• Trump has signed an executive order to revive coal, partly to meet AI power demand in the US.
• It asks for coal to be labeled a "critical mineral" and directs federal agencies to identify coal resources.
• AI data centers require a lot of energy, sometimes requiring new power infrastructure.

President Donald Trump signed an executive order on Tuesday that aims to revive the US coal industry — in part to power soaring demand for AI.

The executive order "pushes for using coal to power new artificial intelligence (AI) data," per a White House announcement.

The measures would see some older coal plants, which were set to be retired, continue to produce electricity — including to power energy-intensive AI data centers across the US.

Encircled by a group of miners at the White House on Tuesday, Trump announced a plan he said "slashes unnecessary regulations that targeted beautiful, clean coal."

The executive order would label coal a "critical mineral," per the White House release. The administration also aims to "promote coal and coal technology exports, facilitate international offtake agreements for US coal, and accelerate the development of coal technology."

Powering the AI boom

A growing number of power-guzzling data centers, which are crucial for running AI services, have contributed to rising electricity demand in the US in recent years.

Elon Musk's startup xAI, for example, is planning to build a supercomputer in Memphis and has applied for a permit to construct a new electrical substation alongside it.

The International Energy Agency estimates that data centers consume 2% to 3% of the world's energy — a figure that's expected to grow in tandem with the mass adoption of AI.

Coal has been considered a stable energy source because it's fairly abundant, and its supporters say it can meet rising power demands for data centers.

As the US looks to bolster its AI ecosystem and encourage juggernauts to develop their technologies on home soil, the administration is touting coal as one way to power the AI boom.

Trump has repeatedly drawn comparisons to China — which has doubled down on its coal power plant production, as well as its renewables industry — and the economic advantages it has gained in the AI race as a result of this.

Big Tech giants such as Google, Meta, and Amazon, which are scrambling to develop ever more advanced AI models, had previously pledged to use low-carbon, renewable energy sources to build their data centers. However, much of the electricity needed for this has so far been generated by fossil fuels.

It may not be smooth sailing to revive the coal industry. Critics say that natural gas, wind, and solar are cheaper alternatives to coal — and less polluting and harmful to the environment. The US Energy Information Administration estimated in February that 93% of new electric-generating capacity set to be added to the US grid this year is expected to come from solar, wind, and batteries — which could make coal less competitive.

"Coal plants are old and dirty, uncompetitive and unreliable," said Kit Kennedy, managing director for power at the Natural Resources Defense Council, in a statement Tuesday. "The Trump administration is stuck in the past, trying to make utility customers pay more for yesterday's energy. Instead, it should be doing all it can to build the electricity grid of the future."

• PyannoteAI has raised $9 million to make voice transcriptions clearer with AI.
• The startup's AI model can be used for tasks such as audio indexing and dubbing.
• BI got an exclusive look at the pitch deck used to secure the fresh funding.

Voice intelligence startup PyannoteAI has just raised $9 million in a seed round led by Crane Ventures and Serena.

The French startup, launched in 2024, has developed an AI model for a concept called "speaker diarization." It's the process of enhancing audio transcriptions to distinguish between different speakers — something that AI has historically struggled with.

"Our mission is to make human and AI interactions natural through voice," Vincent Molina, cofounder and CEO, told Business Insider. "The voice AI industry today mostly focuses on one-to-one conversations between humans and AI. But real-life conversations aren't like that. They're full of multi-speaker situations, overlapping speech, interruptions, and short and chaotic speech turns," he said.

The startup's platform aims to ensure that the AI model not only picks up on an accurate audio transcription but also understands who is speaking and their intonation. The intended result is for AI to preserve the meaning of audio across different languages and use cases.

The model was developed using research conducted by the startup's cofounder, Hervé Bredin, who has authored over 30 papers on the topic.

PyannoteAI's enterprise use cases include indexing large-scale audio for media platforms, streamlining the dubbing process, and transcribing consultations across enterprise and healthcare. It counts Gladia and MediVox as its clients.

Voice AI has gained significant traction in the past year, with the likes of ElevenLabs and PolyAI raising significant rounds in recent months.

