• D.C.
  • BXL
  • Lagos
  • Riyadh
  • Beijing
  • SG
  • D.C.
  • BXL
  • Lagos
Semafor Logo
  • Riyadh
  • Beijing
  • SG


In today’s edition, we look at how ambiguities around tax incentives for compute power could temper ͏‌  ͏‌  ͏‌  ͏‌  ͏‌  ͏‌ 
 
rotating globe
May 29, 2024
semafor

Technology

technology
Sign up for our free email briefings
 
Reed Albergotti
Reed Albergotti

Hi, and welcome back to Semafor Tech.

Cities all over the world have gotten good at attracting tech startups, offering tax incentives and setting up incubators. But there’s a new wrinkle to this trend in the AI era, where the most important new tech companies have massive infrastructure costs.

Suddenly, powerful computers that suck up huge amounts of energy are a serious part of the business equation for firms building the powerful foundation models behind products like Google’s Gemini or OpenAI’s ChatGPT.

That changes the game on incentives. It creates opportunities for countries like the United Arab Emirates, which can throw meaningful funds at resource-hungry tech companies burning through cash.

And it throws a bit of a wrench in the gears for countries like France. Paris has become an AI hub thanks to an abundance of talent and attractive tax incentives. But a startup founder recently told me those financial carrots might not be enough. Spending money on compute doesn’t qualify and there was ambiguity about the specifics of the tax terms around GPUs.

France, though, with its long tradition of socialist leanings, sometimes frowns on these kinds of boosts for well-funded tech companies. You can read more below about how these tensions are playing out there, and how it could threaten its AI boom.

Move Fast/Break Things

➚ MOVE FAST: Roll out. As the first reseller of OpenAI’s product for businesses, PwC is launching ChatGPT Enterprise in a new deal. That will make the accounting and consulting firm Sam Altman’s largest customer.

➘ BREAK THINGS: Walkout. The union representing more than a fifth of Samsung Electronics workers in South Korea is planning its first work stoppage next week. The employees are demanding higher wages at a critical time for the chipmaker as it ramps up production in the AI boom.

Kim Soo-hyeon/Reuters
PostEmail
Artificial Flavor

Meta’s Chief AI Scientist Yann LeCun has always been vocal on X, and is known for feuding with folks online. It was still surprising, however, when he decided to take a shot at Elon Musk on his own social media platform.

Musk boosted a post advertising open roles at xAI, a startup he recently founded that just secured $6 billion in its latest funding round, only to be attacked by LeCun. 

Yann LeCun/X/Screenshot

The exchange set off a flurry of replies from Musk fans and critics, and it even roped in a researcher from Google DeepMind, who stepped in to both defend and promote his employer.

Lucas Beyer/X/Screenshot

AI talent is scarce, and companies often compete to attract and keep the best developers. It’s just a bit amusing and bizarre to see some of that play out openly on social media.

— Katyanna Quach

PostEmail
Reed Albergotti

France is an AI hub, but a wrinkle in tax policy is holding it back

THE SCENE

France’s high price of electricity could temper its surprisingly hot AI boom.

Paris has become a major player in the growing industry, aided by generous tax credits that help startups pay for expensive compute power. But the terms of those incentives might threaten its status.

Besides homegrown powerhouse Mistral, some US-based firms such as Poolside and Raive have moved to France or opened up offices there to take advantage of the country’s math talent, which is helpful in advancing complex AI architectures, and its tax benefits. It’s now home to about 600 AI startups, according to France Digitale, and French President Emmanuel Macron praised the industry last week, saying it would help Europe compete against China and the US.

Nadja Wohlleben/Reuters

The main subsidy that benefits tech companies is called the Crédit d’Impôt Recherche, or the research tax credit, which offers incentives ranging from 30% to 60% of R&D expenses, including employee salaries. The subsidies also effectively cover the first-year salary of PhD students after they graduate, if they work in R&D.

By most measures, the R&D tax credit has been a huge success. For about €6.5 billion in annual cost as of 2018, the French government estimates it will, in the long run, add 0.8% to the country’s GDP and create 60,000 jobs.

