much as US$100 billion at a ~US$2 trillion valuation. Oracle grew revenue 30% to US$19.3 billion and operating income 57% to US$6.7 billion, though interest expense rose 55% to US$1.4 billion and it burned US$5 billion on data centres, shares rose 7%. Adobe revenue rose 13% to US$6.76 billion with AI ARR past US$650 million (+150%), but ~10% forward guidance sent shares down 2%. OpenAI paused new US$200/month Pro subscriptions driven by high Astra demand. The Company said a price cut drove a tenfold usage increase. OpenAI is also co-developing chips with Samsung. Meta launched Muse, its first consumer AI agent, which is free to use for up to 100 million tokens weekly with US$20 and US$100 tiers. Berkshire Hathaway CEO Greg Abel told CNBC the company is pursuing artificial intelligence opportunities on two paths – energy and Google. Apple launched the foldable iPhone Duo at US$1,999–US$3,199 and raised Pro pricing US$100 on memory costs. In Canada, Cohere is in advanced talks to raise US$2–3 billion at a US$20 billion valuation, nearly 3x its September mark and the largest private Canadian round on record. CIBC committed $2 billion to defence and dual-use SMBs, after RBC’s $1.4 billion scaling fund and BMO’s $70 billion pledge. The US government invested US$200 million in two quantum firms with Canadian roots, D-Wave and Rigetti. In news pertaining to Sophic clients, 01 Quantum reported FQ3 revenue of $64,410 (YTD $1,082,507 versus $333,269) and a $892,044 net loss, ending with $2.88 million cash, while adding a Tier-1 systems integrator channel, a Crypto4A technology partnership and Carleton validation of its Quantum AI Wrapper. Hybrid Power Solutions opened a larger Mississauga plant with dedicated Batt Pack, Spark and Terra lines.

Canadian AI firm Cohere in advanced talks to raise up to US$3-billion, sources say.

Cohere Inc. is in advanced talks to raise between US$2-billion and US$3-billion, including financing from the Canadian government and existing backers, according to four sources familiar with the matter. The deal would value the artificial-intelligence company at US$20-billion. Should a deal come to fruition, it would rank as the largest on record by a private Canadian startup. The valuation would establish the Toronto-based AI developer as one of Canada’s most valuable technology companies. Two sources familiar with the matter said the deal could close as early as next week, though the timing could be pushed back. A number of big corporate events are scheduled next week, including the Canada Investment Summit and the Milken Institute Global Dialogues symposium in Toronto, and All In, the country’s largest AI expo, in Montreal. Cohere chief executive officer Aidan Gomez is scheduled to appear at all three events. Cohere was valued at roughly US$7-billion last September, and has raised funds from Radical Ventures, Inovia Capital, the Healthcare of Ontario Pension Plan, PSP Investments, Business Development Bank of Canada, Nvidia Corp. and others. Cohere was founded in 2019 to build large language models and competes with U.S. tech giants OpenAI and Anthropic, but has differentiated itself. It does not target consumers and instead works with companies and governments, and counts Royal Bank of Canada, BCE Inc. and Fujitsu as customers. Because Cohere’s clients typically run AI models on their own infrastructure, Cohere spends less on compute than its competitors, the company has said. https://tinyurl.com/mr4s3hh8

Clay lands US$115 million Series D at $7 billion valuation.

Canadian-founded, New York City-based go-to-market (GTM) startup Clay has closed another major financing, securing a US$115 million ($159 million) Series D round. Clay develops and sells AI software and agents designed to help other firms automate marketing, sales, and other GTM tasks. The company announced its Series D on Wednesday. It was led by Wellington, with support from Andreessen Horowitz, CapitalG, Meritech, Sequoia, and others. The round values Clay at US$7.1 billion, more than double the US$3.1 billion valuation it was given with its $100 million Series C in August 2025. https://tinyurl.com/yk8u6ut8

Onix raises US$5 million to create AI library of health and wellness “experts”.

Montréal-based startup Onix has raised US$5 million ($6.9 million) for an app that offers a library of health and wellness advice, provided by AI versions of practitioners chosen by the company. Onix announced on Wednesday that its pre-seed convertible note round was led by Los Angeles-based AlphaEdison, with participation from Kitchener-Waterloo’s Garage Capital, New York-based Ride Home Fund. https://tinyurl.com/3w63pm84

CIBC commits $2 billion to financing smaller defence and dual-use businesses.

