Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

In the mid 1970s, after the storm of controversy around the Ford Pinto and its fuel tank prone to catching fire, Ford was the first automaker to pull an engineering tool out of the military and aerospace world and set it down on the drawing board of civilian car design. That tool is called FMEA, Failure Mode and Effects Analysis. Its nature is harsh: it forces engineers to list, in advance, every way a component could fail, to score how dangerous each way is, and to deal with the worst risks right there on the blueprint. Ford did not invent FMEA, but Ford is the company that turned it into the common language of the entire auto industry.

Half a century later, that same name sits at the top of the list for how often it has to call cars back to the shop. That is the paradox opening today’s story: a company that understands defect prevention better than almost anyone is now carrying one of the largest recall loads in the industry.

Look the numbers straight in the eye. As of early July 2026, Ford had issued roughly 50 recalls, reaching more than 11.2 million vehicles, the precise figure in the tally being 11,271,100 cars, plus one small recall for 2,633 engine block heaters. Half the year is not even gone, yet that volume already far exceeds the scale most automakers reach in a full year. And it follows a 2025 that already set a grim record: Ford logged around 152 recalls over twelve months, the highest any automaker has ever reached in a single year in the U.S. market.

Behind the total figures are very specific incidents, and what stands out is that they touch parts tied to safety. Nearly 1.7 to 1.9 million vehicles were called back because the reversing camera might not display an image, or worse, display an image flipped upside down, at exactly the moment the driver needs it most. Around 849,310 other vehicles carried a fault where the screen control module could overheat and shut itself off, leaving the driver staring at a black screen while in reverse. Some 741,000 vehicles were recalled over a transmission and parking brake fault that raised the risk of the car rolling away on its own. Other actions targeted the risk of cracked fuel injectors together with a fire hazard. Every line of description reads like an entry in exactly the analysis sheet that FMEA was born to prevent.

But the picture is not all one dark color, and this is the detail easily missed when you only read the sensational headline. Roughly 80% of Ford’s 2026 recalls are software faults, the kind that can be patched via an over the air update, a mobile service visit, or a single trip to the dealer. In other words, most of the 11.2 million figure does not mean 11.2 million cars sitting in a lot or needing costly parts replaced. Ford itself insists its quality program is working: the company doubled its safety team, expanded testing, and tightened software review, arguing that it is catching defects earlier and proactively calling cars back more often, not building worse cars.

So is the 11.2 million figure a sign of a quality machine falling apart, or of a detection system running at full capacity? The answer lies in whether we can distinguish two very different things inside the same number: defects caught after the car has reached the customer, and defects that should have been stopped before the blueprint was ever approved. This article does not judge Ford good or bad. It uses the very tool Ford brought into the industry, FMEA, as a magnifying glass to separate the cost of fixing mistakes from the value of doing it right from the start, a boundary that every business, whether it sells cars or software, is paying for every day without always realizing it.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On June 24, 2026, Micron Technology reported its results for fiscal Q3 2026, and the headline numbers made many people read them twice. Revenue more than quadrupled year over year. Gross margin hit 84.9%, a level that is almost unimaginable for a memory-chip company, an industry famous for its boom-and-bust swings. To appreciate how steep that climb was: just one quarter earlier the gross margin was 74.9%, and a year ago it was a mere 39%. Within four quarters, Micron more than doubled its profit on every dollar of revenue.

The force behind that number is not hard to name: the AI boom. As data centers around the world race to install GPUs to train and run artificial intelligence models, what each of those processors craves most is high-bandwidth memory (HBM, High Bandwidth Memory). This is a premium product line, hard to manufacture, with fat margins, and Micron is one of the very few suppliers capable of producing it at scale. Demand far outstrips supply, and in a market short on goods, whoever holds the production capacity holds pricing power.

But what makes this story worth dissecting through an operations lens is not the phrase “lucky thanks to AI.” It lies in a detail easily overlooked in the report: DRAM days of inventory were kept below 120 days, very tight. In semiconductors, inventory is both an asset and a time bomb. The memory cycle usually ends the same way: companies see high prices and produce aggressively, warehouses fill up, then when demand cools, that mountain of inventory must be dumped, dragging prices and margins down with it. The fact that Micron achieved record margins while keeping inventory lean shows this was not a market windfall, but the result of deliberate capacity-allocation discipline and inventory management.

And yet the paradox occurred: Micron’s stock slid right after the dream-like report was released. A company that had just set a profitability record of its own was turned away by the market in the next trading session. The reason cited by many analysts is that expectations were already too high: the share price had run up beforehand, pricing in a perfect scenario, so even an excellent result no longer had the power to surprise. On top of that sits the perennial fear of the memory industry, the question “how far away is the cycle’s peak?” that always lingers in investors’ minds whenever margins touch historic highs.

This is where two worlds intersect: the operator measures success by margins, days of inventory, and factory efficiency; while the investor measures value by expectations of the future. Micron won outright on the first playing field this quarter, but still took a minus on the second.

This article will not argue whether to buy or sell Micron stock. Instead, it pulls the story apart to answer a sharper question for operators in any industry: what truly created that record margin, what management tool stands behind it, and why “doing operations right” is a necessary condition, but has never been a sufficient one, for winning.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

From June 11 to July 19, 2026, the world is witnessing a special edition of the World Cup: for the first time in history, the tournament is co-hosted by three nations, the United States, Canada, and Mexico. Behind the stadium lights lies a fact few notice: experts call this the most complex sporting logistics operation ever staged.

The scale says it all. The tournament spans 16 host cities (11 in the U.S., 2 in Canada, 3 in Mexico), 16 stadiums, with 48 teams and 104 matches, roughly a 63% increase over the 64 matches in Qatar 2022. Expanding the field from 32 to 48 teams doesn’t merely add matches; it multiplies, almost geometrically, the number of destinations, travel schedules, and coordination points that must be synchronized.

To carry this load, FIFA appointed Rock-it Cargo as its primary logistics partner. By that company’s own forecast, the campaign will mobilize more than 5,000 trucks and vehicles, use about 1 million square feet of warehouse space, and move more than 1 million pounds of equipment throughout the tournament. These are real operational figures, not metaphors.

“Equipment” here is not just balls and nets. It includes broadcasting production systems, stage and staging structures, medical supplies, merchandise, and the technical infrastructure serving each venue. Every shipment must be transported, cleared through customs, installed, then dismantled on a tight schedule so it can move on to the next match site in time.

What makes the problem thorny is not the total volume but the three separate customs regimes. Goods moving among the U.S., Canada, and Mexico must comply with three sets of documentation rules, three tariff frameworks, and three inspection processes. To reduce friction, organizers rely heavily on the ATA Carnet temporary-import mechanism, a kind of “passport for goods” that lets equipment cross borders without triggering full import duties, provided it is re-exported.

Alongside the flow of equipment run quieter but no less strained supply chains. The cold chain and food-and-beverage logistics must serve millions of fans at stadiums, fan zones, and hospitality areas, demanding just-in-time delivery with high precision. Match days bring road closures, restricted-access zones, and security barriers around stadiums, making even finding safe truck parking near host cities a persistent operational challenge.

The final layer of complexity is cross-border security coordination. The agencies of the U.S., Canada, and Mexico operate under different legal frameworks and communication protocols; merging them into a real-time, shared response system spanning two continents is unprecedented in sport.

The official message, then, is not about football. It is this: World Cup 2026 is a live, continental-scale stress test of the capacity of multinational supply chains, customs systems, aviation networks, and digital infrastructure operating simultaneously under one harsh constraint: a match schedule that cannot move.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

The State of Logistics 2026 Report, one of the most respected annual reports in global logistics, has just been released with a striking central message: volatility has become the “new normal” shaping global supply chains, requiring businesses to adapt continuously. The report was released by the Council of Supply Chain Management Professionals (CSCMP) at a press briefing at the Empire State Building, authored by consultancy Kearney and presented by Penske Logistics. This year’s report is titled “Forged in Disruption.”

The report’s core figure is U.S. Business Logistics Costs (USBLC). In 2026, this stood at $2.4 trillion, equal to 7.8% of U.S. GDP. Compared with 2025, when costs were $2.6 trillion and 8.7% of GDP, this is a notable decline. Yet what stands out is that the report does not treat this cost decline as a sign of returning stability; on the contrary, it stresses that the operating environment is more deeply volatile than ever.

Korhan Acar, partner in Kearney’s Strategic Operations practice and lead author of the report, spelled out the nature of the change: “Tariff complexity, geopolitical uncertainty, changing trade flows, AI adoption, and new network designs are becoming permanent features of the logistics landscape.“ In other words, factors once seen as temporary shocks have now become a fixed operating condition every business must live with.

An important argument the report makes is the change in the nature of risk. It argues that external geopolitical shocks and trade policy changes now have a bigger impact on the logistics market than traditional concerns such as demand or capacity. For decades, supply chain managers worried mainly about demand rising or falling, about whether there were enough ships, trucks, warehouses. Now, the variables beyond the firm’s direct control, such as war, tariffs, and political instability, are the strongest shapers of the market. This is a fundamental shift in how operational risk is understood.

Hence the report’s call to action is clear and rather blunt: U.S. companies have stopped waiting for the supply chain to “return to normal.” In recent years, many organizations operated on an implicit assumption that the turbulence was temporary and everything would return to the old stable state. The report states plainly that that assumption is wrong, and only the most successful companies are truly adapting, where continuous adjustment is now a requirement for supply chain resilience.

