Chinese tech giant Huawei has reportedly developed a powerful artificial intelligence chip that could rival high-end processors from US chip maker Nvidia.
The Shenzhen-based Huawei is poised to start testing a new AI chip called the Ascend 910D, and has approached local tech firms, which are slated to receive the first batch of sample chips by late May, The Wall Street Journal reported on April 27, citing people familiar with the matter.
The development is still at an early stage, and a series of tests will be needed to assess the chip’s performance and get it ready for customers.
Huawei is pinning hopes on its latest Ascend AI processor being more powerful than Nvidia’s H100 chip, which was used for AI training in 2022.
Huawei is also poised to ship more than 800,000 earlier model Ascend 910B and 910C chips to customers, including state-owned telecoms operators and private AI developers such as TikTok parent ByteDance.
Beijing has also reportedly encouraged Chinese AI developers to increase purchases of domestic chips as trade tensions between China and the US escalate.
In mid-April, Nvidia stated that it was expecting around $5.5 billion in charges associated with its AI chip inventory due to significant export restrictions imposed by the US government affecting its business with China.
The Trump administration added Nvidia’s H20 chip, its most powerful processor that could be sold to China, to a growing list of semiconductors restricted for sale to the country.
Some key components for AI chips, such as the latest high-bandwidth memory units, have also been restricted for export to China by the US.
Huawei is focusing on building more efficient and faster systems, such as CloudMatrix 384, a computing system unveiled in April, connecting Ascend 910C chips. This would leverage their chip arrays and use brute force rather than making individual processors more powerful.
China seeks self-reliance on AI
Reuters reported on April 26, citing state media reports, that Chinese President Xi Jinping pledged “self-reliance and self-strengthening” to develop AI in the country.
“We must recognise the gaps and redouble our efforts to comprehensively advance technological innovation, industrial development, and AI-empowered applications,” Xi said at a Politburo meeting study session on April 25.
Donald Trump (left) meeting with Xi Jinping (right) in 2018 at the G20. Source: Dan Scavino
“We must continue to strengthen basic research, concentrate our efforts on mastering core technologies such as high-end chips and basic software, and build an independent, controllable, and collaborative artificial intelligence basic software and hardware system,” Xi added.
US President Donald Trump has repeatedly urged Xi to contact him for discussions about a potential trade deal after his administration imposed 145% tariffs on most Chinese goods.
China has stated that it is not having any talks with the US and that the country should “stop creating confusion.”
Elon Musk recently announced on X his plans to visit India this year. (AI image)
Elon Musk’s Tesla to enter India soon? Tesla Inc. is returning reservation fees to Indian customers who had pre-booked its Model 3. This action suggests the Elon Musk-led US electric vehicle manufacturer is preparing to enter India’s automobile market, which ranks as the third-largest globally. According to a Bloomberg report, the company’s correspondence to customers who placed bookings in 2016 says: “We would like to return your reservation fee for the time being. When we finalize our offerings in India, we will reach out in the market again. We hope to see you back with us once we are ready to launch and deliver in your country.” Tesla is processing refunds for these long-standing reservations as the earlier version of the Model 3 is being discontinued. These communications, sent from official Tesla email addresses, indicate the manufacturer’s intention to commence operations in India, following years of discussions regarding the country’s substantial import tariffs, the report said. Also Read | ‘India a very hot market but…’: Elon Musk-led Tesla says 100% car tariffs make customers anxious Elon Musk recently announced on X his plans to visit India this year, coinciding with ongoing trade negotiations between India and the US, which could potentially result in reduced automobile tariffs. A revised tariff framework might influence Tesla’s strategic outlook. The company experienced its first decline in global vehicle deliveries in over ten years, whilst facing significant competition from BYD Co. The presence of Tesla vehicles in India would appeal to the growing affluent middle-class segment, though it could potentially impact local automobile manufacturers who provide employment to numerous workers in their production facilities. Meanwhile, in another significant development Tesla has secured a private office space in Mumbai’s Kurla area, adjacent to the Bandra-Kurla Complex business district, demonstrating its commitment to expand operations in India. Tesla has obtained a 30-seat workspace within a co-working facility at Phoenix Market City with a monthly payment of Rs 3 lakh, according to an ET report. Also Read | Apple tempted by India! In shift away from China, 70-80 million iPhones to be made in India soon amid Trump tariff tensions The company has agreed to a one-year tenancy with a three-month lock-in clause, offering short-term adaptability whilst Tesla develops its India strategy. The arrangement between Tesla India Motor & Energy and the co-working provider EFC was finalised earlier this month. “The new office is likely to support early-stage functions such as local hiring, regulatory coordination, vendor meetings, or backend support for its anticipated entry into the Indian EV market,” according to a source familiar with the situation.