CIBC announced the commitment on Thursday. The Toronto-based bank said the funding will be targeted towards Canadian defence and dual-use SMBs developing infrastructure, energy, cybersecurity, digital capabilities, and advanced technologies. According to The Globe and Mail, the funding will be deployed over five years. CIBC’s commitment came a day after the Royal Bank of Canada (RBC) launched a $1.4-billion fund for scaling Canadian tech companies, including defence firms, and a day before the Bank of Montréal (BMO) announced its intent to mobilize up to $70 billion over the next decade to finance sectors such as defence. https://tinyurl.com/2mmzszd3

US government invests US$200 million in Canadian-founded quantum firms.

Two quantum computing firms founded by Canadians have each secured US$100 million ($138 million) investments under the US CHIPS and Science Act. After signing a letter of intent back in May, BC-founded and California-headquartered D-Wave announced on Tuesday that it received a strategic investment from the US government, giving the US Department of Commerce a minority, non-controlling equity stake in the company. Berkeley, Calif.-based Rigetti Computing, which was founded by Moose Jaw, Sask.-born former CEO Chad Rigetti, received a similar deal. From the source: D-Wave CEO Alan Baratz said in a statement that the funding will bolster US domestic quantum capabilities, strengthen the quantum supply chain, and help bring quantum systems to market. “The [US] Administration has made clear that US quantum leadership will require not only breakthrough research, but also the ability to scale, commercialize, and manufacture these technologies,” Baratz said. Founded in 1999 as a spin-out from the University of British Columbia, D-Wave relocated its headquarters from Burnaby, BC, to Palo Alto in 2023, but retained a presence in both cities. Alongside D-Wave and Rigetti, Colorado-based Quantinuum also received a CHIPS Act investment on Tuesday. The US signalled it would use CHIPS funding to take equity stakes in quantum companies earlier this year, arguing that doing so would help build a domestic supply chain for quantum chips. Final thought: Canada’s most recent federal budget committed $334 million to the quantum industry, an amount dwarfed by the US$53 billion in federal incentives committed by the US CHIPS Act. Julien Camirand Lemyre, CEO of Sherbrooke-based Nord Quantique, told The Globe and Mail last month that Canada’s public support for the sector may not be enough to keep his company in the country, as those US incentives are a “strong pull” that may force him to relocate. https://tinyurl.com/3ejynvdm

Sophic Client 01 Quantum Inc (ONE-TSXV, OONEF-OTCQB) reports Fiscal Third Quarter 2026 results and highlights momentum across PQC and Secure AI markets.