The official message lies not in the cost figure itself, but in a watershed admission for the whole industry: the old normal will not return, volatility is a permanent condition, and resilience together with continuous adaptation has shifted from a backup option to a core capability. This is an event worth analyzing deeply, because it poses a fundamental operational question: if volatility is permanent, by what principle must one operate not just to survive but to prevail?

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

In mid-2026, the pharmaceutical cold chain logistics industry is going through a phase of strong growth and technological transformation. According to compiled figures, the market reached about US$22.75 billion in 2025 and is forecast to reach about US$44.1 billion by 2033, nearly doubling in under a decade. It is one of the fastest-growing logistics segments, reflecting a profound shift of the pharmaceutical industry toward temperature-sensitive products.

The pharmaceutical cold chain is a system for transporting and storing medicines, vaccines, biologics and biomedical products within a strictly controlled temperature range, typically 2 to 8 degrees Celsius, and much lower for some vaccines and cell and gene therapies. The defining feature of these goods is that a single excursion beyond the allowed threshold, called a temperature excursion, can render the entire shipment ineffective, no longer safe to use, and fit only for destruction. Because the value of each biologic lot can reach millions of dollars, the consequences of a small incident are enormous, both financially and for patient safety.

The most important shift lies not only in market size, but in monitoring technology. According to industry analyses, Internet of Things (IoT) sensors can now beam temperature data directly to satellites in real time, turning every refrigerated truck into a “rolling laboratory“ continuously reporting cargo conditions across the whole journey, whether by air, ocean, road or rail. This is a leap from a model of discrete, point-based checks to one of continuous, real-time, end-to-end monitoring.

Along with monitoring, the last mile is becoming a focal point of innovation. Companies such as Zipline and Matternet have developed drone delivery solutions, granted FAA approval for Beyond Visual Line of Sight (BVLOS) operation and noted as compatible with cold chain requirements, serving urgent-medicine delivery in areas hard to reach by road. Meanwhile, the pharmaceutical committee of the International Safe Transport Association has identified last-mile shipping as a focus area in 2026, because it is the most failure-prone stage, especially for direct-to-consumer orders where goods are left at the destination under unpredictable conditions.

The picture is rounded out by the trend of electrification and sustainability. Many major distributors have committed to fully electric delivery fleets by 2030; hybrid or electric refrigerated vehicles are emerging for the last mile, combined with route optimization, load consolidation and intermodal transport to reduce fuel consumption and emissions.

In sum, the message is clear: the pharmaceutical cold chain is not only growing fast in scale, but is being reshaped in its very nature. End-to-end visibility — the ability to see and prove the temperature condition of goods at every moment and every location — is emerging as the industry’s main competitive battleground. The winner is no longer the one with the coldest warehouse, but the one who controls and proves cargo conditions most tightly and transparently.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

Entering mid-2026, a series of logistics industry surveys and reports together sketch a unified picture: warehouse automation is no longer a competitive advantage reserved for the largest corporations, but is becoming a baseline expectation for businesses of nearly every size, from regional e-commerce fulfillment centres to global manufacturing distribution networks. In other words, the question has shifted from “whether to automate the warehouse“ to “how to automate effectively.”

The figures accompanying this trend are notable. According to industry statistics compiled for 2026, about 60% of warehouses plan to increase their automation budgets by 20%, with the focus channelled into three groups: robots, Automated Guided Vehicles (AGVs), and AI-driven operating software. Also according to these surveys, as many as about 92% of warehousing businesses are implementing or planning AI projects in the near term, of which only a small share, roughly more than 6%, have implemented at an extensive level. The picture shows a broad wave of investment, but most are still at the early stage of the journey.

However, the most important shift in thinking in 2026 lies not in the budget figures, but in the strategic focus. Many analysts argue that in 2026, “orchestration“ will become the foundation of warehouse automation strategy. Instead of being designed around a single technology investment, such as buying a fleet of robots or a conveyor system, warehouse facilities will be designed around how humans, robots and equipment intersect and coordinate with one another. The focus shifts from “equipment” to “the coordination between equipment and humans.”

Accompanying this is a practical observation many industry experts emphasize: in warehouse automation in 2026, change management matters more than any single new technology. That is, the factor deciding the success or failure of an automation project is not how advanced the robot is, but whether the organization can redesign the processes, roles and ways of working of its people to mesh with the new technology. This is a maturing of the industry’s awareness: after years of chasing equipment, the focus is returning to the systems problem.

The context of the event is also marked by a concrete milestone. Automate 2026, one of the largest robotics and automation events in North America, is scheduled to take place from June 22 to 25, 2026, at McCormick Place in Chicago. There, companies such as Applied Manufacturing Technologies (AMT), together with robot maker FANUC, will demonstrate mixed-load handling applications, combining a collaborative robot with AI-integrated machine vision for pallet stacking and unstacking operations. This is a vivid illustration of the trend: technology is increasingly available, but true value only appears when it is integrated and orchestrated into a complete workflow.

In sum, the official information of mid-2026 is not a press release from a single company, but a consensus forming across the whole industry: warehouse automation has become the baseline, investment is rising sharply, but the key to success has shifted from owning technology to orchestrating humans and machines into a well-meshed operating system.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 20, 2026, C.H. Robinson (stock symbol NASDAQ: CHRW), one of the world’s largest providers of logistics and supply chain management services, officially announced the launch of “the world’s first AI technology that continually assesses, improves and operates global supply chains.” The product, named Lean AI Engineer, is designed to serve the customer segment using the company’s 4PL Managed Solutions (fourth-party logistics management solutions).

The core of C.H. Robinson’s announcement is that Lean AI Engineer does not operate alone, but works in concert with Lean AI Planner, a tool introduced back in 2025, to form one connected system. In this pairing, Lean AI Planner handles the real-time execution of global shipments, while Lean AI Engineer continuously analyzes data to optimize future freight operations. C.H. Robinson calls this a closed-loop AI framework, meaning a system that closes the loop between execution and improvement on its own, “enhancing a supply chain as it runs.”

Mr. Arun Rajan, Chief Strategy and Innovation Officer of C.H. Robinson, explained the reason this technology was created with a notable observation: supply chains “suffer from the gap between knowing and doing.” According to Rajan, “Tech that sits above or outside of a supply chain can aggregate data, harmonize signals and recommend. But it relies on someone else to execute on the signals and someone else to learn whether those actions worked. Our tech closes the gap, delivering 24/7 premium service with one unified system no one else can match.”

On technical capability, C.H. Robinson states that Lean AI Engineer can assess an entire supply chain within 25 to 30 minutes and identify points for improvement before performance is impacted. The company frames this figure against traditional supply chain assessments, which typically take up to four weeks and only look backward rather than look forward. This is precisely the shift from a backward-looking mindset to a forward-looking and predictive one.

More important than the assessment figures, the technology has already gone into real operation at large scale. According to the announcement, Lean AI is currently autonomously handling 92% of 4PL shipments globally for C.H. Robinson, spanning all four transport modes: trucking, ocean, air and rail. The scope of automation covers the entire lifecycle of a shipment: from the moment an order is created, through the steps of tendering, routing, delivery, exception handling, all the way to carrier payment.

To build this platform, C.H. Robinson says it mobilized about 450 in-house software engineers and data scientists, while creating a proprietary “context layer“ by systematically capturing the institutional knowledge accumulated from the company’s real-world workflows. This is the factor that allows the AI not merely to compute on raw data but to operate according to the “know-how” accumulated over decades.

On real-world results for customers, C.H. Robinson cites two representative cases. One customer cut loads by 17% across 20 sites, achieving annual savings exceeding US$1 million. Another customer, after restructuring shipment flows, reduced loads by 81% and lowered costs by 40%. These figures turn a technology press release into concrete operational evidence.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On June 8, 2026, Iran’s Ambassador to Moscow Kazem Jalali declared in an interview with the newspaper Izvestia (Russia) that the Strait of Hormuz would be reopened, but with new conditions: Iran and Oman would jointly determine a transit fee applied to all commercial vessels passing through the strait, with the fee depending on the type of vessel, the cargo transported, and “geopolitical factors.” This declaration was made on precisely day 100 since the Strait of Hormuz was effectively blockaded on February 28, 2026, after the United States and Israel launched an air campaign against Iran codenamed Operation Epic Fury, destroying military and nuclear facilities, and resulting in the death of Supreme Leader Ali Khamenei. Iran retaliated with missiles and drones targeting Israel, U.S. military bases in the region, and Gulf states allied with the United States. On March 4, 2026, Iran officially declared the Strait of Hormuz “closed,” with the Islamic Revolutionary Guard Corps (IRGC) warning it would attack any vessel attempting to pass, while simultaneously laying sea mines across the shipping lane.

Before the blockade, the Strait of Hormuz carried 25% of global seaborne oil and 20% of liquefied natural gas (LNG), with approximately 3,000 vessels transiting each month, equivalent to roughly 95 vessels per day. As of May 31, 2026, commercial traffic through the strait had fallen to just 11% of pre-crisis levels, approximately 10 vessels per day. The International Energy Agency (abbreviated as IEA) called this “the largest supply disruption in the history of the oil market.”