Molina said his startup had increased investor interest from both Europe and the US. "The entire voice AI landscape is accelerating — across all layers: infrastructure, applications, models, and more," he told BI. "Since we're positioned at the very beginning of the value chain, and because we serve the entire ecosystem, we received strong interest from a wide range of investors."

The funding confirms BI's November report that PyannoteAI was in talks to raise about $10 million.

In addition to Crane Ventures and Serena, PyannoteAI's funding round had participation from Julien Chaumond, the chief technology officer of HuggingFace, and former Meta and OpenAI researcher Alexis Conneau.

With the fresh funding, the startup said it plans to grow its research team.

Check out the pitch deck used to raise the capital, shared exclusively with Business Insider.

• Andrew Ng said that "lazy prompting" can be an efficient way to use AI — in some scenarios.
• Lazy prompting entails giving minimal context to large language models.
• It's the latest in a line of AI techniques, like vibe coding, that's transforming software development.

Sometimes, it's OK to be lazy with your AI prompts thanks to models becoming smarter, the Stanford professor and former Google Brain scientist Andrew Ng has said.

Standard prompting advice is to give large language models, or LLMs, a lot of context and specific instructions for a query.

But Ng said in an X post that "lazy prompting" — when users put information into a model with little context or without explicit instructions — could, in some cases, be a better strategy.

"We add details to the prompt only when they are needed," he said Thursday.

He pointed to developers using LLMs to debug code as an example.

"When debugging code, many developers copy-paste error messages — sometimes pages of them — into an LLM without further instructions," Ng wrote. "Most LLMs are smart enough to figure out that you want them to help understand and propose fixes, so you don't need to explicitly tell them."

The idea is that the LLM can then detect a buggy implementation and quickly provide a good solution without much context.

Developers of foundational models are scrambling to make LLMs more inferential, which means the models go beyond producing output and begin to "reason" and gauge the intent of the prompt.

Ng said that lazy prompting is an "advanced" technique that works best when the LLM has enough preexisting context — and an ability to infer intent. It would only work when users can "iterate quickly using an LLM's web or app interface," he added.

It's not useful if the LLM needs a lot of context to provide a detailed response or if the model is unable to detect hidden bugs or errors, he said.

AI is fast transforming how people code and interact with software. Vibe coding — when people give natural language instructions to AI to write code — has recently swept Silicon Valley and beyond.

Last week, Ng launched a "Vibe Coding 101" short course for newbies who want to learn how to use generative AI tools to write and manage code.

• Trump's sweeping tariffs could hit tech startups hard.
• VCs expect rising costs and fundraising struggles for companies relying on offshore operations.
• Investors worry the tariffs could freeze exit markets as companies hope to IPO or get bought.

President Donald Trump's long-awaited tariffs have arrived — and the US markets are reeling.

The president debuted broad "Liberation Day" tariffs on Wednesday, announcing a 10% baseline tariff on all countries, as well as higher tariffs of up to 50% on imported goods from specific countries. The Trump administration's goal, as they put it, is to manage "highly unbalanced" trade deficits and bring manufacturing and supply chains back into the US.

But the tariffs have plenty of US-based investors and founders worried. The S&P 500 index fell more than 4% after Trump's announcement, with some companies' individual stocks tumbling far further.

Investors and founders told Business Insider the tariffs could have significant consequences for the tech industry, including for startups and venture capital — potentially raising the costs of doing business, stunting exit opportunities, and forcing startups to further tighten their belts.

Here are five ways Trump's tariff policy could affect tech workers and startups.

Tariffs could lead to a broader market correction that makes fundraising harder

President Trump's trade war has rattled stock markets and stoked fears of a recession. That has some in tech worried about how a broader market correction could trickle down to startups.

A tighter economy could force some limited partners to call off investments in venture funds as the value of their public assets tumbles — a phenomenon known as the "denominator effect." This would make it harder for investors to close new funds. They might also whittle down the target size of their funds, recognizing that startups will raise smaller rounds in a funding winter.