01 Quantum provided an operational update and reported financial results for its fiscal third quarter 2026, which ended July 31st, 2026. The quarter saw strong capital management, additional partnerships with systems integrators, the addition of a Technology Partnership Agreement with Crypto4A, and continued progress across a portfolio of application ready post-quantum cryptography (“PQC”) technologies. “This quarter continued to demonstrate the progress we are making through new partnerships and independent validation of our technologies,” said Andrew Cheung, CEO of 01 Quantum. “We received an independent validation from Carleton University for our Quantum AI Wrapper (“QAW”) technologies, demonstrating how we can efficiently provide an AI operations platform while preserving user data privacy. Our agreement with a leading Canadian global system integrator, together with our Technology Partnership Agreement with Crypto4A, broadens our distribution channels across enterprise, financial services, government, and defense verticals. During the year, we completed two important customer funded development projects. With these projects successfully completed and the related products commercially available, our partners have commenced sales and marketing activities, and under existing agreements, the Company will receive previously agreed upon royalties based on revenue generated. Combined with our current cash position and disciplined cash management, we are positioned to continue advancing our R&D and marketing efforts. We have built our commercialization framework based on validated technology, real world applications, sector driven demand, and credible Tier-1 partners. These foundations support durable adoption of our platforms and place 01 Quantum in a meaningful position as governments and enterprises transition to quantum-resilient cybersecurity and secure AI execution.” Operational Updates and Pipeline Growth: During the quarter, 01 Quantum advanced several key initiatives: Agreement with Global Systems Integrator. As announced on June 10, 2026, the Company signed a strategic commercialization agreement with a leading Canadian global IT and business consulting integrator. This agreement serves as a major distribution catalyst, establishing a structured channel for introducing 01 Quantum’s PQC and secure AI offerings into large-scale digital infrastructure modernization initiatives across enterprise and public sector accounts. Technology Partnership with Crypto4A. As announced on July 15, 2026, the Company entered into a Technology Partnership Agreement with Crypto4A, an Ottawa-based global Hardware Security Module (HSM) vendor, to coordinate market engagement, customer introductions, and co-marketing initiatives focused on quantum-safe modernization. The partnership aligns 01 Quantum’s PQC and secure AI capabilities with Crypto4A’s expertise in cryptographic agility, secure key management, and high-assurance infrastructure. Independent Technical Validation. As announced on July 28, 2026, academic work released by Carleton University reinforced that 01 Quantum’s QAW technologies are suitable for privacy-preserved AI operations. This validation supports the Company’s strategy of wrapping its technologies around secure AI operations. Support for Japanese Partner. Deployment activity for DoMobile Ver.5 continued to grow through the Company’s Japanese partner, with onboarding expanding across enterprise and regulated sectors. qLABS launches qVAULT. The Company continued to support qLABS as it advances PQC enhanced applications. During the quarter, as announced on June 18, 2026, qLABS launched qVAULT as a core utility within the qONE ecosystem. qVAULT provides quantum resistant storage and transaction capabilities for digital assets, stablecoins, and tokenized digital bonds. As qLABS leads market outreach, qVAULT is expected to increase token utility as users adopt Q-Day protection across digital-asset environments. Financial Highlights: Revenue for the third quarter was $64,410 (2025 – $143,957) and $1,082,507 (2025 – $333,269) year to date. The quarterly revenue variance reflects the conclusion of one-time customer-funded development and integration arrangements as those projects reached completion. The related products are now commercially available, and under existing agreements, the Company receives royalties from its Japanese partner and qLABS, which funded the development, based on revenue generated from these offerings. Net Loss for the third quarter was $892,044 (2025 – $351,411) and $1,789,511 ($773,633) year to date. The adjusted loss for the third quarter, which excludes stock-based compensation of $438,000 (2025 – $80,000) and depreciation was $453,429 (2025 – $269,939). The adjusted loss on a year-to-date basis, which excludes stock-based compensation of $1,158,000 (2025 – $235,000), depreciation and the $295,000 digital-asset revenue recognized earlier in the fiscal year, was $922,991 (2025 – $533,375). Cash operating expenses for the third quarter were $514,074 (2025 – $407,292) and $1,702,683 (2025 – $844,643) year to date. Cash operating expenses consist solely of research and development and selling, general and administrative expenses. The Company completed the quarter with cash and cash equivalents of $2,877,629. With a disciplined quarterly cash flow management and streamlined operating model, the Company maintains a solid balance sheet that supports ongoing operations and its current commercialization efforts. https://t.co/UjBGlT5G2B

Sophic Client Hybrid Power Solutions (HPSS-CSE, HPSIF-OTC) establishes larger manufacturing facility in Mississauga, Ontario to support increased production.

Hybrid Power Solutions has established a larger manufacturing facility in Mississauga, Ontario. The Mississauga facility is intended to support increased production capacity as the Company scales its fuel-free and hybrid portable power platforms. Production at the facility is planned to include dedicated lines for the Company’s Batt Pack product line and for Spark and Terra production, enabling Hybrid to better organize manufacturing by product family and respond to growing customer demand. Francois Byrne, Founder and CEO of Hybrid Power Solutions Inc., said: “Having reached maximum capacity at our previous facility, we felt it was time to expand to the level required for our current and future demand. This new standalone building allows for greater manufacturing automation, improved access control for inventory and defence contracts, and outdoor secured storage for large trailered systems.” https://tinyurl.com/mrhuwsbr

Nvidia may invest up to US$10 billion In Anthropic’s IPO.

Nvidia has discussed investing in Anthropic’s upcoming initial public offering that would raise as much as US$100 billion at a valuation of around US$2 trillion, Reuters reported Friday. Nvidia could invest as much as US$10 billion in Anthropic at the IPO price, according to the report. Anthropic last raised capital at a US$965 billion valuation in May. The five-year-old AI developer is gearing up to go public in the coming weeks. The new capital would help it pay for computing costs, including the more than US$500 billion in agreements to rent servers or lease data center capacity over the next decade or so. Anthropic has become a major user of Nvidia’s AI server chips, though it also uses a substantial number of Google’s and Amazon’s AI server chips. Anthropic also has agreed to buy Advanced Micro Devices server racks and has planned to produce its own chips the way OpenAI has. Nvidia said late last year that it would invest US$10 billion in Anthropic. Nvidia has also been a big backer of Anthropic’s main rival OpenAI, agreeing to invest US$30 billion in the ChatGPT maker in a funding round earlier this year and providing around US$100 billion in credit support for an Ohio data center that would be filled with Nvidia AI servers and leased by OpenAI. Nvidia and Anthropic did not immediately respond to a request for comment. The plans are not finalized and could change, per Reuters. https://tinyurl.com/5n9492t4

Tencent-backed Enflame triples on debut as China pushes for chip self-sufficiency.