U.S. Secretary of State Marco Rubio immediately rejected Iran’s transit fee proposal: “It can’t happen. It would be unacceptable.” The American response was grounded in the principle of transit passage through international straits as codified in the United Nations Convention on the Law of the Sea (abbreviated as UNCLOS), 1982. Article 38 of UNCLOS establishes the right of freedom of navigation through international straits, and Article 44 explicitly states that states bordering straits “shall not hamper transit passage” and may not suspend this right for any reason. However, a critical legal reality is that Iran has never ratified UNCLOS, creating a legal grey zone regarding whether UNCLOS provisions are binding on Iran.

The 100-day Hormuz blockade was not a straight line but a chain of successive escalations. On April 12, after the Islamabad negotiations between the U.S. and Iran failed, President Trump declared a naval blockade of the Strait of Hormuz, preventing vessels that had paid fees to Iran from passing through. On April 17, Iran announced a temporary pause of the blockade during the Lebanon ceasefire, but just one day later (April 18) Iran re-imposed the blockade, pushing Brent oil prices up 7% in a single trading session. Since then, the strait has remained effectively blockaded until Ambassador Jalali’s declaration on June 8. If Iran succeeds in imposing transit fees, this would be the first precedent in modern maritime history of a nation collecting fees on an international strait, with potential implications for every other strategic strait in the world, from Malacca to Bab el-Mandeb.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

Throughout the history of the copper mining and smelting industry, the financial relationship between mining companies (miners) and smelters has always operated according to one immutable principle: mining companies extract copper concentrate, transport it to smelters, and pay smelters a fee called Treatment and Refining Charges (abbreviated as TC/RC) for the smelter to process the concentrate into refined copper with 99.99% purity. TC is calculated in dollars per tonne of concentrate, RC is calculated in cents per pound of refined copper. This model has existed for decades and has been regarded as the economic foundation of the entire global copper value chain. In 2026, that model was completely reversed.

In January 2026, the annual TC/RC benchmark established through negotiations between the Chilean mining group Antofagasta and major Chinese smelters fell to 0 dollars per tonne and 0 cents per pound, the lowest level ever agreed upon in the history of TC/RC negotiations. Just one year earlier, the benchmark for 2025 was $21.25 per tonne TC and 2.125 cents per pound RC. The decline from $21.25 to $0 within 12 months was already an unprecedented shock. But the spot market went even further: in Q1 2026, spot TC/RC fell to negative $66.40 per tonne, meaning smelters not only received no processing fee, but had to pay mining companies just for the right to purchase copper concentrate to process. This was the first time in the history of the copper smelting industry that the cash flow completely reversed: smelters paying miners instead of miners paying smelters.

The direct cause of this reversal was a structural imbalance between smelting capacity and concentrate supply. China, which accounts for over 50% of global copper smelting capacity, had been continuously building new smelters throughout the past decade. The total primary smelting capacity of the CSPT (China Smelters Purchase Team), comprising China’s largest smelters, reached 9.10 million tonnes per year in 2025 and was projected to increase to 9.61 million tonnes per year in 2026 following the commissioning of the Chifeng Jinjian II plant (capacity 300,000 tonnes per year) and the completion of the Jinguan plant ramp-up phase. Meanwhile, global copper concentrate mine output increased by only approximately 1.4% (equivalent to roughly 500,000 tonnes) in 2026, according to estimates from the International Copper Study Group (abbreviated as ICSG). The result was dozens of smelters competing for a nearly stagnant supply of concentrate, pushing concentrate prices up and processing fees below zero.

But the supply-demand imbalance in copper concentrate was not the sole factor creating the crisis. Two consecutive geopolitical shocks transformed a structural problem into a comprehensive crisis. The first shock occurred on February 28, 2026, when the Strait of Hormuz was blockaded due to military conflict between the United States, Israel, and Iran. Before the blockade, the Strait of Hormuz was a maritime route carrying 25% of global seaborne oil and 20% of liquefied natural gas (LNG). But few realized that Middle Eastern oil refineries also produce 70% of the world’s elemental sulfur, the raw material for producing sulfuric acid, an essential chemical in the hydrometallurgical copper leaching process. Approximately 20% of global copper production depends on hydrometallurgical processes that consume more than 1 tonne of sulfuric acid for every tonne of copper produced. When the Strait of Hormuz was blockaded, approximately 50% of global seaborne sulfur supply was cut off, and spot sulfuric acid prices tripled in some markets.

The second shock came from China. On May 1, 2026, China officially banned sulfuric acid exports, a by-product of pyrometallurgical copper smelting. China is the world’s largest sulfuric acid producer, accounting for over 40% of global output, and in 2025 had exported a record 4.65 million tonnes. The ban was expected to last through the end of 2026. When sulfuric acid supply from the Middle East (via the Strait of Hormuz) and from China (via the export ban) was simultaneously cut, copper mines in Chile, Peru, the Democratic Republic of Congo, Zambia, and Indonesia that depended on imported sulfuric acid for hydrometallurgical processes were all directly affected, creating additional pressure on an already deficit refined copper supply.

The consequences on the copper price market were immediate and dramatic. Copper prices on the London Metal Exchange (abbreviated as LME) reached an all-time high of $14,527.50 per tonne on January 29, 2026. As of June 2, 2026, LME spot copper prices remained at $13,965.50 per tonne. J.P. Morgan Global Research forecast an average copper price of $12,075 per tonne for the full year 2026 and projected that the United States would face a deficit of 330,000 tonnes of refined copper. ICSG forecast a global refined copper market deficit of approximately 150,000 tonnes in 2026. This was not a temporary price fever driven by speculation. This was the result of a structural shift in the global copper supply chain, where demand exceeds supply at every link, from concentrate to processing chemicals to the final refined product.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 14, 2026, Stellantis (operating in the United States under the legal name FCA US, LLC) officially decided to recall 419,035 Jeep Grand Cherokee vehicles from the 2022 to 2026 model years (internal code WL74) and Jeep Grand Cherokee L from the 2023 to 2025 model years (internal code WL75) across the entire United States. The cause of the recall is a software error in the passenger protection system control module, called the Occupant Restraint Controller (abbreviated as ORC), which may cause side curtain airbags to deploy with a delay during a side-impact collision. According to the filing submitted to the National Highway Traffic Safety Administration (abbreviated as NHTSA), the ORC software error produces no warning signs whatsoever before the fault occurs. Only when the error is already present in the system does the airbag warning light on the instrument cluster illuminate, accompanied by a chime alert during each vehicle start-up cycle. This means that hundreds of thousands of Jeep Grand Cherokee drivers may be traveling on the road every day with a side airbag system that is not functioning as designed, completely unaware of the defect, until an actual collision occurs and the airbag deploys a few hundredths of a second later than the optimal moment, enough to turn a minor injury into a serious injury or fatality.

To understand the severity of this defect, one must understand the operating mechanism of side airbags. In a side-impact collision at 50 km/h, the time from impact to the moment the occupant’s head contacts the door glass or roof pillar is approximately 15 to 20 milliseconds. The side curtain airbag must fully deploy within the first 10 to 15 milliseconds after impact to create a protective cushion in time. Any delay, even by just a few milliseconds, can result in the airbag not being fully inflated when the occupant’s head has already contacted a hard surface, significantly reducing its protective effectiveness. The ORC software error in the Jeep Grand Cherokee creates precisely that scenario: the control module processes crash sensor signals slower than designed, causing the airbag activation command to be delayed within a time window where every millisecond has life-or-death significance.

Stellantis announced it would send recall notification letters to all affected vehicle owners between June 11 and June 19, 2026. The remedy is an ORC software update performed at authorized Jeep dealerships, completely free of charge for vehicle owners. However, the gap between the recall decision date (May 14) and the earliest notification letter date (June 11) is nearly one month, a period during which 419,035 vehicles continue to travel on roads with the unfixed airbag software error. NHTSA advises vehicle owners to proactively check the recall status of their vehicles on the official website nhtsa.gov by entering their Vehicle Identification Number (abbreviated as VIN), rather than waiting for the notification letter to arrive.

This is not the first time the Jeep brand under Stellantis has faced a large-scale recall campaign in recent times. Previously, Stellantis had to recall 320,065 Jeep Wrangler 4xe (model years 2020 to 2025) and Jeep Grand Cherokee 4xe (model years 2022 to 2026) plug-in hybrid electric vehicles (abbreviated as PHEV) due to high-voltage battery fire risk. This campaign was later expanded to 375,000 vehicles through three separate recall waves: campaign 23V-787 in October 2023, campaign 24V-720 in September 2024, and campaign 25V-741 in October 2025. NHTSA documented a total of 19 fires and 1 injury related to the 4xe battery defect. During the repair waiting period, Stellantis instructed owners to park their vehicles outdoors, away from structures, and to refrain from charging the battery overnight. When adding the recall of 419,035 vehicles for the ORC airbag software error, the total number of Jeep vehicles recalled within the past three years has surpassed the 794,000-unit mark, an alarming figure for any automotive brand, especially when two entirely different types of defects (battery hardware and airbag software) both target life-safety systems of occupants.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On February 23, 2026, at the Renaissance Phoenix Glendale Hotel in Arizona, more than a thousand process improvement professionals from around the world walked into a conference hall decorated with a large number 25 behind the stage. They came to celebrate a quarter century since ASQ (the American Society for Quality) first organized the Lean and Six Sigma Conference in the early 2000s, transforming a methodology born in semiconductor factories into a global movement with its own stage, its own community, and influence that extended far beyond the borders of manufacturing. This year’s theme was “Leaning into the Future”, a subtle play on words between “lean” (as in lean manufacturing) and “leaning” (tilting forward), accurately reflecting the question that everyone in the auditorium was asking themselves: in the age of artificial intelligence, Digital Twin, and automation, how long can this forty-year-old methodology stand firm, or is it entering the most powerful renaissance in its history?