"If the tariffs stick, and if we see kind of a broader market correction, you're going to see venture going the same way it has over the last two years, where it's difficult to raise capital and to manage funds," said Marcos Fernandez, cofounder and managing partner of Fiat Ventures. "Unless you have a really differentiated model, it's going to be difficult to weather that storm."

The IPO window could close shortly after it opened

Just as startups have begun to queue up for IPO listings, Trump's latest round of tariffs has raised questions about whether companies will shelve their plans to go public.

"If tariffs affect the equity markets and that affects the trading of tech stocks, then that could affect the ability of later-stage companies to go public," said Justin Stevens, founder and chief executive of Overlap Holdings, a frontier-tech-focused venture capital firm.

Fernandez of Fiat Ventures shared that sentiment, saying the backlog of companies seeking to go public is shrinking as they wrestle with value multiples out of line with their own.

Both Klarna and Hinge Health have filed their S-1s to go public.

The stock market's reaction to the tariffs could hit potential mergers and acquisitions in addition to public listings, said Jordan Nof, managing partner of Tusk Venture Partners.

"The big headline number on the Google-Wiz deal, other M&A activity that's just heating up, and the momentum in the IPO market — this really does throw a wet blanket on all of that," he said.

Spiking costs for hardware and manufacturing startups

Investors said startups building hardware or other physical goods could be especially impacted by Trump's new tariffs.

Hardware startups rely on a host of manufacturers that are mostly based in China. Those Chinese manufacturers now face a 54% total effective tariff on goods imported to the US — a combination of a 34% Liberation Day tariff imposed on China and a previous 20% price hike.

Andreas Schwarzenbrunner, a Paris-based general partner at Speedinvest, said industries that depend on complex manufacturing, such as electronics, machinery, and chip production, and hardware-dependent sectors could see spiking costs and supply chain disruptions.

Even healthcare companies could feel the sting of tariffs if they manufacture their drugs and devices offshore, said Sunny Kumar, a partner at Informed Ventures. Those companies "now have to decide whether to pass on these new costs to their customers or absorb them themselves, given the significant time and infrastructure needed to onshore," Kumar said.

However, Overlap Holdings' Stevens maintains that some high-margin businesses in categories such as robotics, life sciences, or space may more easily weather the price hikes. "If you're building a robot, the cost of the metal inputs into the robot is a very small fraction of what you're selling that robot for. And so you have a lot more profit margin to work with," Stevens said.

The growth of US manufacturing may lead to rising demand for global talent

President Trump said on Wednesday, "Jobs and factories will come roaring back" as he rolled out sweeping tariffs. If that happens, the creation of new jobs could exacerbate an already stretched labor market.

The tariffs may also deter some skilled foreign workers from bringing their talents onshore, said Krish Ramadurai, a partner at AIX Ventures, which invests in the software and biotech industries.

"Tariffs signal a closed-door stance that repels global talent," he said. "The US innovation edge depends on openness, and losing that could prove more costly long-term than any tariff."

However, a manufacturing revival could also lead companies to seek out top global talent.

Sophie Alcorn, a business immigration attorney who advises tech workers and startups, predicts that the construction of state-of-the-art factories, such as Apple's new Houston digs and Hyundai's Georgia plant, will actually create job opportunities for high-skilled immigrant workers.

"The technicality of some of these advanced manufacturing processes is going to require a lot of brilliant talent," Alcorn said. "We will certainly need to rely on brilliant people from around the world to help us build them in the US."

Companies may face a pricing conundrum

Companies now face a difficult choice: absorb the increased costs or raise prices for customers.

The companies that are best positioned to pass rising costs onto customers are those making and selling "mission-critical" products, said Tusk Venture Partners' Nof.

"If you're providing a mission-critical product to anyone that's a must-have rather than a nice-to-have — groceries, medications, shelter, things you have to have to survive — you're going to buy that regardless, so you'll be more willing to take an incremental increase in price," Nof said. "Everything else starts to get cut."

The same is true for business software. "You're not going to stop buying software altogether," Nof said, but enterprises will become more discerning about the subscriptions they keep.