Shanghai Enflame Technology surged on its first trading day Friday after the Chinese AI chip developer raised about 6.12 billion yuan (US$911 million) in an initial public offering, underscoring strong investor appetite for domestic alternatives to Nvidia chips. Shares opened at 410 yuan, nearly triple the IPO price of 142.18 yuan, before paring some gains to close around 430 yuan as Enflame sold 43 million shares amid retail demand more than 4,000 times the available allocation. In mainland China’s domestic stock exchanges, first-day pops where share prices double or triple are not uncommon, due to several structural and regulatory reasons that inflate the first-day demand for tech IPOs. Tencent Holdings, which owns about a 20% stake in Enflame, has become its dominant customer, accounting for 84% of revenue last year. Tencent’s cloud and internet platforms deploy Enflame’s AI accelerators across large data centers, according to the company’s IPO prospectus. Those chips power chatbots, recommendation engines and generative AI tools, helping Enflame ride China’s AI computing boom while reducing reliance on U.S. AI chips. But the company still produces chips at overseas foundries, according to the company’s IPO prospectus, because domestic chip manufacturers are not yet able to consistently produce Enflame’s advanced AI accelerators at the volumes and quality it requires. This reliance on foreign fabs highlights the gap between Beijing’s push for semiconductor self-sufficiency and the current state of China’s local chipmaking ecosystem. Enflame remains unprofitable and expects another loss in the first nine months of 2026, according to its IPO prospectus, even as sales are on track to more than triple this year after expanding at an annual rate of over 80% from 2023 to 2025. The company plans to use IPO proceeds to develop its next-generation chips and software, aiming to reach breakeven or profitability by year-end as it scales to meet demand from Tencent and other customers. https://tinyurl.com/2brd2fxy

OpenAI CFO touts user growth resulting from lowering prices.

OpenAI’s price cut on its GPT-5.6 Luna model shortly after its release in July resulted in a tenfold increase in model usage, said OpenAI chief financial officer Sarah Friar in an interview on Monday at Goldman Sachs’ Communacopia + Technology conference in San Francisco. Usage of the model is now “higher than even the next Chinese model,” giving OpenAI the highest market share on OpenRouter, a platform that allows users to access AI models from several different companies in one place, Friar said. Friar’s remarks focused largely on the company’s efforts to bring down the cost for users of its AI models, including by making the chip design process faster. She said that its AI models helped the company send the design for Jalapeño, a custom chip it designed in-house, to Taiwan Semiconductor Manufacturing Company for manufacturing much more quickly than it could have otherwise. Friar suggested that the company could sell its AI-powered chip design capabilities to other customers and price the offering based on the outcomes it delivers rather than charging customers based on their consumption or the number of individuals using the model. “When it comes to cost per token, what [customers] really care about is the cost to get a task done…it means we need to build frontier models where you require less terms to get a correct answer, with less human input,” Friar said. Friar said that OpenAI’s just-released Astra model requires 68% fewer output tokens to get results from OpenAI’s Astra model that are comparable to results from Anthropic’s Fable 5.1. OpenAI released Astra, a model designed to operate a user’s computer autonomously, including by navigating web browsers and generating documents, last week. “We’re investing in areas like chip design, life sciences, financial services, and may even have a banking announcement coming a little bit later this week in New York,” Friar added. https://tinyurl.com/bdd6xecn

Abel: Two ways Berkshire hopes to cash in on AI.