To answer that question, one must go back forty years, to a place more than two thousand kilometers from Phoenix: Motorola’s factory in Schaumburg, Illinois. In 1986, Bill Smith, a senior engineer with more than thirty years of industry experience, presented to CEO Bob Galvin a study that made the entire leadership team pause and reflect. Smith discovered that products which underwent multiple repairs during production had significantly higher field failure rates compared to those that passed through the line without intervention. That finding sounds obvious today, but in the mid-1980s, the prevailing philosophy was “inspect quality at the end of the line”: if the final product met standards, the process that produced it did not matter. Smith completely overturned that thinking. He argued that quality must be built into the process, not inspected into the product, and the only way to achieve that was to reduce process variation to near zero: 3.4 defects per million opportunities. He called that target Six Sigma, named after the statistical symbol representing six standard deviations from the mean.

Together with Mikel Harry, an engineer later known as the principal architect of the methodology, Smith built a four-stage problem-solving framework: Measure, Analyze, Improve, Control, abbreviated as MAIC. This framework was not invented from nothing. It inherited and systematized the quality management principles that W. Edwards Deming, Joseph Juran, and Walter Shewhart had developed since the mid-twentieth century. But Smith and Harry’s contribution was transforming those scattered principles into a structured methodology that could be taught, deployed, and measured in dollars. Just two years after implementation, Motorola won the Malcolm Baldrige National Quality Award in 1988, the most prestigious presidential-level quality award in the United States. By 2005, Motorola announced that Six Sigma had helped the company save a cumulative total of more than $17 billion since inception.

But Six Sigma might have forever remained an internal Motorola program if not for Jack Welch. In 1995, the legendary CEO of General Electric decided to make Six Sigma the core business strategy of the entire corporation, not just in the manufacturing division but in every department, from finance to human resources, from sales to logistics. Welch added a Define stage at the beginning of MAIC to create the complete DMAIC framework known today, ensuring that every project began by clearly defining the problem to be solved and the voice of the customer. GE reported savings of $350 million in 1998 alone, a figure that later exceeded $1 billion per year and accumulated more than $12 billion over five years. More important than the numbers, Welch created an incentive system that no company before had dared to implement: he tied 40% of senior management bonuses to Six Sigma project results, and promotions were reserved only for those who had completed at least one project at the Green Belt level or above. Six Sigma was no longer optional. It was a condition of survival at GE.

Parallel to the statistical revolution in the West, across the Pacific, Toyota’s production system with its Lean philosophy developed by Taiichi Ohno and Shigeo Shingo since the 1950s had proven that eliminating waste and reducing variation were not two different battles but two sides of the same coin. Lean focused on speed and flow: eliminating seven types of waste (Muda), reducing cycle time, and continuously creating value for the customer. Six Sigma focused on precision and consistency: reducing variation, eliminating defects, and bringing every process into statistical control. In the early 2000s, pioneering companies recognized that combining the two approaches would create synergy greater than the sum of both parts, and the term Lean Six Sigma was born. According to a report by iSixSigma Magazine in 2007, Lean Six Sigma had helped Fortune 500 companies save an estimated $427 billion over twenty years, a figure that no other management methodology in history had ever achieved.

Bill Smith died in 1993 from a heart attack at work on the Motorola campus, just seven years after laying the foundation for the methodology that bore his name. He did not live long enough to see GE turn Six Sigma into a global standard, to see hospitals use DMAIC to reduce emergency room wait times, to see banks use Lean to optimize loan approval processes, or to see millions of people around the world call themselves Green Belts and Black Belts like martial arts rankings. But in February 2026, when more than a thousand people stood up in the auditorium in Phoenix to celebrate 25 years of the conference named after the methodology he created, Bill Smith’s legacy needed no introduction.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 22, 2026, a robotics company in San Antonio, Texas did something that virtually no automation vendor has ever attempted: it pointed a camera at its warehouse robots and invited the entire world to watch them work, uninterrupted, for a full eight-hour shift. No edits. No curated highlights. No scripted narration designed to mask the moments when things go wrong. Plus One Robotics broadcast a continuous livestream of its InductOne AI-powered parcel induction system on YouTube and LinkedIn, and over the course of those eight hours, the system completed 19,784 picks at a sustained throughput of 2,488 picks per hour, averaging 1.45 seconds per parcel. More than 950 viewers watched the performance in real time, and the footage remains publicly accessible for anyone who wants to verify the numbers.

The event landed in an industry context that made its transparency deliberately provocative. Just days earlier, Figure AI, the humanoid robotics startup valued at over $39 billion, had streamed its own robots sorting packages in a warehouse, generating widespread media coverage but also significant skepticism about whether the system was operating fully autonomously or receiving hidden human assistance. Plus One Robotics’ livestream was, in effect, an answer to that skepticism: here is a real system, in a real workflow, doing real work, for a real shift, with no place to hide.

Plus One Robotics was founded in 2016 by Erik Nieves, Shaun Edwards, and Paul Hvass. Nieves spent 25 years at Yaskawa Motoman Robotics, rising from trainer to Technology Director for North America, where he was responsible for the company’s technology roadmap and emerging applications. Edwards and Hvass came from Southwest Research Institute (SwRI), and Edwards co-founded ROS-Industrial, the open-source robotics framework that has become a standard in industrial automation research. The company is headquartered at Port San Antonio and has raised approximately $95 million across four funding rounds, including an $8.3 million Series A in November 2018 led by Pritzker Group Venture Capital with participation from Zebra Technologies, a $33 million Series B in April 2021 led by McRock Capital and Translink Capital with BMW i Ventures participating, and a $50 million Series C in March 2023 led by Scale Venture Partners. The investor roster reflects a bet on logistics automation as a category-defining opportunity.

The system demonstrated in the livestream was InductOne, a dual-arm automated parcel induction solution launched in May 2024. InductOne integrates three proprietary technologies: PickOne, Plus One’s AI-powered 3D vision software that uses computer vision and machine learning to identify, classify, and determine optimal grasp points for parcels of varying sizes, shapes, weights, and packaging conditions; Yonder, the company’s remote supervision platform; and integrated pick-and-place conveyors with safety systems and analytics. The system’s published specifications cite sustained pick rates of 2,200 to 2,300 parcels per hour with peak rates reaching 3,300 per hour. The livestream’s sustained rate of 2,488 per hour over eight continuous hours fell comfortably within the upper range of sustained performance and demonstrated the critical metric that separates laboratory demonstrations from operational reality: consistency over time.

The distinction between sustained and peak performance is where the comparison to manual labor becomes most revealing. In parcel induction operations, a trained human operator typically handles 1,200 to 1,800 parcels per hour at a well-designed station. That rate is achievable for the first few hours of a shift, but human physiology imposes a cost that no amount of training or motivation can eliminate: fatigue. By the sixth, seventh, and eighth hour of repetitive lifting, scanning, and placing, manual throughput degrades measurably. Injury rates climb. Error rates increase. The economics of a three-shift, seven-day parcel sortation operation depend on maintaining throughput across the entirety of every shift, and it is precisely in the late hours of late shifts that manual performance deteriorates most significantly. The InductOne system’s 2,488 per hour in hour eight was functionally identical to its rate in hour one, a performance characteristic that no human operator can replicate.

What makes Plus One Robotics’ approach distinctive in the crowded warehouse automation landscape is its Crew Chief model, a human-in-the-loop remote supervision system that the company treats not as a temporary crutch on the path to full autonomy but as a permanent architectural feature. When PickOne’s AI encounters a parcel it cannot confidently identify or grasp, the system sends a request through Yonder to a trained Crew Chief, a remote human supervisor available 24 hours a day, 7 days a week, who clicks the image to guide the robot’s next action. The entire intervention takes seconds, and the AI learns from each interaction, progressively reducing the frequency of human assistance while maintaining pick accuracy above 99 percent. This philosophy stands in deliberate contrast to competitors pursuing full autonomy: Plus One’s position is that reliability at scale matters more than autonomy as an ideal, and that the most economically valuable robotic system is not the one that never needs a human but the one that knows exactly when to ask for help.

The competitive landscape surrounding Plus One validates the scale of the opportunity. The parcel sorting robots market was estimated at $1.07 billion in 2025 and is projected to reach $4.19 billion by 2032, growing at a CAGR of 21.38 percent. Key competitors include Kindred Systems (now part of Ocado Group), offering the INDUCT system; RightHand Robotics, which received a strategic investment from Rockwell Automation in March 2025; Dexterity, which raised $95 million at a $1.65 billion valuation in March 2025 and deploys robots at FedEx, UPS, and GXO; Pickle Robot, which signed a deal with UPS worth up to $120 million for 400 truck-unloading robots; and Ambi Robotics, which launched its PRIME-1 foundation model in January 2025. Covariant, once a leading competitor in AI-powered picking, was effectively acquired by Amazon in August 2024 through a reverse acqui-hire. Plus One’s most prominent customer relationship is with FedEx, which has deployed 12 new InductOne systems at its Memphis Express Hub in addition to 4 existing installations, with further expansion to its Reno, Nevada facility.