Russia launches deadliest attack on Kyiv this year
Headline update to reflect proper location and scope of attack: A previous version of this video misstated today’s attack as the deadliest on Ukraine in months.
President Donald Trump said his meeting in Rome with Ukrainian President Volodymyr Zelenskyy “went well” and expressed disappointment with Russian President Vladimir Putin‘s attacks on Ukraine, urging him to sign a peace deal.
“I want him to stop shooting, sit down and sign a deal,” Trump said of Putin, speaking to reporters at the Morristown Municipal Airport in New Jersey while traveling back to the White House on April 27. “We have the confines of a deal, I believe. And I want him to sign it.”
Trump met with Zelenskyy on the sidelines of Pope Francis‘ funeral. Trump said Zelenskyy told him “he needs more weapons, but he’s been saying that for three years.”
Meanwhile, Trump envoy Steve Witkoff met with Putin in Moscow on April 25 as U.S. officials pressed to halt the three-year conflict, and warned they could walk away from the peace talks.
This is a “very critical week” that will determine whether the Trump administration continues with negotiations over ending the war, Secretary of State Marco Rubio said in an April 27 interview. Rubio told NBC’s “Meet the Press” that he believes the two sides are “close” to reaching a deal and there are “reasons to be optimistic,” but “we’re not close enough.”
“This week is going to be a really important week in which we have to determine whether this is an endeavor that we want to continue to be involved in,” Rubio added.
Trump repeatedly has expressed his frustration with Putin on social media in recent days. He chided Putin in an April social media post after Russia bombed Ukraine’s capital Kyiv.
“I am not happy with the Russian strikes on KYIV. Not necessary, and very bad timing. Vladimir, STOP!” Trump wrote, adding: “Let’s get the Peace Deal DONE!”
On April 26, Trump wrote that there was “no reason for Putin to be shooting missiles into civilian areas, cities and towns, over the last few days” and that “it makes me think that maybe he doesn’t want to stop the war, he’s just tapping me along, and has to be dealt with differently.”
Trump tells European leaders: ‘Let’s see what happens’
Trump, Rubio, and Vice President JD Vance have warned Ukraine that Washington could abandon trying to negotiate a Ukraine-Russia peace settlement if there is no progress on a deal soon.
“If it’s not possible, if we’re so far apart that this is not going to happen, then I think the president is probably at a point where he’s going to say, ‘Well, we’re done,’” Rubio said on April 18.
Trump was asked during a meeting with Norwegian leaders whether the U.S. would stop sending Ukraine weapons and sharing military intelligence with the country if the administration decides to walk away from peace negotiations.
“Let’s see what happens,” Trump said. “I think we’re gonna make a deal. And if I make a deal that will be wonderful, we won’t have to worry about your question.”
As the U.S. works to get a peace deal, where to draw territorial lines is a big sticking point. Trump made it clear in a Time magazine interview that Ukraine will not regain control over Crimea, which Russia seized in 2014.
“Crimea will stay with Russia. And Zelenskyy understands that, and everybody understands that it’s been with them for a long time,” Trump said.
Trump also said Ukraine can’t join NATO.
Trump said he discussed Crimea with Zelenskyy and believed the Ukrainian leader is prepared to give it up.
“I see him as calmer. I think he understands the picture, and I think he wants to make a deal,” Trump said of Zelenskyy.
Contributing: Kim Hjelmgaard and Francesca Chambers, USA TODAY
Browns GM Andrew Berry on why they drafted Shedeur Sanders
Despite already drafting Dillon Gabriel to an already crowded Browns QB room, GM Andrew Berry explains why they couldn’t pass up Shedeur Sander in the fifth round.
While hosting a Twitch livestream on Saturday during the 2025 NFL Draft, Shilo Sanders told his viewers that – after starting the draft process with his dad representing him – he had signed with an agent.
“Dad was our agent,” Sanders said, referring to his brother Shedeur and himself, “but that hasn’t been working out too good. So today I had to sign with an agent.”
The elder Sanders brother appeared to be referring to Shedeur’s draft slide from potential first-round pick to fifth-round selection by the Browns. And though Sanders was smiling as he broke the news on the livestream, he wasn’t entirely joking.
Sanders had indeed “fired” his father and hired NFL agent Drew Rosenhaus to represent him.
Rosenhaus announced Sanders’ signing with a social media post on Saturday evening welcoming the defensive back to the Rosenhaus Sports Representation “family.”