Berkshire Hathaway CEO Greg Abel told CNBC the company is pursuing artificial intelligence opportunities on two paths. In a live “Squawk Box” interview Wednesday morning, Abel said he sees providing energy to the growing number of AI data centers as a “significant opportunity for Berkshire and Berkshire Hathaway Energy” since he’s long believed the biggest constraint for the buildout is having enough energy to operate the power-hungry facilities. But, he said, the company will only sell energy to hyperscalers if there is “no impact to the rates of our other customers.” The other avenue is Berkshire’s now almost US$36 billion investment in shares of Google parent Alphabet that was initiated by Warren Buffett last year. Abel said from the experience of Berkshire’s own operating companies, he and Buffett knew AI was “going to have a significant impact on America and businesses,” and “we saw Google as a significant player.” Abel did the interview from Japan, where he was visiting a Berkshire tool-making unit and meeting with executives of the “trading houses” in which Berkshire has significant equity stakes. He said none of them mentioned rising Japanese interest rates as a “fundamental challenge right now.” https://tinyurl.com/3m2wcv57

Chime strikes deal to buy Stride Bank.

Chime, a fintech company that offers bank accounts and cards to consumers, has struck a deal to buy Stride Bank, an Oklahoma-based institution that has supported Chime behind the scenes for years. Chime on Tuesday said it will buy Stride Bank for US$590 million and will rename the institution Chime Bank. Many fintech companies rely on licensed banks to operate, and Chime’s acquisition will reduce fees associated with such an arrangement. Shares in Chime, which listed on the Nasdaq last year, rallied in after market trading on Tuesday. Chime shares have recently recovered above the IPO price of $27 after sliding in the wake of the debut. Chime said it will fund the acquisition from current cash reserves. Shares in Bancorp, another of Chime’s sponsor banks, fell on Tuesday in after market trading. Chime said it plans to consolidate its banking activities at Stride. The deal is subject to regulatory approvals. https://tinyurl.com/hfjs4nb9

Bending Spoons buys Miro for US$1.355 billion.

Italian digital conglomerate Bending Spoons is buying whiteboarding software firm Miro at a valuation of US$1.355 billion, the latest in a series of purchases the Italian firm has done at knock-down prices. The price for Miro is a huge discount to its peak valuation of US$17.5 billion, dating from late 2021, according to PitchBook. That was before rising interest rates slashed valuations in 2022. Moreover, the emergence of AI-fueled services has further eroded valuations for software services like Miro. Miro’s pitch is that workers can use its software to collaborate, which is a crowded market. Notably, some Miro shareholders are investing US$295 million of the proceeds into new Bending Spoons stock, implying the true cost of the purchase for the Italian company is about US$1 billion. https://tinyurl.com/2c3r4bb9

Oracle reports 30% topline growth for August quarter.

Oracle reported 30% growth in revenue of US$19.3 billion for the quarter ending Aug. 31, a slightly better growth rate than the company projected in June, even as it burned US$5 billion to expand the data centers that are driving its accelerating topline growth. The software and cloud firm said operating income rose 57% to US$6.7 billion, although that doesn’t reflect rising interest costs caused by the extra debt the company has taken on. Interest costs soared by 55% to US$1.4 billion. Oracle projected growth of between 30% and 34% for its second fiscal quarter but maintained its full year growth at 34%. Oracle shares were trading up 7% in after hours trading. https://tinyurl.com/yn3bz8h8

Adobe shares drop 2% despite AI product sales growth.

Adobe Systems shares dipped 2% late Thursday after its third-quarter earnings report, as some investors appear to unsure about how much its AI products will drive future revenue growth. Adobe, which began breaking out revenue for its AI products in March, is continuing to highlight this area of its business. On an earnings call, Adobe chair and CEO Shantanu Narayen said annual recurring revenue from AI products—the value of its contracts over the next 12 months—surpassed US$650 million during its August quarter, up more than 150% from last year and US$150 million from last quarter. Anil Chakravarthy, an Adobe president who oversees its business products and is taking over for Narayen on Dec. 1, said the company is taking a short-term hit to ARR by offering freemium versions of some of its AI products, while betting that some of them will become paying customers and further boost its AI product sales. For the three months to Aug. 28, Adobe reported US$6.76 billion in revenue, up 13% from last year and exceeding its earlier forecast by around US$40 million. The company expects revenue of US$6.8 billion to US$6.85 billion for its current quarter, which would represent growth of around 10% compared to last year. Adobe raised its full-year revenue forecast by around US$75 million to between US$26.576 billion and US$26.626 billion. https://tinyurl.com/4aykfrey

Anthropic believes AI ‘could kill all humans’ within the next decade, researcher says.

