This city is one of the world’s most important AI hubs. It’s not where you’d expect这座城市是全球最重要的AI中心之一。它的位置出乎你的意料。
Geoffrey Hinton, a Nobel Prize-winning computer scientist, took a leap in the 1980s — leaving a well-funded post at Carnegie Mellon University to pursue a little-studied area of machine learning research at the University of Toronto.

Geoffrey Hinton, a Nobel Prize-winning computer scientist, took a leap in the 1980s — leaving a well-funded post at Carnegie Mellon University to pursue a little-studied area of machine learning research at the University of Toronto.
Four decades later, Hinton and that research paved the way for how modern AI systems function, all while changing the face of the Canadian city.
Silicon Valley is the beating heart of the tech industry, but a thriving AI ecosystem has developed around Hinton in Toronto.
Tech giants including Google, Nvidia, Meta and Microsoft, along with a slew of startups, have opened offices or research labs within miles of the University of Toronto. Even non-tech firms are now eyeing the city for its AI expertise: pharmaceutical company Sanofi this year announced a nearly $300 million investment in a Toronto AI Center of Excellence.
Toronto now boasts the world’s third best tech talent market, ahead of New York City but behind San Francisco and Seattle, according to an annual analysis by commercial real estate firm CBRE. That’s thanks to educational institutions like the University of Toronto, and public-private research institutes like the Vector Institute, which Hinton co-founded to help commercialize AI breakthroughs.
“I was lucky because very few people were working on this 20 years ago, and so smart young students would come to Toronto … and they’ve now turned into a whole generation of AI leaders,” Hinton told CNN in an interview in Toronto.
Canadian Prime Minister Mark Carney is now aiming to boost the local AI ecosystem with a national AI strategy that includes more than $2 billion in spending on AI research, education and efforts to increase adoption, as well as new safety guardrails.
Toronto’s growing AI power comes as trade relations between the United States and Canada have deteriorated. A trade war has exploded in recent weeks, with Canadian officials seeking to protect the country’s industries. It may be yet another reason why the country’s tech investments are important.
“I don’t think we can afford for Canada to get all this technology from the States or from China,” Hinton said.
Among Toronto’s new generation of AI leaders is Nick Frosst, who studied under Hinton at the University of Toronto and was one of the first hires at Hinton’s Google AI lab in Toronto in 2017. Frosst then went on to co-found the $7 billion enterprise AI firm Cohere.
Frosst decided against the traditional route of moving to Silicon Valley to start his company for a simple reason: He likes living in Toronto.
“There’s a great tech ecosystem, there’s also a great arts community, there’s also a great academic community, there’s a great business community,” said Frosst, who also fronts Toronto-based indie rock band Good Kid. “All of it is here.”
Many other founders in the city feel the same.
Nadtochiy/iStockphoto/Getty Images
Gennady Pekhimenko, who sold his startup CentML to Nvidia last year and now serves as senior director of AI software for the chip giant, said Toronto’s international food scene serves as a useful recruiting tool for out-of-town talent.
Tomi Poutanen, founder and CEO at AI healthcare firm Signal 1, mentioned the benefits of Canada’s healthcare system and Toronto’s famed “Hospital Row.”
Hinton added that, unlike the United States under the Trump administration, Canada welcomes immigration, which tech firms have long relied on as a talent pipeline. One of Pekhimenko’s first hires at CentML was an Amazon engineer who was unable to renew his H1-B visa to stay in the United States. He could, however, get a Canadian work permit.
Friendlier immigration policies make it easier to attract international talent who likely have a knack for thinking about how to serve a global market, said Jordan Jacobs, managing partner at investment firm Radical Ventures.
Toronto also benefits from close proximity to major markets such as New York, Washington, DC and Chicago.
“You can get on a flight that’s 90 minutes long and reach 200 million people,” Jacobs said.
San Francisco is still the undisputed center of the tech industry, with around 20% more tech workers than Toronto, according to CBRE. Tech firms are also making a big push back into the city center after many dispersed during the Covid-19 pandemic.
Still, builders in Toronto’s AI ecosystem say they benefit from being headquartered outside the Silicon Valley bubble.
Nick Iwanyshyn/The Canadian Press/AP
Canada just announced new tariffs on US goods. Here’s how the growing trade war could hurt Americans
The cost of living is generally less expensive in Toronto, thanks in part to lower home prices, socialized healthcare and good public schools, local founders say. That means high-quality talent can come cheaper, too. (Although many Toronto AI firms also have a California presence.)
Toronto also doesn’t struggle with high staff turnover rates in the same way as San Francisco, according to Ray Reddy, a former vice president of Shopify who is building a new startup in stealth.
In Silicon Valley, “even people who are very mission driven are always looking over their shoulder, like, is there something better there?” said Reddy, who previously worked at Google’s Silicon Valley headquarters.