Sanders didn’t hear his name called during the 2025 NFL Draft. He signed a contract with the Tampa Bay Buccaneers as an undrafted free agent shortly after the draft ended.
He started 19 games at safety for the Colorado Buffaloes across the 2023 and 2024 seasons after two years at South Carolina and two years at Jackson State. In his time in Boulder, Sanders tallied 134 tackles with five forced fumbles, one sack and an interception that he returned for a touchdown.
Simply sign up to the Media myFT Digest — delivered directly to your inbox.
Music creation start-up Splice has agreed to acquire the orchestral sound studio Spitfire Audio for about $50mn, as the Goldman Sachs backed company looks to buy up audio samples and technology to build songs using artificial intelligence.
London-based Spitfire Audio owns a deep library of digital orchestral sounds — short recordings of violins, cellos, drums and basses that composers can use to create music. Its palette of sounds, such as a $299 drum kit recorded by composer Hans Zimmer, have been used in Radiohead songs and in the scores for films such as Moonlight.
Splice is a platform that offers audio samples — such as vocal hooks, drum beats and guitar riffs — for anyone, ranging from novices to professional producers, to create music. The group was founded in 2013 and has grown quickly as it benefits from soaring demand for music creation tools. Splice is profitable, reaching annual revenue of more than $100mn and 600,000 paying subscribers, according to people familiar with the matter.
Splice and Spitfire declined to disclose the price of the deal. A person familiar with the matter said it was about $50mn.
The rise of platforms like Splice reflects the changing nature of music production. As software production tools have become more sophisticated, anyone can create music from their bedroom. Splice and Spitfire offer sounds that are often just a few seconds long, in repeated loops that can be layered with other instrumentals or vocals to build out a song.
Hit songs have used samples from the Splice library that can be downloaded for as little as $13. The samples include the guitar riff used in Sabrina Carpenter’s “Espresso” — the guitar loop and drum sounds were created by Vaughn Oliver, a DJ whose samples are popular on Splice. Other hits, such as Doja Cat’s “Say So” and Playboi Carti’s “OPM Babi”, have also used Splice samples.
Splice was valued at nearly $500mn in a 2021 fundraising led by Goldman Sachs and entrepreneur Matt Pincus’s investment group MUSIC.
Spitfire has a library of “super high-end, incredibly luxurious rich sounds, recording with the BBC Symphony Orchestra or Abbey Roads Studio”, said Splice chief executive Kakul Srivastava. She also wanted to acquire the underlying audio technology and talent that Spitfire has built, she said.
Splice has been rolling out AI technology to help musicians build songs, using algorithms to suggest samples based on a genre, mood or even singing into your mobile phone. Pincus says this AI element has led to “explosive” growth for Splice in the past few months, calling it a “land grab” in the still nascent market.
All of the sounds on Splice are initially created by humans. The group has its own in-house artists who record samples, but it also scours the world for cutting-edge trends from different genres, such as funk artists in São Paulo or K-Pop musicians in Seoul. Splice pays these artists a percentage of its subscription revenue.
Srivastava says Splice is looking to do something fundamentally different from the prompt-based AI song generators that have popped up in the past few years and fuelled backlash in the music industry.
“Simplistic tools where you type in a prompt and a song comes out? Most musicians do not want to make music that way,” she said. “But AI will enable [artists] to do things they could not do today. They could use string quartets from Spitfire, but you might want to invent your own instrument. You can start with a particular sound and merge instruments together to get a novel sound that has never been heard before . . . we are still figuring it out.”
In the April 27 Rock & Roll Hall of Fame-themed episode, the contestants had a vast catalog of songs to perform, with their assignment being to sing a track from one of the nearly 400 Rock Hall inductees. “You’ve Got a Friend” singer James Taylor took on a mentorship role.
In a bold move, Desmond Roberts sang The Commodores in front of Lionel Richie – and taking a risk paid off, as Richie and Luke Bryan insinuated Desmond outdid the former frontman with his cover. Carrie Underwood also stepped outside her comfort zone, hesitantly providing critical feedback (but always couching it with praise, of course).
Desmond Roberts ‘tried to take my song,’ Lionel Richie claims
As Desmond rehearsed with Taylor, he admitted in an interview, “It’s a risk to play (‘Jesus is Love’) in front of Lionel.”
He started at the piano and made his way to center stage, where he let out several consecutive notes that showed off his head voice. His rendition was so successful that Desmond had Richie feeling self-conscious.
“You found some octaves that I didn’t sing,” Richie said. “You tried to take my song. That is a very hard song to sing, and you did it so well.”
Bryan cheekily agreed: “The whole time I was like, ‘Lionel didn’t sing it like that.’” He added that he didn’t believe the falsettos were necessary, but ended the compliment sandwich with “You’ve got one of the most dynamic voices I’ve ever heard.”