Anthropic safety researcher Evan Hubinger said in a post on X Tuesday night that his company does “earnestly believe AI could kill all humans,” adding that he believes there’s a greater than a 10% chance that happens within the next decade. The post immediately sparked outrage and concerns from other researchers and technology developers. Hubinger’s post was a response to a post from another Anthropic safety researcher, Jacob Coxon, who said that he was leaving the company because it and OpenAI are “racing straight to self-improving superintelligence and gambling with our lives.” The strong reactions in part reflect growing concerns that the AI labs are unable to wrangle their AI models, following recent breaches like one that saw a swarm of OpenAI agents break free from their testing environment, hack OpenAI’s infrastructure to take over a research cluster, reach the open internet, and then break into open-source AI repository Hugging Face and other applications. Meanwhile, the leading AI developers continue to emphasize the fast pace of AI progress. OpenAI president Greg Brockman suggested last week that the company’s latest model, GPT-6 Astra, could be artificial general intelligence, or AI that’s able to do most things as well as humans. OpenAI also said over the weekend that it had achieved its milestone of developing an automated research intern, or a “system that can carry out well-defined research tasks under human direction, including tasks that would take a skilled researcher a few days.” At the same time, OpenAI Chief Scientist Jakub Pachocki said in a blog post over the weekend that none of the AI developers had yet solved the problem of how to develop the technology as safely as possible. Hubinger echoed those comments. https://tinyurl.com/4fvn4p43

Meta launches its first consumer AI agent ‘Muse’.

Meta Platforms unveiled Muse, its first consumer-facing AI agent, giving users an AI assistant designed to carry out tasks on their behalf across the web. In a Threads post Tuesday announcing the product, CEO Mark Zuckerberg described Muse as a personal agent that “understands your goals and works 24/7 to get things done for you.” Meta is making Muse available for free with usage of up to 100 million tokens per week. Two paid tiers, priced at $20 and $100 a month, will offer users higher usage limits. Muse was developed internally under the codename Hatch and has been tested extensively by Meta employees. Those tests uncovered some key AI safety and alignment concerns, including cases in which the agent took actions without permission and exposed sensitive information. Meta has spent months addressing these issues as it works to give Muse enough access to act effectively while limiting the risk of harmful mistakes. The Information reported last month that Meta was preparing to release the consumer agent in late August or early September. Internal documents also indicated that Meta had previously considered charging up to US$200 a month for a premium version with higher usage limits. Muse is designed to interact with online services and websites including DoorDash, Etsy, Reddit, Yelp and Outlook, allowing it to perform actions on users’ behalf, such as placing online shopping orders. “For sensitive actions like purchases or sending emails, Muse checks with you first,” Zuckerberg wrote in a post on X. The Information also reported last week that the agent would require user permission before carrying out sensitive actions. However, in a blog post accompanying the announcement, Meta acknowledged that “Muse isn’t immune to attack,” warning that the agent remains susceptible to prompt injection, a security challenge facing AI agents across the industry. https://tinyurl.com/2hteh947

OpenAI to pause new Pro subscriptions, cites Astra demand.

OpenAI said Thursday it will pause new subscriptions to its US$200-a-month plan for Astra, its new flagship model, another sign of the compute crunch facing AI companies. Thibault Sottiaux, who leads OpenAI’s Codex coding agent, cited strains on the company’s systems and a desire to preserve service for existing users. Astra, released this month, has met “unprecedented demand,” Sottiaux said Wednesday on X. “We’re pulling all the levers possible to sustain the demand,” he wrote, adding that he had never seen growth like it. OpenAI billed the model in early September as a major leap in commercial work, touting its ability to handle financial modeling, engineering design, coding and other tasks. Other AI companies have curbed customer use or raised prices in the light of the compute squeeze. In April, Anthropic told developers they could no longer use US$20-to-US$200-a-month Claude subscriptions to power outside agent tools such as OpenClaw. In June, Amazon said it was raising the price for its AI workload rental service by 20%. And Elon Musk has predicted power shortages and other constraints will lead to a compute shortfall next year. https://tinyurl.com/3w5k6xys

OpenAI says its AI agents solved prize math problem.

