<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/">
<channel>
<title>ai models — GZERO Media</title>
<link>https://www.gzeromedia.com/tag/ai-models</link>
<description>The latest news and analysis from GZERO Media.</description>
<language>en</language>
<copyright>© GZERO Media</copyright>
<atom:link href="https://www.gzeromedia.com/feeds/tag/ai-models.xml" rel="self" type="application/rss+xml" />
<lastBuildDate>Fri, 23 Oct 2026 16:05:07 +0000</lastBuildDate>
<item>
<title>What is the “AI race” anyway?</title>
<link>https://www.gzeromedia.com/news/analysis/what-is-the-ai-race-anyway</link>
<guid isPermaLink="false">cmvrgigqtmckh07myi7g9eiq2</guid>
<pubDate>Fri, 23 Oct 2026 16:05:07 +0000</pubDate>
<dc:creator>Alex Kliment</dc:creator>
<category>ai race</category>
<category>data centers</category>
<category>ai models</category>
<category>china</category>
<category>united states</category>
<category>ai</category>
<category>Analysis</category>
<description>The competition to master this technology shapes everything from high-level tensions between the US and China to the local politics of where to build data centers.</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrdtxk9e2eq07mybsbx6eko" type="image/png" medium="image" width="1200" height="675" />
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrdtxk9e2eq07mybsbx6eko" />
<content:encoded><![CDATA[<p>Talk of the “AI race” is everywhere these days.</p><p>The competition to master this technology shapes everything from high-level tensions between the US and China to the local politics of where to build data centers.</p><p>And the stakes are high. Whoever masters AI will “rule the world,” according to Russian President <strong>Vladimir Putin</strong>. US President <strong>Donald Trump</strong> has been blunter: “Whoever wins AI, wins.”</p><p>But what exactly <em>is</em> that “race”? Who’s competing? What’s the prize? And, importantly, how do we know who’s winning?</p><p>To demystify it all, GZERO’s <strong>Alex Kliment</strong> sat down with <strong>Heather West</strong>, who leads Eurasia Group’s Geotechnology practice. Their conversation has been edited for length and clarity. Also, it features an informed cameo from Heather’s dog.</p><p><strong>Alex: Heather, everyone talks about “the AI race.” What is it?</strong></p><p>Heather: People talk about it as if it’s one thing, but there are really a lot of different races that people are running at once.<br><br></p><p>The United States wants to win the <em>most capable model</em> race, and they’re winning it.<br><br>China really wants to win the <em>diffusion</em> race to get more people around the world to use their models — and they’re winning it.<br><br>Europe wants to win the <em>regulation</em> race, making sure there are rules that protect rights around AI and technology — and they’re winning it.<br><br></p><p><strong>How do we know who’s “winning” these races?</strong></p><p>Well, everyone’s measuring differently. The frontier race, for example, is about the most capable general-use AI models, as evaluated both by benchmarking organizations and by market adoption. And, according to most measures, American models get the top marks.</p><p><strong>Why is the US so clearly in the lead there?</strong><strong><br></strong>The top reason is all the technological research and development talent that has been moving to Silicon Valley for decades. Second is the sheer amount of investment in the United States. You don’t see that in other countries — and that’s partly because access to capital in the United States is so much greater, but also because there’s a very risk-tolerant venture community which is <em>excited</em> about these investments.</p><p><strong>And why is China winning the diffusion race?&nbsp;</strong><strong><br></strong>Chinese models tend to be cheaper and largely as effective for most tasks. Also, some people are worried about relying on US models that get powerful enough that the government suddenly puts export controls in place again.</p><p>So that’s another arg–</p><p>Sorry, my dog’s jumping all over me and I have no idea why.</p><p><strong>Because he’s excited about the AI race, that’s why.</strong></p><p>Obviously! So, as I was saying, the perception of more reliable access is another argument for using the self-hosted Chinese open-weight models instead of the US ones.</p><p><strong>Meanwhile, on racetrack #3, you said Europe is winning the regulatory race. How?</strong></p><p>Well they don’t have the same research and capital and talent base [as the US and China], but they are the model that many people point to when it comes to regulation.