Tech Funding News has been tracking superintelligence phenomena from the inside for two years: the funding rounds, the researcher defections, the compute deals. What follows is an attempt to work out which of these bets could plausibly pay off, and which are pricing reputation rather than results.
Global AI investment hit $243.9 billion in 2025, more than half of all venture dollars deployed worldwide, according to PitchBook. TFN’s own count of 2025’s biggest European rounds, Helsing and Mistral among them, told a similar story, and a separate TFN tally of the year’s ten AI mega-rounds put SSI and Recursive Superintelligence on the same list as OpenAI and Anthropic.
Here are the seven labs defining that race, listed alphabetically, and the questions each one still has to answer.
Aleph Alpha
Founders: Jonas Andrulis, Samuel Weinbach
Founded year: 2019
Total funding: $519.64M
Germany’s Aleph Alpha built its identity around sovereign AI for enterprises and public-sector customers years before “sovereignty” became a competitive category. It raised $500 million in a 2023 Series B backed by Schwarz Group, Bosch, SAP, HPE, and Burda.
That positioning became its exit strategy: in April 2026, Aleph Alpha agreed to merge with Canada’s Cohere, forming a transatlantic entity valued at around $20 billion, with Schwarz Group committing a further $600 million, as TFN reported.
Structurally, it is closer to an acquisition than a merger. Cohere shareholders take roughly 90% of the combined company. But the deal gives Aleph Alpha’s European public-sector relationships a bigger platform than it could have built on its own.
AMI Labs
Founders: Yann LeCun, Alexandre LeBrun, Laurent Solly, Pascale Fung, Saining Xie, Michael Rabbat
Founded year: 2025
Total funding: $1.03B
LeCun left his post as Meta’s chief AI scientist in late 2025 to build what he has long argued the industry is missing: world models, AI systems that build internal representations of physical reality instead of predicting the next word in a sentence.
AMI Labs, the Paris-based company, raised $1.03 billion at a $3.5 billion pre-money valuation in March 2026, backed by NVIDIA, Samsung, Temasek, and Jeff Bezos, in what TFN reported at the time was one of Europe’s biggest seed rounds ever.
LeCun has since joined 224 Ventures alongside DeepMind’s Oriol Vinyals, a sign that he is now backing the world-model thesis beyond his own company, as TFN covered here.
AMI’s early commercial foothold is in healthcare, through co-founder Alexandre LeBrun’s prior startup, Nabla.
Ineffable Intelligence
Founder: David Silver
Founded year: 2025
Total funding: $1.1B
Silver spent over a decade at Google DeepMind, leading the reinforcement learning research behind AlphaGo and AlphaZero before launching Ineffable in London. The company raised $1.1 billion in seed funding at a $5.1 billion valuation, which TFNreported at the time as one of the largest seed rounds Europe had seen.
Silver’s underlying thesis, laid out in an April 2025 paper he co-authored with Richard Sutton, argues that AI needs to learn from its own experience rather than human-generated datasets, what the pair call the “era of experience.”
That points toward robotics and scientific research as the most plausible early proving grounds, well before anything resembling a product.
Meta Superintelligence Labs
Leadership: Mark Zuckerberg, Alexandr Wang, Nat Friedman
Founded year: 2025
Total funding: N/A
Unlike the independent labs, Meta Superintelligence Labs does not need to raise a cent before it starts training models. It draws directly on Meta’s balance sheet.
Zuckerberg built the unit after growing frustrated with Llama’s reception, and the deal that defined it was not a funding round but an acquisition-adjacent investment: $14.3 billion for 49% of Scale AI, bringing founder Alexandr Wang in to lead the new lab, as TFN reported when the deal closed.
Scale itself has since moved on, hiring a new chief executive with a security and government background as it tries to diversify beyond its biggest shareholder and customer, TFN noted here. MSL is the one lab on this list whose survival is not existential for its parent company. That is both its biggest advantage and the reason it is easiest to underestimate.
Mistral AI
Founders: Arthur Mensch, Guillaume Lample, Timothée Lacroix
Founded year: 2023
Total funding: $4.01B
Mistral AI has taken the opposite route from the American labs: open-weight models and European sovereignty as the product, not an afterthought. Its €1.7 billion Series C, led by ASML, valued the Paris-based company at €11.7 billion, making it one of Europe’s most valuable independent AI labs, TFN reported.
Since then, it has kept building sovereign infrastructure rather than chasing a bigger foundation-model lead: a €1.2 billion investment in a Swedish data centre, an $830 million debt raise for a French compute cluster, and, as of June 2026, early talks to roughly double its valuation to €20 billion.
Mistral argues that Europe does not need to outscale OpenAI. It only needs to be the credible option when GDPR and the EU AI Act make American providers a harder sell for governments and regulated industries.
Recursive Superintelligence
Founders: Richard Socher, Tim Rocktäschel, Josh Tobin, Jeff Clune, Tim Shi, Alexey Dosovitskiy, Caiming Xiong, Yuandong Tian
Founded year: 2025
Total funding: $650M
Socher, Salesforce’s former chief scientist, and seven co-founders drawn from OpenAI, Google DeepMind, Meta AI, and Uber AI want to automate the AI research process itself: evaluation, data selection, training, even research direction.
Funding escalated fast, from a $500 million round at a $4 billion pre-money valuation to $650 million at $4.65 billion, backed by GV, Greycroft, NVIDIA, and AMD. TFN also reported its $410 million AWS compute commitment, part of a wider arms race in dedicated infrastructure for self-improving systems.
Recursive Superintelligence’s own roadmap talks about training a system with the research capacity of “50,000 doctors.” Whether recursive improvement actually compounds, or just burns compute faster, is the question critics keep raising, and Recursive keeps deferring.
Safe Superintelligence (SSI)
Founders: Ilya Sutskever, Daniel Levy, Daniel Gross
Founded year: 2024
Total funding: $7B
Safe Superintelligence, or SSI, the company Sutskever built after leaving his post as OpenAI’s chief scientist, was designed around a single premise: skip commercial products entirely and go straight for safe superintelligence.
It is one of the most closely watched labs in the industry precisely because it says almost nothing publicly. The company raised $1 billion in 2024, added $2 billion in 2025 at a $32 billion valuation, and took a further $5 billion from NVIDIA in July 2026. TFN reported that the deal pairs equity with priority access to NVIDIA’s next-generation Vera Rubin chips.
That takes SSI’s total funding to roughly $7 billion for a company with, by most reports, around 50 employees and no published research papers. TFN broke down the earlier $2 billion round here. Sutskever’s track record alone is apparently enough to justify pricing a company on conviction rather than output. Twice now, it has worked.