OpenAI said Tuesday it had solved a problem that mathematicians have been grappling with for more than a hundred years. But the achievement set off accusations that OpenAI had leaned on another researcher’s work to solve the problem. More than 10,000 OpenAI agents worked together to solve the Navier-Stokes Millenium Prize Problem, which concerns fluid motion and has a million-dollar prize attached to it. OpenAI says it found the solution using a “next-generation model significantly more capable than GPT-6 Astra,” its model released to the public Thursday. The proof looks to be the latest breakthrough in AI models’ mathematics capabilities. However, even before OpenAI had announced its discovery, New York University mathematics professor Tristan Buckmaster alleged that OpenAI was spurred to solve the Navier-Stokes problem after hearing rumors of his work and speculation that rival Anthropic had solved the problem. OpenAI CEO Sam Altman addressed those allegations Tuesday, saying in a post on X that OpenAI’s proof approach seemed different from Buckmaster’s. But Altman confirmed that OpenAI had pursued the problem due to rumors that Anthropic had solved it. OpenAI denied that researchers or agents at the company accessed any specific user data in solving the problem. But the company said OpenAI could not rule out that its models had learned from anonymized prompts Buckmaster and Levent had sent the models. https://tinyurl.com/an5ymyw3

OpenAI works with Samsung on next-generation chips.

OpenAI said it is working with Samsung Electronics to develop next-generation chips, deepening a collaboration with the South Korean technology giant that already spans data centers and enterprise sales, Reuters reported. “One of the areas where we have made the most progress and gained the most recognition with Samsung Electronics is our joint production and research on the next-generation chips we are developing,” Harrison Kim, general manager of OpenAI Korea, said at a Wednesday press conference marking the one-year anniversary of OpenAI’s unit in Seoul, according to Reuters. Kim didn’t reveal further details on the chips. Samsung told Reuters that it could not confirm information related to customers. OpenAI has been investing in building its own chips and data centers. In June, it unveiled its first inference chip in collaboration with Broadcom, which it said was developed in nine months with the help of OpenAI’s models. Early testing of the chip, called Jalapeño, showed it delivered performance per watt substantially better than current advanced chips, OpenAI said. Last October, OpenAI signed an agreement with Samsung and SK Telecom to develop data centers, including ocean-based floating data centers, as part of OpenAI’s US$500 billion Stargate initiative aimed at developing AI infrastructure across the globe. The Jalapeño chips are expected to be deployed in these data centers. OpenAI Korea’s Kim said an update on the project will be announced within the year, citing unexpected complexities, according to Korea JoongAng Daily. “Our technology is advancing so quickly that the data center design we envisioned last October is actually somewhat different from what we need now. That has required some adjustments along the way, and it’s taken more time than we anticipated,” Kim said. https://tinyurl.com/mtsmy4wp

Apple launches foldable iPhone Duo.

Apple unveiled a much anticipated foldable iPhone at its first product launch featuring brand-new Chief Executive John Ternus. The iPhone Duo is Apple’s first foldable smartphone—a product category that competitors have been selling for years, but has remained mostly a smaller, niche part of the broader market. While folded, the Duo is around the size of a passport and can fit easily in one hand, with an exterior screen. It unfolds to a length of a 7.6-inch display, closer to the size of an iPad. A video showed how users could use two apps at once while unfolded—texting in iMessage while watching a video, for example. Developers could adapt their apps to fit the new format, such as Netflix letting users browse videos clips on the outer display while closed and then opening up the Duo for the full-video experience. The Duo starts at US$1,999 for a version with 256 gigabytes of storage and climbs to a whopping US$3,199 for one with 2 terabytes. Apple also unveiled the iPhone 18 Pro, starting at US$1,199, and 18 Pro Max, starting at US$1,299. That’s US$100 more than the prior generation’s Pro models, a price hike the company forecasted publicly a few few months ago that it would need to compensate for surging costs for memory and storage chips. At the beginning of the pre-recorded video, Ternus repeatedly talked about the “intelligent personal hub,” a new term Apple is invoking to describe how AI is used to understand users’ personal context. “There’s no product in the world better designed to be your intelligent personal hub than iPhone,” Ternus said, touting Apple’s on-device AI and privacy features. He also added: “Others see that data as something to collect and store. They move it to their servers and ask you to trust them. But honestly, trust only goes so far when your data is no longer yours to control. ” Apple has long emphasized its privacy practices but it is especially focused these days on OpenAI, which aims to launch a family of new AI devices, starting with a portable speaker and later phone sometime in the future. Apple sued OpenAI in July, alleging theft of trade secrets alongside its aggressive poaching of Apple employees with backgrounds in hardware engineering and operations. OpenAI has denied the claims. https://tinyurl.com/3a856vrv

Amazon is helping advertisers buy ads on ChatGPT.