Carrie Underwood issues challenges to Canaan James Hill, Mattie Pruitt
In a rare moment, Underwood offered a well-couched rebuke of aspiring preacher Canaan James Hill, who took the stage with Aretha Franklin’s “Mary Don’t You Weep.”
“I kind of think you cheated the system a little on the Rock & Roll Hall of Fame,” she said.
Bryan agreed with Underwood, though he noted, “We love the gospel stuff.”
Earlier in the episode, Mattie Pruitt, this season’s youngest contender, unleashed her signature rasp in “Piece Of My Heart” by Big Brother and the Holding Company (featuring Janis Joplin). Underwood presented Mattie with a challenge: “If there’s anything I can say to you, (it) is you’ve gotta find your sass. … You need some sass, girl, come on!”
Who went home on ‘American Idol’?
Despite the unanimous praise he’d received, Desmond did not receive enough votes to crack the Top 12. Powerhouse singer Amanda Barise, no longer a self-described underdog, was also sent home.
The foresight gap is not a failure of technology—it is a challenge of institutional alignment.
Shutterstock
By Ioannis Ioannou, Associate Professor of Strategy and Entrepreneurship at London Business School
In parts of East Africa, AI tools are increasingly able to predict rainfall patterns, crop failure, and soil degradation. Satellite imagery and machine learning models inform planting schedules and flag pest risks early. These tools—developed by agritech startups and increasingly integrated into the procurement strategies of multinational agribusinesses—offer resilience through foresight. But while global firms adjust sourcing and hedge exposure, smallholder farmers—who produce around a third of the world’s food—often lack the means to act on the same insights. Irrigation, credit, and institutional support are scarce. Foresight is not the constraint. Capacity is.
This asymmetry points to a deeper challenge: the growing gap between those who can anticipate disruption and those with the means to respond. Artificial intelligence is reshaping how we assess climate risk, optimise resource use, and navigate supply chain volatility. But it may also widen existing disparities—less by harming low-capacity actors directly than by accelerating the adaptive advantage of those already ahead. As foresight becomes central to sustainability strategy, the question shifts from who can see risk to who can act on it—and who is positioned to do so more effectively than others.
At the heart of this shift lies what I call the ‘foresight gap’: the widening distance between insight and agency. The issue is not data availability. Many actors—municipalities, farming cooperatives, suppliers—now have access to forecasts, dashboards, and models. The issue is that insight, in the absence of financing, technical tools, or enabling institutions, often leaves actors aware of risk but unable to manage it. The result is uneven resilience, where some accelerate while others struggle to keep pace.
AI could reinforce this pattern. Corporates with advanced modelling reconfigure procurement, redirect investment, and hedge operations. Suppliers in more vulnerable environments shoulder the consequences. Risk moves, but does not diminish. This is a form of selective adaptation: those with resources fortify their position, while others absorb the shock. Over time, this dynamic risks undermining both equity and stability. A transition that reallocates risk without building shared capacity creates fragility at the system level.
Agriculture illustrates this dynamic, but the pattern extends far beyond it. Cities with the means to invest in AI-integrated infrastructure planning are improving energy efficiency and emergency response. Others, particularly in the Global South, operate with outdated systems and limited technical capacity. In insurance, AI is reshaping climate risk pricing, pushing premiums upward or removing coverage in high-risk areas. And in supply chains, predictive analytics allow firms to reroute around disruption, while those at the frontline remain vulnerable.
These shifts introduce a systemic risk that is often underestimated. When only some actors adapt, the costs of disruption cascade across sectors and geographies. Fragility at the margins—among smallholders, subcontractors, or overstretched public agencies—can ripple outward. Consider the 2022 floods in Pakistan: extreme weather forced global retailers to shift orders and reroute logistics, but smaller suppliers faced months of operational paralysis and loss of income. Without adequate capacity to absorb shocks at every level, the system as a whole becomes more brittle. Concentrated resilience cannot ensure collective stability.
This reveals a core tension. AI is often promoted as a force for inclusion, but without governance aligned to that purpose, it risks doing the opposite. The foresight gap is not a glitch. It reflects underlying disparities in capital, capability, and institutional design. Without deliberate attention to how foresight is distributed—and how action is enabled—the gap will grow.
What does a just transition look like under these conditions? The conventional framing has focused on costs, benefits, and protections—especially for workers and communities. These remain important. But in the context of AI, justice must also mean access to adaptive capacity. The transition cannot depend solely on who already has the tools to respond. It must support others to acquire them. That is not only a matter of fairness. It is a condition for managing shared risk in a connected world.