But a major gray area covers one of the biggest, if not the biggest, expense of many AI companies, which is the training of AI models with costly and powerful GPUs. If a company buys GPUs for R&D purposes, it gets a 60% tax credit. If they rent the GPUs from a cloud provider, the subsidy doesn’t apply, tax experts say.

But buying the GPUs and receiving the tax credit isn’t so simple. That’s because vague language in the measure makes it unclear whether those GPUs have to stay in France, where electricity costs are prohibitively expensive to train AI models, a task that needs a lot of energy.

Ultimately, the parliament would have to decide on whether and how to clarify the rules, but nothing has been finalized, according to French officials.

That leaves startups weighing whether the tax credits, which one venture capitalist noted were so valuable that they amounted to another funding round, are worth incurring high electricity expenses.

Some startups and advisors to French companies say that the law as it is written should cover the costs of training AI models, no matter who owns the GPUs or where they are located.

The AI startup Mistral, which has become a kind of symbol of the AI renaissance happening there, has chosen to train its AI models with cloud providers outside the country. That includes Microsoft, which inked a partnership with Mistral earlier this year to provide supercomputing infrastructure.

Two people with knowledge of the matter said Mistral’s decision to use outside compute resources, rather than build its own infrastructure inside France, was at least partly due to tax policy. Mistral did not respond to a request for comment.

Microsoft earlier this week announced that it was investing €4 billion in cloud and AI infrastructure in France, bringing up to 25,000 of the most advanced GPUs to the country by the end of 2025. A Microsoft spokesman said it won’t receive tax subsidies for the investment.

Reed’s view on what else France could do to attract AI startups. →

PostEmail
Mixed Signals

Introducing Mixed Signals, a new podcast from Semafor Media presented by Think with Google. Co-hosted by Semafor’s own Ben Smith, and renowned podcaster and journalist Nayeema Raza, every Friday, Mixed Signals pulls back the curtain on the week’s key stories around media, revealing how money, access, culture, and politics shape everything you read, watch, and hear.

Whether you’re a media insider or simply curious about what drives today’s headlines, Mixed Signals is the perfect addition to your media diet. Listen wherever you get your podcasts.

PostEmail
Semafor Stat

The number of foreign smartphones that were shipped to China last month, according to Bloomberg. Most of those devices are iPhones, which have been declining as consumers take to mobiles built by Huawei instead. Apple, however, looks like it may be staging a comeback, partly fueled by price cuts.

PostEmail
Friends of Semafor

Meet Patent Drop, a bi-weekly newsletter that uncovers where tech titans (like Apple, Meta, Tesla) are investing their time and money by scouring the U.S. Patent & Trademark website. Stay ahead of the curve with proprietary news and analysis of innovation at its earliest stage. Subscribe for free today.

PostEmail
What We’re Tracking
Justin Sullivan/Getty Images

OpenAI has been in the news lately more for its drama than its products. Yesterday, the firm announced it had formed a Safety and Security Committee to advise the board “on critical safety and security decisions for [its] projects and operations,” after critics questioned CEO Sam Altman’s leadership following some missteps.

When OpenAI’s strict non-disclosure agreement that prevented past employees from speaking out about their previous employer came to light, Altman said he wasn’t aware that the company threatened to take away their vested equity if they didn’t sign it. But in documents shared with Vox, he had signed off on them. OpenAI later changed its rules, promising not to take away employees’ vested equity.

Then there was the exodus of Ilya Sutskever, its ex-chief scientist, and Jan Leike, the co-lead of its Superalignment team (Leike has joined rival Anthropic), which came just before the Scarlett Johansson blowup. Those concerns about Altman’s stewardship of OpenAI haven’t gone away with the formation of the new committee, which is raising fresh suspicions because it’s being led by Altman and three other board members.

The OpenAI boss has been accused of being manipulative and untruthful before and the latest controversies have only raised doubts on whether he can be trusted to develop AI safely.

–— Katyanna Quach

PostEmail
Hot on Semafor
PostEmail