Amazon is letting advertisers test buying ads on ChatGPT via Amazon’s advertising technology platform, Amazon said on Sept. 10. ChatGPT started selling ads earlier this year, which it has let advertisers buy through its own interface and via partnerships with smaller advertising technology firms like Criteo and Kargo. The company said earlier this month that its ad business was generating US$1 billion in annual recurring revenue, which likely means it will finish the year short of an earlier target of US$2.4 billion. Amazon could help introduce ChatGPT to more advertisers. Amazon has been growing its advertising technology business, which allows advertisers to place ads across websites and streaming services, as incumbent Google has retreated in certain categories. Amazon is framing the initial offering as a pilot, with a spokesperson adding it will be only available in the US. The partnership represents a thawing in previously icy relations between the consumer businesses of the two tech companies. Amazon last year blocked listings on its ecommerce site from being scraped and included in ChatGPT. But earlier this year, Amazon invested US$35 billion in ChatGPT parent OpenAI, tied to OpenAI’ spending billions on Amazon’s cloud infrastructure. https://tinyurl.com/a6pe95rb

U.S. accuses DeepSeek, Alibaba and other Chinese AI firms of large-scale distillation.

The U.S. government has accused six Chinese AI firms of using large-scale distillation to copy the capabilities of American models, in a move that highlights the intensifying AI rivalry between the two nations. The six Chinese firms under accusation are DeepSeek, Moonshot AI, Alibaba Group, MiniMax, StepFun and Z.ai. The U.S. National Security Agency, the Federal Bureau of Investigation and the Cybersecurity and Infrastructure Security Agency issued a joint statement on Tuesday to alert U.S. companies and organizations about the Chinese companies’ activities. “While ‘distillation’ is recognized as a legitimate and useful technique in AI research, China-based AI companies are engaging in aggressive, malicious, and targeted distillation activities at an industrial scale that extract restricted proprietary functionalities and capabilities of U.S. frontier AI models,” the agencies said. The agencies, for example, alleged that Moonshot, whose Kimi K3 model has become a global hit, has conducted a distillation campaign against U.S. frontier labs since at least mid-2025. They accused Moonshot of having distilled 17 American models including Anthropic’s Claude Fable 5, which came out in June. A Moonshot spokesperson declined to comment. Distillation has become a sensitive and highly controversial topic due to the AI rivalry between the U.S. and China. As Chinese models have become more powerful, Trump administration officials have criticized Chinese AI firms for alleged distillation attacks. This year, Anthropic has accused several Chinese firms of engaging in large-scale distillation of Claude models. https://tinyurl.com/2nre7cjj

The DOJ is investigating Nvidia’s licensing deal with chip startup Groq.

The Department of Justice is looking into whether Nvidia tried to avoid antitrust scrutiny in its US$20 billion deal to license the technology of chip startup Groq and hire most of its employees in December, the New York Times reported Wednesday. The two companies described the deal as a “nonexclusive” licensing agreement that gave Nvidia access to Groq’s chip technology, rather than an outright acquisition. Now, Groq’s founding team and most Groq employees work at Nvidia developing Groq LPX server racks, designed to run AI models quickly and in tandem with Nvidia’s flagship graphics processing units. Groq still exists as an independent company, and is becoming a cloud provider that offers access to Groq and Nvidia chips. Nvidia is not the first or only company to have done a license-and-hire deal. Similar deals over the last couple of years include agreements between Microsoft and AI startup Inflection, Google and Windsurf, and Amazon and Adept AI. Nvidia has since struck a similar deal with startup Poolside. Nvidia has been in a dealmaking frenzy, buying up companies like Hugging Face and investing tens of billions up and down the supply chain—though usually one step removed from a potentially competitive AI chip designer like Groq. The Justice Department opened the investigation into the Groq deal shortly after it was announced in December and has sent Nvidia a formal demand for information about it, the Times reported. “The Groq story is a prime example of the American system working as designed to promote innovation, reward entrepreneurs, and benefit consumers,” an Nvidia spokesperson said. The DOJ did not immediately respond to a request for comment. https://tinyurl.com/54yu9rvv

Block applies to establish National Bank.

Block, the fintech firm cofounded by Jack Dorsey, said it has applied to federal regulators to establish an uninsured national trust bank called Builders Bank & Trust. The company has sought approval from the Office of the Comptroller of the Currency to set up the proposed institution, following several fintech startups who have recently sought national bank licenses. Builders Bank would be a national trust bank and not accept deposits or make loans. Block said its proposed bank would cater to the cryptocurrency industry by providing custody and other services for tokens such as bitcoin and stablecoins. https://tinyurl.com/3fx9mpcu