This perspective has direct implications for institutions. First, we need investment in public foresight infrastructure. Predictive tools must be designed and deployed with broad usability in mind. That includes open-access climate models, data collaboratives, and analytics suited to under-resourced settings. National adaptation plans and resilience strategies should be grounded in intelligence that reflects real-world constraints. Crucially, this infrastructure must not remain siloed in ministries or multilateral organisations. It must be made accessible to those on the front lines of disruption—local governments, cooperatives, and civic organisations with the credibility to act and the trust to mobilise.
Second, we need AI-to-action partnerships. Companies using AI to manage their exposure should contribute to the adaptive capabilities of suppliers, contractors, and local communities. This is not a philanthropic gesture. It is a practical approach to reducing risk concentration across the value chain. Some firms are already exploring shared data platforms with suppliers or financing adaptation initiatives as part of broader ESG-linked targets. But these efforts remain isolated. What is needed is a shift in mindset—from risk extraction to risk co-management.
Third, we need to update our understanding of fiduciary responsibility. Boards and investors must ask whether AI-enabled strategies are building system-level resilience or simply reinforcing firm-level protection. Are companies redistributing risk across weaker links, or investing in broader capacity? These are strategic questions, not compliance matters. Fiduciary duty, in this context, should include attention to how foresight tools influence the allocation of risk—and whether they contribute to long-term value creation that is stable, inclusive, and credible.
It’s also worth acknowledging that AI is not only a predictive tool. Generative models and large language systems are increasingly shaping how knowledge is accessed, decisions are supported, and strategies are refined. These technologies may also expand forms of adaptive capacity, especially when designed for public benefit or integrated into frontline decision-making. But the existence of such potential does not negate the foresight gap. It changes its contours. As AI becomes more embedded in corporate strategy and public infrastructure, the question remains: who can use these tools meaningfully—and under what conditions?
The trajectory of AI in sustainability is not predetermined. It will reflect choices about governance, design, and accountability. The foresight gap is not a failure of technology—it is a challenge of institutional alignment.
AI is often assessed by its ability to predict, automate, or optimise. But in the context of sustainability, a different question matters more: does it support a transition that builds resilience across the system? Concentrated foresight without distributed capacity creates brittleness. Resilience must extend beyond the firm, the sector, or the region.
Those shaping AI’s role in the transition—corporate leaders, investors, regulators—are defining more than tools. They are shaping trajectories. The key issue is no longer whether AI can enhance foresight. That question has been settled. The real question is whether we align that foresight with the capacity to act—broadly, deliberately, and with urgency.
That is where leadership will be measured. Not by who saw risk first, but by who ensured others were ready when it arrived.
Ioannis Ioannou is an associate professor of strategy and entrepreneurship at London Business School. His research focuses on corporate sustainability and the strategic integration of ESG issues by companies and capital markets.
Watch Mark Carney sworn in as Canadian prime minister
Mark Carney, who has never held political office, was sworn in as prime minister of Canada.
As Canadians prepare to vote for prime minister in an election where President Donald Trump looms large, Secretary of State Marco Rubio reiterated the president’s desire to make Canada a U.S. state.
“They’ll have their elections this week,” Rubio said about Canada on NBC’s “Meet the Press”. “They’re going to have a new leader, and we’ll deal with a new leadership of Canada. There are many things we work with cooperatively on Canada on, but we actually don’t like the way they treated us when it comes to trade.”
Pressed by “Meet the Press” host Kristen Welker whether Trump still wants to make Canada part of the United States, Rubio responded: “I think the president has stated repeatedly he thinks Canada would be better off as a state. I mean, he has said that based on what he was told by the previous prime minister, who said Canada can’t survive unless it treats the U.S. unfairly in trade.”
Trump began pressuring Canada on trade issues before taking office and repeatedly said the country should become a new U.S. state. He recently told Time magazine that he’s serious about Canada becoming part of the United States.
“I’m really not trolling,” Trump told Time. “Canada is an interesting case.” He complained about the U.S. trade deficit with Canada and said the United States doesn’t need Canadian products.
“And I say the only way this thing really works is for Canada to become a state,” he added.
Trump’s comments are roiling Canadian politics ahead of the April 28 vote for prime minister. The Liberal Party was trailing in polling when Prime Minister Justin Trudeau announced in January that he would resign.
But the latest CBC polling average shows Liberal Prime Minister Mark Carney with a slight edge in the race over Conservative Party leader Pierre Poilievre. Trump’s trade policies and comments about becoming a U.S. state provoked a backlash in Canada and have become a major issue in the race for prime minister.