The company reached a milestone on April 9, 2026 that contextualizes the livestream’s significance: Plus One’s global fleet surpassed 2 billion successful picks. The first billion took eight years. The second billion took just two, illustrating the exponential scaling curve that characterizes successful automation deployments.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 12, 2026, the Bureau of Labor Statistics released the Consumer Price Index data for April that confirmed what American households had already been feeling at checkout counters for weeks: food prices were accelerating again. The headline CPI registered 3.8% year-over-year, the highest since May 2023, with a 0.6% month-over-month increase that signaled inflation was not merely persisting but re-intensifying. Within that headline number, the food-at-home index, the measure that tracks what consumers pay at grocery stores rather than restaurants, rose 0.7% in a single month, the largest monthly gain since August 2022, pushing the year-over-year grocery inflation rate to approximately 2.9%. Five of six major grocery store food categories posted gains in April. But the aggregate numbers, while concerning, obscure the violence happening at the category level. Tomato prices surged approximately 39.7% year-over-year, the fastest inflation of any food-at-home item the BLS tracks, driven by a catastrophic Florida freeze that destroyed up to 80% of the state’s crop, compounded by weather and disease damage to Mexican production and a 17% tariff on Mexican tomato imports. Fresh vegetables overall climbed 11.5% year-over-year. Beef rose 2.7% in April alone, with the USDA forecasting full-year beef inflation of +9.4% and acknowledging the range could reach as high as 16.6%. Nonalcoholic beverages, heavily weighted toward coffee, increased 5.1% year-over-year. Sugar and sweets hit 8.1% annual inflation. The only major category offering consumers relief was eggs, which fell 44.7% year-over-year from March 2025’s avian influenza-driven spike, a decline that was mathematically dramatic but emotionally irrelevant to shoppers watching nearly everything else in their cart cost more than it did a year ago.

The forces driving this acceleration are not domestic in origin, or rather, they are not only domestic. They arrive from two directions simultaneously, and their convergence in the same quarter is what makes the spring of 2026 unlike any inflationary episode since the pandemic peak of 2022. The first force is tariff policy. On March 12, 2025, the United States imposed 25% Section 232 tariffs on steel and aluminum imports from all countries. On June 4, 2025, those tariffs were doubled to 50% on imports from every country except the United Kingdom. Then on April 6, 2026, a new proclamation extended tariff coverage to the full customs value of covered metal articles and derivatives, not just the metal content, with a tiered structure: 50% for products made almost entirely of covered metals, 25% for equipment and components substantially made of them, 15% transitional for certain metal-intensive equipment through December 2027, and 10% for products made with U.S.-origin metal. The impact on food is neither abstract nor indirect. Aluminum can warehouse prices rose 10.3% year-over-year. Tinplate steel spot prices climbed 18.2% year-over-year. Steel mill products increased 20.7%. Aluminum mill shapes surged 33%, the largest annual increase since early 2022. For a food can manufacturer, tariffs now account for an estimated 3 to 12% of total production costs depending on product type and sourcing. Conagra Brands, one of America’s largest packaged food companies, disclosed that it expects tariffs and related inflation to add approximately 3% to its cost of goods sold, translating to more than $200 million in additional annual costs in 2026. Packaging typically represents 7 to 10% of a food product’s final retail price, but for heavily packaged goods like canned soups, canned vegetables, aluminum-tray frozen meals, and foil-wrapped snack bars, the percentage runs higher, meaning the tariff’s fingerprint touches a significant portion of the American grocery aisle.

The second force is geopolitical disruption, specifically the Strait of Hormuz crisis triggered on February 28, 2026, when the United States and Israel launched Operation Epic Fury against Iran, and Iran responded by blocking commercial shipping through the strait. The Hormuz chokepoint carries 20% of the world’s daily oil supply and 20% of global LNG, but its relevance to food inflation extends far beyond energy pricing. Approximately 30 to 35% of global urea exports and 20 to 30% of global ammonia exports transit the strait, meaning that up to 40% of globally traded urea, the world’s most widely used nitrogen fertilizer, was effectively cut off from international markets at precisely the moment the Northern Hemisphere planting season demanded peak supply in March and April. The International Food Policy Research Institute and the Carnegie Endowment both warned that the fertilizer disruption would flow through to food prices with a lag of three to six months, meaning the full impact on harvest yields and food costs will not be felt until late 2026 and into 2027. Global fertilizer prices are estimated to have risen 15 to 20% during the first half of 2026. Meanwhile, the energy shock hit food transportation immediately. Diesel prices surged from $3.89 to $5.37 per gallon in just two weeks during early March, a 38% spike that carriers translated into surcharges within days. Major carriers announced fuel surcharges of 8% to 26.5%. The USDA has documented that transportation costs represent approximately 5% of the retail cost of food consumed at home, and that the correlation between diesel prices and truck transportation costs carries a coefficient of 0.64, meaning diesel increases flow through to food freight costs with meaningful statistical reliability. For food vendors operating on 1 to 4% net margins, a logistics cost increase of 3.5% can exceed the entire profit on a shipment, forcing either price increases to consumers or margin destruction for producers.

The consumer response is already visible in behavioral data. Food insecurity in the United States rose to 14.2% of households in 2025, up 1.7 percentage points from 2024’s 13.7%, and the April 2026 inflation acceleration is expected to push that figure higher when the next annual survey is completed. The lowest-income quintile of American households already spends 33% of pre-tax income on food, compared to 6.4% for the highest quintile, meaning identical percentage price increases impose radically different burdens across the income spectrum. Consumers are adapting through two primary mechanisms. First, private label substitution: U.S. store brand sales reached a record $282.8 billion in 2025, growing 3.3% over 2024, with private labels now commanding 21.3% dollar share and 23.5% unit share of grocery sales. 72% of consumers report buying private label over national brands, and 32% switched to lower-priced brands in early 2026. Second, channel shifting: food-away-from-home spending, which the USDA forecasts will inflate at +4.6% versus +1.7% for food-at-home in 2026, has driven a reported 23% decline in out-of-home dining occasions as households redirect meals back to the kitchen. The USDA’s official full-year 2026 food price outlook projects +3.1% overall food inflation, with food-at-home at +2.5% and food-away-from-home at +3.7%, figures that were published before the April CPI data revealed the acceleration now underway. Independent analysts warn that if tariff escalation continues and Hormuz-linked energy and fertilizer costs persist through summer, grocery inflation could reach 4 to 4.5% by year-end, approaching levels not seen since 2023 and reviving memories of the 11.4% food-at-home inflation that defined the cost-of-living crisis of 2022.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 7, 2026, DHL Express officially launched a tool that no other global express logistics company had ever deployed at scale: an AI-powered customs vision system that allows shippers to photograph their items with a smartphone camera and receive a customs-compliant product description within seconds. The tool, integrated directly into DHL Express’s digital shipping platforms, uses server-side computer vision to analyze the uploaded image, identify the object’s material composition, function, and category, then generate a structured description aligned with international customs classification standards. Rather than requiring customers to interpret complex tariff codes or draft technical descriptions themselves, the system translates a single photograph into the precise language that customs authorities expect to see on commercial invoices and electronic declarations. DHL Express confirmed the tool is live across eight markets at launch: Canada, Germany, Hong Kong, the Netherlands, Singapore, South Africa, Spain, and the United Arab Emirates, with plans to expand to additional countries throughout the remainder of 2026.

The technology operates entirely on DHL’s cloud infrastructure rather than on the customer’s device, a deliberate architectural choice that allows the company to update and retrain the underlying model without requiring users to download new software. Dirk Olufs, Executive Vice President and Global Chief Information Officer of DHL Express, described the tool as part of DHL’s broader strategy to embed artificial intelligence into every customer touchpoint where friction traditionally exists. Enna Zarate, Senior Vice President of Digital Customer Solutions, emphasized that the vision system was trained not only on conventional product photography but also on synthetic images and region-specific samples to ensure the model performs accurately across the diverse goods flowing through DHL’s network in 220+ countries and territories. This training methodology addresses a fundamental challenge in customs classification: the same physical object can look dramatically different depending on packaging, lighting, angle, and regional manufacturing variations, yet it must map to the same harmonized system code regardless of where it enters the customs pipeline.

The problem this tool targets is neither new nor small. Industry data consistently shows that approximately 80% of customs clearance delays originate from errors on commercial invoices and packing lists, specifically inaccurate, vague, or incomplete product descriptions that force customs officers to halt shipments for manual review. For cross-border e-commerce shippers, many of whom are small businesses without dedicated trade compliance teams, describing products in customs-acceptable language has long been one of the most intimidating barriers to international expansion. The consequences of getting it wrong extend beyond delay: US Customs and Border Protection reported that its penalty revenue grew 163.9% in fiscal year 2024, reflecting both stricter enforcement and the sheer volume of non-compliant declarations flowing through automated screening systems. On the commercial side, research from the Baymard Institute found that 48% of online shopping cart abandonments occur because of unexpected costs revealed at checkout, including duties and taxes that sellers failed to calculate accurately because their product descriptions did not map cleanly to tariff classifications. That 48% represents an estimated $260 billion in recoverable revenue globally, a figure that underscores how customs description accuracy is not merely a compliance issue but a direct revenue lever for any business selling across borders.

