President Donald Trump on Saturday proposed a new path to end the Russia-Ukraine war: a complete NATO ban on Russian oil purchases and tariffs of 50% to 100% on China for its imports of Moscow’s petroleum.

Trump said NATO’s current commitment to victory has been “far less than 100%,” and described alliance members’ continued oil trade with Russia as “shocking.”

“[Buying Russian oil] greatly weakens your negotiating position, and bargaining power, over Russia,” Trump posted on his social media platform, calling for energy sanctions paired with punitive tariffs that he said would “break [China’s] grip” over Moscow.

The suggestion arrives after Russian drones crossed into Polish airspace last week — a move U.S. Secretary of State Marco Rubio called “dangerous” and Britain answered with new sanctions on Russian shipping and suppliers. However, Trump’s hardline proposal is raising questions about whether another layer of restrictions would yield a breakthrough, or whether Moscow’s proven ability to evade sanctions would make the effort another symbolic gesture.

A Long Trail of Sanctions

Since Russia’s full-scale invasion of Ukraine in February 2022, the U.S. and its allies have imposed one of the most extensive sanctions regimes in modern history. Washington and Brussels targeted Russia’s central bank reserves, cut major lenders off from SWIFT, restricted technology exports, and capped the price of Russian crude at $60 per barrel.

The goal was to starve the Kremlin of revenue. Yet Russia’s economy has adapted. Moscow rerouted oil flows to Asia, with China and India emerging as lifelines. According to the International Energy Agency, the two countries now buy more than 70% of Russia’s seaborne crude. Turkey and other middlemen have also played roles, creating back channels that blunt Western measures.

As a result, Russia’s oil revenues remain resilient. The ruble has wavered but avoided collapse, and the Kremlin has continued to fund its war machine despite Western predictions that sanctions would bring it to its knees.

India, China, and the Limits of Tariffs

Trump highlighted China’s importance as Russia’s patron, suggesting tariffs of up to 100% could sever Moscow’s financial safety net. But Washington has already tried punishing India for its energy ties to Russia. The Trump administration earlier this year slapped a 50% tariff on Indian goods, citing New Delhi’s large-scale imports of Russian crude.

The move changed little. India continued buying, often at discounted rates, refining the crude and even re-exporting some products back to Europe. For New Delhi, Russian oil has proven too crucial to abandon — an economic reality that raises doubts about whether higher tariffs on China would truly alter Moscow’s fortunes.

Russia’s Defiance of Past Pressure

This pattern echoes earlier sanction campaigns. When the U.S. and its allies imposed restrictions on Russia after its 2014 annexation of Crimea, Moscow absorbed the shock and pivoted eastward. It deepened ties with Beijing, expanded trade in non-dollar currencies, and fortified domestic industries.

Even the unprecedented 2022-2023 sanctions package — broader and deeper than anything Russia had faced before — failed to cripple the Kremlin. Instead, Moscow adapted by exploiting gaps in enforcement, turning to countries outside the Western alliance, and leaning on its vast energy resources.

A Risk of Diminishing Returns

Trump’s proposal of banning Russian oil within NATO and doubling down on tariffs against China reflects a belief that only maximum pressure can change Putin’s calculus. But the record of the past two years suggests that every new measure carries diminishing returns.

Russia has shown an ability to reroute trade, build shadow fleets of tankers, and create alternative payment systems. China and India, driven by their own strategic and economic interests, have resisted U.S. pressure and kept buying. That resilience underscores a sobering reality: even if NATO bans Russian oil entirely, the barrels may still flow elsewhere, softening the intended blow.

Meanwhile, higher tariffs on China risk igniting another round of tit-for-tat trade wars that could hit U.S. and European economies harder than Russia’s. Earlier this year, Trump raised Chinese tariffs to 145%, prompting Beijing to retaliate with 125% duties on U.S. goods, effectively freezing trade until both sides negotiated a reduction. Another escalation could send shockwaves through global markets.

A Strategy in Question

At the U.N. Security Council last week, acting U.S. Ambassador Dorothy Shea pledged America “will defend every inch of NATO territory,” insisting Moscow must not mistake restraint for weakness. Britain tightened sanctions on vessels and suppliers. G7 ministers urged a “unified front” to cut off Putin’s war revenues.

