Tech leaders react to DeepSeek’s rapid rise
Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn more
If you haven’t heard, there’s a new AI star in town: DeepSeeksubsidiary of Hong Kong-based quant High-Flyer Capital Management, has sent shockwaves through Silicon Valley and around the world with their own release earlier this week of a major new open-source reasoning model, DeepSeek R1this matches The most powerful OpenAI o1 model available — and at a fraction of the cost to consumers and to the company itself (when they train it).
Although the arrival of DeepSeek R1 has already changed the consistently topsy-turvy, fast-moving, intensely competitive market for new AI models – in the previous months, OpenAI was battling with Anthropic and Google for the most powerful proprietary models available, while Meta platforms often came up with “close enough ” open source competitors — the difference this time is that the company behind the hot model is based in China, the geopolitical “frenemy” of the US, and whose tech sector has until now been widely seen as inferior to that of Silicon Valley.
As such, it has caused no shortage of hand-wringing and existentialism from US and Western Bloc techies suddenly questioning OpenAI and the overall grand tech strategy of throwing more money and more compute (GPUs, GPUs, the powerful chips for games typically used to train AI models) to the problem of inventing increasingly powerful models.
Still, some Western tech leaders have had a largely positive public response to DeepSeek’s rapid ascent.
Marc Andreessen, co-inventor of the pioneering Mosaic web browser, co-founder of the Netscape browser company and current general partner at renowned Andreessen Horowitz (a16z) venture capital firm, posted on X today: “Deepseek R1 is one of the most amazing and impressive breakthroughs I’ve ever seen — and as an open source, profound gift to the world (robot emoji, greeting emoji).”
Jan Lekun, the chief artificial intelligence scientist in Meta’s Fundamental Artificial Intelligence Research (FAIR) department, posted on his LinkedIn account:
“For people who see DeepSeek’s presentation and think:
“China surpasses US in AI.”
You are reading this wrong.
The correct reading is:
“Open source models outperform proprietary ones.”
DeepSeek benefits from open research and open source (eg PyTorch and Llama from Meta)
They came up with new ideas and built on other people’s work.
Because their work is published and open source, anyone can benefit from it.
That’s the power of open research and open source.”
And even Mark “Zook” Zuckerberg, founder and CEO of Meta AI, seemed to be trying to counter DeepSeek’s rise with his own post on Facebook promising that a new version of Facebook’s Llama family of open-source AI models will be the “leading modern model” when it’s released sometime this year. like he placed it:
“This will be a defining year for AI. In 2025 I expect Meta AI to be the leading assistant serving more than 1 billion people, Llama 4 will become the leading modern model, and we will build an AI engineer who will begin to contribute increasing amounts of code to our R&D efforts. To ensure this, Meta is building a 2GW+ data center that is so large it will cover a significant portion of Manhattan. We will bring ~1GW of computing online in 25 years. and we’ll end the year with more than 1.3 million GPUs. We plan to invest $60-65 billion in capital expenditures this year, while significantly growing our AI teams, and we have the capital to continue investing in the coming years. This is a huge effort, and in the coming years it will drive our core products and business, unlock historic innovations, and extend America’s technology leadership. Let’s go build!“
He even shared a graphic showing the 2 gigawatt data center mentioned in his post superimposed over Manhattan:
It’s clear that while he supports his commitment to open source AI, Zuck isn’t convinced that DeepSeek’s approach to optimizing for efficiency while using far fewer GPUs than big labs is right for Meta or the future of AI.
But with American companies raising and/or are spending record amounts on new AI infrastructure that many pundits have noted is rapidly depreciating (due to hardware/chip and software advances), the question remains which vision of the future will ultimately win out to become the world’s dominant AI provider. Or maybe there will always be multiple models, each with a smaller market share? Stay tuned because this competition is getting deeper and fiercer than ever.
English