DHL Express enters this AI customs space not as a newcomer to artificial intelligence but as an organization that has been layering intelligent automation across its operations for several years. In 2022, the company deployed OptiCarton, an AI system that optimizes box sizing to reduce dimensional waste. In 2024, DHL launched myDHLi GenAI for its freight division and introduced generative AI tools across DHL Supply Chain warehouses. The customs vision tool represents a different vector of AI application, however: instead of optimizing internal operations, it places intelligence directly in the hands of the customer at the moment a shipping decision is made. For DHL Express, which reported €24.43 billion in revenue and €3.16 billion in EBIT for fiscal year 2025 while its time-definite international daily shipment volumes declined 9.4%, the strategic logic is clear. In a market where volume is compressing, reducing the friction that prevents potential shippers from completing cross-border transactions is not an innovation luxury but a growth imperative.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 13, 2026, a press release from two companies that most warehouse workers have never heard of working together quietly redrew the boundary between what machines can do and what humans must do inside a logistics facility. SAP, the German enterprise software giant whose systems touch approximately 77 percent of the world’s transaction revenue, announced that it had deployed fully autonomous, AI powered robots inside its own active logistics warehouse at St. Leon-Rot, Germany. The robots were not prototypes confined to a demonstration zone. They were not tele-operated by engineers watching from a control room. They were performing box folding, packaging, and shipping fulfillment tasks entirely on their own, integrated directly into SAP’s live Logistics Management (LGM) cloud native system, receiving work orders from the same software that manages inventory for some of the largest corporations on Earth. The technology partner was Cyberwave, a robotics AI company whose platform enables organizations to train, fine tune, and deploy robots capable of handling high variability tasks in real world environments without requiring deep robotics expertise. SAP’s head of warehouse and shipping operations stated: “Physical AI is no longer a concept. It is delivering real value today.” The deployment was already delivering measurable throughput improvements and freeing human workers from repetitive, physically demanding tasks.

This was not SAP’s first step into physical robotics. In January 2026, SAP had partnered with BITZER, a global leader in heating and cooling technologies, and NEURA Robotics to run a proof of concept under the name Project Embodied AI. In that pilot, NEURA’s 4NE1, one of Europe’s most advanced humanoid robots, was virtually trained on NVIDIA Isaac Sim software before being deployed to perform pick tasks independently at a BITZER warehouse. The robot demonstrated the ability to create material orders automatically, operate 24/7 to meet shifting demand, and connect directly to SAP Extended Warehouse Management (EWM) with no costly middleware required. Dr. Lukasz Ostrowski, head of Embodied AI and Robotics at SAP, called the BITZER pilot a decisive step forward. Four months later, the Cyberwave deployment at St. Leon-Rot moved from proof of concept to live production, marking the moment SAP began operationalizing advanced robotics within its own facilities rather than merely demonstrating them at a partner site.

The significance of the May 13 announcement lies not in the tasks themselves, box folding and packaging are among the most mundane activities in any warehouse, but in the architecture that made them possible. Traditional warehouse automation requires months of engineering, custom programming for each task, fixed conveyance paths, and extensive reconfiguration whenever product types or packaging formats change. The Cyberwave platform inverts this approach entirely. Operators collect training data using intuitive demonstration interfaces, essentially showing the robot what to do rather than coding it step by step. The platform then uses Vision Language Action (VLA) models and Reinforcement Learning (RL) to produce robot policies that generalize across object types, orientations, and workflow variations. A robot trained to fold one type of box can adapt to a different size or material without reprogramming. The entire integration, from robot training to live operation, was completed using SAP Business Technology Platform (BTP) and the SAP Embodied AI Service, which translates logistics tasks from SAP’s enterprise system into precise robot commands. What previously required weeks of setup was accomplished in a matter of minutes. Cyberwave’s approach reduces training time from weeks to hours, and non expert operators can teach robots new tasks through simple demonstrations. The robots operate with real time feedback loops that allow continuous refinement as warehouse conditions evolve, meaning the system does not merely execute instructions but learns and improves while working.

This deployment arrives at a moment when the global warehouse labor crisis has moved from inconvenient to existential. Data from 2026 shows that labor shortages affect 77 percent of warehouses and manufacturing facilities according to a survey of more than 800 U.S. manufacturing companies, driving operational costs 15 to 25 percent above industry averages. The quit rate within the first 90 days of warehouse employment reaches 30 to 35 percent, meaning roughly one in three new hires leaves before completing their first quarter. Annual turnover for warehouse staff averages 36 percent, creating a perpetual cycle of recruiting, training, and onboarding that consumes management attention and capital. Hiring costs in logistics have risen 22 percent year over year, while demand for supply chain talent is growing at 31 percent annually according to Talent Traction, far outstripping the available labor pool. In the United States, the Bureau of Labor Statistics reported total transportation and warehousing employment of approximately 6.5 million workers as of January 2026, down 1.8 percent from the previous year. The American Trucking Associations projects a driver shortfall of 174,000 by the end of 2026. The numbers describe an industry that is structurally unable to staff its existing operations, let alone scale to meet the relentless growth of e-commerce fulfillment volumes.

What makes the SAP Cyberwave deployment different from the hundreds of warehouse robotics announcements that have preceded it is the combination of three elements that have never before appeared together in a single operational system. The first is enterprise grade integration: the robots are wired directly into SAP’s logistics management software, the same platform used by thousands of companies worldwide. When SAP’s system generates a shipping order, the robot receives it, executes it, and reports completion through the same data pipeline as a human worker. The second is generalization without reprogramming: Cyberwave’s VLA and RL models allow robots to handle the variability that has historically defeated automation, different box sizes, irregular product shapes, changing packaging specifications, without stopping for reconfiguration. The third is speed of deployment: the minutes not months integration timeline means that the technology is accessible not only to companies with dedicated robotics engineering teams but potentially to any organization running SAP, a population that includes over 440,000 customers across 180 countries, approximately 80 percent of which are small to medium sized enterprises. SAP is not merely automating its own warehouse. It is building the infrastructure layer that could enable its entire customer base to deploy autonomous robots with the same ease as activating a new software module. The warehouse at St. Leon-Rot is small. The robots are performing tasks that skilled humans can still complete faster today. But the architectural pattern, enterprise software generating tasks, AI translating tasks into robot commands, robots executing and learning in real time, represents a structural shift whose implications will unfold across the next decade.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On the morning of May 2, 2026, at exactly 3:00 a.m. Eastern Time, every screen at every Spirit Airlines gate across the United States went dark. The airline that had spent two decades rewriting the rules of budget travel cancelled its entire schedule, locked its customer service counters, and began what executives called an orderly wind down. Seventeen thousand workers lost their jobs before sunrise, including 14,000 direct employees and thousands of contractors whose livelihoods depended on Spirit’s operations. Millions of passengers holding tickets for the weeks and months ahead woke up to find their itineraries worthless, with the airline instructing customers not to come to the airport. Southwest Airlines alone flew more than 20,000 stranded Spirit passengers by late Saturday afternoon, while United, Delta, JetBlue, and Southwest capped rescue fares at roughly $200 one way. Spirit became the first major American carrier to collapse in twenty five years, and the speed of its fall, from operational airline to historical footnote in a single overnight memo, shocked even the analysts who had been predicting trouble for months. But Spirit was not the cause of the crisis. It was the canary.

The real story began 11,000 kilometers away, in a narrow waterway between Iran and Oman where roughly one fifth of the world’s oil supply passes every day. When the United States and Israel launched military operations against Iran in late February 2026, the Strait of Hormuz effectively closed to commercial tanker traffic. Within weeks, the cascading effects reached every airport fuel depot on the planet. Jet fuel prices surged past $4 per gallon at the U.S. Gulf Coast by April, an increase of more than 80 percent in barely two months. European jet fuel hit a record $1,838 per metric tonne at the start of April, more than double the $831 level recorded before the conflict began. Middle East refineries supply around 75 percent of Europe’s jet fuel, and the war severed that pipeline almost overnight. For airlines operating on margins that rarely exceed five percent in a good year, the math turned brutal within weeks.

The industry’s response was immediate and sweeping. Data from aviation analytics firm Cirium showed that global carriers removed approximately 13,000 flights and two million passenger seats from May 2026 schedules within just eleven days of the price spike becoming undeniable. Lufthansa announced the cancellation of 20,000 flights between May and October, a decision calculated to conserve roughly 40,000 metric tonnes of jet fuel across the network. The group no longer offered flights to three destinations entirely: Bydgoszcz and Rzeszów in Poland and Stavanger in Norway. Lufthansa disclosed that it expected to absorb an additional €1.7 billion in fuel costs for 2026. KLM cancelled 160 intra-European flights, equivalent to eighty round trip services from Amsterdam Schiphol, representing just under one percent of its European schedule for May but targeting routes that were no longer financially viable at the new fuel price, including high frequency connections to London and Düsseldorf. Qatar Airways, whose hub sits within visual range of the conflict zone, slashed two million seats from its June through October schedule. Emirates cut 700,000 seats and Etihad Airways removed 450,000 over the same period. In North America, United Airlines trimmed its previously planned summer capacity by approximately five percent, while Delta cut about three and a half percent. Air Canada, Air France, SAS, and British Airways followed with reductions of their own, each calibrated to the airline’s specific fuel exposure and hedging position.