San Francisco is often characterized by its hype and unbridled enthusiasm around AI, but Frosst says there’s an openness to friendly debate in Toronto that feels unique.
Frosst offered his own relationship with friend and mentor Hinton as an example: the two have long disagreed on core questions, like whether building AI constitutes creating a new type of being or simply a tool. But they still regularly play chess together, and Frosst will bring up the debate in hopes of throwing Hinton off. (Hinton’s game is normally undeterred, he said.)
“I think it goes back to … Toronto and Canada not being a monoculture,” Frosst said. “It isn’t just tech workers here, it’s a bit of everything, and I think that leads to innovation, it leads to people thinking about what isn’t popular.”
Already, Canada’s digital sector contributes more than $140 billion to the country’s GDP, according to the government . And the local AI ecosystem has other benefits for the country, according to Hinton.
“If AI causes a lot of unemployment, we’ll want the profits that are made by replacing workers to stay within Canada, so they can be taxed within Canada to support the replaced workers,” Hinton said.
What’s more, governments and companies around the world are increasingly looking to diversify away from buying advanced technology solely from the United States. That’s created an “ongoing advantage” for Canada-based firms, Frosst said.
Cohere received an influx of new customer inquiries after the White House in June put export controls on Anthropic’s advanced AI models.
“The flag of Canada is one that builds trust,” Frosst said. “We’re looking to increase the resiliency and independence of anybody who uses our technology, and broadly I think Canada, as a country, is interested in doing the same for all of our partners.”
And while the most well-known tech investors are in Silicon Valley, being located elsewhere is no longer an obstacle to fundraising.
Raquel Urtasun left her role running Uber’s autonomous driving lab in Toronto to start self-driving semitruck company Waabi five years ago. She founded the company in Toronto, although it’s piloting its trucks mainly in the American southwest. Earlier this year, Waabi raised a more than $1 billion Series C, the largest funding round in Canadian history.
“At the end of the day, what attracts investors is: what is going to be the top company that is leading a massive market,” Urtasun told CNN. “Whether we are in Canada, whether we are in the Silicon Valley, it didn’t really matter.”
20 世纪 80 年代,诺贝尔奖得主、计算机科学家杰弗里·辛顿做出了一个重大决定——他放弃了卡内基梅隆大学待遇优渥的职位,前往多伦多大学从事机器学习这一鲜为人知的研究领域。