However, Trump’s approach — punishing China, and indirectly India — faces the same obstacle that has haunted Western sanctions for years: Russia’s survival hinges on partners seemingly outside NATO’s reach. Unless Beijing or New Delhi reverse course, analysts warn, more tariffs and bans may tighten the noose only slightly, without achieving the knockout blow Trump envisions.

As the war drags on, geopolitical analysts believe that NATO faces a stark dilemma. Push harder on sanctions and risk collateral economic fallout — or accept that Russia’s oil leverage, and its allies in Asia, have blunted Western tools of economic warfare.

A decade ago, cloud infrastructure was the invisible scaffolding that allowed the internet’s biggest companies to scale. Today, training data is playing a similar role for artificial intelligence. The providers who can reliably supply it — much like AWS, Azure, and Google Cloud did for the internet — are quickly becoming the indispensable backbone of the industry.

Into this arena steps Micro1, a three-year-old startup that has just raised a $35 million Series A round at a $500 million valuation, betting it can be to AI data what cloud was to software.

The round was led by 01 Advisors, the venture capital firm founded by former Twitter executives Dick Costolo and Adam Bain. As part of the deal, Bain is also joining Micro1’s board alongside Joshua Browder, founder of AI legal assistant DoNotPay.

A Market Shaken by Scale AI’s Fallout

The data labeling market has been roiled since Meta invested $14 billion in Scale AI and hired its CEO. Concerned about their research pipelines being too closely tied to Meta, companies like OpenAI and Google began cutting ties with Scale AI. The company insists it does not share client data with Meta, but the perception of risk left AI labs searching for alternatives.

That opened the door for competitors such as Mercor, Surge, and now Micro1. And unlike Scale AI, which pioneered paying relatively low wages for global contractor labor, Micro1 says it’s focused on the next frontier: high-quality domain expertise.

Growth Against Bigger Rivals

Despite being only three years old, Micro1 is scaling quickly. Its 24-year-old CEO, Ali Ansari, told TechCrunch that the startup’s annual recurring revenue has surged from $7 million at the start of 2025 to $50 million now. Its clients include Microsoft and several Fortune 100 companies.

Still, the numbers pale compared to its larger rivals. Mercor generates $450 million in ARR, while Surge earned $1.2 billion in 2024. Yet the trajectory suggests Micro1 is carving out a spot in a market that may ultimately resemble the cloud sector — a handful of giants supporting countless smaller but essential providers.

Ansari argues that AI labs no longer just want bulk labeling at the cheapest possible cost. To make models smarter, they need training data from senior engineers, doctors, professional writers, and even university professors. To meet that demand, Micro1 built Zara, an AI-powered recruiter that interviews and vets candidates. The system has already onboarded thousands of experts, including faculty from Stanford and Harvard, and Micro1 says it plans to add hundreds more every week.

“Really the only way models are now learning is through net new human data. Micro1 is at the core of providing that data to all frontier labs, while moving at speeds I’ve never seen before,” Bain said.

Micro1 isn’t stopping at labeling. The startup is moving into AI training environments — simulated workspaces where AI agents can practice real-world tasks. Analysts say this could become as important as cloud servers were for the early internet, because it allows companies to scale model training without being constrained by static datasets.

Just as no single cloud provider controls the entire internet, no single data provider can meet all of an AI lab’s needs. OpenAI, Anthropic, Meta, and Google all work with multiple partners, and likely will continue to do so. That fragmentation, combined with the sheer demand for data, means there’s room for startups like Micro1 to grow even as bigger competitors dominate.

A Cloud-Like Trajectory

The parallel to cloud computing is striking. A decade ago, AWS, Azure, and Google Cloud turned server capacity into an essential service, giving software startups the foundation to grow without building infrastructure from scratch. Now, companies like Micro1, Surge, and Mercor are doing the same with training data, offering AI labs a way to scale without building vast labeling operations in-house.

The difference is that while cloud became highly consolidated, the training data space remains messy and fragmented. Analysts suggest it could eventually consolidate in a similar fashion — a few giants handling the bulk of the market, while smaller firms survive by specializing in niche domains.

For now, though, the prize is still wide open. With a fresh $35 million in its pocket and a board stacked with high-profile backers, Micro1 is betting it can turn its early momentum into long-term staying power — just as AWS once did in a very different but equally foundational market.