What Is the AI Bio Threat Open Letter About?
The AI bio threat open letter, signed by Demis Hassabis, Sam Altman, Dario Amodei, Alex Wang, and dozens of other high-profile leaders across AI, tech policy, nucleic acid synthesis, and biotech, calls on the US government to mandate screening and recordkeeping for nucleic acid synthesis orders. In plain terms: they want companies that chemically print DNA and RNA sequences to be legally required to check whether a customer is ordering something that looks suspiciously like a dangerous virus — and to keep records of who ordered what. It sounds like common sense. The alarming part is that this isn't already the law.
The letter, covered in detail by TBPN's Brandon Guerell, isn't the typical AI doom marketing you might expect. It's not claiming a frontier model just unlocked the ability to generate pandemic-level pathogens in a single prompt. It's a more grounded, specific, and arguably more urgent ask: close the regulatory gap in the nucleic acid synthesis industry before AI makes that gap catastrophically exploitable.
Can AI Actually Be Used to Create Dangerous Viruses?
To understand why this letter matters, you need a quick history lesson. In 1981, researchers published the complete primary structure of the polio virus genome in the journal Nature — essentially open-sourcing the blueprint for one of the most feared diseases in modern history. At the time, this was a scientific achievement. In retrospect, it was a warning shot.
By 2002, researchers demonstrated just how dangerous that kind of openness could be. Using only the publicly available sequence data — no physical virus sample required — they chemically synthesized short DNA fragments, assembled them into a full-length copy of the polio virus genome, and used that DNA to produce infectious viral RNA. They recreated polio from a text file. Three years later in 2005, scientists used similar techniques to reconstruct the 1918 Spanish flu virus, which killed 675,000 Americans and carried a 2–3% mortality rate among those infected.
The implication is stark: you no longer need a physical sample of a dangerous virus to recreate it. You just need the code. And as AI systems become more capable at designing and optimizing biological sequences, the barrier to reconstructing — or even engineering new — dangerous pathogens keeps dropping. AI excels at exactly this kind of problem: tight feedback loops, verifiable outputs, pattern recognition across massive datasets. The same properties that made AI so powerful in cybersecurity make it a potential accelerant in bioweapon development.
This is why the great houses of AI, as Guerell put it, have united behind the bio threat. It's not hypothetical doom. It's a documented capability that's been real since 2002 and is getting more accessible every year.
What Is Nucleic Acid Synthesis Screening and Why Does It Matter?
Nucleic acid synthesis companies are essentially biological printing services. You send them a sequence — a string of A, T, G, and C (or A, U, G, C for RNA) — and they chemically manufacture it and ship it back to you. These services are essential for legitimate research: drug development, gene therapy, vaccine design, and much more.
But the same services could theoretically be used to print the building blocks of dangerous pathogens. Screening means checking incoming orders against known dangerous sequences — flagging anything that looks like a pathogen of concern before it gets manufactured and shipped. Recordkeeping means maintaining logs of who ordered what, creating accountability and an audit trail.
The letter's signatories are asking the government to make this mandatory. Right now, compliance is voluntary, and as the discussion on TBPN made clear, voluntary compliance has real limits. The customer verification, the sequence screening, the recordkeeping — all of it should be required by law, not encouraged by industry guidelines.
What Is the International Gene Synthesis Consortium?
The International Gene Synthesis Consortium (IGSC) was established in 2009 as an industry-led effort to address exactly this problem. Member companies commit to screening synthesis orders for dangerous sequences and verifying customer legitimacy. By current estimates, roughly 80% of commercial synthesis volume has opted into the IGSC framework.
That sounds reassuring until you dig into the details. Membership is voluntary. Reporting is self-reported. There's no government verification of the numbers, no independent audit confirming that member companies are actually following through on their commitments. The 80% figure comes from the organization itself, and the government isn't cross-referencing it against its own data.
And then there's that remaining 20%. In a domain where a single bad actor with access to the right sequence and the right synthesis equipment could cause mass casualties, 20% unaccounted-for synthesis capacity is not a rounding error. The letter's signatories are essentially saying: 2009 was a good start, but it's 2025 now, and we have AI. It's time to finish the job with binding regulation.
HHS also has guidance around nucleic acid synthesis risks, but again — it's voluntary. The through-line of this entire issue is that the industry has been relying on good faith, and good faith isn't an adequate biosecurity strategy.
Is Biotech Back? Why AI Is Reigniting the Sector
The signing of this letter also reflects something broader happening in biotech right now: the sector is experiencing a genuine renaissance, driven largely by AI. Just 14 months ago, biotech investors were openly questioning why anyone would allocate to the asset class. Returns had been poor, exits had dried up, and the narrative had collapsed.
That narrative has flipped. Isomorphic Labs spun out of DeepMind. Brian Armstrong founded New Limit. Retro Biosciences is making moves. Altos Labs is attracting serious capital. Anthropic acquired Coefficient Bio. Jensen Huang and Larry Ellison at Oracle are both actively involved in biotech initiatives. The convergence of AI capabilities with biological research is creating new categories of companies that didn't exist five years ago.
The deal flow is smaller than AI or semiconductors in raw dollar terms — we're not talking about trillion-dollar IPOs — but the momentum is real and accelerating. And unlike pure software AI plays, biotech companies tend to be more acquisition-friendly, with a culture of flipping companies at meaningful multiples across multiple exits. The power law dynamics may be slightly less extreme than in pure AI, but the opportunity is enormous.
Ramp Raises $750M at $44B Valuation: What's Driving the Surge?
Shifting gears to fintech: Ramp just closed a $750 million funding round at a $44 billion valuation. The headline number is impressive, but the standout data point is this — the last time Ramp was growing this fast, it was one-hundred-and-twentieth the size. That's not a typo. The growth rate they're currently sustaining is comparable to their earliest hyper-growth phase, but at a scale 120 times larger.
For context, Ramp is now worth more than PayPal on a market cap basis, despite PayPal generating $32 billion in annual revenue. That comparison has generated significant debate, but the core argument for Ramp's premium is simple: momentum. PayPal is widely seen as having negative momentum — a legacy business struggling to find its next act. Ramp has the opposite problem: it's growing faster than it can explain.
Ramp CEO Eric Glyman also published an essay framing the company's mission around what he calls the "three pillars" of finance: people, vendors, and now tokens. His argument is that AI-generated spend — tokens consumed by AI systems — represents a new category of financial activity that existing corporate finance infrastructure wasn't built to handle. Ramp's thesis is that managing this quadrillion-token blind spot is the next major frontier in business finance.
Strip away the framework and the core value proposition is timeless: who spent what, was it worth it, and what's the bill next month? That's been the question at the heart of all business finance since ancient times. Ramp's bet is that answering it for the AI era is a multi-hundred-billion-dollar opportunity.
Why Is Benchmark's New Growth Fund Such a Big Deal?
Benchmark, long considered the last major pure venture capital firm at the top tier, has raised $2 billion across two new funds — including its first-ever dedicated growth fund. For an institution that built its reputation on disciplined early-stage investing and fierce focus, this represents a meaningful strategic evolution.
The firm's addition of Ev Randall, who brings a background spanning Bond Capital and Founders Fund, positions them well for growth-stage investing. Whether this signals a broader shift in Benchmark's identity or a pragmatic response to market conditions remains to be seen. But for the venture ecosystem, the last holdout of pure early-stage orthodoxy is now playing the full stack. That's a signal worth paying attention to.
