Gensyn uses blockchain to connect machine learning researchers with compute power. Check out the 11-slide pitch deck it used to land $6.5 million.

Gensyn cofounders Dr Ben Fielding and Harry Grieve (l-r)
  • A startup connecting cryptocurrency miners with machine learning researchers raised $6.5 million.
  • London-based Gensyn claims it will help researchers access the compute power they need to train AI.
  • The cofounders walked Insider through the 11-slide pitch deck they used to raise the cash. 

When Ben Fielding and Harry Grieve met at the kick-off weekend of an accelerator program in March 2020, they were told it would be the last time they’d see each other for some time.

Grieve and Fielding had joined Entrepreneur First’s six-month program, where would-be founders mix with one another in the hope of launching a startup together.

The pair clicked immediately and hunkered down throughout the pandemic while working on their new business Gensyn.

The startup connects machine learning researchers with the computing power they need to train AI models. Gensysn has just raised $6.5 million, following on from a previously unannounced $1.1 million pre-seed raise.

Researchers often need to conduct heavy calculations that are essential for everything using AI from self-driving cars to computer-generated images but often struggle to secure the computer power to do so.

Gensyn aims to address this using blockchain. People can offer up their computer power, commonly a graphics processing unit (GPU), to the blockchain and get paid for crunching machine learning training data submitted by researchers.

The startup will have its own token that is mined similarly to cryptocurrencies by users solving the complex calculations required by researchers. Once solved, miners then add their solution to the blockchain to earn rewards. Gensyn claims its system will help address the enormous amounts of wasted energy associated with cryptocurrency mining.

“Training these models is super high value from a societal perspective,” Grieve told Insider.

The Genysn cofounder aims to tap into the market for deep learning compute, which he believes to be worth $181 billion. Deep learning is a complex subsection of machine learning, which involves a series of algorithms that mimic the way the human brain works.

Researchers currently rely on stockpiling costly equipment or renting power from the likes of AWS or Microsoft Azure. Eventually, this becomes too expensive and poses a risk to research, the company said – something Fielding experienced while undertaking a PhD in machine learning.

“If you have a GPU and you want to use it to mine cryptocurrency, you might want to mine Ethereum,” Grieve said. “Ethereum – for a standard middle of the range video GPU – pays about 34 cents an hour, when we did our calculations. That same GPU has been sold on AWS for $2 an hour. There’s a disconnect there.”

Gensyn uses a proof-of-stake method to process calculations and add them to the blockchain, and has established its own system for verifying work has been done. This means that miners put up tokens that they already own as collateral and are incentivized to complete the task correctly because their existing tokens are otherwise at stake.

The startup, which is currently pre-launch, will use the fresh funds to build out its platform and hire its first employees.

The round was led by Web3 investor Eden Block, with participation from Galaxy Digital Ventures, Maven 11, Coinfund, Hypersphere, and Zee Prime. 7percent Ventures and Counterview Capital led the pre-seed, with participation from Entrepreneur First and id4 Ventures.

Grieve and Fielding talked us through their 11-slide pitch deck below:

The title slide puts Gensyn’s vision front and center


Gensyn’s end goal to unlock compute power “for the frontiers of artificial intelligence” is front and center of their deck, which was sent to investors ahead of their first meeting alongside a litepaper. 

“That kind of goes all the way back to the beginning of our time Entrepreneur First,” Grieve said. 

The deck starts by outlining the current demand


“We’re big fans of the problem-solution format of decks,” Grieve said. “It’s just like clear for us.”

The graph shows that the complexity of a neural network – a series of algorithms used to power AI – is doubling every three months. 

For this reason, Gensyn priced the demand for more computer power as a “huge opportunity.”

The next slide outlines the problem – and Gensyn’s opportunity


This slide shows supply is not keeping up with demand and outlines the reasons why, from stockpiling chips to compute power being wasted. 

Web3 VCs were often unfamiliar with the machine learning space, the cofounders said. They understood the solution as a computational problem, without the AI component. 

“It’s a kind of weird marriage of worlds,” Grieve said. “The fact that it was connected to a specific market – machine learning, which is super high value is like – was a separate thing to the technical discussion.”


Zooming in on wasted compute power


Grieve said that VCs were bullish on “Web2 people coming to Web3,” which may work in the startup’s favour.

“They need Web2 people to come over because, if they don’t, they don’t get adoption.”

Gensyn sees it having the “greatest effect” on wasted compute power. As well as individuals being able to connect to the network, the company sees data centers offering up idle hardware. 


Laying out the solution


Gensyn provided an overview of its solution, detailing its core features. 

The cofounders said they were selective with their pitches, only spending time with Web3-savvy investors. For this reason, they opted not to break down jargon like “L1 trustless protocol.”

Being able to trust and verify the work done on the blockchain is not a new issue, Fielding said. 

“We were pretty disciplined in who we wanted to talk to. It was a lot of fundamental infrastructure web3 VCs,” he added. “A lot of them have seen potential solutions in the past so a lot of this detail was accessible to them. Whereas, in another raise in another space, we wouldn’t have included this much technical detail. It made sense for who we were targeting.”

Grieve added that they turned down “a lot” of non-Web3 investors otherwise they would have “had to create two decks and split messaging.” 

Gensyn claims to be faster than the “best competitor”


Slide six goes deeper on Gensyn’s verification process.

Other networks replicate work to establish a consensus and ensure a calculation is completely correct but this is not needed with Gensyn’s process, the cofounders said. This makes Gensyn’s process more time and energy-efficient. 

The company claims it was 1,350% faster than “the best competition” when running a model that turns handwritten numbers into digits. 

This redacted slide shows how Gensyn stacks up against its competition


This redacted slide revealed how Gensyn stacks up compared with competitors. Intended as an “ah ha” moment for investors who understand the market, the supply shortage, and the verification issues, the redacted two-by-two chart showed how the company’s solution related to everything else out there. 

“It put us next to names that they’d heard of, and the degree to which we were higher scale and lower cost, theoretically, at least,” Grieve said. 

The benefit of this is to contextualize Gensyn versus the rest of the market, including both Web3 and Web2 companies. 

The end goal is to be infrastructure others build on


Gensyn said it has a community of around 350 people, which is larger than stated in the deck, and “a lot of them are keen to use the product when it is out.”

The end goal is to build out an ecosystem where third-party tools are used on top of Gensyn’s software, which would act as infrastructure. 

People can train models directly on Gensyn, the company said, but “more likely than not” they’re going to use it indirectly, via something else that uses it. 


Clearly outlining the benefits of the solution


The cofounders plan to split the business in two: Gensyn Limited will develop the technology and the Gensyn Foundation will represent the interests of the protocol.

Governance of the technology will be “on-chain,’ which means token holders vote on how it is run by staking tokens.



Presenting the founders and why they are the right people to build the solution


Gensyn’s penultimate slide offers a context as to why they are the right team for the job and who has backed them so far, including both pre-seed investors and angels.

The cofounders said they removed two slides, one detailing its roadmap and another explaining what the money would be used for. Expecting to grow the team, Fielding and Grieve listed target candidates and salaries. 

The final slide is a reminder of how to get in touch with the company