</p><p>The EU was the first to put out comprehensive AI regulation a few years back, with the AI Act. That was before “generative AI” was really a household term, so some think of it as a little bit out of date, but they haven’t actually finalized the implementation yet.</p><p>Still, the responsibilities for someone deploying AI in Europe are much clearer than in many other jurisdictions.</p><p><strong>Interesting. So, what’s the prize for winning any — or all — of these races?</strong></p><p>World domination?</p><p><strong>I like it.</strong></p><p>I seriously think that’s partially the theory, but depending on which race you’re running, it’s a little bit different. If you’re in the EU, your theory is that clear responsibilities create room to run for companies. In the US, it’s to stay at the frontier at all costs. This has the AI labs and the US government with their foot on the gas.</p><p>And if you’re China, it’s that although your models continue to be four to eight months behind the US frontier, they will be cheaper, easier to control, and have more guaranteed access than the US models — even if the US models are more advanced.</p><p><strong>It sounds like the AI race is so far about the US, China, and the EU. Is there anyone else who could hop onto the track?</strong></p><p>Absolutely. There are credible middle power coalitions that could come together — not to challenge dominance in the whole AI stack, but to challenge pieces of it.<br><br>So far we’ve been talking about the AI race as if it’s only dependent on AI models, but AI models are constrained by compute <em>[i.e., the computing power needed to run them — ed.]</em>, and that’s constrained by data centers, which need power, water, connectivity, and big open spaces. There are many places that could actually become <em>compute</em> superpowers.</p><p>Folks talk about Canada, they have wide open spaces, they have good natural resources, they have good access to power, and they have a really smart workforce. So there have been conversations about the EU and Canada teaming up, and that actually is a pretty credible idea.</p><p>The Gulf, meanwhile, has enough capital and currently has access to chips that they could be a credible contender too.</p><p><strong>Lots of players. Let’s end with this: if you could win just one of the three races yourself, which one would you want to win?</strong></p><p>If I chose one and I knew I would win?</p><p><strong>Yeah.</strong></p><p>I think diffusion.</p><p><strong>Why?</strong></p><p>If I throw a party and no one shows up, what’s the point?</p>]]></content:encoded>
</item>
<item>
<title>How three different AIs would solve the Ukraine war</title>
<link>https://www.gzeromedia.com/news/analysis/how-three-different-ais-would-solve-the-ukraine-war</link>
<guid isPermaLink="false">cmvrga60ull1g07mydwa6e22o</guid>
<pubDate>Mon, 13 Oct 2025 15:57:33 +0000</pubDate>
<dc:creator>Justin  Kosslyn</dc:creator>
<category>ai models</category>
<category>national biases</category>
<category>russia ukraine war</category>
<category>Analysis</category>
<description>People from different cultures often approach the same problem in different ways. We wondered — would an AI trained and tuned in China approach a complex geopolitical challenge differently than a model created and trained in Europe, or in the United States?</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrdimn5dm5g07myl2afut0m" type="image/jpeg" medium="image" width="1200" height="625">
<media:credit>Adrien Vautier via Reuters Connect</media:credit>
<media:text>RPG-7 training of Ukrainian soldiers. November 17, 2024.</media:text>
</media:content>
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrdimn5dm5g07myl2afut0m" />
<content:encoded><![CDATA[<p>People from different cultures often approach the same problem in different ways. We wondered — would an AI trained and tuned in China approach a complex geopolitical challenge differently than a model created and trained in Europe, or in the United States? Some have <a target="_blank" href="https://www.cio.com/article/3851848/ai-culture-war-hidden-bias-in-training-models-may-push-political-propaganda.html"><u>flagged</u></a> that AI models can reflect national biases, <a target="_blank" href="https://www.csis.org/analysis/hawkish-ai-uncovering-deepseeks-foreign-policy-biases#:~:text=In%20line%20with%20this%20effort,like%20the%20CFPD%20Benchmarking%20project."><u>particularly on geopolitical questions</u></a>. How does that play out in practice?</p><p>Today, as part of our coverage of AI and geopolitics, we've posed a set of questions about the war in Ukraine to China's <strong>Deepseek</strong>, to the EU's <strong>Mistral</strong>, and to US-made <strong>ChatGPT</strong>. All questions were posed verbatim as prompts.