Antithesis, a Northern Virginia startup that bills itself as infrastructure for software that never fails, raised a $105 million Series A led by Jane Street, a bet that stress testing distributed systems is as important for blockchains as it is for high-speed commerce.
The company’s platform uses deterministic simulation testing, running large-scale, production-like simulations, to reveal the types of edge cases that can exploit in live networks, Antithesis said in a press release Wednesday.
When a failure occurs, Antithesis said it can reproduce the error exactly, helping engineers isolate problems without the usual limbo of being unable to reproduce, a familiar weakness for cryptographic protocols where small glitches can cause instability in the chain.
Other investors in the funding round include Amplify Venture Partners, Spark Capital, Tamarack Global, First In Ventures, Teamworthy Ventures and Hyperion Capital, along with the likes of Patrick Collison, Dwarkesh Patel and Sholto Douglas, the company said.
Antithesis has leaned on the credibility of cryptocurrencies, saying that the Ethereum network used its simulations before The Merge to model extreme conditions and detect vulnerabilities before the proof-of-stake transition.
The company also cited finance, artificial intelligence, blockchain and data infrastructure clients, and said revenue has increased more than 12-fold over the past two years.
Antithesis said it will use the proceeds to expand engineering, increase automation, scale go-to-market globally and drive distribution through channels including AWS Marketplace.
AI agents are now capable enough to identify exploitable weaknesses in smart contracts, and attackers could already use them to weaponize those flaws, according to new research from Anthropic’s Fellows program.
The researchers said that as AI models become cheaper and more capable, automated hacking could spread from decentralized finance (DeFi) vulnerabilities to a broader range of software and critical infrastructure bugs.
Read more: Anthropic Research Shows AI Agents Are Getting Closer to DeFi’s Real Attack Capability
