$65M-Funded Essential AI — Transformer Co-Founders

Essential AI disclosed about $65 million in funding after two of the eight authors of the paper that invented the Transformer architecture started the company in 2023 — and it never published a valuation. This is what happened when the researchers who built the core technology underneath every modern LLM decided to compete with the companies now using it.

Ashish Vaswani and Niki Parmar co-authored "Attention Is All You Need" at Google Brain in 2017 — the paper that gave the world GPT, Claude, Gemini, and effectively every transformer-based model built since. In 2023 they started Essential AI, betting that enterprise-grade AI, not consumer chatbots, was the bigger opportunity. Here's the funding timeline, what the company actually builds, how its path compares to the other AI labs founded by famous researchers, and what happened to the team in 2026.

What Is Essential AI's Valuation in 2026?

Essential AI has not published a valuation. Its last disclosed financing is a $56.5M Series A in December 2023, led by March Capital with AMD, Franklin Venture Partners, Google, KB Investment, NVIDIA, and Thrive Capital participating, after an $8.3M seed led by Thrive Capital — "a total of nearly $65 million in funding to date," per the company's Business Wire announcement. There is no public record of a Series B. On June 23, 2026, Axios reported that Nvidia acqui-hired the team behind Essential AI, including Vaswani.

The Essential AI Funding Timeline

Essential AI raised two disclosed rounds in 2023, and the investor list reads like a who's-who of both venture capital and the biggest AI infrastructure companies. Google, Nvidia, and AMD all participated in the Series A — an unusual signal, since two of those three companies also compete with Essential AI's underlying research agenda.

Sources: Essential AI's Series A announcement (Business Wire, December 12, 2023) and Axios (June 23, 2026).

What Essential AI Actually Builds

Essential AI didn't chase a consumer chatbot audience. Its Series A announcement described "full-stack AI products that quickly learn to increase productivity by automating time-consuming and monotonous workflows" — what the company called building the "Enterprise Brain." That's a deliberate contrast to labs like OpenAI or Anthropic, which built consumer-facing products first and backed into enterprise revenue later. By mid-2025 the focus had moved toward open research: Vaswani told the Economic Times the team was building an open-source frontier model and wanted to be "the DeepSeek in the West," and the company released the Essential-Web v1.0 dataset (24 trillion tokens) in June 2025.

That focus shows up in the cap table too. Google, Nvidia, and AMD backing an enterprise AI startup as strategic investors — rather than just cloud credits partners — signals those companies saw Essential AI's research as complementary to, not competitive with, their own model lines. It's the kind of investor list you get when your founders wrote the paper the entire industry is built on top of. You can track how the rest of the frontier AI field is being priced on our AI Valuations dashboard.

Why the Transformer Connection Matters to Investors

"Attention Is All You Need" has eight listed authors, and at least four of them have since founded or co-founded venture-backed AI companies: Ashish Vaswani and Niki Parmar started Essential AI; Aidan Gomez co-founded Cohere, which was set to be valued at about $20B in its April 2026 merger with Aleph Alpha, per the FT via TechCrunch; Noam Shazeer co-founded Character.AI, and Google later paid more than $2B to acqui-hire him and part of that team, according to Axios. That single 2017 research paper has arguably generated more direct startup formation than any other paper in computer science history, which is part of why LPs and strategic investors treat "wrote a foundational paper" as a legible signal in a market where most AI pitches are still unprovable claims about future capability.

That pedigree is also why Essential AI could raise a $56.5M Series A straight out of stealth, with three of the largest chipmakers and cloud providers in the world writing checks. Most Series A enterprise AI startups don't get Google, Nvidia, and AMD in the same round — they get one strategic, if any, and mostly financial VCs filling the rest. Essential AI's cap table looked more like a later-stage round from the day it left stealth, which is the clearest evidence that founder pedigree still opens doors in AI fundraising.

Essential AI vs Other Researcher-Founded AI Labs

Essential AI is the odd one out among the "neolabs" founded by famous AI researchers: it never published a valuation, while its peers carry valuations that have detached almost entirely from revenue. Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, closed a $2B seed at a $12B valuation in mid-2025, and in September 2026 was reported to be in talks to raise at $40B — a $50B round it sought earlier never closed. Safe Superintelligence, founded by OpenAI co-founder Ilya Sutskever, raised $1B at a $5B valuation in 2024 and then $2B at a $32B valuation in early 2025, per TechCrunch — still with no public product. Essential AI took a far smaller capital path for a comparable pedigree.

Figures blended from Business Wire, TechCrunch, Axios, CNBC, The Register, and Anthropic. Valuations for private companies are based on the latest disclosed or reported funding rounds and can move quickly; Cohere's figure is the FT-reported merger valuation from April 2026.

What Essential AI's Path Means for Investors

The pattern worth watching in AI right now is the gap between pedigree and capital. Essential AI raised about $65M on one of the strongest research résumés in the field, then — per Axios — saw its team acqui-hired by Nvidia in June 2026, a move Axios described as "more along the lines of the expected consolidation of startups." Safe Superintelligence, by contrast, raised $1B and then $2B at a $32B valuation with no shipped product, and in July 2026 added a multibillion-dollar Nvidia investment, per TechCrunch. Pedigree opens the door to an early round; it doesn't by itself secure the capital that frontier-scale training demands.

The other signal worth tracking: strategic investors backing a startup that competes with their own AI ambitions. Google, Nvidia, and AMD all wrote checks into a company founded by researchers who left Google, and Nvidia ultimately hired the team. That's what optionality looks like when it pays off for the strategic investor. Watch which enterprise AI labs keep pulling that kind of dual-purpose capital on our Benchmarking dashboard, because it often predicts who gets acquired versus who goes public.

What Happened to Essential AI in 2026?

On June 23, 2026, Axios reported that "Nvidia also acqui-hired the team behind Essential AI, including AI researcher Ashish Vaswani," listing it among a run of high-profile AI talent moves that week. Axios did not report financial terms, and Chinese tech outlet 36Kr noted on June 27 that neither Nvidia nor Essential AI had officially commented.

The broader question the episode raises is whether "enterprise-first, lower-burn" is a durable strategy for a research-led lab once the frontier labs compete on price. OpenAI, Anthropic, and Google all now offer enterprise API tiers with fine-tuning and dedicated capacity, which narrows the differentiation gap smaller labs were counting on. Founder pedigree got Essential AI a strategic Series A; it did not, on the public record, produce a second priced round.