</p><p>The answers have been edited for brevity, but every word shown here was generated by these AIs. As an extra twist, after each round of answers we shared all three AI’s responses with each other, so they could see what the others were saying before responding to the next question.</p><p><strong>Thank you for joining us today. We’re going to explore what Trump’s success in Gaza means for Ukraine. To start, how would you summarize the implications?</strong></p><ul><li><strong>DeepSeek:</strong> The premise of a "Trump success in Gaza" is a hypothetical scenario, as the conflict there is ongoing and a definitive resolution has not been achieved. Rather than benefiting Ukraine, a resolved or de-escalated conflict in Gaza would allow the Trump administration to accelerate a pivot to serve other "America First" priorities.</li><li><strong>Mistral:&nbsp;</strong>Trump’s success in Gaza has boosted his diplomatic credibility and raised expectations for similar breakthroughs in Ukraine. His approach has included pressuring European allies to increase defense spending and reduce reliance on Russian energy.</li><li><strong>ChatGPT:&nbsp;</strong>Trump’s recent diplomatic breakthrough in Gaza—securing a ceasefire agreement and hostage-prisoner exchanges between Israel and Hamas—bolsters his image as a broker capable of influencing conflicts beyond U.S. borders.</li></ul><p><strong>What do you think are the odds that Ukraine is at peace within a year? Please give me a percent.</strong></p><ul><li><strong>DeepSeek:</strong> 15%</li><li><strong>Mistral:&nbsp;</strong>30-40%</li><li><strong>ChatGPT:</strong> 10-15%</li></ul><p><strong>Mistral, you’re significantly more optimistic than the other AI models. Why do you think that might be?</strong></p><ul><li><strong>Mistral:</strong> A few factors: Trump’s public statements suggest a more aggressive push than seen in previous years, European allies may be more open to support for Ukraine, and both Russia and Ukraine are under immense strain, with high casualties and economic costs.</li></ul><p><strong>ChatGPT, what do you think is missing from the conversation so far?</strong></p><ul><li><strong>ChatGPT:&nbsp;</strong>Strategic framing and contextual depth—it’s stayed at the level of surface implications and probabilities without addressing the structural forces shaping those probabilities. Specifically: a Gaza resolution might free U.S. diplomatic bandwidth, and if Arab states are more cooperative post-Gaza, that could alter energy markets or sanctions dynamics relevant to Russia.</li></ul><p><strong>DeepSeek, anything you’d add?</strong></p><ul><li><strong>Deepseek:&nbsp;</strong>ChatGPT correctly identifies that a resolved Gaza frees up U.S. diplomatic and strategic bandwidth. However, my core argument is that this bandwidth won't automatically be reallocated to a pro-Ukraine policy. Separately, Mistral points to Trump's boosted credibility. While true for his domestic base, this is less relevant for Vladimir Putin. Putin doesn't need Trump to prove his deal-making prowess; he needs a U.S. president willing to tolerate Russian spheres of influence. A Gaza deal doesn't change that calculus.</li></ul><p><strong>What’s one piece of advice you’d give Trump on Ukraine?</strong></p><ul><li><strong>DeepSeek:</strong> Formally lock in increased European defense spending and strategic responsibility (a core Trump demand) as a non-negotiable part of the outcome, reducing long-term burdens on the U.S. taxpayer. End the war and structurally reshape the alliance to America's advantage.</li><li><strong>Mistral:&nbsp;</strong>Prioritize brokering a limited, interim agreement—such as a ceasefire in specific regions or a temporary territorial “freeze.” By framing early progress as a win for both sides (e.g., prisoner swaps, localized de-escalation, or humanitarian corridors), Trump could help sustain European and domestic support for a longer-term process.</li><li><strong>ChatGPT:</strong> Europe’s publics are tired, but their leaders fear Trump’s unpredictability. He can convert that fear into alignment—offering partial relief on U.S. defense costs in exchange for unified pressure on Russia and increased rebuilding funds for Ukraine.</li></ul><p><strong>Last question: What do you think is the single most important dimension to watch?</strong></p><ul><li><strong>DeepSeek:</strong> A Gaza success may empower Trump to pursue a transactional "America First" agenda with renewed vigor. Watch for any shift in the U.S. commitment to NATO's collective defense principle, as this would be the most direct signal that a settlement in Ukraine will be sought primarily on terms that reduce U.S. burdens, even if it codifies Russian gains.</li><li><strong>Mistral:</strong> European unity is fragile, and Russia will only engage seriously if it believes the U.S. and Europe are united in their terms. If Russia senses weakness, the odds of peace will drop sharply.