The financial transmission mechanism extended far beyond cancelled flights. Air France KLM initially imposed a €50 fuel surcharge on long haul round trip tickets in March, then doubled it to €100 by April 13, with North American routes carrying a separate €70 levy. Short and medium haul flights incurred a more moderate €10 per round trip. SunExpress, the joint venture between Turkish Airlines and Lufthansa, introduced a €10 surcharge per passenger per flight segment on all Turkey to mainland Europe routes effective May 1. Across the board, the average international airfare from the United States reached $1,101 in the last week of April according to Kayak data, a 16 percent increase compared to the same period the previous year. Domestic fares inside the United States climbed even more steeply, rising 24 percent year on year. Analysts at multiple firms noted that airlines were not merely passing through fuel costs but using the crisis to reprice structurally, testing whether consumers would absorb permanently higher fare levels even after fuel eventually retreated. The total economic burden of flight disruptions in 2026 was estimated at $34 billion when airline losses, passenger expenses averaging $500 to $1,200 per disruption event, and broader productivity impacts were combined, a figure that industry observers described as an annual floor rather than an anomaly.

What made this crisis different from previous fuel shocks was its compound nature. The 1973 oil embargo raised prices but left supply chains largely intact. The 2008 spike was driven by speculation and demand, forces that corrected themselves within months. The 2026 fuel crisis combined a physical supply disruption at the world’s most critical chokepoint with an active military conflict whose duration no analyst could predict, layered on top of an aviation industry still carrying structural debt from COVID 19 restructuring. Airlines that had spent four years rebuilding route networks and rehiring crews suddenly found themselves dismantling that work in a matter of weeks. The result was not a temporary scheduling adjustment but what IATA characterized as a potential reshaping of global aviation capacity for the remainder of the decade.

The numbers tell a story of an industry bleeding from multiple wounds simultaneously. Beyond the two million seats lost in May, carriers had already removed an additional 9.3 million seats from summer schedules covering June through September according to Cirium. For destinations across Southern Europe, Southeast Asia, and the Caribbean that depend on summer tourism revenue, those missing seats translated directly into missing visitors, missing hotel bookings, and missing economic activity. On May 7, Spain, Italy, and France issued official alerts warning citizens and inbound tourists to expect reduced flight availability and elevated prices throughout the peak season, prompted by what authorities described as a 30 percent drop in European jet fuel supplies. The domino had fallen in the Strait of Hormuz, and by May 2026, its reverberations were being felt in airport terminals, tourism ministries, and household budgets on every continent.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On April 22, 2026, the National Highway Traffic Safety Administration (NHTSA) opened recall campaign 26V255000 covering exactly 173 Tesla Cybertruck vehicles from model years 2024 through 2026, all equipped with 18-inch steel wheels. The defect: brake rotor stud holes may develop cracks under normal driving conditions, specifically from on-road disturbances and cornering forces, which could allow an entire wheel stud to separate from the hub assembly, creating a risk of loss of vehicle control. Tesla service centers were notified on April 24, 2026, and affected owners will receive mailed notifications on or about June 20, 2026.

The technical root cause, as described in the recall filing, traces to the mounting geometry between the hub and bearing assembly. Tesla’s internal investigation began in August 2025, prompted by engineering analysis that identified a stress concentration pattern in the rotor’s stud hole area under operational loads. The combination of the 18-inch steel wheel’s weight distribution, the Cybertruck’s substantial curb weight of approximately 6,600 pounds, and the geometric relationship between the rotor mounting surface and the wheel hub created conditions in which repetitive loading cycles could initiate and propagate fatigue cracks around the stud attachment points. Tesla reported awareness of three warranty claims that may be related to this failure condition but stated it was unaware of any crashes, injuries, or fatalities.

The remedy is a physical recall, not an over-the-air software update, a distinction worth noting for a company that has relied heavily on remote fixes for previous recalls. Tesla will replace all four brake rotors, wheel hubs, and lug nuts on each affected vehicle with redesigned components featuring modified geometry that increases the contact area between the rotor and hub, thereby reducing stress concentration under operational loads. The repair is provided at no cost to owners.

By the metrics of automotive safety, this is a minor recall. 173 vehicles. No injuries. No crashes. A known root cause with a straightforward engineering fix. In any other context, it would merit a brief notice in trade publications and nothing more.

But the number 173 is where this recall transforms from a routine safety action into one of the most revealing data points about Tesla’s product strategy in recent years, because 173 is not just the number of vehicles affected by a brake rotor defect. It is the total number of Cybertruck Long Range Rear-Wheel-Drive units that Tesla produced and sold during the variant’s entire commercial existence. The recall, by covering every single RWD Cybertruck ever built, inadvertently disclosed what Tesla had never publicly confirmed: that the most affordable version of its most polarizing vehicle was a commercial failure of extraordinary proportions.

The Cybertruck Long Range RWD was launched in April 2025 at a price of $69,990, positioned as the entry-level pathway into the Cybertruck lineup. The strategic logic was straightforward: after establishing the Cybertruck’s brand presence with the higher-priced All-Wheel Drive and Cyberbeast trims, Tesla would expand the addressable market by offering a more affordable single-motor variant that could qualify for the $7,500 federal EV tax credit, bringing the effective price below the psychologically significant $65,000 threshold. On paper, this was a textbook market penetration strategy, the same approach that had succeeded spectacularly with the Model 3 and Model Y.

In practice, the RWD Cybertruck lasted five months. By September 2025, Tesla had removed the variant from its online configurator, and production ceased entirely by November 2025. The vehicles covered by the recall were manufactured between March 21, 2024, and November 25, 2025, but the RWD-specific production run occupied only the final months of that window. Fewer than 250 units were reportedly built in total, of which 173 were delivered to customers, meaning that Tesla could not sell even the limited quantity it had produced.

The failure of the RWD variant cannot be attributed to a single cause but rather to a convergence of product positioning errors that, from an operational perspective, reveal systematic weaknesses in Tesla’s demand forecasting and product portfolio management processes. At $69,990, the RWD Cybertruck was nominally the cheapest variant, but the price point still positioned it $10,000 to $20,000 above competing electric pickups from Ford (F-150 Lightning) and Rivian (R1T), while offering fewer capabilities. The single-motor rear-wheel-drive configuration in a vehicle weighing over three tons struck many potential buyers as functionally inadequate for a truck marketed on its ruggedness and off-road capability. The feature compromises necessary to reach the $69,990 price point, which reportedly included reductions in interior appointments and certain technology features, created a product that satisfied neither the performance-oriented early adopters who had already purchased the premium trims nor the value-conscious buyers who found the price still too high relative to alternatives.

This recall also lands in the broader context of a Cybertruck program in significant commercial distress. Total Cybertruck sales fell from 38,965 units in 2024 to 20,237 units in 2025, a decline of 48.1% that represented the steepest year-over-year drop of any electric vehicle in the United States market on a volume basis. Fourth-quarter 2025 sales collapsed to 4,140 units, down 68.1% from the 12,991 sold in the same quarter of the prior year. An Electrek investigation in April 2026 revealed that SpaceX had purchased 1,279 Cybertrucks in Q4, raising questions about how much of the reported sales volume reflected genuine consumer demand versus intra-Musk-ecosystem purchasing. The Cybertruck, which has been subject to more than ten separate recall campaigns since deliveries began in late 2023, addressing defects ranging from accelerator pedal assemblies detaching (3,878 units, April 2024) to windshield wiper failure (11,688 units, June 2024) to exterior trim panels falling off (46,096 units, 2025) to front parking lights exceeding brightness regulations (63,000 units, 2025), now carries a quality reputation that stands in stark contrast to the manufacturing excellence that Tesla’s operational narrative has long projected.

The following sections will examine what the Cybertruck RWD failure reveals about demand sensing, product-market fit validation, and the operational cost of launching variants without adequate market signal confirmation, principles that extend well beyond Tesla to any manufacturing organization making high-stakes product portfolio decisions.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On May 4, 2026, Amazon officially launched Amazon Supply Chain Services (ASCS), opening its entire end-to-end logistics infrastructure to businesses of all types and sizes across all industries. This announcement marks a fundamental shift in how the world’s largest e-commerce company positions itself: no longer just a retailer that built logistics to serve its own operations, but now a full-service logistics provider offering its infrastructure as a standalone commercial product to any enterprise, anywhere.

The scope of ASCS covers the complete supply chain spectrum. Amazon’s transportation network spans ocean freight, air freight, ground freight, and rail freight, supported by a fleet of over 80,000 trailers, more than 24,000 intermodal containers, and over 100 cargo aircraft. The service portfolio includes full truckload (FTL), less-than-truckload (LTL), and intermodal transport; air freight; inbound shipping from China to the United States including customs clearance; 2-to-5-day parcel shipping; bulk storage and distribution; and seven-day-a-week delivery services. Businesses also gain access to Amazon’s proprietary AI forecasting models and its vast supply chain data set, which help optimize inventory placement based on demand patterns.

Among the first enterprises to adopt ASCS are four major brands across different industries. Procter & Gamble, the consumer goods giant, is using Amazon’s freight services to transport raw materials to production facilities and move finished goods across its distribution network. 3M, the industrial and manufacturing conglomerate, is leveraging Amazon’s freight capabilities to move products from its manufacturing sites to distribution centers worldwide. Lands’ End, the apparel and home goods retailer, and American Eagle Outfitters, the fashion retail chain, have also signed on as early adopters. The service targets industries including healthcare, automotive, manufacturing, and retail, signaling Amazon’s ambition to serve far beyond its traditional e-commerce base.