四十年后,欣顿和他的研究为现代人工智能系统的运行方式铺平了道路,同时也改变了这座加拿大城市的面貌。
硅谷是科技产业的中心,但在多伦多,欣顿周围已经发展出一个蓬勃发展的AI生态系统。
包括谷歌、英伟达、Meta和微软在内的科技巨头,以及众多初创公司,都在多伦多大学附近数英里范围内设立了办事处或研究实验室。就连非科技公司也开始关注这座城市的人工智能技术:制药公司赛诺菲今年宣布投资近3亿美元,在多伦多建立人工智能卓越中心。
根据商业地产公司世邦魏理仕(CBRE)的年度分析报告,多伦多如今拥有全球第三大科技人才市场,领先于纽约,但落后于旧金山和西雅图。这要归功于多伦多大学等教育机构,以及像Vector Institute这样的公私合作研究机构。Hinton联合创办了Vector Institute,旨在帮助人工智能突破性成果实现商业化。
“我很幸运,因为 20 年前很少有人从事这方面的工作,所以很多聪明的年轻学生会来到多伦多……而他们现在已经成长为整整一代人工智能领域的领导者,”辛顿在多伦多接受 CNN 采访时说道。
加拿大总理马克·卡尼现在正致力于通过一项国家人工智能战略来促进当地的人工智能生态系统发展,该战略包括投入超过 20 亿美元用于人工智能研究、教育和推广应用,以及新的安全保障措施。
多伦多人工智能实力的崛起正值美加贸易关系恶化之际。近几周来,贸易战愈演愈烈,加拿大官员正努力保护本国产业。这或许正是加拿大科技投资如此重要的原因之一。
“我认为加拿大负担不起从美国或中国获取所有这些技术,”欣顿说。
多伦多新一代人工智能领导者之一是尼克·弗罗斯特,他曾在多伦多大学师从辛顿,并于 2017 年成为辛顿在多伦多的谷歌人工智能实验室的首批员工之一。之后,弗罗斯特联合创办了市值 70 亿美元的企业人工智能公司 Cohere。
弗罗斯特没有选择传统的硅谷创业路线,原因很简单:他喜欢住在多伦多。
“这里有很棒的科技生态系统,很棒的艺术社区,很棒的学术社区,很棒的商业社区,”弗罗斯特说道,他同时也是多伦多独立摇滚乐队Good Kid的主唱。“这里应有尽有。”
城里许多其他创业者也有同感。
Nadtochiy/iStockphoto/Getty Images
去年将自己的创业公司 CentML 出售给英伟达,现在担任这家芯片巨头人工智能软件高级总监的 Gennady Pekhimenko 表示,多伦多的国际美食氛围是吸引外地人才的有效途径。
AI 医疗保健公司 Signal 1 的创始人兼首席执行官 Tomi Poutanen 提到了加拿大医疗保健系统的优势以及多伦多著名的“医院街”。
欣顿补充说,与特朗普政府领导下的美国不同,加拿大欢迎移民,而科技公司长期以来一直依赖移民作为人才来源。佩希缅科在CentML的首批员工之一是一位亚马逊工程师,他无法续签H1-B签证留在美国。但他可以获得加拿大工作许可。
投资公司 Radical Ventures 的管理合伙人 Jordan Jacobs 表示,更友好的移民政策更容易吸引国际人才,这些人才可能擅长思考如何服务全球市场。
多伦多还受益于其靠近纽约、华盛顿特区和芝加哥等主要市场的地理位置。
“你可以乘坐90分钟的航班,触及2亿人,”雅各布斯说。
据世邦魏理仕(CBRE)的数据显示,旧金山仍然是无可争议的科技产业中心,其科技从业人员比多伦多多出约20%。在新冠疫情期间许多科技公司分散办公后,目前它们正大力回流市中心。
不过,多伦多人工智能生态系统的建设者们表示,总部设在硅谷泡沫之外对他们有利。
Nick Iwanyshyn/加拿大通讯社/美联社
加拿大刚刚宣布对美国商品加征新关税。这场不断升级的贸易战可能会如何损害美国人的利益?
多伦多的生活成本普遍较低,部分原因是房价较低、全民医疗保健和优质的公立学校,当地创业者表示。这意味着高质量人才的成本也更低。(尽管许多多伦多人工智能公司在加州也有分支机构。)
据 Shopify 前副总裁 Ray Reddy 称,多伦多不像旧金山那样面临员工流动率高的问题。Ray Reddy 目前正在秘密筹建一家新的创业公司。
曾在谷歌硅谷总部工作的雷迪说,在硅谷,“即使是那些目标明确、充满干劲的人,也总是会回头看看,那里是不是有更好的机会。”
旧金山经常以对人工智能的炒作和无拘无束的热情而闻名,但弗罗斯特表示,多伦多对友好辩论的开放态度是独一无二的。
弗罗斯特以自己与朋友兼导师辛顿的关系为例:两人长期以来在一些核心问题上存在分歧,例如,人工智能的构建究竟是创造了一种新型生命体,还是仅仅是一种工具。但他们仍然经常一起下棋,弗罗斯特会提起这些争论,希望能扰乱辛顿的思路。(他说,辛顿的棋艺通常不会受到影响。)
弗罗斯特说:“我认为这可以追溯到……多伦多和加拿大并非单一文化。这里不只有科技工作者,各种人群都有,我认为这会促进创新,会促使人们思考那些不流行的东西。”
据加拿大政府统计,加拿大数字产业目前已为该国GDP贡献超过1400亿美元。欣顿表示,本地人工智能生态系统也为加拿大带来了其他益处。
“如果人工智能导致大量失业,我们希望通过取代工人而获得的利润留在加拿大,这样就可以在加拿大境内征税,以支持被取代的工人,”欣顿说。
此外,世界各地的政府和企业正日益寻求减少对美国先进技术的单一依赖,实现技术多元化。弗罗斯特表示,这为加拿大企业创造了“持续优势”。
6 月白宫对 Anthropic 的先进人工智能模型实施出口管制后,Cohere 收到了大量新客户的咨询。
弗罗斯特说:“加拿大国旗象征着信任。我们希望增强所有使用我们技术的人的韧性和独立性,而且我认为,加拿大作为一个国家,也同样希望为我们所有的合作伙伴做同样的事情。”
虽然最知名的科技投资者都集中在硅谷,但身处其他地方已不再是筹集资金的障碍。
五年前,拉奎尔·乌尔塔松 (Raquel Urtasun) 辞去了她在多伦多运营 Uber 自动驾驶实验室的职务,创办了自动驾驶半挂卡车公司 Waabi。该公司虽然在多伦多成立,但其卡车主要在美国西南部进行试点。今年早些时候,Waabi 完成了超过 10 亿美元的 C 轮融资,这是加拿大历史上规模最大的一轮融资。
“归根结底,吸引投资者的是:哪家公司能够成为引领庞大市场的龙头企业,”乌尔塔松告诉CNN。“无论是在加拿大还是在硅谷,其实都无关紧要。”