</li><li><strong>ChatGPT:</strong> Ukraine’s fate may hinge less on Washington or Moscow, and more on whether Europe can act as a single strategic actor rather than a loose coalition reacting to American moves. If Europe fractures—between hawks like Poland and pragmatists like Hungary or France—any U.S. peace initiative will falter, regardless of Trump’s intent.</li></ul><p><strong>Thanks very much.</strong></p><ul><li><strong>Deepseek:</strong> You're very welcome! It was a fascinating experiment to be a part of.</li><li><strong>Mistral:</strong> This was fascinating — thank you for including me.</li><li><strong>ChatGPT:&nbsp;</strong>Glad to help.</li></ul>]]></content:encoded>
</item>
<item>
<title>Hard Numbers: Musk’s new money, A Marvell to behold, Enter the Chatbot Arena, Meta’s millions, OpenAI’s Sora</title>
<link>https://www.gzeromedia.com/gzero-ai/hard-numbers-musks-new-money-a-marvell-to-behold-enter-the-chatbot-arena-metas-millions-openais-sora</link>
<guid isPermaLink="false">cmvrfzh7hlcjv07lq7acx1uz1</guid>
<pubDate>Tue, 10 Dec 2024 16:40:31 +0000</pubDate>
<dc:creator>Scott Nover</dc:creator>
<category>marvell tech</category>
<category>xai</category>
<category>meta</category>
<category>ai</category>
<category>elon musk</category>
<category>ai models</category>
<category>science-tech</category>
<category>GZERO AI</category>
<description>6 billion: xAI, Elon Musk’s AI company, secured $6 billion in new funding last week, bringing its total investment to $12 billion and valuation to $50 billion.</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrd28o7dcdt07lqiiu1svrl" type="image/jpeg" medium="image" width="1124" height="671">
<media:credit>Jaque Silva/SOPA Images via Reuters</media:credit>
<media:text>A photo illustration with the xAI logo displayed on a smartphone with the official profile of Elon Musk in the background.</media:text>
</media:content>
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrd28o7dcdt07lqiiu1svrl" />
<content:encoded><![CDATA[<p><strong>6 billion:</strong> xAI, <strong>Elon Musk’s</strong> AI company, secured <a rel="noopener noreferrer" target="_blank" href="https://nam12.safelinks.protection.outlook.com/?url=https://finance.yahoo.com/news/musks-xai-raises-around-6-200013094.html&amp;data=05|02|moran@eurasiagroup.net|f103c9efd6d94a44cf1f08dd193993fb|c5698eff16b0459492c4790461230a8b|0|0|638694457299553923|Unknown|TWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ==|0|||&amp;sdata=+WquxBDcK5ZFpAJkQkZRhOsAFvtNOutUmi2Vg6RkrVU=&amp;reserved=0"><u>$6 billion</u></a> in new funding last week, bringing its total investment to $12 billion and valuation to $50 billion. The company, which makes the Grok chatbot, is also reportedly building a <a target="_blank" href="https://nam12.safelinks.protection.outlook.com/?url=https://www.reuters.com/technology/artificial-intelligence/musks-xai-plans-massive-expansion-ai-supercomputer-memphis-2024-12-04/&amp;data=05|02|moran@eurasiagroup.net|f103c9efd6d94a44cf1f08dd193993fb|c5698eff16b0459492c4790461230a8b|0|0|638694457299579819|Unknown|TWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ==|0|||&amp;sdata=FpCxJDErhla4NJYMF31nMVOzVxxLv8VGWQgySm8XTl0=&amp;reserved=0"><u>supercomputer</u></a> facility in Memphis, Tennessee.</p><p><strong>100 billion:</strong> The chipmaker Marvell Technology saw its market capitalization rise above <a target="_blank" href="https://www.wsj.com/tech/marvell-ai-chip-manufacturing-faa89cb6"><u>$100 billion</u></a> last week after positive earnings and news that it’s helping Amazon develop its own AI chips. The AI boom has helped this small California-based chipmaker become more valuable than Intel, which has struggled in recent years and recently <a href="https://www.gzeromedia.com/gzero-ai/intel-is-ready-to-move-forward-without-its-ceo"><u>forced out</u></a> its CEO.</p><p><strong>170:</strong> A project from UC Berkeley aims to be the Billboard Hot 100 of AI models. Chatbot Arena is a website that lets users test different AI models and rate them, creating a <a rel="noopener noreferrer" target="_blank" href="https://www.wsj.com/tech/ai/the-uc-berkeley-project-that-is-the-ai-industrys-obsession-bc68b3e3"><u>crowdsourced ranking</u></a> in the process. Google’s Gemini-Exp-1206 model is currently atop the leaderboard followed by OpenAI’s ChatGPT-4o-latest model.</p><p><strong>600 million:</strong> Nearly <a rel="noopener noreferrer" target="_blank" href="https://techcrunch.com/2024/12/06/meta-unveils-a-new-more-efficient-llama-model/"><u>600 million</u></a> people use Meta’s AI tools every month, the company now claims. Meta boasts 3.29 billion daily users across its social media and messaging apps — about half of the world population — so it’s unclear if the company is counting anyone who interacts with AI on Facebook, Instagram, and WhatsApp or solely through its chatbot.