According to Amazon’s data, sellers using its fully managed supply chain option report an average 20% higher sales conversion, and the company claims transportation costs up to 25% lower than alternatives. Additionally, businesses consolidating into a single inventory pool through Amazon’s distribution network can reduce total inventory requirements by 20% on average. These figures, while representing Amazon’s own claims and will require independent verification over time, illustrate the scale advantage that Amazon believes its network delivers.

The market reaction was immediate and severe. On the day of the announcement, FedEx Corp. (NYSE: FDX) tumbled 9.4% to approximately $359, its worst trading session in over a year. United Parcel Service (NYSE: UPS) dropped 9.7% to approximately $97. GXO Logistics (NYSE: GXO), the contract logistics provider, cratered 16.5% on contract-loss fears, on pace for the stock’s worst day ever. C.H. Robinson Worldwide (NASDAQ: CHRW), the freight brokerage giant, sank 8.7%. In contrast, Amazon shares rose 1.2%. The Dow Jones Transportation Average fell into bear market territory, a signal that the broader transportation sector views Amazon’s entry as a structural threat rather than a temporary disruption.

Morgan Stanley analyst Ravi Shanker described the launch as a potential “watershed moment for North American freight transportation companies.” Bloomberg headlined its coverage: “FedEx, UPS Shares Tumble on Amazon’s ‘Watershed’ Logistics Move.” The unresolved question, as Shanker noted, is whether ASCS represents a structural reset for the entire listed logistics complex, or whether the entrenched customer relationships and air freight scale that FedEx and UPS still control prove more durable than the market’s initial reaction suggests. It is worth noting that UPS had already been preparing for reduced Amazon volumes: CEO Carol Tomé stated that “upon completion of the Amazon glide-down, 2026 will be an inflection point,” and the company had already eliminated approximately 48,000 positions while closing 93 facilities in 2025.

The most significant competitive disruption in the global third-party logistics (3PL) industry in decades. Amazon, which spent over two decades and hundreds of billions of dollars building a logistics network originally designed to deliver its own packages, is now offering that same network as a commercial product to the very businesses that FedEx, UPS, and traditional 3PL providers have served for generations. The implications for supply chain strategy, competitive dynamics, and operational model design across every industry are substantial, and will be explored in the following sections.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On April 15, 2026, Snap Inc., the parent company of the social media platform Snapchat, officially announced a sweeping reduction in force that would eliminate approximately 1,000 full time positions, representing roughly 16% of its global workforce. The announcement came directly from CEO Evan Spiegel through an internal memo distributed to all employees, in which he framed the decision not as a retreat or a sign of weakness but as a deliberate strategic realignment designed to position the company for what he described as a fundamentally new era of technology driven work. In addition to the 1,000 positions being eliminated, Snap confirmed that at least 300 open roles that were actively being recruited for would be immediately closed, effectively removing nearly 1,300 potential positions from the company’s organizational structure in a single stroke.

What made this announcement particularly striking was not merely the scale of the cuts but the rationale Spiegel articulated to justify them. In his memo, the CEO stated explicitly that advances in artificial intelligence had enabled Snap’s remaining workforce to operate at a significantly higher velocity, reducing what he called “repetitive work” and allowing smaller teams to accomplish what previously required much larger groups. The most remarkable data point to emerge from the announcement was Snap’s disclosure that more than 65% of its new code was now being generated through AI powered tools, a figure that sent shockwaves through the technology industry and beyond because it represented one of the most concrete and quantifiable admissions by a major technology company that AI was not merely augmenting human work but actively and measurably replacing it at scale.

The financial dimensions of the restructuring were equally significant. Snap projected that the workforce reduction would deliver more than $500 million in annualized cost savings by the second half of 2026, a figure that underscored just how substantial the financial burden of the eliminated positions had been. The company estimated that it would incur one time severance charges ranging from $95 million to $130 million to execute the layoffs, meaning that the restructuring would pay for itself within roughly three months of full implementation. For employees affected by the cuts in the United States, Snap committed to providing four months of severance pay, continued healthcare coverage, accelerated equity vesting, and access to career transition support services, a package that was broadly viewed as relatively generous compared to industry norms during the current wave of technology sector layoffs.

The market’s response to the announcement was immediate and unambiguous. Snap’s stock price surged approximately 7% on the day of the announcement, a reaction that revealed the financial community’s overwhelming approval of the cost cutting measures and its belief that the company was making the right strategic trade by reducing headcount in favor of AI driven efficiency. This positive market reaction stood in sharp contrast to the human reality of the announcement, creating a tension that would become one of the defining narratives of the story: the same decision that destroyed 1,000 livelihoods simultaneously created billions of dollars in perceived shareholder value.

What made the April 15 announcement even more consequential was the fact that it did not occur in isolation. Just weeks earlier, in March 2026, Snap had already executed a separate round of cuts that eliminated approximately 1,300 employees. When combined with the April reduction, this meant that Snap had shed more than 2,300 positions in approximately six weeks, reducing its workforce from roughly 5,261 employees at the end of 2025 to fewer than 3,000 in a matter of weeks. The speed and depth of this contraction were remarkable even by the standards of the current technology industry downturn, which has seen more than 96,000 tech workers lose their positions across 95 companies in 2026 alone.

The timing of the announcement also carried significant strategic implications. Snap’s restructuring came just days before Meta Platforms announced its own plan to cut 8,000 employees (10% of its workforce) to fund AI infrastructure spending projected at $115 to $135 billion for 2026, and in the same month that the broader technology sector continued to grapple with the fundamental question of how aggressively to replace human labor with artificial intelligence systems. Spiegel’s memo positioned Snap not as a company in crisis but as a company that had arrived at the future ahead of its peers, one that had already proven through its 65% AI code generation rate that the substitution of human intellectual labor with machine generated output was not a theoretical possibility but an operational reality already embedded in its daily workflows.

The CFO departure that followed the layoff announcement added another layer of complexity to the story. Industry analysts noted that the timing of the CFO’s exit, coming so closely after the restructuring announcement, suggested potential internal disagreements about the pace and scope of the transformation Spiegel was driving. Whether the departure reflected philosophical differences about the company’s direction or simply the natural turnover that accompanies major organizational change remained a subject of speculation, but it reinforced the impression that Snap was undergoing not merely a workforce reduction but a fundamental reimagining of what a technology company’s operating model should look like in the age of generative AI.

Share

Read more

Subscribe now

This is the bilingual post in English and Vietnamese. Vietnamese is below.

Đây là bài viết song ngữ Anh-Việt. Tiếng Việt ở bên dưới.

English

PART 1 – OFFICIAL INFORMATION

On the morning of April 27, 2026, Microsoft and OpenAI simultaneously published joint statements announcing a fundamental restructuring of the partnership that has defined the global artificial intelligence landscape since 2019. Microsoft will no longer hold exclusive rights to sell, distribute, or resell OpenAI’s AI models. The exclusivity clause that had given Microsoft sole commercial access to the most powerful AI technology on the planet has been dissolved. OpenAI is now free to sell its products across any cloud provider, including Amazon Web Services and Google Cloud Platform.

The restructuring touches every dimension of the relationship. Microsoft’s license to OpenAI’s intellectual property continues through 2032 but is now explicitly non exclusive. Revenue share payments from OpenAI to Microsoft continue through 2030 but are now subject to a total cap not previously in place. Microsoft will no longer pay a revenue share to OpenAI on products it resells on Azure. The deal also eliminates the controversial AGI clause, which would have automatically transferred OpenAI’s intellectual property to Microsoft upon achieving artificial general intelligence.

The partnership began in July 2019 with a $1 billion investment. Between 2021 and early 2023, Microsoft deployed approximately $2 billion more. In January 2023, weeks after ChatGPT surpassed 100 million users, Microsoft committed roughly $10 billion. The total $13 billion gave Microsoft a 26.79% stake, now worth approximately $228 billion at OpenAI’s $852 billion valuation, a 17.6x return. OpenAI’s revenue trajectory has been extraordinary: $2 billion in 2023, $6 billion in 2024, $20 billion in 2025, reaching approximately $25 billion annualized by February 2026. CEO Sam Altman forecasts revenue exceeding $280 billion by 2030.

The restructuring was triggered by OpenAI’s aggressive diversification. In November 2025, OpenAI signed a $38 billion cloud deal with AWS. In February 2026, Amazon invested $50 billion in OpenAI, with OpenAI committing to spend $100 billion on AWS over eight years and deploying 2 gigawatts of Amazon’s Trainium chips for its enterprise platform Frontier. An internal OpenAI memo stated Microsoft had “limited our ability” to reach enterprise clients. Microsoft reportedly considered legal action but chose negotiation, recognizing that enforcing exclusivity against a partner valued at $852 billion would be commercially destructive.

Microsoft shares declined on the announcement, though analysts noted the company retains its equity stake, primary cloud partner status with first ship rights on Azure, and deep integration across Copilot, GitHub, and Office 365. The AI cloud market, previously structured as a near duopoly between Microsoft with OpenAI and Amazon with Anthropic, now enters a more fluid competitive phase.

Share

Read more