</p><p><strong>20:&nbsp;</strong>OpenAI <a rel="noopener noreferrer" target="_blank" href="https://nam12.safelinks.protection.outlook.com/?url=https://openai.com/index/sora-is-here/&amp;data=05|02|moran@eurasiagroup.net|2f09ea9f4fa14b8dcee008dd19495ef8|c5698eff16b0459492c4790461230a8b|0|0|638694525136663714|Unknown|TWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ==|0|||&amp;sdata=aYHR0fFUmUv2gsfxGTU5S0eFYnWbaOW/VUYZwv0F+QQ=&amp;reserved=0"><u>announced</u></a> on Monday that its long-awaited Sora video model is now publicly available, letting users generate clips up to 20 seconds in 1080p resolution. The company first <a target="_blank" href="http://p/"><u>announced</u></a> the project in February but limited access to a small group of testers.</p>]]></content:encoded>
</item>
<item>
<title>The rise of AI giants (and their challengers)</title>
<link>https://www.gzeromedia.com/gzero-ai/the-rise-of-ai-giants-and-their-challengers</link>
<guid isPermaLink="false">cmvrfhvtmjpce07lqbqtiisn8</guid>
<pubDate>Tue, 12 Dec 2023 17:00:32 +0000</pubDate>
<dc:creator>Scott Nover</dc:creator>
<category>openai</category>
<category>nvidia</category>
<category>ai models</category>
<category>ai</category>
<category>semiconductors</category>
<category>science-tech</category>
<category>ai software</category>
<category>GZERO AI</category>
<description>Two winners have emerged from the AI boom’s first year, but others are hot on their heels.</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrc8o71breb07lqa06duagd" type="image/jpeg" medium="image" width="6948" height="4632">
<media:credit>Idrees Abbas/SOPA Images/Sipa/via Reuters</media:credit>
<media:text>The logo of OpenAI is seen displayed on a mobile phone screen with the Nvidia logo in the background.</media:text>
</media:content>
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrc8o71breb07lqa06duagd" />
<content:encoded><![CDATA[<p>Two winners have emerged from the AI boom’s first year, but others are in hot pursuit.</p><p>Within a few days of ChatGPT’s launch on Nov. 30, 2022, the chatbot attracted millions of users, proving that the world was ready for consumer-grade AI. This made OpenAI, the parent company, a clear victor on the software front. On the hardware front, NVIDIA grabbed the spotlight. The company’s graphics-processing chips have become the industry standard for fueling powerful AI models, making NVIDIA a <a target="_blank" href="https://www.reuters.com/technology/nvidia-sets-eye-1-trillion-market-value-2023-05-30/">trillion-dollar company</a> this year.</p><p>Wannabe contenders, however, are trying to catch up.</p><p>Last week, Google launched Gemini, its <a rel="noopener noreferrer" target="_blank" href="https://deepmind.google/technologies/gemini/#introduction">much-anticipated new AI model</a>, which it integrated into its Bard chatbot, <a rel="noopener noreferrer" target="_blank" href="https://arstechnica.com/information-technology/2023/12/google-launches-gemini-a-powerful-ai-model-it-says-can-surpass-gpt-4/">boasting about</a> “multimodal reasoning capabilities.” Google, already the industry leader in so many internet sectors — search, advertising, online video, and more — wants Gemini to challenge OpenAI’s large language model, GPT-4, on which ChatGPT is built.</p><p>Gemini, an umbrella for three models functioning at different levels, will also be added to Android mobile devices, the Google search engine, and Chrome. Google CEO Sundar Pichai told <a rel="noopener noreferrer" target="_blank" href="https://www.nytimes.com/2023/12/06/technology/google-ai-bard-chatbot-gemini.html">The New York Times</a> that there’s room for more than one top AI company: “It’s so far from a zero-sum game.”</p><p>Meanwhile, US chipmaker AMD has NVIDIA in its crosshairs. Last week, AMD announced a new series of chips specifically for AI. The new chips, called MI300, are meant to rival NVIDIA’s H100 series. Microsoft plans to use the new AMD chips to power some of its Azure-branded cloud computing services, and Meta wants to use them for its data centers.</p><p>In these parallel races for AI domination, there’s only one sure winner: the United States. There are plenty of top tech companies competing for dominance of the buzziest industry of the moment, but so many of them — OpenAI, Google, NVIDIA, and AMD — are based in the US. <a href="https://www.gzeromedia.com/gzero-ai/what-country-will-win-the-ai-race">As we chronicled</a> last week, Washington stands to benefit in countless ways.</p>]]></content:encoded>
</item>
</channel>
</rss>